Full text: XVIIth ISPRS Congress (Part B4)

3.1 Data used 3.4 Cross correlation between image 
; ; chips of stereo pair 
An experiment with the following P 
SPOT/HRV stereo pair of images Was It 
is X 
conduct ed. expected that 
planimetrically corrected Stereo pair 
shows a good cross correlation. Widely 
distributed 8 image chips consists of 32 
3.2 GCP location X 32 pixels were extracted from the 
: Y stereo pair of SPOT/HRV. Cross 
The following three GCPs with correlations between the image chips of 
almost same height (around 600 m above the stereo pair are shown in figure 2 
sea level) were extracted from About 0 to 98% of cross correlation was 
topological map. A portion of SPOT/HRV increased after the planimetrie 
image and topoglaphic locations of the correction. 
GCPs are shown in figure 1. Study area 
is stuated in a montaintanuous region of 3. 5. DEM estimation accuracy 
Kiso-Komagatake in the middle of Honshu 
in Japan. Height in this area ranges 9 GCPs were used for evaluati 
ion of 
from about 600 to 2500 m above sea : : : 
level. an effect of the planimetric correction. 
Table 3 shows DEM derived from 
topological map and estimated DEM with 
3.3 Planimetric correction ; 3 : : 
and without planimetric correction. 
Based on the aforementioned  Affine 
transformation with the coefficients 
derived from the GCPs, planimetric 
distortion was removed. Assuming the 
following relationship between line 
pixel coordinates for before and after 
the planimetric correction, (L,P) and 
(La Pa), planimetric distortion was 
evaluated with the selected 9 GCPs for 
evaluation. 
RMS error of estimated DEM for the 
proposed method was reduced by one 14th 
compared to without planimetric 
correction. 
4. Concluding remarks 
The proposed method is simple and 
effective for improvement of DEM 
estimation accuracy. Just before DEM 
estimation, Affine transformation with 6 
  
La=al+a2L+a3P+a7L”2 +a8LP +a9P 2 
= S ^ 4) coefficients derived from 3 GCPs with 
p ALS GRFE LOL SHE LIERAGLZPCH ( almost same hieght is applied to the 
a ala stereo pair of images. In this case 
^ 5 292 
+a16L 3+al7L ZP+al8LP with SPOT/HRV, the experimental results 
show that planimetric fidelity is 
remarkably improved, cross correlation 
between image chips of stereo pair is 
Table 2 shows the coefficients of 
the above equation for before and after 
the planimetric correction. 
Table 1 Data used 
  
  
Acquisition Pointing GRS Process Remarks 
date angle level 
Dec. 8’ 86 13. 6deg. 327-278 1b Kiso-Komagatake in 
Dec. 12' 86 20.2deg. 327-218 1b Japan 
  
Table 2 A comparison of the coefficients of the Affine 
transformation for system corrected and planimetrically 
corrected SPOT/HRV data 
  
  
System corrected data Planimetrically corrected data 
al —2.16 E 6 6.45 E 5 
a2 2.04 E 3 =b. 31 BE 2 
a3 5.44 E 2 —-2:.26 E 2 
a4 4.96 E 6 -—4.29 E 5 
ab —4.64 E 3 3.01 E 2 
a6 z1.26 E 3 1.96 E 2 
a7 -6.51 E 1 1.48 E-1 
a8 —-3.11 E-1 1.16 E-1 
a9 =6. 11 E-2 3.10 E-2 
a10 1.48 E 0 =5. T7 E-2 
all 7.21 E-1 —-1.09 E-1 
a12 1.42 E-1 —2.32 E-2 
a13 7.63. E-5 —1. 70 E-5 
al4 3.17 E-5 —7,59 E-6 
a15 2.08 E-5 1.36 E-5 
a16 —1.78 E-4 3. 89 E-6 
al’ =7.33 E-5 9. 85 E—-6 
a18 -6.22 E-5 1.03 E-5 
  
858 
 
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.