ce DTM 
ed areas 
2.6 m, a 
sidered, 
In both 
3 RMS), 
ATION 
  
  
rategy. 
of five 
d form, 
cept the 
> image 
egies 5 
ts were 
2s (flat, 
ie point 
2d form 
1 of the 
curacy 
M thin- 
ftware 
still in 
d. The 
>. Each 
h pass 
ber of 
d steep 
for thé 1 : 
zy files : uM 
of the | | 
Others Fig. 3. Orthoimage with overlaid contours (with 30 m interval) and tick marks every 400 m. Data are derived Ii 
ley are from the DPW 770 results. Small rectangles show regions of big blunders, and the big rectangle shows 
re not the region without big blunders where a separate accuracy analysis was performed. 
  
  
  
  
ategies i 
ing is | 
ers for 
eletiön Table 1. Statistics of height differences between manual and automatic measurements. 
ke into 
tips of Version Number of Time! required (sec) Maximum Average (m) | RMS (m) 
uld be points compared | / number of match points | absolute (m) 
  
  
  
> steep 
not be 
given 
ching. DPW 770 3997 16.5 0.15 1.08 
them (region without big blunders) 
sitive. 
"ies as VirtuoZo 8594 860 / 467,000 103.5 0.27 4.13 | 
ble to (whole image) 
nation ki 
nd the VirtuoZo 3997 14.6 0.21 0.96 | 
(region without big blunders) 
DPW 770 (whole image) 8728/ 5340 / 462,000 87.0/ 0.28/ 3.33/ 
all points / points with code? > 32 356,000 70.7 0.28 1.38 
  
  
  
  
  
tation 
  
  
  
  
  
  
  
1 The time refers to matching only (for VirtuoZo it also includes the generation of the image pyramid). For DPW 770 the 
elapsed time is given, for VirtuoZo the CPU time. For VirtuoZo the times for epipolar resampling and DTM grid interpo- 
os lation were 340 sec and 110 sec respectively. Latter included 480,000 output points and was used only for timing pur- I, 
ig poses. | 
sus 2 Quality code provided by DPW. Points with code > 32 are considered reliable. 1 
  
107 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996