1155 
ma 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
rational functions are nearly 5.5cm for optimum sampling, 5cm for over sampling and 15cm for under sampling. 
t I 
îf# 
Y at- 
/++>>////+ 
+ T + T + + + + + 
fjßUU 
4 cm 
4 cm 
20 cm 
(a) (b) (c) 
Figure 11. Planimetric error vectors of first order rational functions in a) optimum sampling b) over sampling c) under sampling 
(a) 
(c) 
Figure 12. Height error vectors of first order rational functions in a) optimum sampling b) over sampling c) under sampling 
3.6 Results of space data 
Because of there wasn’t enough control points, we have done 
the experiments for only under sampling. The results were 
shown in table 7. Number of points are 30 GCPs and 14 check 
points. Regards to table 7 we can conclude: 
• Accuracy of first and second order rational functions 
are more than third orders. 
P 2 * P 4 
Accuracies in is much more than 
P2 = P 
Direct rational functions are 
inverse ones 
jjüjâj 
than 
N.O.GCPs 
N.O.CKPs 
RMSECKPXYZ RMSECNPXYZ 
(m)(m) 
lorderRFM(P2#>4) 
39 
14 
14.349 
10.642 
2orderRFM(P2^P4) 
39 
14 
26.717 
12.933 
3orderRFM(P2^P4) 
39 
14 
59.088 
35.115 
2orderRFM(P2=P4) 
39 
14 
68.162 
24.941 
3orderRFM(P2=P4) 
39 
14 
136.356 
61.126 
Inverse lorderRFM 
(P2#>4) 
39 
14 
16.698 
12.543 
Inverse 2orderRFM 
(P2f?4) 
39 
14 
34.412 
235.479 
Inverse 3orderRFM 
(P2^P4) 
39 
14 
125.155 
249.016 
Table 6. Results obtained from space data in under sampling 
Planimetric and height error vectors were shown in figures 13 
and 14. For seeing error vectors more obvious, some of these 
vector were eliminated.