The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
When we used optimum sampling we had GCPs in all places 
that there was high slope change. 3Ddistribution of control 
points were shown in figure 3. Number of control points are 56 
and number of check points are 40. Model accuracies are shown 
in table 1. Regards to this table we can conclude: 
Figure 3. Ground control points distribution in optimum 
sampling 
• When denominator of rational functions aren’t equal 
P2 P4 
( ) the accuracies are more than when they 
are equal. 
• Higher degree rational functions have more accuracy 
than low degrees. 
• The accuracy of direct rational functions is more than 
the inverse one. But third order inverse rational 
functions have high accuracy too. 
When are using oversampling there are GCPs in all places that 
there are high change slopes and in places between them. 
Number of control points are 88 and number of check points are 
8. Distribution of control points were shown in figure 4. 
Results of table 2 show that when we used oversampling the 
accuracies of the models are a little more from when we used 
optimum sampling but these differences aren’t high. 
Figure 4. Ground control points distribution in over sampling 
N.O.GCPs 
N.O.CKPs 
RMSECKPXYZ 
(cm) 
RMSECNPXYZ 
(cm) 
1 orderRFM(P2^P4) 
56 
40 
5.946 
6.178 
2orderRFM(P2^P4) 
56 
40 
1.261 
0.617 
3orderRFM(P2#>4) 
56 
40 
1.349e-4 
2.797e-5 
2orderRFM(P2=P4) 
56 
40 
2.354 
1.871 
3 orderRFM(P2=P4) 
56 
40 
0.503 
0.308 
Inverse lorderRFM 
(P2#>4) 
56 
40 
487.040 
471.593 
Inverse 2orderRFM 
(P2#>4) 
56 
40 
1.548 
1.343 
Inverse 3orderRFM 
(P2#>4) 
56 
40 
0.0466 
0.010 
Table 1. Results obtained from simulated data in optimum sampling 
N.O.GCPs 
N.O.CKPs 
RMSECKPXYZ 
(cm) 
RMSECNPXYZ 
(cm) 
lorderRFM(P2#>4) 
88 
8 
6.987 
5.862 
2orderRFM(P2^P4) 
88 
8 
0.732 
0.616 
3orderRFM(P2^P4) 
88 
8 
7.04e-5 
3.57e-5 
2orderRFM(P2=P4) 
88 
8 
2.040 
1.891 
3orderRFM(P2=P4) 
88 
8 
0.609 
0.358 
Inverse lorderRFM 
(P2#>4) 
88 
8 
361.667 
453.722 
Inverse 2orderRFM 
(P2#>4) 
88 
8 
1.890 
1.342 
Inverse 3orderRFM 
(P2#P4) 
88 
8 
0.025 
0.014 
Table 2. Results obtained from simulated data in over sampling