Pix4D Introduction Guide
Pix4D Introduction Guide
October 2013
Plan
Introduction
Stereoscopy/calibration theory
Processing steps
Step1: Calibration + exterior orientation
Step2: Point densification
Step3: DSM and orthomosaic
Introduction
Pix4D converts thousands of images into
3D Point Cloud
3D Digital Surface Model
Orthomosaic
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + exterior orientation
Step2: point densification
Step3: DSM and orthomosaic
From 2D to 3D
From 3D to 2D
From 2D to 3D
One camera
Far appears smaller
than near on image
Two images
Triangulate to get
sense of depth
2D Keypoint
2D Keypoint
Camera model
(focal length,
principal point,
lense distortion)
External orientation
(camera position &
rotation)
Camera model
(focal length,
principal point,
lense distortion)
above 1 meter
below 1 centimeter
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + exterior orientation
Step2: point densification
Step3: DSM and orthomosaic
Step 1: initial
processing
Images
=> calibrate cameras +
exterior orientation
Step 2: point
densification
Calibrated cameras
=> point clouds
Point clouds
=> DSM and orthomosaic
n
n
s
n
tim
ing
ing
tio
CP
r
ing
tio
p
h
tio
ge
a
e
d
c
G
o
c
a
t
a
c
/
t
l
n
i
a
r
l
i
f
if
e
tr
ne
PS
ma
ble
nt
c im
ns
ex
e
od
i
i
G
s
c
e
t
g
o
i
t
a
m
d
n
a
P
in
os
oin
ra
tio
os
int
SM
po
p
m
e
a
o
y
y
D
m
o
c
m
P
Ke
Ke
rth
tho
olo
Ca
r
O
e
O
G
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + external orientation
Step2: point densification
Step3: DSM and orthomosaic
e
int
o
p
ing
n
tio
c
ra
xt
y
Ke
ts
oin
h
atc
yp
Ke
Average of 6'000
matches per pair
of images
CP
pti
lo
od
m
a
me
a
C
600'000
matches =>
150'000 3D
Points for 100
images project
S/G
o
Ge
G
on
i
t
a
loc
Average of GPS
OR
Min 3 GCP
>1000 matches
75% overlap
enough matches
<100 matches
20% overlap
not enough matches
Extract Keypoints
Good image:
rocks,bushes,dirt
buildings, urban
more than 10MP
extract > 10'000 keypoints
Difficult images:
Match keypoints
Easy to match
Hard to match
Impossible to match
Flight plan
Ideal flight plan
low overlap
multiple images at same location
corridor mapping
high difference in altitude (>2xGSD)
Model Georeference
Project size, orientation and position unknown
GPS tags
approximate position,orientation,size
good synchronisation, high number of GPS tags => better approximation
~10 meter shift in random direction
GCP
Geolocation: accuracy
GCPs
well distributed in dataset, visible on the images (GSD)
5-8 usually enough for 1000 images
Quality report shows reprojection error
Good calibration + GCP: error at most 1-2 times
the GSD
Verification points
Error is minimized at GCPs: not optimal to assess global
accuracy
Verification points are not used during optimization
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + exterior orientation
Step2: point densification
Step3: DSM and orthomosaic
Point densification
Step 2: point densification
Calibrated cameras + exterior
orientation => cloud of points
on
ti
a
c
i
if
ns
td
oin
ng
ri
lte
i
f
t
in
Po
Example:
8 millions 3D
points
2 millions filtered
and triangulated
3D points
Point densification
Visual content
3D points computed where there is visual content
improves quality
2.5D only
Smoothing
Global Accuracy
Calibration+Georeference+Point Cloud
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + exterior orientation
Step2: point densification
Step3: DSM and orthomosaic
ion
t
era
M
DS
ge
ing
ag
m
i
ic
sa
Or
mo
o
h
t
Or
th
ic
sa
o
om
nd
ble
DSM generation
DSM accuracy
Interpolation at place with no visual information
Global Accuracy
Calibration+Georeference+Point
Cloud+Interpolation
Photo Stitching VS
Orthorectification
Orthorectification
UAV image
with
perspective
Orthorectify image
1) calibrate camera
2) project image on 3D model
3) generate ortho image by looking
from above
Computed 3D model
Undistorted orthoimage
Same as satellite
Roof correct size
Orthorectification artefacts
3D model not always perfect
Edges of buildings
Small details
trees (branches smaller than pixels)
lamppost, fences
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + exterior orientation
Step2: point densification
Step3: DSM and orthomosaic
Generated files
Quality report
Camera position & parameters
Raw point cloud
(ply,las)
Filtered point cloud (ply/las)
regular point cloud
DSM in tiff
individual ortho and
planar
orthomosaic in
tif, KML tiles
Step 1: initial
processing
Step 2: point
densification
n
n
s
n
tim
ing
ing
tio
CP
r
ing
tio
p
h
tio
ge
a
e
d
c
G
o
c
a
t
a
c
/
t
l
n
i
a
r
l
i
f
if
e
tr
ne
PS
ma
ble
nt
c im
ns
ex
e
od
i
i
G
s
c
e
t
g
o
i
t
a
m
d
n
a
P
in
os
oin
ra
tio
os
int
SM
po
p
m
e
a
o
y
y
D
m
o
c
m
P
Ke
Ke
rth
tho
olo
Ca
r
O
e
O
G
Generated files
Pix4uav outputs
Format
Possible use
Examples of compatible
software
Raster orthomosaic
geoTIFF (.tiff)
KML tiles (.png/.kml)
Area overview
Digitize buildings
Annotate areas
Overlay in GIS package
Analyze spectral bands
ArcGIS
Global Mapper
QuantumGIS
AutoCAD
Google Earth
Undistorted images
TIFF (.tiff)
Stereo Viewing
LPS
Individual ortho/planar
geoTIFF (.tiff)
Seamline editing
OrthoVista
3D
point cloud
Visualization
Surface Editing
DSM generation
DTM generation
ArcGIS
Global Mapper
AutoCAD
Quick Terrain Reader
3D Reshaper
Trimble RealWorks Viewer
geoTIFF (.tiff)
Analyze surface
Measure volumes
Generate contour lines
Draw breaklines
ArcGIS
Global Mapper
QuantumGIS
Quick Terrain Reader
Wavefront (.obj)
AutoCAD
Bentley Pointools View
CC Viewer
3D Reshaper
Editing
Analyzing
Visualization
3D point cloud
Raster DSM
3D Mesh
best (VRMesh)
possible
possible
best (ArcGIS/Global
Mapper)
good (ArcGIS/Global
Mapper)
Plan
Intro
Stereoscopy/calibration theory
Processing steps
Step1: calibration + exterior orientation
Step2: point densification
Step3: DSM and orthomosaic
3)
geotags and camera model selection
- automatic geotag, or using text file
- automatic camera model selection,
or using database
- at least three valid geotag required
2) Start button
Steps in red if not yet
processed. Checkbox
allows run of individual
step.
127 images
12 MP Canon Ixus
quadcore laptop
Initial Processing:
Option: rapid,full
Point densification:
Option: density
Orthomosaic/DSM:
Option: resolution
optimal: 28 min
Choose high
tolerance if forest
2) List of images
use top icons to sort
closest images
use right click to remove
measurements
4) Import/export
clicked points without
coordinates
5) If initial processing is
already done, bundle
with GCP allows for
faster re-processing
5) Blend mosaic
2) Select ortho/planar
to improve visual
aspect
3) Choose most
appropriate image in
the list (use keyboard
up/down arrow).
Good to remove
blurry images/ moving
cars
1) Select a cell by
clicking
Thank you