FACULTY OF ARCHITECTURE, PLANNING AND SURVEYING

CENTRE OF STUDIES FOR SURVEYING, SCIENCE AND GEOMATICS

LAB REPORT 1
GLS 696: UAV IMAGE PROCESSING

CLASS:
RAP2208B

PREPARED FOR:
SR. KAPT KHAIRIL AFENDY BIN HASHIM

PREPARED BY:
NOOR NASUHA BT KAMARUDIN 2015149883
FATIN SHAHIRAH BT MOHD SAFIEE 2015110779
NUR NASUHAH BT ABDULLAH 2015111175
SITI NURSYAZANA BT ZULKIFLI 2015180243

Lab Report 1 UAV | 1

Table of Contents
CHAPTER 1 INTRODUCTION .......................................................................................................................... 3
1.1 Unmanned Aerial Vehicle (UAV) ................................................................................................... 3
1.2 Component of Unmanned Aerial Vehicle (UAV) ........................................................................... 3
1.3 Processing Software ...................................................................................................................... 4
CHAPTER 2 AGISOFT PHOTOSCAN ................................................................................................................ 5
2.1 Introduction .................................................................................................................................. 5
2.2 Processing Flow Chart ................................................................................................................... 5
2.3 Agisoft Image Processing .............................................................................................................. 6
2.3.1 Add Photo ............................................................................................................................. 6
2.3.2 Photo Alignment ................................................................................................................... 7
2.3.3 Build Dense Cloud ................................................................................................................. 8
2.3.4 Build Mesh ............................................................................................................................ 9
2.3.5 Build Texture ....................................................................................................................... 11
2.3.6 Build Tiled Model ................................................................................................................ 12
2.3.7 Build Digital Elevation Model (DEM)................................................................................... 12
2.3.8 Build Orthomosaic............................................................................................................... 13
2.3.9 Export image results ........................................................................................................... 15
CHAPTER 3 PRODUCTION OF MAP ............................................................................................................. 16
3.1 Introduction ................................................................................................................................ 16
3.2 Add data ...................................................................................................................................... 16
3.3 Data editing ................................................................................................................................. 17
3.3 Projection System ....................................................................................................................... 18
3.3 Map Element ............................................................................................................................... 19
3.4 Export Map.................................................................................................................................. 23
3.5 Output of DEM and Orthophoto map ......................................................................................... 24
CHAPTER 4 RESULTS.................................................................................................................................... 25
4.1 Introduction ................................................................................................................................ 25
4.2 Root Mean Square Error (RMSE) ................................................................................................ 25
APPENDICES ................................................................................................................................................ 26

Lab Report 1 UAV | 2

 CHAPTER 1 INTRODUCTION

1.1 Unmanned Aerial Vehicle (UAV)

Active technological development has filled rapid growth in the number of

Unmanned Aerial Vehicle (UAV) platform being deployed around the globe. Improved
sensors and enhanced image processing techniques have consolidated and confirmed
UAVS as the technology of choice. Many jurisdictions have regulated stringent
restrictions on flying UAVs in numerous scenarios where they could be great of value.
Despite the regulation, UAV or drone has rapidly become the tool of choice for both the
environmental science and surveying communities.

However, UAVs are employed in high risk environments. The ability to sense
and avoid obstacles, both natural and man-made, and to rebuild its flight path is an
important feature that UAVs should possess, and the corresponding algorithms should
be imbedded in their guidance and control systems. In many cases, UAVs are described
either as a single air vehicle (with associated surveillance sensors) or a UAV system,
which usually consists of three to six air vehicles, a ground control station, and support
equipment.

1.2 Component of Unmanned Aerial Vehicle (UAV)

Unmanned Aerial Vehicle (UAV) carry payloads (such as a camera) that weigh
considerably less than an adult human, and as a result it can be considerably smaller. All
drone parts and components are vital to a smooth and safe flight. Knowing the parts of a
UAV will give user extra confidence while flying.

Users will also know which components to inspect on a regular basis and the
UAV parts which are easy to replace or upgrade. If have any flight problems, then
knowing what each UAV component does will assist you greatly in getting to the
bottom of any flying issues. The UAV component below shows overview of all the
UAV components can be found on consumer and professional UAV.

Lab Report 1 UAV | 3

 Body Power supply Platform

Flight Control Software Sensor

Communications

Figure 1.1: UAV components

1.3 Processing Software

Agisoft PhotoScan is photogrammetry software developed by Russia-based

Agisoft that allows user to generate point clouds data, 3D models and their
corresponding textures from a series of photos. It’s an easy to use and affordable
alternative to laser scan solutions.

Figure 1.2: Agisoft PhotoScan

Lab Report 1 UAV | 4

 CHAPTER 2 AGISOFT PHOTOSCAN

2.1 Introduction

This chapter elaborate methodology composing of definite procedures used to

identify and analyze the provided source about Agisoft PhotoScan. This section also
describes the emphases on the processing approaches along with UiTM Arau aerial
photo which were applied for this image processing operation.

2.2 Processing Flow Chart

7. Build Digital
1. Add UiTM Arau
6. Build Tiled Model Elevation Model
aerial image
(DEM)

8. Build
2. Photo Alignment 5.Generate texture
Orthomosaic

3. Generate Dense 4. Build Mesh (3D 11. Export image

Cloud polygonal model) results

Figure 2.1: Agisoft PhotoScan methodology

Lab Report 1 UAV | 5

2.3 Agisoft Image Processing

2.3.1 Add Photo

Browse to the folder contains the needed processing image dataset into Agisoft
PhotoScan software. It has 293 aerial images of UiTM Arau, Perlis taken in 2004. Select
all the images by clicking Add Photos command on Workflow menu and click Open.

Figure 2.2: Add Photos command

Reference workspace shows the images information loaded with longitude,

latitude and altitude inputs. Photo pane view will now appear on aerial image chosen.
The visible blue dots in model view represent camera positions.

Figure 2.3: Agisoft PhotoScan Workspace view

Lab Report 1 UAV | 6

2.3.2 Photo Alignment

Succeeding from the Add photos command, to align photos open workflow
menu and click align photos command. The parameters set for this processing as shown
below. After the alignment is finished, orientations of camera positions on photos were
refined and a sparse point cloud will be displayed on model view.

Figure 2.4: Align photos command and parameter

Figure 2.5: Sparse point cloud

Table 2.1: An overview of parameters for align photo

Parameter Description
Accuracy - As it suits with current computer speed processing however
Medium High accuracy offers the best results.
Pair Chosen for match processing. Agisoft PhotoScan detected
preselction - match feature only on those photos which overlap according
Reference to the reference data.

Lab Report 1 UAV | 7

2.3.3 Build Dense Cloud

Generation and visualization of a dense point cloud model can be achieved in

Agisoft PhotoScan environment. Based on the estimated camera positions, the program
calculates depth information for each camera to be combined into a single dense point
cloud. It tends to produce additional dense point clouds, which are of almost the same
density, if not denser, as LIDAR point clouds. A dense point cloud can be edited and
classified within the software or exported to an external tool for further analysis.

Figure 2.6: Build dense cloud command and parameter

Table 2.2: An overview of parameters for build dense cloud

Parameter Description
Quality - Higher quality settings can be used to obtain more detailed
Medium and accurate geometry, but they require longer time for
processing. Interpretation of the quality parameters here is
similar to that of accuracy settings given in Photo Alignment
section.
Additional - At the stage of dense point cloud generation reconstruction
Aggressive Agisoft PhotoScan calculates depth maps for every image.
To sort out the noisy or badly impaired images, Agisoft
PhotoScan has several built-in filtering algorithms that filter
among the outliers.

Lab Report 1 UAV | 8

 Figure 2.7: Output of build dense cloud

2.3.4 Build Mesh

When the dense point cloud has been generated it is possible to generate
polygonal mesh model based on the certain point classes. It specifies the source for the
mesh generation procedure. Sparse cloud can be used for fast 3D model generation
based solely on the sparse point cloud. Face count option represent the optimal number
of polygons for a mesh in corresponding level of detail. Select Build Mesh command
from the Workflow menu to produce DEM on ground, point class should be used for
mesh reconstruction. Then, followed the parameters set below.

Figure 2.8: Build mesh command and parameter

Lab Report 1 UAV | 9

Table 2.3: An overview of parameters for build mesh
Parameter Description
Surface – To optimized for modelling of planar surfaces, such as
Height field terrains or base reliefs. It is suitable for aerial photography
processing as it requires lower amount of memory and
allows for larger data sets processing.
Source data Setting of Dense Cloud is chosen because as it will generate
– Dense high quality output based on the previously reconstructed
cloud dense point cloud even though it will result in longer
processing time.
Face count - It specifies the maximum number of polygons in the final
Medium mesh. Suggested values of High, Medium and Low are
calculated based on the number of points in the previously
generated dense point cloud.
Interpolation This mode in Agisoft PhotoScan will interpolate some
- Enable surface areas within a circle of a certain radius around every
dense cloud point. As a result some holes can be
automatically covered. Yet some holes can still be present on
the model and are to be filled at the post-processing step.

Figure 2.9: Output of build mesh

Lab Report 1 UAV | 10

2.3.5 Build Texture

This step can be considered to opt out in in the orthophoto export workflow, but
it might be necessary to inspect a textured model before exporting it or it might be
useful for precise marker placement. Orthophoto mapping mode sets the whole object
surface is textured in the orthographic projection. Blending mode in build texture shows
how pixel values from different photos will be combined in the final texture.

Figure 2.10: Build texture command and parameter

Table 2.4: An overview of parameters for build texture

Parameter Description
Mapping The Orthophoto mapping mode produces even more
mode - compact texture representation than the Adaptive orthophoto
Orthophoto mode at the expense of texture quality in vertical regions.
Blending Implies two-step approach where it does blending of low
mode - frequency component for overlapping images to avoid
Mosaic seamline problem such as weighted average. Weight being
dependent on a number of parameters including proximity of
the pixel in question to the center of the image.
Texture Shows the width & height of the texture atlas in pixels
size/count -
4096

Lab Report 1 UAV | 11

2.3.6 Build Tiled Model

Tiled model produce a detailed textured mesh from the dense point cloud.
Hierarchical tiles are textured from the UiTM Arau, Perlis aerial imagery. On the
textured model with higher polygon count such as 1 000 000 - 10 000 000 it allows for
responsive visualisation of large area 3D models in high resolution. While polygon
count below from the recommended values is very rough and inaccurate might need to
rebuild the mesh basing on the dense point cloud. Both pixel size (m) and tile size are
generated by Agisoft PhotoScan environment.

Figure 2.11: Build tiled model command and parameter

2.3.7 Build Digital Elevation Model (DEM)

Following the tiled model process, Build DEM command allows to generate and
visualize the model type. A DEM represents a surface model as a regular grid of height
values. DEM in this project were rasterized from a dense point cloud while other
options are sparse point cloud or a mesh. Most accurate DEM output is calculated based
on dense point cloud data. Ensure that the coordinate system for the model is set before
building DEM operation
Agisoft PhotoScan also enables to perform DEM-based point, distance, area,
volume measurements as well as generate cross-sections for a part of the scene selected
by the user. Additionally, contour lines can be calculated for the model and depicted
either over DEM or Orthomosaic in Ortho view within software environment. DEM can
be calculated for referenced models only.

Lab Report 1 UAV | 12

 Figure 2.12: Build DEM command and parameter

Table 2.5: An overview of parameters for build DEM

Parameter Description
Coordinate At this step coordinate system for the model as set earlier in
system the WGS 84.
WGS84
Source data Mesh surface type allows to create orthomosaic for less
– Dense common like orthomosaic generation for facades of the
cloud buildings or other models that might be not referenced at all.
Interpolation The highest frequency component is blended independently
-Enabled along the seamline only, each further step away from the
seamline resulting in a less number of domains being subject
from several frequency domains.

2.3.8 Build Orthomosaic

Agisoft PhotoScan allows to perform orthomosaic seamline editing for better

visual results. The most common application is aerial photographic survey data such as
UiTM Arau aerial image and reconstructed model processing, but it may be also became
handy when a detailed view of the object on image is required. Orthomosaic export is
normally used for generation of high resolution imagery based on the source of loaded
photos.
Orthomosaic creation based on DEM data is especially efficient for aerial survey
data processing allowing for time saving on mesh generation step. Coordinate system
for the Orthomosaic referencing was set for Geographic in the Build Orthomosaic
dialog box. The desired type of surface for orthomosaic generation process is DEM
while Mosaic (default) on blending mode. Generated orthomosaic can be reviewed in

Lab Report 1 UAV | 13

Ortho mode similar to the DEM. It can be opened in this view mode by double clicking
on the orthomosaic label in the Workspace pane.

Figure 3.13: Build orthomosaic command and parameter

Table 2.6: An overview of parameters for orthomosaic

Parameters of Orthomosaic
Parameter Description
Projection - The default projection set in the Ground Control Settings is
Geographic used
Surface - Mesh surface type allows to create orthomosaic for less
Mesh common like orthomosaic generation for facades of the
buildings or other models that might be not referenced at all.
Blending The highest frequency component is blended independently
mode - along the seamline only, each further step away from the
Mosaic seamline resulting in a less number of domains being subject
from several frequency domains.
Pixel size Default value (maximum effective resolution is shown by
(deg) default) of the UiTM Arau aerial original images.
Region Setup boundaries were tick for the model's part that should
be projected and presented as orthophoto.

Lab Report 1 UAV | 14

 Figure 2.14: Results of orthomosaic

2.3.9 Export image results

After the polygonal model reconstruction and GCP verification on UiTM Arau,
Perlis corrected image the results need to be export using the standard export
instruments. In the Export Model dialog specify the coordinate system and indicate
export parameters applicable to the selected file type

Lab Report 1 UAV | 15

 CHAPTER 3 PRODUCTION OF MAP

3.1 Introduction

This chapter elaborates production of maps composing of definite procedures

used to processing the visual results about ArcMap. This section also describes the
emphases on the processing approaches along with UiTM Arau aerial photo which were
applied for this production map operation.

3.2 Add data

To start ArcMap, choose Start the software of ArcMap 10.3.

1. When the first dialog box comes up, go to My Templates – Blank Map and press
OK.

Figure 3.1: ArcMap start view

2. Choose Customize – Toolbars, and make sure that Standard and Tools are
checked.

Lab Report 1 UAV | 16

 3. On the left side of the screen, you should see your Table of Contents area – right
now it should only say "Layers". If you do not see this separate area, choose
Windows – Table of Contents.
4. Below is the Standard Toolbar:
5. Click on the "plus" sign to add data.

Figure 3.2: ArcMap menu

3.3 Data editing

6. Click on Symbology tab to change the colour ramp.

Figure 3.3: ArcMap layer properties

Lab Report 1 UAV | 17

3.3 Projection System

7. Projection is one of the important element in mapping. Therefore, if the

projection has not been set up, the coordinate system should be chosen.
8. Right-click on Layers in the Table of Contents (this is the data frame) and go to
Properties (or choose View – Properties).
9. Click on the Coordinate System tab.

Figure 3.4: ArcMap data frame properties

10. Choose the type of the coordinate system you want.

11. Click OK when done.

Lab Report 1 UAV | 18

3.3 Map Element

12. When you create a map, you should include the map itself, plus a title, a legend,
a scale, and a north arrow. You should also provide the name of the
cartographer, a date, and acknowledgements about data sources.
13. Click view and select layout view.

Figure 3.5: ArcMap layout view

14. Set up the Page properties. Choose File – Page and Print Setup.

Figure 3.5: Page and print setup in ArcMap

Lab Report 1 UAV | 19

15. The layout tools work on the layout as if you were zooming in and out of the paper
itself, while the same tools on the Tools toolbar work on the data inside the data
frame.

Figure 3.6: Layout toolbar in ArcMap

16. To resize and move the data frame, use the Select Elements pointer to select the
frame first.
17. Use the small grab points on the corners and sides to resize.
18. Place the cursor anywhere over the data frame to move it.
19. Insert North Arrow.

Figure 3.7: North Arrow option in ArcMap

20. Insert Scale Bar.

Lab Report 1 UAV | 20

 Figure 3.8: Scale bar option in ArcMap

21. Insert Title.

Figure 3.9: Insert title in ArcMap

22. Insert Legend. Legend can be modified by double-clicking on it to bring up the legend
properties. For example in the Legend Properties, you can tell it NOT to show the word
“Legend” or you can change that word to something else.

Lab Report 1 UAV | 21

 Figure 3.10: Legend wizard in ArcMap

23. Insert Dynamic Text such as Author’s name, date saved and coordinate system into
the map layout.

Figure 3.11: Insert text in ArcMap

Lab Report 1 UAV | 22

3.4 Export Map

24. When the map are done, it can be exported into other formats such as PDF, png.,
JPEG and etc.

Figure 3.12: Export map

Lab Report 1 UAV | 23

3.5 Output of DEM and Orthophoto map

Figure 3.13: Digital Elevation Model (DEM) map

Figure 3.12: Orthophoto map

Lab Report 1 UAV | 24

 CHAPTER 4 RESULTS

4.1 Introduction

This chapter elaborates output composing of UiTM Arau aerial photo which were applied for Orthophoto map and
DEM map with GCPs specifications list in Appendix which is the final output for this lab report also accuracy assessment
report shown below.

4.2 Root Mean Square Error (RMSE)

GCP OBSERVATION AGISOFT RESIDUAL

Easting Northing Height Easting Northing Height Easting Northing Height
Column1 Column2 Column3 Column4 Column5 Column6 Column7 Column8 Column9 Column10
1 100.280204 6.446 26.459 100.280204 6.446 26.459 0 0 0
2 100.280119 6.447269 28.922 100.280119 6.447269 28.922 0 0 0
3 100.280088 6.449169 32.192 100.280088 6.449169 32.192 0 0 0
4 100.275947 6.445694 18.611 100.275947 6.445694 18.611 0 0 0
5 100.27554 6.446992 17.62 100.27554 6.446992 17.62 0 0 0
6 100.275802 6.448867 23.835 100.275802 6.448867 23.835 0 0 0
7 100.277654 6.447839 35.542 100.277654 6.447839 35.542 0 0 0
8 100.278033 6.446911 34.535 100.278033 6.446911 34.535 0 0 0

Lab Report 1 UAV | 25

APPENDICES

Lab Report 1 UAV | 26