912
December 2016
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
3, is a comparison of the AT
results from the UAS mount-
ed cameras with the results
from the UltraCam Falcon
and how they relate to AS-
PRS accuracy standards.
It is evident that the Sony
performs poorly in Z at the
lower altitudes with limited
control while the large for-
mat sensor performs best at
all altitudes. However, Fig-
ure 4 highlights that all of
the UAS sensors generate
7cm or better vertical clas-
sification in the ASPRS ac-
curacy specification, making
the data acceptable for creat-
ing 1-foot contours, with the
Table 2: Automatic camera calibration results for UAS datasets from Pix4Dmapper Pro.
Camera Altitude GCPs Camera
Optimization
Focal Shift (mm) PP x shift (mm) PP y shift (mm) Average 3D points per
image
Canon 200
2
4.86%
0.973
0.012
0.121
5518
Canon 200
3
4.85%
0.971
0.013
0.122
5551
Canon 200
5
4.83%
0.968
0.012
0.123
5531
Canon 400
2
4.82%
0.964
0.005
0.139
10620
Canon 400
3
4.81%
0.960
0.006
0.141
10588
Canon 400
5
4.78%
0.957
0.005
0.141
10626
Sony
200
2
3.48%
0.626
-0.019
-0.066
9786
Sony
200
3
3.21%
0.546
-0.007
-0.073
9786
Sony
200
5
3.29%
0.579
-0.011
-0.068
9784
Sony
400
2
0.29%
-0.104
-0.098
0.029
12214
Sony
400
3
0.38%
-0.136
-0.071
0.400
13082
Sony
400
5
0.06%
-0.022
-0.081
-0.001
12211
Table 3: Comparison of RMSE of AT of UAS Datasets from Pix4D and UCFp from Inpho.
Flight
GSD
DX in cm
DY in cm
DZ in cm
(cm)
2GCP
RMSE
3GCP
RMSE
5GCP
RMSE
2GCP
RMSE
3GCP
RMSE
5GCP
RMSE
2GCP
RMSE
3GCP
RMSE
5GCP
RMSE
Sony 200ft
0.7
5.141
2.722
2.695
3.199
2.031
2.302
37.619 20.686 3.939
Sony 400ft
1.5
2.555
1.866
1.398
2.848
1.449
1.532
19.233 21.539 6.116
Canon 200ft
1.4
4.875
3.211
2.739
4.726
4.456
2.916
21.369 20.770 6.995
Canon 400ft
2.8
3.021
1.825
1.283
1.323
0.857
1.351
10.011 12.296 5.314
UCFp 1387ft
2.0
1.900
1.668
1.000
1.700
1.400
0.700
4.100
2.700
2.900
UCFp 2560ft
4.0
2.200
1.200
1.200
2.000
2.100
1.300
3.800
3.700
3.500
Figure 4: RMSE (cm) expressed in ASPRS accuracy standards.
