The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008
h Value
Matrix
Singularity
(Y/N)
Iteration
Number
CkP Point
Position
Error
(pixel)
h Value
Matrix
Singularity
(Y/N)
Iteration
Number
CkP Point
Position Error
(pixel)
<0.004
Y
Not
Convergence
N
0.05
N
5
3.2425
0.004
N
25
13.266
0.06
N
6
3.256
0.008
N
14
4.159
0.1
N
10
3.272
0.009
N
3
3.305
0.2
N
20
3.626
0.01
N
3
3.293
0.3
N
20
4.806
0.03
N
4
3.2498
0.4
N
27
7.030
0.04
N
4
3.2476
0.8
N
30
14.699
Table 1 Positioning Accuracy with different h values
3.2 Calculation of RPC
The test imagery is QuickBird imagery of Shanghai area, of
which the collection time is February 15, 2004, the coordinates
of lower left comer ( 31.14796 ° , 121.424732 ° ) ,the
coordinate of upper right corer (31.299428° , 121.61359
) . The image collection azimuth is 355.3 degree; the
elevation angle is 68.2 degree. 50 GCP distributed evenly were
Figure 2 GCP Distribution
selected from the overall 139 surveyed points to calculate the
initial RPC, the distribution of which was shown in figure 2. 26
points were selected randomly from the rest surveyed points as
CkP for accuracy analysis, of which the distribution was as
figure 3.
Figure 3 CKP Distribution
Taking h = 0.05 into formula (10) to get 80 RPC parameters, we can get the image coordinated by adding the parameters and
ground points (P,L, H) of the 26 CkP into formula (1). The differences between calculated image coordinates and corresponding
image points of CkP were shown in table 2:
Results and Conclusion of tests:
Error Type
Accuracy
(pixel)
Error Type
Accuracy
(pixel)
Absolute
Error
Max Row Error
4.092
Relative Error
Row Error
1.059
Row Mean Square
Deviation
1.864
Column Error
2.055
Max Column Error
7.060
Point Position Error
2.312
Column Mean Square
Deviation
2.653
Point Position Error
3.243
Table 2 The positioning accuracy of TD method
1290