irt B3. Istanbul 2004 004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
round control pons From this table we find the more the GCPs in same subsection 
1 parallel flight strips Anchor GCP/ Pass Point length, the more accurate the results. 
agery were obtained 
‘igure 3). More than 
el strips are ensured. 
ts for different block 
d in triangulation and 
GCP/Pass point 3) Different GCP configuration 
We tested several cases of GCP configurations in different 
subsection length and GCP number, similar conclusion to frame 
aero images. In general, GCP on corner of blocks or strips can 
easily produce accurate results. 
x PS Rar Pam ar X Tones à 3.2 Multiple Strips 
2nd logical stri 5 ; > E ; : 
Ist logical strip e P ord logical strip Same test measures as single strip were conducted for the three 
parallel strips in these strategies: 
ntinues stereo image 
, railways, rivers and ; i ; LB : 
EUN s IG Figure 4. eee CORRER of continues € Different section number 
CER subsection approach. a. 
acy these single strips PP € Different GCP number 
  
  
rols, we tested these ak oe = : 
® Different GCP configuration 
  
Total 66 GCPs, 1189 pass points and tie points were measured in 
semi-auto matching method. Table 3 and 4 list the obtained 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Section RMS X(m) | RMS_Y (m) RMS Z results in different subsection length and different GCP number. 
Length (m) 
ë : F .183 : 
dle strip of the three 20000 0:077 3g Gs Section | RMS X(m) | RMS_Y (m) | RMS_Z 
results based on these 30000 0.08 0.136 0.178 Length (m) 
78 pass points were 
xd over 10 km length 40000 0.099 0.18 0.183 20000 0.066 0.083 0.179 
& 30000 
50000 0.105 0.182 0.18 3 0.076 0.099 0.193 
> C 40000 
js ran 100000 0.182 0.217 0.219 0.082 0.12 0.179 
observed data errors 
> W : : 2 50000 
1ce satisfactory results ha ga Qn 0.213 0.079 0.13 0.172 
distortions cased by 100000 0.158 0.189 0.243 
roblem we designed a ues : A : ; 
wing characters: Table 1. Statistics associated with different section number. Whole 0.224 0.205 0.254 
he complex distortion 
image strips Into a From this table we find the shorter the subsection, the more Table 3. Statistics associated with different section number. 
es (Figure 4). Such a accurate the results. 
to implement. The 
s 2) Different G ; 
e can then be me ded ) Hrn OCT number From this table we find the shorter the subsection, the more 
a high accuracy level. To investigate the influence of GCP number to the final results accurate the results. 
we tested several cases for the single strip in different section 
l, the continuity of the 
number. Table 2 lists statistics associated with different GCP 
d using the concept of 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
other words, the whole number for 30000 lines length subcetion. GCP RMS X(m) | RMS_Y(m) | RMS Z 
d from the pieces, will Number (m) 
he framework of the RE ; 35 
GCP RMS X(m) | RMS Y(m) | RMSZ 2 GCP A 0.123 0.134 0.224 
Number = 3 (m) Sep H 0.114 0.145 0.403 | 
a GP 0.106 0.295 0.287 n 
TT 2 GCP A 0.209 0.335 0.1806 3 GCP A 0.198 0.241 0.223 i 
2 CP B 0.154 0.41 0.191 3 GCP B 0.115 0.166 0.252 j 
3 GCP C 0.132 0.295 0.241 4 GCP 0.103 0.145 0.398 | 
3 GCP A 0.158 0.339 0.187 5 GCP 0.107 0.155 0.264 n 
3S GCP-B 0.128 0.359 0.171 6 GCP 0.107 0.14 0.309 | 
4 GCP 0.113 0.304 0.239 8 GCP 0.1 0.162 0.249 | 
S GCP 0.118 0,338 . 0.217 ]i GCP 0.11 0.141 0.252 i 
6_GCP 0.127 0.302 0.202 12 GCP 0.108 0.129 0.221 | 
8 GCP 0.107 0.261 0.195 All GCP 0.076 0.099 0.193 | 
1 GCP 0.106 0.192 0.163 | 
E XE iE "a = Table 4. Statistics associated with different GCP number. | 
From this table we find the more the GCPs in same subsection N 
Table 2. Statistics associated with different GCP number. length, the more accurate the results. !