DJI TERRA

Operation Guide
V4.2 2024.07
Updated In: July 2024
Applicable version: This Operation Guide is prepared for DJI Terra v4.2 and later versions.
Operations may vary between this version and earlier versions.
Document description: This document only briefly describes certain key functions and is not
a product manual. For functions not mentioned in this document, please download the user
manual from the DJI Terra official website to find out more:
https://www.dji.com/downloads/products/dji-terra

Change Log
Version Date Revision
v1.0 February 2022 First version created
1. Replaced operation screenshots with version 3.5.0.
2. Modified GCP import method to version 3.5.0.
3. Added description for Seven-parameter calculation.
4. Added description for Vertical Datum import.
5. Added description for Light Uniformity/Haze Reduction of 2D
v2.0 August 2022
orthophotos.
6. Added description for 3D Reconstruction: ROI/Block Splitting function.
7. Added description for Custom Model Origin function.
8. Added description for Merged Output function.
9. Modified description for error prompts.
1. Added description for third-party cameras compatible with
Visible Light Reconstruction.
2. Added description for Smooth Point Cloud function.
3. Added the LiDAR Point Cloud Accuracy Check function.
v3.6 February 2023 4. Added the Detailed Inspection function supported by Matrice
30 and Mavic 3 Enterprise series.
5. Modified description for Light Uniformity/Haze Reduction function.
6. Modified description for Distance to Ground/Subject in
Reconstruction Settings.
1. Updated the description of the correspondence between
point cloud reconstruction scale and computer performance.
2. Updated the description of the licenses of the online version.
3. Added the description of the semi-automatic GCP marking,
"Reduce Model To", and Work with DJI Modify features for 3D
reconstruction; added the description of the Ground Point
Classification feature; added description of the Digital Evaluation
Model (DEM), Contour, and Grid of Points & TIN features;
v4.2 July 2024 updated the description of Region of Interest (ROI) or blocks.
4. Updated the description of Worker Device Management for
cluster reconstruction.
5. Added the description of new features for point cloud
reconstruction, including Cloud PPK Service (only for Chinese
Mainland), Substation Reconstruction scenario, Ground Point
Classification, Generate DEM, Contour, and profiles.
6. Optimized the section of backend software for model application.
7. Added instructions on uploading project error data and logs.
Table of Contents

Introduction to DJI Terra 3

Product Overview 3
Product Highlights 3
Configuration Requirements 3
Software Version 5
Reminders for First Use 5
Download Links 5

Visible Light Reconstruction 6

Function Overview 6
Importing Data 6
Aerotriangulation 18
2D Reconstruction 28
3D Models 41
Reconstruction Mission Management 52
FAQ 52

Cluster Reconstruction 53
Function Overview 53
Operating a Worker Device 53
Operating a Control Device 55
Cluster Reconstruction Process 57

2D Multispectral Reconstruction 58
Function Overview 58
Importing Data 58
Radiometric Correction (Optional) 59
Output Indices 60
2D Multispectral Map File Format and Storage Path 61

LiDAR Point Cloud Processing 62

Function Overview 62
Importing Data 62
Base Station Center Point Settings (Optional) 63
Point Cloud Density 64
Scenarios 65
Advanced Settings 68
Start Processing 69
Viewing Output 70
LiDAR Point Cloud File Format and Storage Path 72

© 2024 DJI All Rights Reserved 1

DJI Terra Operation Guide

Detailed Inspection 73
Function Overview 73
Importing Data 73
Route Setting and Planning 74
Route Export and Execution 80

DJI Modify 81
Introduction to DJI Modify 81

Appendix - Error Alerts and Operating Suggestions 84

2 © 2024 DJI All Rights Reserved

Introduction to DJI Terra
Product Overview
DJI Terra is a PC application developed by SZ DJI Technology Co., Ltd. and it focuses on 2D
Orthography and 3D Model Reconstruction while supporting functions such as 2D Visible Light
Reconstruction, LiDAR Point Cloud Processing, and Detailed Inspection. It is a one-stop solution
that helps enterprises fully enhance their operational efficiency, providing a complete suite of
reconstruction features for verticals such as surveying, power generation, emergency rescue,
construction, transportation, and agriculture.

Product Highlights
1. High processing efficiency: DJI Terra’s standalone reconstruction processing speed is 3-5
times that of other mainstream software. For cluster reconstruction the processing efficiency
increases several fold.
2. Excellent reconstruction quality: Superb modeling that can recreate minute structures
from close-up photography data. High reconstruction accuracy down to the centimeter level
without ground control points.
3. Large throughput 64G RAM enables processing of up to 25,000 photos for standalone
reconstruction and 400,000 photos for cluster reconstruction.
4. Supports cluster reconstruction: 2D and 3D reconstruction support parallel processing of
all PC network clusters within a local area network, increasing efficiency by multiple times.
5. Easy to operate: Operation is easy without complex parameter settings. Easily accessible by
beginners.

Configuration Requirements
The data-processing computer must be installed with an NVIDIA graphics card and Windows 7
or higher, while meeting the following hardware specifications:

2D Map Reconstruction /
LiDAR Point Cloud
Hardware Real-time 2D Mapping 3D Model Reconstruction /
Processing
Real-time 3D Mapping
CPU i5 or later
GPU NVIDIA graphics cards with a compute capability of no less than 6.1*
VRAM No less than 4GB
RAM No less than 8GB No less than 32GB
HDD 200GB Free (basic requirement) or SSD+200GB Free (better)

* For DJI Terra earlier than v3.9.0, real-time 2D mapping does not require NVIDIA graphics cards with a
computing power of 6.1 or above.

© 2024 DJI All Rights Reserved 3

DJI Terra Operation Guide

The table below shows how RAM size determines the maximum photo processing quantity for
2D and 3D reconstruction:

Standalone Computation Cluster Computation

RAM Max Processing Quantity Max Processing Quantity
(Photos) (Photos)
32GB 12,800 192,000
48GB 19,200 288,000
64GB 25,600 384,000
128GB 51,200 768,000

* RAM refers to the available RAM of the DJI Terra software (available RAM = total RAM - RAM used
by the system and other software). Before reconstruction, you may check the remaining RAM by
accessing Windows Task Manager.
* Maximum photo processing quantity is unrelated to the resolution of a single image. That means the
maximum processing quantities for P1 and P4R are the same.
* The RAM for cluster computation refers to the available RAM of the control device.

The table below shows how RAM determines the maximum data size for LiDAR point cloud
processing:

RAM Max Raw Point Cloud Size

16GB 16GB
32GB 32GB
64GB 64GB
128GB 128GB

* Raw point cloud size shall be calculated based on the total LDR and LDR~1 file size in the imported
file, and is not affected by the JPG format of any photo.
* If the RAM is too small, you may split a big project into multiple tasks to be processed, then merge
the multiple LAS files using third-party point cloud analysis software.
* LiDAR point cloud processing only supports standalone computation and not cluster computation.

4 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Software Version
DJI Terra comes in four versions: Agriculture, Pro, Electricity, and Cluster. The paid features
supported by each version are shown below. Note that cluster computation is available only in the
Cluster version. In the Education version, visible light reconstruction missions support no more
than 500 photos and point cloud missions support up to 8 GB of raw point cloud data. The same
software is used by all the versions. The availability of paid features on each version is determined
by their respective license.

Reminders for First Use

There are two types of product licenses for DJI Terra: online and offline. For online licenses, paid
features can only be used if the license is verified on a computer that remains connected to the
Internet. Offline licenses cater to scenarios requiring greater information security, where paid
features are available even when the equipment is offline (the device used for modeling does
not need to be connected to the internet, as long as an online computer is used to generate the
license). Online licenses can be activated at https://license.dji.com/, while offline licenses can be
activated at https://terra-license.djiservice.org/.
For details relating to software installation, offline license activation and other relevant matters,
please refer to the “For First-time Users” and “Guidelines for DJI Terra Offline License” chapters
in “Preparation Before Using DJI Terra”, available on the DJI Terra official website.

Download Links
The latest software and manual as well as past software versions can be downloaded from the
DJI Terra official website: https://www.dji.com/cn/dji-terra?site=brandsite&from=nav
© 2024 DJI All Rights Reserved 5
Visible Light Reconstruction
Function Overview
Visible light reconstruction includes 2D and 3D reconstruction. In 2D reconstruction, Digital
Surface Models (DSM) and Digital Orthophoto Maps (DOM) of an area are generated using images
captured by a drone, based on the principles of photogrammetry; while 3D reconstruction relies
on photogrammetry, multiple view geometry in computer vision, and computer graphics to
generate a 3D model or 3D point cloud model of a scene using drone-captured images.
You may obtain high-precision 2D maps or 3D models through the visible light reconstruction
feature, for a wide range of uses including topographical mapping, engineering survey and
maintenance, geological disaster investigation, fire and disaster rescue, national land survey,
urban planning, cultural heritage conservation, and crop protection.
The basic workflow of 2D reconstruction is as follows: Import data -> Aerotriangulation -> 2D
reconstruction
The basic workflow of 3D reconstruction is as follows: Import data -> Aerotriangulation -> 3D
reconstruction
Aerotriangulation is an essential step in 2D and 3D reconstruction. It can be done separately or
together with 2D or 3D reconstruction. The detailed data processing process is as follows:

Importing Data
Create Reconstruction Mission
Open and log into DJI Terra, click “New Mission” in the lower left corner and select “Visible Light”
mission type.

Add Photos
1. Add raw images through either of the following methods:
Click , then select photos to add from the computer. You may press “Ctrl + A” to select all
photos to import.
Click , then select the folder in which the desired photos are located. If there are sub-
folders, all images in them will be added automatically.
Note that the folder path name must not include special characters, such as #. Otherwise,
the mark view on the GCP page cannot be displayed.
6 © 2024 DJI All Rights Reserved
 DJI Terra Operation Guide

2. Camera POS Display

After adding images, the icon will appear on the top right corner of the map. The location
corresponding to each photo will be displayed as a round dot on the 2D map.

3. Photo Management
Click on the right of an image to manage photos. The photos are displayed in groups
according to their folder. Expand the list of each group to check and manage photos.

4. Select Photos for Designated Area

To retain or remove photos from a designated area, perform the following on the photo
management interface:
a. Add boundary points
Click the icon, then use the left button of your mouse to add boundary points on the
map to drag-select an area. If you have previously set a KML area, you may click the
icon to import the KML file, and the points contained in the file will become boundary
points forming a drag-selected area.

© 2024 DJI All Rights Reserved 7

DJI Terra Operation Guide

b. Edit Boundary Points

Click and select a boundary point with the left button of your mouse, press and drag the
left button to adjust the position of the boundary point, and click on a boundary to insert
a new boundary point. Click to remove the current selected boundary point, and click
to remove all boundary points.
c. After selecting an area, press the right button of your mouse. In the pop-up menu, select
photos to remove from within or outside the designated area.

Thereafter, click to return to the reconstruction page.

5. View Camera Information

If the imported photos were captured by a DJI Enterprise drone with a third-party camera
that is compatible with DJI Terra, a green check mark will appear which indicates that the
camera information is complete and includes all details such as the focal length and the
principal point. Click on the “eye” icon to check the camera details:

8 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

6. Input Camera Information

If the imported photos were not captured by a DJI Enterprise drone or with a camera that
is compatible with DJI Terra, then a yellow exclamation mark will appear next to “Camera
information”. When this happens, processing the data directly will result in accuracy loss and
even aerotriangulation errors. In this situation, click the “eye” icon to input the camera details
manually.

© 2024 DJI All Rights Reserved 9

DJI Terra Operation Guide

If you are using a single-lens camera, then input the relevant information directly on the
“Camera Type” page. After entering the name of your camera, press the Enter key and then
input a 35mm equivalent focal length.

This will greatly increase the aerotriangulation accuracy. To achieve greater accuracy, we
recommend clicking "Advanced Settings" then enter these parameters: F, CX, CY, K1, K2, K3,
K4, P1, and P2.

Camera parameter details are usually provided in full by their manufacturers. If the F value is
not provided by a manufacturer, an approximate value can be calculated using the following
formula:

F = (Focal Length/Sensor Size) x Width or F = Physical Focal Length (mm) / Pixel Size (μm)

10 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

This formula is for reference only. Please follow parameters provided by the camera
manufacturer.

Example of a camera parameters file from a camera

manufacturer (the data shown are not real)

© 2024 DJI All Rights Reserved 11

DJI Terra Operation Guide

Fill in the information based on the above camera details as follows:

If the imported photos were captured by multiple cameras, such as a five-lens oblique
camera, the photos shot by each camera should be saved in separate folders. Do not save
data from different cameras into the same folder. Select “Folder” and input the parameters
of each camera separately.
After entering the details of the camera/each camera, click Apply to save them. Expand the
“Camera Name” list to select and use any of the saved camera details.
By clicking the exclamation mark button on the right of “Camera Name”, then clicking "View
"and expanding the camera details list, you can import or export any camera details already
saved in the list, for use with other camera details in DJI Terra.

12 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

To ensure information accuracy, we recommend that you first enter the initial camera details
provided by the manufacturer, then perform an aerotriangulation computation using DJI
Terra and export the aerotriangulation information. Thereafter, update the camera details
according to the optimized parameters in the aerotriangulation report. Using these intrinsic
parameters will yield better modeling results.

Below is a screenshot of part of the camera parameters from an aerotriangulation report. F

value represents the focal length.

Import Image POS Data

Image POS data record the geographical location, attitude and other positioning details of a
photo. Accurate POS data can help improve reconstruction speed and output accuracy. The POS
of some third-party cameras are separate from their images, and needs to be imported through
an additional step. This step is not required for data captured by a DJI drone and payload (such
as Phantom 4 RTK and Zenmuse P1), as their POS will be written into the photos.
Certain third-party cameras do not write POS into their images, in which case you may use the
import image POS function to match the photos with their POS. If output based on the local
coordinate system is required, you can use a coordinate conversion tool to convert raw image
POS into local coordinate POS, then import it. The steps are as follows:

© 2024 DJI All Rights Reserved 13

DJI Terra Operation Guide

1. Prepare a POS data file (shown below) according to the import format requirements
for image POS data. DJI Terra supports the import of data in TXT and CSV format. The
data must at least contain the image name (an absolute path with a .jpg suffix), latitude
(X/E), longitude (Y/N), altitude (Z/U), and euler angle. For .txt files, the comma (,), period (.),
semicolon (;), space, and tab can be used as a separator. Please ensure the image name in
POS corresponds uniquely to the imported image data name.

To convert the coordinates of the intrinsic POS data of an image, click “Export POS
Data” on the right of “Image POS Data” to export the POS, convert it using a third-party
coordinate conversion tool (such as Coord), then import it back.

2. Click “Import POS Data” on the right of “Image POS Data”, then select the POS data file you
wish to import. Note that if the photos do not contain POS, the software page will still not
display any POS point even after POS has been imported. However, the imported POS data
will be used during reconstruction. If the photos have intrinsic POS, the original data will be
overwritten by any new imported and converted POS.

3. In “File Format”, click the import data format and set “Lines to Skip from Top”, “Decimal
Separator”, and “Column Separator”.

14 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

The “Define Data Column” window will display data based on your settings in “File Format”.
“Lines to Skip from Top” is for deleting the titles and sample lines of the data file.
“Decimal Separator” allows you to define how decimal points should be displayed (which
varies between countries).
With “Column Separator”, you can set the separator to be used between the columns in the file.

4. In “Data Properties”, set “POS Data Coordinate System” and “Geoid Settings”. For a special
coordinate system, you may select an arbitrary coordinate system. For an elevation system
that does not exist in the system, you may select “Default” (ellipsoidal height) in “Geoid
Settings”.
5. With “Height offset”, the overall height can be increased or reduced. Vertical irregularity
between ellipsoidal height and altitude in a small scope can be treated as a constant value.
By setting this parameter, you can enable quick adjustment of ellipsoidal height to altitude.
6. In “Euler angle”, you may select the image attitude details. The details supported by DJI Terra
are Yaw, Pitch, Roll and those in the Omega, Phi, and Kappa format. If attitude information is
not applicable, you may select “None”.
7. “POS Data Accuracy” enables you to set the accuracy of your image POS data. If you opt
for DJI Terra’s default accuracy, the software will determine based on the photos’ XMP
information whether each photo was captured in RTK status. If so, the default horizontal and
vertical accuracy shall be 0.03 m and 0.06 m respectively; otherwise they shall be 2 m and
10 m respectively. If you are using a third-party camera or importing PPK post-differential
results, define a custom accuracy and set the accuracy options for each data column.
8. In “Define Data Column”, you can select the corresponding item for each data column, then
click “Import” at the bottom to import the POS data.

© 2024 DJI All Rights Reserved 15

DJI Terra Operation Guide

Define Data Column Total Rows 25

Undefined Undefined Undefined Undefined Undefined Undefined Undefined

Name Latitude Longtitude Altitude Yaw Pitch Roll

31.75678992 123.103553535 23.13423423 -179.9633456 -89.476576546 0

31.75689223 123.103689989 23.36525445 -179.9663678 -89.637746546 0

31.75646378 123.103469989 23.26243534 -179.9645676 -89.987976789 0

31.75711232 123.103359979 23.56464565 -179.9879788 -89.768798546 0

31.75699889 123.103345346 23.67648356 -179.5789868 -89.786586465 0

31.75677782 123.103398979 23.13443545 -179.9885777 -89.989605456 0

31.75669682 123.103422424 23.89876454 -179.9898599 -89.786965764 0

111.10000008 123.100000009 23.10000008 -123.1000007 -1.1000000009 -12.10000008
31.75683357 123.103447722 23.36754364 -123.1000007 -1.1000000009 -12.10000008

31.75692240 123.103528574 23.98764533 -123.1000007 -1.1000000009 -12.10000008

Photo name, latitude (X/E), longitude (Y/N), and altitude (Z/U) are required.
111.10000008 123.100000009 23.10000008 -123.1000007 -1.1000000009 -12.10000008
You cannot select the same
111.10000008
definition
123.100000009
for multiple
23.10000008
columns.
-123.1000007 -1.1000000009 -12.10000008

9. After import is completed, click “View POS Data” on the right of “Image POS Data” to
check whether the POS data has been imported properly.

16 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

10. After confirming the import was successful, proceed to the next step.
For the POS import function, you may also refer to the tutorial video “DJI Terra - Image POS
Data Import”

Using the PPK Result Files

Phantom 4 RTK
For Phantom 4 RTK, the Cloud PPK service on the remote control app can be used to perform
PPK computation. For details on the PPK service, you may refer to the “Cloud PPK Service”
chapter in “Phantom 4 RTK - User Manual”. DJI Terra can automatically identify cloud PPK
computation results.

Follow these steps:

1. Copy the file of the cloud PPK computation results on your Phantom 4 RTK (named “result.
csv”) into the image folder.
2. Create a mission and add the photos.
3. The POS data in the PPK result file will be automatically imported into the software with the
photos. Click icon on the right of “Image POS Data” to view and edit the POS data in the PPK
result file.
4. If you wish to use the POS data included in the photos for reconstruction, remove the PPK
result file from the image folder before adding photos.

Zenmuse P1 or Other Third-Party Payload

For Zenmuse P1 or other third-party payload, a third-party PPK computation software may be
used to perform the computation. After obtaining the output, you can then import them in the
same manner as in the previous section, “Import image POS Data”.
After PPK computation is complete, if the image solutions are fixed, you may set the horizontal
and vertical POS accuracy to 0.03 m and 0.06 m respectively, which can enhance processing
efficiency and accuracy substantially. If the solutions are not fixed, the horizontal and vertical
accuracy can be set to 2 m and 10 m respectively.
The PPK files for M300’s PSDK have not been made open-source, therefore any third-party PSDK
payload mounted on M300 is unable to use the PPK function. Zenmuse P1 will save the satellite
observation source file required for PPK post-processing. You may use third-party software to
perform PPK post-differentiation on P1. For details, please refer to “Zenmuse P1 - User Manual”.

© 2024 DJI All Rights Reserved 17

DJI Terra Operation Guide

Aerotriangulation
Aerotriangulation refers to the photogrammetric process of computing the camera’s attitude
at the time of image formation and the spare point cloud of the target based on the spatial and
geometric correlation between the image point and the target. After aerotriangulation, you can
quickly determine whether the quality of the raw data meets the project delivery requirements
and whether any image needs to be added or removed. Aerotriangulation is required prior to
both 2D and 3D reconstruction.

Aerotriangulation Parameter Settings

1. Scenarios
Different computation methods apply to different scenarios. You may choose a suitable
scenario according to your shooting method. Explanation for different options:
Normal: Applicable to most scenarios, including oblique photography and orthophotography.
Circle: Applicable to circling scenarios and especially for reconstructions of small and vertical
objects, such as base stations, towers, and wind-driven generators.
Power Lines: Applicable only to visible-light cameras (such as a Phantom 4 RTK) when
shooting power lines in a vertical zigzag manner.

2. Computation Method
If the computer has permission for cluster computation, you can choose either standalone
or cluster computation here (for details on cluster computation, please refer to Chapter 4
“Cluster Reconstruction”). If the computer only has permission for standalone computation,
no “Computation Method” options will be available.

3. Advanced Settings
a. Feature Point Density
High: More feature points are extracted in single images. Suitable for scenarios requiring
higher output accuracy and more refined effects.
Low: Fewer feature points are extracted in single images. Suitable for scenarios requiring
faster output.

18 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

b. Distance to Ground/Subject
This refers to the distance between the camera and the subject while capturing data.
If multiple distances exist, the shortest one will be applied. This parameter guides
aerotriangulation block splitting. The greater the distance to ground/subject, the slower
the aerotriangulation computation.
c. XML Format
You may choose to output in XML format, namely with “ContextCapture” and “Blocks
Exchange”. For the coordinate system, we recommend keeping it consistent with that for
2D and 3D reconstruction. Most model editing software requires this file.

XML files only support projected coordinate systems and not geographic coordinate
systems.

Ground Control Points (GCPs)

GCPs are ground marking points with distinctive features and geographical coordinates that are
clearly identifiable on an image. The geographical coordinates of a GCP can be obtained through
measuring technology such as GPS, RTK, or with an electronic total station. Then link the GCP
with the image where it was captured, by marking the GCP using software. GCPs are divided
into control points and check points. Control points are used for optimizing aerotriangulation
accuracy and increasing modeling precision, as well as conversion into a local coordinate system
or the 85 elevation system. Check points are used for checking and evaluating aerotriangulation
accuracy in a quantitative manner.

© 2024 DJI All Rights Reserved 19

DJI Terra Operation Guide

When conducting 2D or 3D reconstruction, you can import GCPs after adding images, to
enhance the aerotriangulation accuracy and robustness, check the accuracy, and convert the
aerotriangulation results into the designated GCP coordinate system, so as to increase the
accuracy of the reconstruction output.

Layout of GCPs

1. Preparing GCP Files

a. Before using the GCP function, prepare a GCP file which must contain the following
information: the GCP name, latitude (X/E), longitude (Y/N), altitude (Z/U), horizontal
accuracy (optional), and vertical accuracy (optional). Each item must be separated by a
space or separator. Note that for a projected GCP, “X” refers to the easting value which
is typically a 6 or 8-figure number (with a zone number); while “Y” refers to the northing
value which is typically a 7-figure number. Be sure not to confuse “X” and “Y”.

GCPs expressed in latitudes and longitudes (left table), and in a projected coordinate system (right table)

b. Click “GCP Management” to enter the GCP management page, which includes the GCP
list, GCP information, photo gallery, aerial triangulation view, and marking view. After an
image is selected in the photo gallery, its marking view will appear on the left of the aerial
triangulation view, as shown below. You may add GCPs and mark points on this page to
perform aerotriangulation computation and optimization.

20 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Marking View Aerial Triangulation View

GCP List

GCP
Information

Photo
Gallery

2. Importing GCPs
a. On the GCP list, click “Import GCP File” to import a GCP file.

© 2024 DJI All Rights Reserved 21

DJI Terra Operation Guide

b. After the GCP is imported, define the file format first. If the first line is not coordinate
information such as data format description, the first line should be skipped; then set the
decimal separator, where care should be taken to differentiate between “.” which is used
by most countriesand “,” as used by some; finally, define the column separator, which can
be a comma (,), space ( ), semicolon (;), and tab; if there are multiple separators, you may
select “Treat combined separators as one” to separate the items correctly.
c. After defining the data format, select the GCP coordinate system and elevation system.
If the POS altitude uses the ellipsoidal height and the GCP altitude uses the National
Vertical Datum 1985, or the GCP uses a local coordinate system, then “arbitrary
coordinate system” should be selected.
d. Lastly, in “Define Data Column”, expand the menu to set the definitions corresponding to
each data column, then click “Import”.

e. If you are marking a GCP with other equipment, you may export the entire mark file then
click “Import mark file” to import the mark file in JSON format.

3. Editing GCPs
a. If you need to add or delete a GCP, click “+” or “-” on the GCP list. Press and hold the “Ctrl”
button to select multiple GCPs, and press and hold the “Shift” button to select all GCPs
between two mouse clicks.
b. Select a point in the GCP list to edit its GCP information at the bottom. If you are setting
a GCP as control point or check point, edit the horizontal accuracy, vertical accuracy, and
set the coordinate values matching the GCP coordinate system.
c. Before marking the points, click the “Aerotriangulation” button first to perform
aerotriangulation processing on the image. After this is completed, the aerotriangulation
computation results will appear on the left, including the camera’s POS and point cloud.

22 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

4. Marking Optimization
Marking is the process of linking the geographical coordinates of a GCP collected on the field
with the photo where this point is seen. Whether it is a control point or check point, marking
is required in order for it to take effect.
a. Before marking a point, we recommend clicking the “Aerotriangulation” button first to
perform aerotriangualtion on the photo. After this is completed, the estimated location
of the GCP will be more accurate. You may skip aerotriangulation and mark a point
directly. However, the estimated GCP location may be inaccurate, with more time needed
to locate the point.
For a special coordinate system, the marking process is as follows: aerotriangulation —
import GCP file — adjust the coordinate system as “arbitrary coordinate system” — mark
the point — disable “Constrain with Image POS Data” — optimize.
For a known coordinate system and where the elevation system is consistent to when
the drone was collecting the data, the marking process is as follows: aerotriangulation —
import GCP file — select “Known Coordinate System” as the GCP coordinate system —
mark the point — leave “Constrain with Image POS Data” enabled — optimize.
b. Select any GCP and enable “Only display photos with GCP” on the right of the photo
gallery. Click an image in the photo gallery that contains this GCP, and its marking view
will appear on the left. The blue crosshairs above it indicates the estimated location of
the selected GCP’s projection onto this image.

c. On the marking view, press and hold the left button of your mouse to drag the image,
and roll the scroll wheel to zoom in and out of the image. Click the image and use the
crosshairs to mark the GCP’s actual location on the image. The mark is displayed as a
green cross on the marking view and photo gallery thumbnail. Meanwhile, a check mark
will appear on the top right corner of the thumbnail to indicate it is a mark.
d. Click the “Delete” icon above the marking view to delete the mark information on the
image.

© 2024 DJI All Rights Reserved 23

DJI Terra Operation Guide

e. For a GCP, after marking of the third image is completed, the estimated location of the
blue crosshairs will be updated in real time according to changes of the mark location.
“Reprojection error” and “3D point error” below the GCP information will also be updated.
f. “Reprojection error” and “3D point error” can be used to determine errors in mark accuracy
and raw POS accuracy. Depending on the error size, the color of the number will change
to green, yellow, or red. If the error suddenly becomes larger after the marking of a photo,
please check if the wrong location has been marked. For an area, we recommend using at
least five evenly distributed control points, with at least eight mark images for each control
point (for data from five lenses, it is preferable to have at least five mark images for each
lens). Scatter the image locations as much as possible, and keep the mark locations away
from the edge of the images. If the estimated location of a newly added photo basically
matches its actual location, the GCP does not require marking.

g. If “Constrain with Image POS Data” is enabled, the horizontal and vertical accuracy in the
initial POS of an RTK image become 0.03 m and 0.06 m respectively. Together with the
GCP, this initial POS will constrain aerotriangulation.
• If the POS and GCP are under the same coordinate system and elevation system, we
recommend enabling this button as it will greatly increase reconstruction efficiency and
accuracy.
• If the GCP is under a local coordinate system, we recommend selecting “arbitrary
coordinate system” for the GCP coordinate system and disabling the “Constrain with
Image POS Data” button.
• If you set the 85 elevation system as the GCP coordinate system and a known
coordinate system as the POS coordinate system, we recommend that you enable
"Constrain with Image POS Data" after importing a GCP file and selecting the 85
elevation vertical coordinate system.
• If you are using a local coordinate system, have created a PRJ file for a local coordinate
system, and are defining the GCP’s coordinate system using an imported PRJ, we
recommend disabling the “Constrain with Image POS Data” button.

24 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

h. After completing the marking of all GCPs, click the “Optimize” button to perform
aerotriangulation computation. Thereafter, an aerotriangulation report will be generated,
and the aerotriangulation area on the left will also be updated as the optimized results. An
aerotriangulation report focuses on errors in control points or check points as well as the
overall error. If an error is too large, the accuracy fails to meet the requirements. In this
case, you will need to re-mark the point with a greater error or increase the GCP count.

i. By selecting a GCP, you can view the post-optimization reprojection and 3D point errors
in the GCP information at the bottom. You can also view details of the control point/check
point errors highlighted in the aerotriangulation report.
j. Click the “Export GCP” button to export control point and mark information as a JSON file
for use in other missions.
k. After ensuring the accuracy is correct, return to the mission home page to perform the
next step.
DJI Terra supports data processing without GCPs as well as skipping the GCP marking
step, where you may click “Aerotriangulation” directly and wait for the aerotriangulation
process to complete, then click “Quality Report” to view the output quality.
For the relevant GCP functions, you may also refer to the tutorial video: “DJI Terra - GCP
Management”.

© 2024 DJI All Rights Reserved 25

DJI Terra Operation Guide

26 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

An aerotriangulation quality report provides the following key parameters:

1. Calibrated images: This refers to the number of images involved in the aerotriangulation
computation. If the number of calibrated images is less than the images imported, it means
some images could not be used for the computation, perhaps because certain areas
captured were without texture or had little texture (such as bodies of water or snow-covered
scenes); or the shooting angles and resolution for these areas shot differed too much from
the other data. If output is incomplete due to such reasons, you may need to recapture shots
on the field.
2. Connected components: If the number of connected components is greater than 1, there is
no perfect aerotriangulation result in the regional network, which may affect the accuracy. It
is recommended to check the image data.
3. Reprojection error: It is the error between the ground point projected back onto the
image and the corresponding point on the image, an important metric for evaluating
aerotriangulation accuracy.
4. Georeferencing RMSE: This refers to the RMSE between the computed image location and
the location recorded in the image. This parameter reflects the relative accuracy of the initial
POS: the smaller its value, the higher the accuracy.
5. Image RTK status: For images with fixed solutions, the accuracy is set as centimeter-level.
The more images with fixed solutions, the better (the accuracy for floating solutions and
single solutions are set as decimeter and meter-level respectively; if there are no solutions,
it means no RTK positioning computation was performed). If all the images have fixed
solutions, this will ensure the accuracy without GCP under the POS coordinate system is
at the centimeter level. If images with fixed solutions only account for a small portion, the
absolute accuracy of the output will be poorer. You will need to add suitable GCPs to ensure
a higher absolute accuracy.
6. Camera calibration information. This focuses on the comparison between the initial camera
focal length/cx/cy and the camera focal length/cx/cy after aerotriangulation optimization.
Generally, the difference before and after optimization does not exceed 50 pixels. If the
difference is large, you may troubleshoot using the following method:
a. If the focal length differs greatly before and after optimization, and the images used for
reconstruction are facing the same direction (for instance, the photos are all orthographic
or facing a building facade), then increase the number of photos shot from other angles
(such as oblique photos);
b. If the cx and cy differ greatly before and after optimization, check whether the sensor
direction was changed when the images were collected (for instance, the aircraft turned
its heading to collect data during a survey mission).

© 2024 DJI All Rights Reserved 27

DJI Terra Operation Guide

2D Reconstruction

Click the “2D Map” button and set the relevant parameters.
1. Setting the reconstruction resolution: “High” means the original resolution, “Medium” means
1/2 of the original resolution (i.e. the image length and width are 1/2 of the original), and “Low”
means 1/3 of the original resolution (i.e. the image length and width are 1/3 of the original).
For example: If the resolution of the original shots is 6000×6000, that would be the high
resolution; the medium resolution would be 3000×3000 and the low resolution 2000×2000.
2. Selecting the mapping scene: Whether in the city or a village, “Urban” scene should always be
selected for survey missions. “Field” and “Fruit Tree” scenes apply to DJI agricultural drones.
For hilly terrain, reconstruction using “Field” and “Fruit Tree” settings may result in dislocation
or distortion.
3. Selecting the computation method: If the computer has permission for cluster computation,
you can choose either standalone or cluster computation to perform the computation. If
the computer only has permission for standalone computation, no such options will be
available. For the relevant settings for cluster reconstruction, please refer to the “Cluster
Reconstruction” chapter.

28 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Region of interest (ROI)

While performing 2D/3D reconstruction, you may select and model an ROI after adding photos.
Generating models of only within an ROI saves time and increases efficiency.
Note that ROI modeling can only be done after aerotriangulation is completed. After finishing
aerotriangulation, click “Region of interest” in Advanced Settings to enter the region of interest
edit page.

1. Defining an ROI: You may define an ROI through the following four methods. The coordinate
system used here is consistent with that in Output Coordinate System Settings.
• Click “KML Import” to convert the dots in the KML file into boundary points of the ROI.
• Input the min/max latitude, longitude and height or XYZ values of the ROI, then click
Apply to set the ROI.
• For the “Reset Region” setting, tap “Auto” or “Max Region” and the software will
automatically generate the ROI. Auto: Automatically calculates an appropriate rectangular
area based on the distribution of the aerotriangular point clouds. Max Region: A
rectangular area that covers all the aerotriangular point clouds.
• Tap at the top of the screen to enter editing mode, then tap on the relevant location
on the map to manually add boundary points, and input the height value in the height
text box to set the ROI.

2. Translating an ROI: Tap to enter Translation Mode and drag the defined ROI to translate
it.
3. Editing a ROI: Tap to enter editing mode. Tap on the relevant location on the map to
manually add boundary points. Drag the boundary points to adjust the location and the
area shape. Select a boundary point then tap to remove it. Tap to remove all boundary
points. Tap to exit editing mode.
4. Other information and settings
a. When the region of interest is rectangular, its length, width, and height will be displayed
at the top of the page.
b. Camera location display: Display/hide the camera location for added photos.

© 2024 DJI All Rights Reserved 29

DJI Terra Operation Guide

c. Display zone: Display/hide the defined ROI.

d. If you are performing 2D and 3D reconstruction at the same time, clicking “Copy ROI to
3D Reconstruction” will copy the ROI into the 3D reconstruction mission.

30 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Output coordinate system - Known coordinate system

During 2D and 3D reconstruction, you may set the output coordinate system after adding
photos. If your photos do not contain POS data, the default coordinate system will be an
“arbitrary coordinate system”. If the added photos contain POS data, the default coordinate
system for 2D reconstruction will be the UTM projected coordinate system for the mission’s
location. Note that when a GCP has been marked, the output coordinate system must be the
same as that for the GCP, otherwise the output’s coordinates will not match the GCP’s.
You may set a known coordinate system by either importing a PRJ file or searching for it in the
DJI Terra coordinate system library.
1. Importing a PRJ file: Search and download the .prj file for the desired coordinate system on
the website https://spatialreference.org. Then click in DJI Terra to import. If it is a custom
coordinate system, you may download another public PRJ file, then modify five parameters
as follows: a - target coordinate system name; b - seven parameters; c - target ellipsoidal
central meridian; d - target ellipsoidal east additive constant; e - target ellipsoidal north
additive constant. For example:

2. Searching: From the drop-down menus of Horizontal Settings and Vertical Datum Settings,
select “Horizontal Coordinate System Database” and “Vertical Coordinate System Database”.
Enter the coordinate system name or authorization code, select the corresponding search
result, then click “OK”. The WGS 84 is the most commonly used UTM projected coordinate

© 2024 DJI All Rights Reserved 31

DJI Terra Operation Guide

system in the world, DJI Terra will automatically calculate the UTM zone number where the
current data is located. Users can check whether the zone number is correct according to the
following formula: (longitude)/6+31 and then round the result. For example: if the longitude
of data is 116.3241, based on the formula (116/6+31=50.38375), the WGS84 UTM zone
number is 50N. If users need to use other projected coordinate systems, log in to https://
epsg.io/ to query the EPSG zone number according to the country and region or the latitude
and longitude coordinates, and then search for it and apply in DJI Terra.

32 © 2024 DJI All Rights Reserved

DJI Terra Operation Guide

© 2024 DJI All Rights Reserved 33

DJI Terra Operation Guide

Output Coordinate System - Arbitrary Coordinate System - Seven-Parameter

Transformation
Seven parameters enable the conversion between any two coordinate systems. When a survey
needs to produce output based on a local coordinate system, seven-parameter transformation
would normally be used to convert geodetic coordinate systems such as WGS 84 or CGCS 2000
to a local coordinate system. In the 2D map and 3D model functions on DJI Terra, you may select
“Seven-Parameter Transformation” or “Calculate Seven Parameters” under “Arbitrary Coordinate
System” to generate output based on a local coordinate system. The steps are as follows:
1. Create a data file according to the data import format required by seven-parameter
transformation (as shown below). DJI Terra supports the import of data in TXT or CSV format.
The data information and format requirements are as follows:
Coordinate name, latitude (B, degree format), longitude (L, degree format), altitude (H), east
(x), north (y), and height (z)

Data items in a TXT file shall be separated by commas (,) and tabs. Please ensure at
least 3 sets of coordinates are used in seven-parameter calculation.

2. In “2D Map"/"3D Model”, click “Advanced Settings”, select “Arbitrary Coordinate System”, and
click “ Seven-Parameter Transformation Settings”. In the “Seven-Parameter Transformation”
pop-up window, click “>Calculate Seven Parameters” to expand the menu, click “Import”, then
select the prepared data in Step 1 to import it.

34 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

3. After importing the data, check whether B/L/H and x/y/z correspond correctly to the order
of “Latitude/Longitude/Height” and “East/North/Height”. Note that the x/east coordinate
is usually 6 or 8-digit, while the y/north coordinate is normally 7-digit. Headings or other
irrelevant information can also be removed using “-” in the “Operation” menu.

4. After clicking “Calculate”, you will see the results of the seven-parameter calculation and its
residual. Preferably, the residual should be smaller than 0.5 of the expected output accuracy
(if the expected accuracy is 5 cm, the recommended residual should be smaller than 0.025 m),
so that the transformation result is more accurate. If the residual is too large, you may use “-”
to remove the gross error, then click “Calculate” again. After ensuring the residual meets the
requirement, click “Apply” to complete the calculation and apply the seven parameters.

© 2024 DJI All Rights Reserved 35

DJI Terra Operation Guide

5. If you will be processing data in the same area in the future, you may click “Export” directly
to save the seven parameters in the JSON format. On your next processing, you can then
import and reapply the file directly without having to calculate again.

Similar to the GCPs, coordinates used for seven-parameter calculation should be

distributed evenly around the measured area and do not necessarily have to be in the
area. However, they must not be concentrated in a small zone within the measured
area, otherwise the seven parameters would still result in significant offset when far
from the computed area despite having produced a high residual.

Map Grid
When the data of a raw image is too large, the DOM/DSM TIF image generated may be too
big or too slow to load when being imported into third-party software. In these situations,
we recommend using the Map Grid function where you can divide a large TIF file into several
smaller ones.
1. Enable the “Map Grid” button, set the maximum grid length in pixel.

2. The software will split the DOM/DSM output into blocks as shown below (using 5,000 pixel as
an example).

36 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

• Map grid output will not replace the original DOM or DSM image. Both files will co-
exist.
• The resulting files will be saved in the output folder of the corresponding mission: 1)
Mission name\map\dsm_tiles; 2) Mission name\map\result_tiles.
• The minimum output grid length is 1,000 pixel.
• Output files bigger than 4GB will contain BigTIFF parameters, while those smaller than
4GB will not (certain third-party software does not support BigTIFF images, in which
case the grid length should be reduced).

After setting the relevant parameters, click “Start Reconstruction” to begin 2D reconstruction.

2D Map File Format and Storage Path

2D map files are saved to this path by default. You can change the cache directory in Settings.
C:\Users\< Computer user name >\Documents\DJI\DJI Terra\<DJI account name >\< Mission
name >\map\
You can also press the shortcut keys “Ctrl+Alt+F” on the reconstruction page to open the folder
of the current mission. An output file contains the following key information:
result.tif: Digital Orthophoto Map (DOM), which is the main output of 2D reconstruction.
dsm.tif: Digital Surface Model (DSM), which is a geoid file for the mission area with each pixel
containing its latitude, longitude, and altitude.
gsddsm.tif: a DSM with a 5 MP resolution after downsampling, which can be imported for use
during terrain following with Matrice 300 or Phantom 4 RTK.
Digital folder (such as 12-21): map tile data, used for displaying 2D models in DJI Terra and that
adopts the same tile classification as Google. Map tiles are standard. If they need to be used on
a third-party platform, they can be called based on the tile calling convention.
“result_tiles” folder: the folder in which map grid split results of orthophotos are saved.
“dsm_tiles” folder: the folder in which map grid split results of geoid files are saved.

© 2024 DJI All Rights Reserved 37

DJI Terra Operation Guide

There will also be a “.temp” folder in the output folder. It is larger in size and stores intermediate
files from model reconstruction. These files be required if you wish to perform additional
actions such as creating a new file format or modifying a coordinate system during or after
a reconstruction. If no other actions are necessary after reconstruction, you may delete the
intermediate files manually to free up storage space.

Haze Reduction
For hazy scenes, a “misty” effect may occur on orthophotos. In such circumstances, you may
enable the “Haze Reduction” feature to optimize the issue.

38 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Before enabling “Haze Reduction” After enabling “Haze Reduction”

Light Uniformity
In strong lighting environment, a mirror-like reflection effect will be produced due to glossy
surfaces of land or crop canopy, leading to light on one side and shade on the other, and
reconstruction output will have stripes with gradients. When this happens, enabling the “Light
Uniformity” feature will even out the overall light and color of the image. Users are unable to
enable “Haze Reduction” and “Light Uniformity” at the same time.

Before enabling “Light Uniformity” After enabling “Light Uniformity”

© 2024 DJI All Rights Reserved 39

DJI Terra Operation Guide

Quality Report for 2D Reconstruction

After the model is reconstructed, click “Quality Report” to view its overall condition. You can
check the output’s resolution, covered area, reconstruction time, and other details in the report.
Note that the entire 2D reconstruction duration should be the total time taken for these four
steps: aerotriangulation; image distortion correction and color correction; densification; and
TDOM generation.

40 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

3D Models
Enable the “3D Model” button and set the relevant parameters.

1. Setting Reconstruction Resolution

“High” means the original resolution, “Medium” means 1/2 of the original resolution (i.e.
the image length and width are 1/2 of the original), and “Low” means 1/4 of the original
resolution (i.e. the image length and width are 1/4 of the original). For example: if the
resolution of the original shots is 6000×6000, that would be the high resolution; the medium
resolution would be 3000×3000 and the low resolution 1500×1500.
2. Selecting Appropriate Mapping Scene
Normal: Applicable to most scenarios, including oblique photography and orthophotography.
Circle: Applicable to circling scenarios and especially for reconstructions of small and vertical
objects, such as base stations, towers, and wind-driven generators.
Power Lines: Applicable to scenarios where power lines are to be captured with visible light
and their point clouds will be reconstructed. Note that for power line scenes, the software
will only generate point clouds and not 3D models. Power line scenarios are only supported
by the Electricity and Cluster versions.
3. Selecting Computation Method
When using the Cluster version, you may choose cluster computation for the reconstruction,
which will greatly increase processing efficiency and volume. For the relevant settings for
cluster reconstruction, please refer to the “Cluster Reconstruction” chapter. If the computer
only has permission for standalone computation, no such options will be available.

© 2024 DJI All Rights Reserved 41

DJI Terra Operation Guide

3D Output Formats
DJI Terra's 3D output contains the following formats:

Model
1. B3DM: Default format for display in DJI Terra (LOD model format, suitable for display in
Cesium)
2. OSGB: LOD model format
3. PLY: Non-LOD model format
4. OBJ: Non-LOD model format
5. S3MB: SuperMap GIS LOD model format
6. I3S: Non-LOD model format
LOD (Level of Detail) applies to models with multiple layers of detail. It involves saving
a model in pyramid form and as multiple small tiles. In most cases, browsing LOD
models is smoother on a WebGIS platform such as DJI FlightHub 2.

Refine Water Surface: This feature automatically identifies any body of water within the
measured area and flattens the model to create an even water surface.

"Refine Water Surface" disabled "Refine Water Surface" enabled

42 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Reduce Model To: Use this feature to cut the size of the model by reducing model triangles
to a set proportion. To facilitate the online browsing of the model and other downstream
operations, we recommend that you retain at least 20% of triangles.

Comparison of meshes based on different proportions set for model reconstruction

Point Cloud
a. PNTS: Default format for display in DJI Terra (LOD point cloud format, suitable for display in
Cesium)
b. LAS: ASPRS LASer, 3D point cloud format, v1.2
c. S3MB: SuperMap LOD point cloud format
d. PLY: Non-LOD point cloud format
e. PCD: Non-LOD point cloud format
Merged Point Cloud Output: By selecting LAS, PLY or PCD as the point cloud format, you can
enable the "Merged Point Cloud Output" feature, which would merge the point cloud output
into a single file.

More Output

Workflow of more 3D reconstruction output

After the generation of a dense point cloud based on aerotriangulation, 3D reconstruction steps
include classifying ground points, generating DEM and contour lines based on the ground point
cloud, and generating a grid of points and TIN data based on ground points.

© 2024 DJI All Rights Reserved 43

DJI Terra Operation Guide

1. Ground Point Classification: If this feature is enabled, ground points are differentiated from
non-ground points. Ground points can be used for generating a topographic map for the
mapping area.
Ground point classification algorithm: DJI Terra uses the specified building dimensions to
set evenly distributed seed points in the mapping area and then connects the seed points to
form an initial triangle network. Whether a laser point is a ground point can be determined
by calculating the angle formed between the laser point and the vertex of the triangle
where it is located and the distance from the laser point to the plane where the triangle is
located. If the angle and the distance are smaller than the values set for the Iteration Angle
and Iteration Distance parameters, the point is identified as a ground point and the triangle
mesh is updated. This process is repeated until all ground points in the mapping area are
identified.
1.1 Ground Type: DJI Terra provides preset parameters for different types of terrain such as
Flat Ground, Gentle Slope, and Steep Slope. The default parameters suit most common
scenarios. If a mapping area has large undulations, you can increase the iteration angle
and distance as needed.
1.2 Building Max Diagonal: Set this parameter based on the diagonal length of the roof
plane of the largest building in the mapping area. If this parameter is set to a small value,
rooftops may be recognized as ground.

2. DEM: If this feature is enabled, DEM data in TIFF format will be generated based on the point
cloud data of ground points. The following scales are supported for DEM results: 1:500,
1:1000, and 1:2000. You can also set a DEM result resolution based on the GSD.

44 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

3. Contour: This feature is used to generate contour lines from DEM. The generated contour
lines can be in DXF or SHP format. The contour lines will be automatically smoothed when
they are generated.
To generate contour lines, set the following parameters:
a. Interval: Difference in elevation between adjacent contour lines.
b. Datum: Elevation of contour lines is calculated based on the elevation datum and interval.
c. Elevation Annotation Radius: Used to generate elevation points for topographic mapping.
d. If "Discard Invalid Contour Lines" is enabled, contour lines shorter than the set length will
be discarded.

© 2024 DJI All Rights Reserved 45

DJI Terra Operation Guide

4. Grid of Points & TIN: If this feature is enabled, a grid of points and TIN data will be generated
based on the ground points from the point cloud. A triangular irregular network (TIN) is an
elevation data structure that uses a network of multiple connected triangles to represent a
surface. A TIN consists of a series of data points that have x, y, and z coordinates and divides
a geographic space into a number of non-overlapping connected triangles, often called
facets. The TIN model is suitable for representing irregular changes in terrain and is able to
use a number of triangles to improve accuracy when representing drastic changes in terrain.
Each triangle node is an elevation point. These nodes comprise the grid of points.
a. Standard Mode: Extract point grids from ground points in the original point cloud at a
fixed distance. More details of original point cloud will be saved in standard mode.
b. Smart Mode: Extract key points from the original point clouds to reduce the number
of points in the grid, which is more compatible with downstream software, while still
ensuring ground reconstruction effects.
c. Sample Distance: Minimum sample distance in grid of points. Increasing the sample
distance helps reduce the size of grid of points and TIN files.
d. Max Number of Points: Maximum number of points of grid in smart mode. Reducing
the number of points helps reduce the size of grid of points and TIN files. By default, the
maximum number of points is equal to the number of images divided by 500, rounded
up to an integer, then multiplied by 3,000.

46 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

ROI/Block Splitting
For details on ROI settings, please refer to the “2D Reconstruction - Region of Interest” chapter.
To ensure efficiency, the software will set the maximum block size based on the computer’s
RAM. However, certain scenes (such as for importing third-party model editing software) require
the block size to be sufficiently small, in which case a custom block size will have to be set. DJI
Terra provides the following block splitting methods:
1. Auto: maximizes the blocks automatically based on the device’s current available RAM.

2. Custom RAM: You input the maximum RAM allocated for processing a single block, and the
software splits the blocks according to the input RAM value.

© 2024 DJI All Rights Reserved 47

DJI Terra Operation Guide

3. Custom Side Length: The software divides the blocks based on the side length input by you,
and the software will show the RAM and block count required for the block splitting within
the current distance.

4. Custom Splitting Origin: This enables you to define a custom coordinate system and
coordinate values for the splitting origin. To update a model, you can set a common splitting
origin and replace the corresponding blocks. This function must be used together with
“Custom Model Origin” to achieve a satisfactory model update.

48 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

5. Click “Display blocks” to enable/disable the block splitting effects.

The smaller the blocks, the slower the reconstruction. Use this feature after
performing Aerotriangulation.

Output Coordinate System

Please refer to the “2D Reconstruction - Output Coordinate System” chapter.

Custom Model Origin

Take the OSGB format for example, in order to minimize its size, each OSGB file contains
coordinates that are relative to the model origin. When multiple missions need to be displayed on
a model display platform, the model origins of the missions can be set as the same coordinates.
When a model needs to be partially updated, you can set the new origin as the old origin and apply
the same splitting origin. This way the blocks to be updated can be copied and replaced directly.

© 2024 DJI All Rights Reserved 49

DJI Terra Operation Guide

The model’s origin coordinate system must be the same as its output coordinate system. After
completing the settings, begin the reconstruction. The model origins of output in OSGB, OBJ
and PLY format will be set as the coordinates of the input point, and the origin’s location will be
written into the “metadata.xml” file:

3D Reconstruction File Format and Storage Path

3D reconstruction resulting files are saved to this path by default. You can change the cache
directory in settings, and can also press the shortcut keys “Ctrl+Alt+F” on the reconstruction
page to open the folder of the current mission.
C:\Users\< Computer user name >\Documents\DJI\DJI Terra\< DJI account name >\< Mission
name >\models\pc\0
Once a format is selected, the folder storing outputs in that format will usually be named as
“terra_XXX” (XXX being the model format). For example, a folder named “terra_osgbs” would
store 3D models in the OSGB format.

50 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

There will also be a “.temp” folder in the output folder. This folder is larger in size and stores
intermediate files from model reconstruction, including image outputs after distortion
correction. These files be required if you wish to perform additional actions such as creating a
new file format or modifying a coordinate system during or after a reconstruction. If no other
actions are necessary after reconstruction, you may delete the intermediate files manually to
free up storage space. This will not affect the output.

If 2D and 3D reconstruction are required at the same time, you may enable both the
“2D Reconstruction” and “3D Reconstruction” buttons, and the software will perform
2D reconstruction first before 3D reconstruction.
The ROI, coordinate systems and other parameters for 2D and 3D reconstruction can
be set separately.

Quality Report for 3D Recontruction

After the model is reconstructed, click "Quality Report" to view its overall condition. You can
check the various parameter settings for the model reconstruction, the time when project data
is generated, and other details in the report.

© 2024 DJI All Rights Reserved 51

DJI Terra Operation Guide

Reconstruction Mission Management

1. After starting a reconstruction mission, you can click on the home page to enter the
reconstruction mission management page.
2. Select standalone computation, and the reconstruction mission queue will appear at the top
of the page, where you can check each mission’s name, reconstruction type, and progress
status. You can pause or cancel a mission on the action bar. Click and drag to sort the
missions.
3. Select cluster computation, and a local network worker device list will appear at the bottom
of the page. It is similar to the cluster reconstruction device list displayed on the cluster
reconstruction settings page. If a worker device status error happens, click the “Release”
button on the action bar to release the worker device. A released worker device will no
longer participate in the current reconstruction.

For other frequently asked questions on reconstruction, please refer to the cluster
reconstruction section in “DJI Terra FAQ”. on our official website.

FAQ
For frequently asked questions on reconstruction, please refer to “DJI Terra FAQ” on our official
website.

52 © 2024 DJI All Rights Reserved

Cluster Reconstruction
Function Overview
Cluster reconstruction means the processing of a mission using multiple computers at the
same time. It is able to handle larger missions with efficiency that is several times higher than
normal reconstruction. The basic process flow for cluster reconstruction is as follows: Open
the cluster computation engine on the worker device -> set up cluster control management on
the control device, and select the devices to participate in the computation -> select “Cluster
Reconstruction” for the computation method during reconstruction -> start the mission as a
normal 2D/3D reconstruction.
For how to perform cluster reconstruction, you may refer to the tutorial video “DJI Terra - Using
the Cluster Version”. Note that the tutorial was produced based DJI Terra v3.1 and the UI may
differ from that of the v3.6 version.

Preparation before performing cluster reconstruction

1. Server room configurations
2. Equipment configurations
3. Equipment parameter configurations
Reference: “Preparation Before Using DJI Terra”.

The Cluster version involves the concepts of “control device” and “worker device”.
Control device: It runs the desktop client for DJI Terra Cluster, and supports functions such
as setting cluster reconstruction parameters, selecting worker devices, and viewing mission
statuses. As the scheduling platform for an entire reconstruction mission, reconstruction
missions are managed on the control device. The license for the Cluster version is bound to
the control device. Worker devices can be changed any time as needed, while the aftersales
unbinding process would be required to change a control device.
Worker device: Other equipment for which a control device may allocate reconstruction
missions after a cluster computation engine has been enabled in a local area network.

Operating a Worker Device

A worker device is a computation engine. Its relevant programs are installed automatically
together with the DJI Terra software. You can find the worker device program, “DJI TerraEngine.
exe”, in the “Cluster” folder under the installation directory or in “Start programs > DJI product”,
and enable it as shown below.

© 2024 DJI All Rights Reserved 53

DJI Terra Operation Guide

1. Local temporary directory: For temporary storage during computation by worker devices.
Please set the directory in the local directory of the worker devices and ensure there is
sufficient disk space. If a worker device is also enabled for the control device, do not place
the local temporary directory and the shared directory on the same location to avoid read-
write conflict.
2. Enable/Disable: During cluster reconstruction, a worker device must be enabled in order to
be found and used by a control device in the same local network.
3. Log: The print window for the worker device log. If no error has been reported on the control
device, you may ignore the process log.
4. Language switch: You can switch between different languages with the button at the top
right of the page.
5. Automatically launch after powering on: Enabling this will launch this worker device program
automatically when the computer starts.
6. One to many: A high-spec computer may be able to support multiple worker devices at the
same time, boosting cluster reconstruction efficiency. To perform this, the computer must
meet certain graphics card memory and RAM specifications. If the weakest worker device
in the cluster has a graphics card memory of 4G while the most powerful worker device has
12G, and the RAM of the most powerful machine is also greater than that of the weakest
machine by 3 times, then you may enable 3 worker devices simultaneously on the most
powerful machine.

54 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Operating a Control Device

After ensuring the Cluster version license has been exchanged and imported, and that you
are logged in offline, click on the top right corner of the DJI Terra home page to enter the
cluster reconstruction device list. Note this icon is only visible on devices bound with the Cluster
version. It will not show on devices with permission for the standalone version only.

Shared file storage directory and settings: We recommend setting this on a Network Attached
Storage (NAS). For how to perform that, please refer to “Preparation Before Using DJI Terra”.

• The image path and shared file path must be on the same storage device. The shared
file storage directory must use the same path as the directory for adding photos
during reconstruction. For instance, they must both use network paths or drive letter
paths.
• When using a drive letter path for a shared file storage directory, the mapped drive
letter on the worker device corresponding to the network’s location must be the same
as that on the control device. For instance, if the control device uses the Z drive, then
all the worker devices should be mapped as Z drives.

© 2024 DJI All Rights Reserved 55

DJI Terra Operation Guide

Local Network Worker Device List

After entering the list, the software will automatically search for all enabled worker devices in
the current local area network. The list will display the computer names, statuses, and algorithm
versions for the found local network worker devices. Click to refresh the worker device
search results and statuses. For details on how to enable worker devices, please refer to the
later section on operating a worker device.
1. Computer name: Displays the computer name for the worker device.
2. Status: Displays the current status of the worker device (busy or idle). When idle, the worker
device can be selected as a working device. If you encounter any worker device error (e.g.
constantly showing busy status while not in use), click on the right of the status to reset
the worker device to idle status.
3. Worker device information: Includes RAM, VRAM, CPU and GPU utilization, CPU and GPU
models, and IP address. Worker device information shows you the running status and
information about all worker devices.
4. Release: Click "Release" to release the current worker device to the LAN. When a worker
device in the cluster has errors or is in idle state, you can release this worker device and add
another worker device to the cluster.
5. Add to Current Device: You can add a worker device to the cluster to which the current
device belongs before a cluster mission starts or during a cluster mission.
6. Algorithm version: Displays the algorithm version number for the worker device. If it is in
blue, it means the version number is consistent with the control device; while red indicates
they are different. The algorithm version number must be consistent in order for the worker
device to be selected as a working device.

56 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

7. Check the box before the serial number to select the worker device as a working device.
That means the worker device can be used during cluster reconstruction. After completing
the setting, click “Apply” to save the settings. The software license for the Cluster version
only binds the control device, while worker devices can be changed at any time as
needed. Licensed quantity refers to the number of worker devices that can be selected
simultaneously for the same control device. For instance, if you have purchased DJI Terra
Cluster that supports 3 worker devices, then you can only select up to 3 devices to participate
in the reconstruction even if it can detect 10 worker devices.

Cluster Reconstruction Process

1. Open “DJITerraEngine.exe” on the worker device, set the local storage directory, enable the
cluster computation engine, ensure each worker device can communicate with the control
device and access the shared storage directory.
2. Open the cluster management page on the control device, set the shared storage directory,
and select the worker devices to participate in the computation.
3. Start reconstruction as per the process of a 2D/3D reconstruction mission, and select “Cluster
computation” for the computation method.

You can set whether to use cluster computation for aerotriangulation, 2D

reconstruction, and 3D reconstruction respectively.

4. Check mission progress and worker device status in reconstruction mission management.
5. If the connection with the shared file storage directory is maintained after completing cluster
reconstruction, you can view the reconstruction results in DJI Terra and the output files in
the corresponding folders.

© 2024 DJI All Rights Reserved 57

2D Multispectral Reconstruction
Function Overview
2D Multispectral Reconstruction only supports 2D multispectral map reconstruction using
photos captured by a P4 Multispectral. Besides generating DOM and DSM for 2D reconstruction,
it can also output vegetation indices such as NDVI and NDRE, as well as composite images for
each multispectral band. To output relevant results based on the unit of reflectivity, you can
also modify the multispectral data using the “Radiometric Correction” feature.
The basic process of multispectral reconstruction is: Import data -> Radiometric correction
(optional) -> Aerotriangulation -> 2D multispectral reconstruction

Importing Data
Create Reconstruction Mission
Open and log into DJI Terra, click “New Mission” on the bottom left corner and select
“Multispectral” as the mission type.

Add Photos
You may add image files or folders. The images imported must contain visible light photos in
the JPG format and multispectral single-band photos in the TIF format. Every 6 photos shall
form a group, with the last digit of 0-6 corresponding to visible light, red, green, blue, red edge,
near infrared respectively. Therefore, the number of photos imported must be a multiple of
6; otherwise, please check if any data is lost or missing. Multispectral reconstruction relies on
visible light photos to perform aerotriangulation and then corrects the other multispectral
photos based on the overall offset of the camera’s structure using the aerotriangulation results.
Therefore, the images imported must contain visible light photos. If they are non-visible light
photos such as real-time NDVI images, the reconstruction will fail.

58 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Reconstruction successful Reconstruction fails due to real-time

with visible light photos NDVI images

After data is imported, the vegetation index output feature will check the image data. If all
appear in green, it means the imported multispectral data works properly and the vegetation
indices can be output successfully. If an item is red, it means the data for some bands is missing
and you need to check and re-import the data.

Radiometric Correction (Optional)

Most of the settings for multispectral reconstruction are similar to visible light reconstruction.
Generally, you may set the relevant parameters based on the 2D reconstruction process. The
biggest difference with 2D multispectral reconstruction is the addition of the “Radiometric
Correction” module in 2D Map > Advanced Settings. Note that this module is optional. If you
need to output single-band maps with reflectivity as the results or conduct index research other
than on vegetation indices output by DJI Terra, you may use “Radiometric Correction” first to
generate more accurate multispectral output.
Radiometric correction reduces the impact of factors such as weather and time, helping to
output more reliable index values for the reconstruction mission. After performing radiometric
correction on the multispectral photos, you may reconstruct them into 2D multispectral maps
with reflectivity as the output results. Click the “Radiometric Correction” button in “Advanced
Settings” to enter the radiometric correction page and perform the relevant steps. After
completing radiometric correction, return to the reconstruction page to start the mission. The
final reconstruction result will incorporate the effect of the radiometric correction.
Before performing radiometric correction, you need to shoot several calibration board photos
with known reflectivity using a DJI Phantom 4 Multispectral. For how to conduct field operations,
please refer to the article “Handy Tips for Using DJI Phantom 4 Multispectral”.

© 2024 DJI All Rights Reserved 59

DJI Terra Operation Guide

Importing Calibration Board Data

For radiometric correction, import at least one set of calibration board photos for different
bands. You may import up to 3 sets of photos, namely Calibration Board 1, 2, and 3 respectively.
1. Select the corresponding calibration board group: Calibration Board 1, 2 or 3.
2. Click “Import Calibration Photo” and select photos for the blue, green, red, red edge, and
near-infrared bands with calibration boards, and import them.
3. Click a photo for a certain band, and it will appear in the area on the left.
4. Click or to replace or delete the current photo.
5. Input the corresponding reflectivity index in the text box for each band.

Marking Calibration Board Zones

1. Left-click your mouse on the photo on the left to manually add points to mark the
boundaries of the area where the calibration board is located.
2. Drag the boundary points to adjust their locations.
3. Click to remove all boundary points from the current photo.
4. Mark the calibration board boundaries of the photos for all bands.
5. The boundaries of a calibration board zone will be automatically applied to the next group of
calibration board photos; if this results in any deviation, drag the boundary points to adjust
accordingly.

Output Indices
DJI Terra supports the output of five vegetation indices: NDVI, GNDVI, NDRE, LCI, and OSAVI.
Each requires image data for a different band. The software will display the indices that can be
calculated based on the bands covered by the photos imported by you. If any band is missing, a
reminder will appear below the output index. After completing reconstruction, you may click the
corresponding index button on the map to view the map reconstruction result for that index.

60 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

If a band is missing in any one group of imported photos, the photos for that band will
be incomplete. You can view their details in the photo list.

2D Multispectral Map File Format and Storage Path

The reconstruction results for 2D multispectral maps contain grid data in the GeoTIFF
format, which can be used by third-party software compatible with the GeoTIFF format. The
reconstruction results will include multispectral maps corresponding to each index, narrowband
maps for each band, and RGB 2D orthomaps. These files are saved to the following paths. You
can change the cache directory in settings, and can also press the shortcut keys “Ctrl+Alt+F” on
the reconstruction page to open the folder of the current mission.
The default storage path for multispectral index maps:
C:\Users\< Computer user name >\Documents\DJI\DJI Terra\< DJI account name >\< Mission
name >\ map\ index_map
The default storage path for narrowband maps for each band and RGB 2D orthomaps:
C:\Users\< Computer user name >\Documents\DJI\DJI Terra\< DJI account name >\< Mission
name >\map
Where “result.tif” refers to a RGB 2D orthomap, while “result_XXX.tif” refers to a narrowband
map for the band corresponding to “XXX”.

© 2024 DJI All Rights Reserved 61

LiDAR Point Cloud Processing
Function Overview
DJI Terra only supports LiDAR processing of data from Zenmuse L1 and L2. It can process raw
LiDAR files from L1 into 3D point clouds in the LAS format. You may import the generated LAS
3D files into point cloud analyzing software for use in surveys or power industry applications.
The basic workflow for LiDAR data processing is: Import data -> Set relevant parameters -> Start
reconstruction

Importing Data
Create Reconstruction Mission
Open and log into DJI Terra, click “New Mission” on the bottom left corner and select “LiDAR
Point Cloud” as the mission type.

Adding Data
Click and select a folder named based on data collection time to add and import LiDAR
point cloud data. The folder should contain files with the following suffixes: CLC, CLI, CMI, IMU,
LDR, RTB, RTK, RTL and RTS. If there are JPG photos, you can generate true-color point clouds;
otherwise you will still be able to rereconstruct point clouds. Multiple groups of LiDAR missions
may be placed in a big folder, which can be imported when adding data.
If the RTK mode is not used during a flight and you have set up your own D-RTK2 base station
or third-party base station, you may rename the file after the RTB file in the folder under the
point cloud data file directory, and copy the file to that folder. For more details, please refer to
Zenmuse L2 User Manual - Base Station Satellite Data Acquisition.

62 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Base Station Center Point Settings (Optional)

Usage scenario:
1. If the base station is set up at an unknown point, you can obtain the precise coordinates of
the base station’s central point using third-party software and the static PPP convergence
method. Then import the coordinates into DJI Terra to replace the original self-convergent
coordinates of the base station.
2. If the base station is set up at a known point and its central point coordinates have not
been set during operation, you can define the base station coordinates as known point
coordinates before software computation.

1. After importing the data, click Base Station Center Point Settings.
2. Enter the Base Station Center Point Settings, and click to search for and set the base station’s
central point coordinate system as the one used when setting up the base station’s central
point during operation.
3. The data list displays the serial number, folder name, and central point latitude (X), longitude
(Y) and altitude (Z).
4. Individual edit: Input data in the fields for central point latitude, longitude, and altitude to
modify.
Note that individual edit:
a. Supports cut and paste;
b. Supports the pasting of latitude, longitude, and altitude together, leaving the software to
auto-fill the data (separate the values with a space and tab);
c. Click “Reset” to reset the base station data.
5. Batch edit: Check the box before the serial number and input/paste the data into the fields
for latitude (X), longitude (Y), and latitude (Z), then click “Batch Edit”.
6. Click “Save” to replace the original base station coordinates.

© 2024 DJI All Rights Reserved 63

DJI Terra Operation Guide

Point Cloud Density

The Point Cloud Density feature can be used to thin point cloud data and reduce point cloud
density.
1. Select point cloud density “By Percentage”: High means the original sampling rate, where
100% of the point clouds are used for processing. This produces the highest quality and is
the most time-consuming. Medium point cloud density uses 25% of the point clouds for
processing. It creates medium quality and takes a moderate amount of time. Low point
cloud density uses 6.25% of the point clouds for processing. This method is the least time-
consuming and generates sparse point clouds.
2. Select point cloud density “By Distance”: You can input any value between 5-50 cm to perform
downsampling on the point clouds evenly. The value represents the average point distance.
The larger the value, the sparser the point clouds and the faster the processing. Perform
operations based on mission’s requirements and avoid over-scattering the point clouds.

64 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Scenarios
Select usage scenario

1. Point Cloud Processing: High-accuracy point cloud data will be generated for this scenario.
For most cases of processing Zenmuse L1 or Zenmuse L2 data, select this scenario.
2. LiDAR Calibration: In case of color overlap with Zenmuse L1 or L2 data, you can select this
scenario to recalibrate intrinsic LiDAR parameters for better effects. After processing is
finished, click "Export Calibration File", save the file to the root directory of the microSD card,
and insert it into the LiDAR. Once powered on, the device automatically calibrates using
the calibration file and all computation for subsequently collected data are based on the
calibrated parameters.
3. Substation Reconstruction: In addition to point cloud data, 3D model data in standard format
will be generated if this scenario is selected. For scenarios involving many small, hollow
structures, such as substations, select "Substation Reconstruction" for better reconstruction
results than the traditional visible light mode can provide.

Point Cloud Processing

© 2024 DJI All Rights Reserved 65

DJI Terra Operation Guide

1. Point Cloud Effective Distance Range: If the points captured by the LiDAR exceed the effective
distance range, the points will be filtered out during processing. You can set an effective
point cloud distance if you need to reconstruct a nearer target area but will inevitably
capture the background in the distance.

2. Optimize Point Cloud Accuracy: If enabled, the software will optimize the adjustment of point
clouds scanned at different times during point cloud processing to improve their overall
accuracy.
3. Smooth Point Cloud: This feature reduces the point cloud thickness to remove discrete
noise, makes local structure clearer, and improves the effects of ground point extraction. It is
recommended to enable this feature in most cases.
4. Ground Point Classification: If this feature is enabled, ground points are differentiated from
non-ground points. Ground points can be used for generating a topographic map for the
mapping area.
Ground point classification algorithm: DJI Terra uses the specified building dimensions to
set evenly distributed seed points in the mapping area and then connects the seed points to
form an initial triangle network. Whether a laser point is a ground point can be determined
by calculating the angle formed between the laser point and the vertex of the triangle where
it is located and the distance from the laser point to the plane where the triangle is located. If
the angle and the distance are smaller than the values set for the Iteration Angle and Iteration
Distance parameters, the point is identified as a ground point and the triangle mesh is
updated. This process is repeated until all ground points in the mapping area are identified.
4.1 Ground Type: DJI Terra provides preset parameters for different types of terrain such as
Flat Ground, Gentle Slope, and Steep Slope. The default parameters suit most common
scenarios. If a mapping area has large undulations, you can increase the iteration angle
and distance as needed.
4.2 Building Max Diagonal: Set this parameter based on the diagonal length of the roof
plane of the largest building in the mapping area. If this parameter is set to a small value,
rooftops may be recognized as ground.

66 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

5. Generate DEM: If this feature is enabled, DEM data in TIFF format will be generated based on
the point cloud data of ground points. The following scales are supported for DEM results:
1:500, 1:1000, and 1:2000. You can also set a DEM result resolution based on the GSD.
6. Contour: This feature is used to generate contour lines from DEM. The generated contour
lines can be in DXF or SHP format. The contour lines will be automatically smoothed when
they are generated.
To generate contour lines, set the following parameters:
a. Interval: Difference in elevation between adjacent contour lines.
b. Datum: Elevation of contour lines is calculated based on the elevation datum and interval.
c. Elevation Annotation Radius: Used to generate elevation points for topographic mapping.
d. If "Discard Invalid Contour Lines" is enabled, contour lines shorter than the set length will
be discarded.

Substation Reconstruction

When "Substation Reconstruction" is selected, 3D model data in standard format will be

generated in addition to point cloud data. In scenarios involving many small, hollow structures,
such as substations, selecting "Substation Reconstruction" produces better reconstruction
results than the traditional visible light mode does.
If you enable "3D Mesh", 3D model data in standard format will be generated. You can set
Quality to Medium, High, or Ultra-High as needed. You can "Reduce Model To" a set proportion
from 5% to 100%. Model triangles will be reduced to the set proportion to cut the size of the
reconstruction file, which facilitates the online browsing of the model and other downstream
operations. It recommended to retain at least 20% of triangles.

© 2024 DJI All Rights Reserved 67

DJI Terra Operation Guide

Advanced Settings
1. Accuracy Control and Check: You can import control points or check points to improve or
check point cloud accuracy. Click "Advanced Settings" to show all the settings. Click the
import icon next to "Accuracy Control and Check".

68 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Similar to importing GCPs for aerotriangulation, select the control point or check point file, set
the correct coordinate system, adjust the data name, latitude (X/E), longitude (Y/N), and altitude
(Z/N), then click "Import".

2. Output Coordinate System: Refer to the section on output coordinate system in Chapter 3 for
details. Note that most backend point cloud analyzing software does not support the import
of point cloud files with a geographic coordinate system. Therefore, DJI Terra generates its
output in the UTM projected coordinate system by default.
3. Point Cloud: DJI Terra can output 3D point clouds in the following formats:
a. PNTS: Default format for display in DJI Terra (LOD point cloud format, suitable for display
in Cesium)
b. LAS: ASPRS LASer, 3D point cloud format, v1.2
c. S3MB: SuperMap LOD point cloud format
d. PLY: Non-LOD point cloud format
e. PCD: Non-LOD point cloud format
4. Merged Output: By selecting LAS, PLY or PCD as the point cloud format, you can enable the
"Merged Output" feature, which would merge the point cloud output into a single file.

Start Processing
Click “Start Processing” and the progress bar at the bottom will show the processing progress.
During processing, you can click “Stop” and the software will save your current progress. If
you resume processing after pausing, the software will backtrack a section before the saved
progress point and continue the processing.

You can start multiple point cloud processing missions. Before the first started mission
is completed, the other missions will be in the queue. Once the previous mission is
completed, the next started mission will be processed.

© 2024 DJI All Rights Reserved 69

DJI Terra Operation Guide

Viewing Output
After completing reconstruction, you can translate, zoom in or out of, or rotate the output, in
the same manner as 3D reconstruction models. You can choose to display the output differently
at the bottom of the page.
1. RGB: Displays based on true color.
2. Reflectivity: Displays the corresponding color based on the object’s reflectivity, in the scale of
0-255. The range of 0-150 corresponds to diffuse objects with a reflectivity of 0-100%, while
151-255 corresponds to fully reflective objects.
3. Height: Displays different colors depending on the cloud point’s height.
4. Return: When double return or triple return is selected during data collection, different
colors are displayed based on the return information received by the point cloud.
5. Type: If the "Ground Point Classification" feature is enabled, you can click this button to view
only ground points or only non-ground points.
6. DEM: Click this button to view DEM results. Only 2D models can be displayed.

Profile: Profiles help you quickly identify elements such as vegetation and ground. Click the profile
icon to open the Profile panel. Select a desired part on the model and draw a profile line by
clicking your mouse. At the top of the Profile panel, set the profile width. The default profile width
is 5 m. The measuring tool allows you to measure straight distances, horizontal distances, vertical
distances, slopes, and other information. The Gridlines feature provides quick access to valid data.

70 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Click "Quality Report" to view and save the report in PDF format. The report contains
reconstruction parameters, mission parameters, accuracy parameters, and output parameters.

© 2024 DJI All Rights Reserved 71

DJI Terra Operation Guide

LiDAR Point Cloud File Format and Storage Path

LiDAR point clouds are saved to this path by default. You can change the cache directory in
Settings. They can also press the shortcut keys “Ctrl+Alt+F” on the reconstruction page to open
the folder of the current mission.
C:\Users\< Computer user name >\Documents\DJI\DJI Terra\< DJI account name >\< Mission
name >\< lidars>
In the output folder, the most important files are the 3D point clouds in LAS format and the
flight trajectory files in OUT format. The LAS point clouds generated by DJI Terra are standard
LiDAR output from aircraft-mounted LiDARs, with a version number of v1.2. Most backend
software supports the direct import of such output. LAS point clouds record information such
as the 3D point coordinates, RGB colors, reflectivity, time, number of returns, which return the
3D point belongs to, total number of points for each return, and scanning angles.
“_sbet.out” indicates a post-processing trajectory file. It records the trajectory information after
adjustment calculation. You may import it into third-party software to view the trajectory, or
use third-party software to perform a secondary adjustment processing. For more information,
please refer to documentation about DJI Zenmuse L1 or DJI Zenmuse L2 available on DJI's
official website: L1-Trajectory File Format Specification (description of the SBET and SMRMSG
formats) or DJI Zenmuse L2 Operation Guide.

72 © 2024 DJI All Rights Reserved

Detailed Inspection
Function Overview
Designed for inspection of power lines and transmission towers, this feature can be used
based on 3D models or point clouds generated by DJI Terra or by importing third-party LAS
point clouds. You can select a target to automatically generate waypoints and a flight route, for
automatic offline detailed inspection. When planning an inspection route, DJI Terra will display
a simulated camera transmission image corresponding to the waypoints, while using the
flight route safety check function to create more accurate waypoints and a more efficient and
safer route. Generated flight routes can be imported directly into the remote controller to be
executed automatically.
The basic process for Detailed Inspection is as shown below:

Importing Data
New Mission
Open and log into DJI Terra, click “New Mission” at the bottom left corner, then select “Detailed
Inspection” as the mission type.

© 2024 DJI All Rights Reserved 73

DJI Terra Operation Guide

Selecting Aircraft and Adding Data

After creating and naming a detailed inspection mission, select the aircraft for the mission on
the left of the interface, click to expand the menu, and select Phantom 4 RTK, Matrice 300 RTK,
Mavic 2 Enterprise Advanced, Matrice 30 series, or Mavic 3 Enterprise series. After selecting
the aircraft, view the list of models that can be imported on the left of the interface. Select
the desired mission, check that it is correct and click “Confirm” on the lower right corner to
complete the import.

If there are multiple models, you can click the icon and quickly look up the name of the
mission that needs to be imported. To import a third-party LAS point cloud model, click the
icon.

A third-party LAS point cloud model must contain and have the right coordinate system
selected, otherwise the import will fail. Point clouds without a coordinate system are
not supported.

Route Setting and Planning

Route Setting
Before planning, we recommend setting the global parameters for the flight route first to
increase planning efficiency. For example, if the global “Shooting Distance” is set at 3 m, the
distance between the aircraft and the shooting location will be automatically adjusted to 3 m
during the creation of any subsequent inspection point. The specific steps are as follows:
1. Aircraft yaw: We recommend using “Set Waypoint Separately”. Selecting waypoints in the
direction of the route may increase the aircraft’s yaw rotation time, thus lowering inspection
efficiency.
2. Finishing Action: Defines the action executed by the aircraft when the mission is completed.
When “Return to Home” is active, the return altitude and logic will follow the parameter
settings in the app.

74 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

3. Photo Ratio: Apply default setting.

4. Shooting Distance: If a shooting distance has been set in “Flight Route Settings”, it
automatically overwrites the shooting distance for every waypoint in “Waypoint Settings”.
You may reset each waypoint as needed in “Waypoint settings”.
5. Flight Route Speed: The aircraft’s flight speed when executing a mission on a flight route.
If a flight speed has been set in “Flight Route Settings”, it automatically overwrites the flight
speed for every waypoint in “Waypoint Settings”. You may reset each waypoint as needed in
“Waypoint settings”.
6. Initial Speed: The aircraft’s speed upon reaching its first waypoint and during its return after
completing the mission. This can be set by you.
7. Safety Distance: The Safety Distance between the flight route and the model. When the
distance between a flight route section and the model is shorter than the set Safety Distance,
the section will appear red and show the shortest distance between the flight route and
model.

Inspection Waypoint Settings

Flight routes involving power lines and transmission towers can vary widely. You can plan the
route according to factors such as shooting order, efficiency, and flight safety. This example
below is based on the inspection method for cat-head type towers described in the “Drone
Photography Manual for Inspection of Overhead Transmission Towers (Sealed Version)”.

© 2024 DJI All Rights Reserved 75

DJI Terra Operation Guide

1. Preparation: DJI Terra’s Detailed Inspection feature can simulate the images seen by the
aircraft. After adjusting the POV with your cursor, left-click a shooting location to create a
waypoint, which will be automatically generated based on the parameters, shooting angles
and other information in “Global Settings”. To modify a waypoint, you can select it (which will
turn blue) and then adjust its location, shooting distance, and flight speed.
A waypoint’s location can be adjusted with the “up”, “down”, “left”, and “right” keys on the
keyboard, or by dragging a blue waypoint or red location point. To switch between two
adjacent waypoints, click “<” or “>” or press “Ctrl+ ← ” or “Ctrl+ → ” on the keyboard.

2. Setting the Entry Point: Before starting an inspection mission, set an entry point to ensure
your aircraft is placed at a preset location and can enter the inspection route safely.
Therefore, an entry point should preferably be a more conspicuous location (example: a
central location facing the transmission tower), so that the pilot can easily judge whether the
aircraft’s correctly located.

76 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

(A conspicuous entry point makes it easier for the pilot to judge the aircraft’s location.

3. Setting Inspection Waypoints: According to inspection requirements, inspection waypoints

should cover the entire tower, tower head, tower body, base, voltage transmitter and
receiver points, insulators, left insulator and hanging point, middle insulator and hanging
point, and right insulator and hanging point.
a. Each inspection waypoint will automatically include a “Take Photo” waypoint action. You
can add or remove any waypoint action if it is an entry or exit point or if you have other
requirements.

b. “Camera Facing Target” is automatically enabled for every inspection waypoint, to ensure
the aircraft places the target at the center of the image. You can disable the feature
manually and set custom aircraft yaw and gimbal pitch angles.

© 2024 DJI All Rights Reserved 77

DJI Terra Operation Guide

c. If you are using a Matrice 300 RTK mounted with a H20 series payload, “Zoom” can also
be set for each waypoint for easier optical zoom adjustment. While adjusting “Zoom”, DJI
Terra can simulate changes on the aircraft image.

4. Setting Transition Points: Since the flight route between any two waypoints is linear,
transition points must be added in order to avoid collision risk. They are often added at
corners or turning points. As transition points are only for adjusting the aircraft’s flight
attitude, you can remove “Take Photo” from their waypoint actions.

Transition Point

Remove Take
Photo Action

78 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

Flight Route Safety Check

By setting the “Safety Distance”, you can judge whether a flight route has any risk of collision.
The flight route will appear green if it is safe. However, when the distance between a flight
route section and the model is shorter than the set safety distance, the section will appear red
and show the shortest distance between the flight route and model. In the example below,
assuming the safe distance is 1.5 m, DJI Terra is automatically judging whether this flight route
section has any collision risk. A warning alert appears on the top left corner of the interface,
while the locations and values with the shortest distance are marked in red (the distance being
1.45 m in this example). When this happens, you should adjust the corresponding waypoint
locations on that section to bring them within the safety distance.

© 2024 DJI All Rights Reserved 79

DJI Terra Operation Guide

Route Export and Execution

After you complete planning the route and setting the parameters, save the mission. Click the
“Export KML” icon on the top right corner of the interface to export the KML flight route
mission and save it in the designated folder. The exported KML flight route file can be imported
into the DJI Pilot app to perform a flight.
1. Open the DJI Pilot app, select “Flight Mission”, click “Import KML” and “Waypoint Flight”, then
find the KML file exported from DJI Terra in the corresponding directory.

2. After loading the relevant mission, check the waypoint parameters again before clicking the
icon on the left to upload the flight route and execute the inspection mission.

The aircraft model number for the detailed inspection settings must be the same as
the actual aircraft being used, otherwise the simulated preview and actual shot images
would differ, thus affecting the inspection effects.

80 © 2024 DJI All Rights Reserved

DJI Modify
Introduction to DJI Modify
Due to a variety of factors including the environment and light, drone lens distortion, and
modeling algorithms, 3D models acquired by oblique photogrammetry are inevitably prone
to weak textures and holes, water surface modeling failures, insufficient overlap, local model
distortion, and other issues.

DJI Modify is DJI's first intelligent 3D model editing software. It features a streamlined and
intuitive interface that makes it simple to operate. It accurately identifies floating parts, holes,
and vehicles in the models. Mesh can be repaired with a single click, with true-to-life textures
generated quickly, significantly simplifying the model editing process.
With one click, DJI Terra can launch DJI Modify. DJI Terra can also easily generate files for model
editing with another click. These simple steps make the subsequent model editing process quick
and convenient.

© 2024 DJI All Rights Reserved 81

DJI Terra Operation Guide

Work with DJI Modify

For a visible light reconstruction project that has not proceeded to aerotriangulation and
subsequent steps, if you select "Work with DJI Modify" in DJI Terra, an aerotriangulation file in
XML format will be generated. The "3D Model" option will be automatically enabled and model
files in B3DM and PLY formats will be generated. If no 3D model block splitting rule is set, DJI
Terra will automatically perform block splitting within the RAM capacity of 32 GB. If you have
special block splitting requirements, return to "3D Model" > "Advanced Settings" to modify the
block splitting rule. If you manually split the model into blocks, it is recommended to keep the
size of the memory used by each single block within 32 GB. If the size of a block is too large,
editing the model will take up too much memory.

82 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

After reconstruction is completed, you can click "DJI Modify" in the project menu to directly
open this 3D model in DJI Modify and start editing the model.

If you need to copy model output from computer A to computer B for model editing,
you must first find the following files in the folder of the 3D project in computer A:
XML aerotriangulation file (path: project folder-AT), image distortion correction file
(path: project folder--models-pc-0-temp-ImageCorrection-Undistort), B3DM model
output file (path: project folder--models-pc-0-terra_b3dms), and PLY model output file
(path: project folder-models-pc-0-terra_ply). Copy these files to computer B. Launch
DJI Modify, create a project, and start model editing.

© 2024 DJI All Rights Reserved 83

Appendix - Error Alerts and Operating Suggestions
If an error is reported, DJI Terra will provide corresponding assistance. Most issues can be
resolved by following the guide. You may also look up the error code in the attached table to
search for a solution.

For an issue that cannot be resolved, click “Export Log” to save the mission’s log and send it to
the relevant aftersales personnel for troubleshooting.

DJI Terra - Reconstruction Error Alerts and Operating Suggestions (Updated)

Error
Error Alert Operating Suggestion
Code
General
No usage permission
0001 1. Use the license containing this feature.
for this function
NVIDIA graphics
1. Check if graphics card of your computer meets the
0004 card (4GB or above)
requirements.
required
1. Check if license is bound successfully.
1001 License error
2. Make sure license is within validity period.
1. Check task JSON file input and make sure each keyword
is completely consistent with sample document and all
1002 JSON file read error
necessary fields exist.
2. Check message prompt of last output of log file.

84 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

1. If error prompt appears at beginning of reconstruction,

check if graphics card of your computer meets the
requirements. Refer to DJI Terra FAQ Purchase and
Preparation for Computer Configuration Requirements). If
NVIDIA graphics card graphics card meets the requirements, download graphics
1003
not detected card driver from NVIDIA’s official website and upgrade it to
latest version.
2. If error prompt appears after reconstruction had started
for a period of time, create a region of interest to remove
edge region and try reconstruction again.
1004 DMV file read error 1. Try aerotriangulation again.
1. Check if there is sufficient space available in disk where
cache directory is located.
File or file folder 2. For cluster reconstruction, check local area network
1005
creation error connection and if there is sufficient space available in disk
where shared directory and local temporary directory of
worker device are located.
Aerotriangulation
block splitting error.
Check available RAM 1. Check available RAM or reduce Distance to Ground/
1007
or reduce distance Subject.
from ground to
camera/subjects
1. Make sure imported image file is not corrupted. Open
image file folder and enable thumbnail view. Check
if image thumbnails are all black, non-displayable,
overexposed, or underexposed.
2. Make sure not to import image thumbnails.
3. Make sure there is zone of mutual visibility between
Aerotriangulation
2001 images and overlap ratio is greater than the required
computation failed
minimum.
4. Make sure to define correct coordinate system for added
GCPs and imported custom Image POS Data, and check
if GCP marks, ordering of latitude and longitude, POS
accuracy, and control point accuracy are correct (the
accuracy cannot be 0.0).
1. Check if there is sufficient space available in disk where
cache directory is located.
Distortion correction
2002 2. For cluster reconstruction, check local area network
error
connection and if there is sufficient space available in disk
where shared directory is located.
1. Check if graphics card has a memory greater than 4 GB.
Point cloud generation
2003 2. Create a region of interest to remove edge region and try
error
reconstruction again.

© 2024 DJI All Rights Reserved 85

DJI Terra Operation Guide

1. Check if there is sufficient space available in disk

where cache directory and shared directory for cluster
Network
2004 reconstruction are located.
reconstruction failed
2. Create a region of interest to remove edge region and try
reconstruction again.
1. Create a region of interest to remove edge region and try
2005 Scene divide error
reconstruction again.
1. Check if there is sufficient space available in disk where
cache directory is located.
2. For cluster reconstruction, check local area network
connection and if there is sufficient space available in disk
2006 LOD error
where shared directory and local temporary directory of
worker device are located.
3. Create a region of interest to remove edge region and try
reconstruction again.
1. Check if there is sufficient space available in disk where
cache directory is located.
2. Create a region of interest to remove edge region and try
reconstruction again.
3. For cluster reconstruction, check local area network
2007 MVS error connection and if there is sufficient space available in disk
where shared directory is located.
4. Check if graphics card has a memory greater than 4 GB.
5. Refer to DJI Terra Quality Report for Aerotriangulation and
check whether optimized intrinsic camera parameters are
significantly different from the initial.
Failed to generate XML 1. Make sure image data is within coverage of XML output
2008
file coordinate system.
Semantic identification
3001 1. Check if graphics card has a memory greater than 4 GB.
error
1. Open Task Manager and check memory usage. End tasks
Digital Surface Model
that occupy too much memory.
3002 (DSM) generation
2. Check if there is sufficient space available in disk where
failed
cache directory is located.
Orthorectification 1. Check if there is sufficient space available in disk where
3003
failed cache directory is located.
Orthophoto mosaic 1. Check if there is sufficient space available in disk where
3004
failed cache directory is located.

86 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

1. Check if there is sufficient space available in disk where

cache directory is located.
2. For cluster reconstruction, check local area network
Generating map tile
3005 connection and if there is sufficient space available in disk
failed
where shared directory is located.
3. Create a region of interest to remove edge region and try
reconstruction again.
1. Check if there is sufficient space available in disk where
cache directory is located.
Generating 2D map
3006 2. For cluster reconstruction, check local area network
report failed
connection and if there is sufficient space available in disk
where shared directory is located.
1. Check if there is sufficient space available in disk where
cache directory is located.
3007 Writing image failed 2. For cluster reconstruction, check local area network
connection and if there is sufficient space available in disk
where shared directory is located.
1. Check if there is sufficient space available in disk where
cache directory is located.
Projecting 2. For cluster reconstruction, check local area network
3008 reconstruction output connection and if there is sufficient space available in disk
failed where shared directory is located.
3. The region where reconstruction data is located is not
within coverage of the set projected coordinate system.
1. Make sure vertical gimbal angle is greater than 15° (35° for
v3.1.0 and later versions).
2. If image data is not of above type, check aerotriangulation
result. Try the following steps if there is error in
aerotriangulation output:
a. Make sure imported image file is not corrupted. Open
image file folder and enable thumbnail view. Check
if image thumbnails are all black, non-displayable,
overexposed, or underexposed;
3009 Invalid image data
b. Make sure not to import image thumbnails;
c. Make sure there is zone of mutual visibility between
images and overlap ratio is greater than the required
minimum;
d. Make sure to define correct coordinate system for
added GCPs and imported custom Image POS Data,
and check if GCP marks, ordering of latitude and
longitude, POS accuracy, and control point accuracy are
correct (the accuracy cannot be 0.0).

© 2024 DJI All Rights Reserved 87

DJI Terra Operation Guide

1. Check if there is sufficient space available in disk where

cache directory is located.
Preprocessing image
3010 2. For cluster reconstruction, check local area network
failed
connection and if there is sufficient space available in disk
where shared directory is located.
1. Open Task Manager and check memory usage. End tasks
that occupy too much memory
2. Create a region of interest to remove edge region and try
reconstruction again.
4002 Insufficient memory
3. For LiDAR point cloud processing, reduce point cloud
density and try again or adjust point cloud effective
distance to save resources used for processing invalid
regions.
1. Open Task Manager and check graphics card memory
Insufficient graphics usage. End tasks that occupy too much memory.
4003
card memory 2. Create a region of interest to remove edge region and try
reconstruction again.
1. Check if coordinate system is set correctly.
2. If image does not contain POS information, you are unable
to set known coordinate system for output.
3. Make sure to set arbitrary coordinate system for output if
Output coordinate
5001 GCPs from arbitrary coordinate system are added.
system error
4. Custom PRJ file contains special elevation system not
supported by DJI Terra at the moment.
5. The region where reconstruction data is located is not
within coverage of the set elevation system.
Quality Report export 1. Check if mission folder and report folder where quality
5002
error report is stored have write permission enabled.
1. Check if there is sufficient space available in disk where
cache directory is located.
6001
2. For cluster reconstruction, check local area network
/ Error
connection and if there is sufficient space available in disk
6002
where shared directory and local temporary directory of
worker device are located.
LiDAR Point Cloud Reconstruction
Original data missing
8001 LDR type file or file 1. Check if original data is missing LDR type file.
path error
CLI file missing in
8002 original data or file 1. Check if original data is missing CLI type file.
path error

88 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

POS data error. Check

base station data and
RTK data, and perform 1. Check if base station data and RTK data are missing.
8007
calibration flight 2. Perform calibration flight before and after data collection.
before and after data
collection.
Optimizing LiDAR
1. Increase overlap ratio of collected data and process data
8008 point cloud accuracy
again.
failed
File read and write 1. Check if there is sufficient space available in disk where
8009
error cache directory is located.
CLC file missing in
8010 original data or file 1. Check if CLC file is missing in original data.
path error
Zenmuse L1
calibration failed.
8012 1. Make sure to import at least 4 images.
Insufficient number of
images in original data
Zenmuse L1
calibration failed.
Different device data
1. Different device data or several subfolders in the same
8015 or several subfolders
imported folder currently not supported.
in the same imported
folder currently not
supported
Parsing base station
8020 data file to RENIX 1. Contact DJI Support for assistance.
format error
Attitude initialization 1. Make sure to perform calibration flight before and after
8022
failed data collection
Optimizing LiDAR
1. Increase overlap ratio of collected data and process data
8027 point cloud accuracy
again.
failed
File read and write 1. Check if there is sufficient space available in disk where
8028
error cache directory is located.
File path error or CLC
8029 file missing in original 1. Check if CLC file is missing in original data.
data
Original data CLC file
8030 1. Check if CLC file is missing in original data.
format error
Zenmuse L1
calibration failed.
8031 1. Make sure to import at least 3 images.
Insufficient number of
images in original data

© 2024 DJI All Rights Reserved 89

DJI Terra Operation Guide

Base station data file

missing in original
data. Check if any of
1. Check if any of the following files are missing: RTB, OEM,
8102 the following files are
RINEX, RTCM 3, and UBX.
missing: RTB, OEM,
RINEX, RTCM 3, and
UBX
RTK file missing in
8104 original data or file 1. Check if RTK file is missing in original data.
path error
Parsing base station
data file failed. Make
sure file is in one of 1. Check if base station file is in one of the following formats:
8108
the following formats: RTB, OEM, RINEX, RTCM 3, and UBX.
RTB, OEM, RINEX,
RTCM 3, or UBX
IMU file missing in
8122 original data or file 1. Check if IMU file is missing in original data.
path error
RTL file missing in
8124 original data or file 1. Check if RTL file is missing in original data.
path error
Attitude initialization
failed. Make sure to
1. Make sure to perform calibration flight before and after
8133 perform calibration
data collection.
flight before and after
data collection
Cluster Reconstruction
1. Unable to sync original images. Check local area network
JPEG required for
1006 connection and if there is sufficient space available in disk
reconstruction
where shared directory is located.
1. Make sure worker devices on local area network are
enabled and available.
2. Refresh worker device list in cluster reconstruction device
No available worker
7001 list of control device. Make sure that worker devices can be
device
searched, then select devices and apply for reconstruction.
3. If firewall is enabled, make sure that DJIPicMapGen.exe
has access to local area network (LAN) in firewall settings.
Unable to visit
control device shared 1. Check local area network connection and make sure
7002
directory. Check control device has access to shared directory.
device connection
Worker device 1. Check and make sure local area network connection is
7003
disconnected stable. Replace network cable if issue recurs frequently.

90 © 2024 DJI All Rights Reserved

 DJI Terra Operation Guide

1. Check if shared directory set by control device is accessible

7004 File folder not found for worker devices. Mapped drive letter path is not
available at the moment.
Unable to write file
7005 1. Check if shared folder has write permission enabled.
folder
Port for searching 1. Check if control device has write permission for shared
7703
worker device error directory.

If issue cannot be solved, contact DJI Support.

© 2024 DJI All Rights Reserved 91

This content is subject to change.
Download the latest version from:
http://www.dji.com/dji-terra

If you have any questions about this document, please contact

DJI by sending a message to: [email protected]

Copyright © 2024 DJI All Rights Reserved