04 
MS 
Y) 
"» 
3, 
  
  
HA] 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Contessa Entellina - QuickBird 
10 coeff. 
RMS GCPs RMS CPs 
N.GCP [XY RMS | X RMS Y RMS N.CP [XYRMS | XRMS | Y RMS 
19 1.90 0.98 2.65 15 2.88 3.44 5.69 
25 1.93 1.08 3.41 15 3.08 3.43 5.55 
30 1.90 1.40 3.20 15 3.12 3.97 5.34 
45 2.28 2.08 3.14 15 2.01 2.26 4.31 
55 2.36 2.17 3.45 15 1.96 2.25 4.14 
20 coeff. 
RMS GCPs RMS CPs 
N. GCP | XY RMS X RMS Y RMS N. CP XY RMS X RMS Y RMS 
39 1.06 0.85 1.76 15 2.17 2.17 8.40 
45 1.03 0.93 1.74 15 2.02 3.18 7.80 
55 1.88 1.52 2.75 15 2.67 4.81 10.91 
  
  
  
  
  
  
  
  
Table 5. GCPs and CPs RMS relative to the QuickBird 
image of Contessa Entellina 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Scopello - QuickBird 
10 coeff. 
+ 
RMS GCPs RMS CPs 
N.GCP [xv RMsS | XRMS | Y RMS N.CP |XYRMS | XRMS | YRMS 
19 1.84 2.45 0.55 15 3.80 5.81 1.07 
25 1.82 2.55 0.86 15 3.59 5.71 1.10 
30 1.65 2.47 0.91 15 4.26 6.14 1.34 
39 1.75 2.82 0.90 15 5.07 7.42 0.90 
45 2.09 3.40 0.93 15 4.63 6.89 0.82 
20 coeff. 
RMS GCPs RMS CPs 
N.GCP [XY RMS | XRMS | Y RMS N.CP |XYRMS | XRMS | YRMS 
39 0.48 0.78 0.50 15 2.28 3.79 1.22 
45 0.56 0.93 0.54 15 2.08 3.49 1.22 
  
  
  
  
  
  
  
  
Table 6. GCPs and CPs RMS relative to the QuickBird 
image of Scopello 
Caselle Torinese - QuickBird 
Toutin Model 
RMS GCPs 
  
Table 7. GCPs and CPs RMS relative to the QuickBird Basic 
image of Caselle Torinese 
All these tables show that in all the examined cases acceptable 
results were achieved, and the geometric accuracy of 
orthorettified images is quite similar to map tolerance of a 
traditional map scale of 1:5000. 
Regarding the QuickBird image of Palermo, comparable results 
were not obtained: this behavior is probably due to distortions 
affecting image and to a pre-processing phase in which the 
product was irreparably damaged. In this case, a geometrical 
accuracy higher than the circular error of the image, was not 
obtained. 
The possibility to obtain a product compatible with the above- 
mentioned scale was tested on a QuickBird Basic image centred 
on Torino *Caselle* Airport. This image was submitted only to 
the radiometric corrections, so that it was possible to use the 
Toutin model. By using Toutin model an orthorectification with 
support of 9 GCPs was carried out; this orthorectification has 
led to excellent results, totally consistent with a map scale of 1: 
5000 (Tab. 7). 
7. ANALYSIS OF RESULTS 
A first order of control was carried out by the visual overlap 
between the ortophotos and the Official Maps. 
The figure 8, regarding a subset of Palermo IKONOS image, 
shows an example of this control. 
597 
  
Figure 8. Overlay between the Municipal Map of Palermo and 
the orthophoto obtained by the IKONOS Palermo image 
The last control regards the comparison between the 
coordinates of some points extracted by the orthophoto and the 
coordinates of the same points extracted by a reference map. In 
this regards 30 couples of points for the IKONOS Geo images 
and 20 couples of points for the QuickBird Standard images 
were considered. 
The figures 9, 10, 11 and 12 show some examples with charts 
describing the frequency histogram of the couples of points 
observed. 
  
Ikonos Contessa 10 coefficients 
  
= 
> 
o 
= 
o 
3 
y 
o 
rz 
o 
> 
= 
= 
© 
m 
ce C c ies c c c e c c 
e. w- o5 (09,5 G! 8 6.5 
ce ce -— vr CN Cx €v2 e = <I 
Class partition (0.50 m) 
  
  
  
Figure 9. Frequency histogram relative to IKONOS 
image of Contessa Entellina 
  
lkonos Palermo 10 coefficients 
  
eo 
a 
> 
o 
£z 
o 
= 
U 
o 
du 
ue 
o 
> 
= 
© 
2 
a 
Q OQ. Q 92 9 Q Q' . cuc e 2. 5 
S 0D oO 0 O n 0.10.0 0 EE 5 OG 
CY ALAIN Oo mios. 50 D dd o 
= += = = = 
Class partition (0.50 m) 
  
  
  
Figure 10. Frequency histogram relative to IKONOS 
image of Palermo