The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B3b. Beijing 2008 
637 
In order to obtain the standard volume of gravel mound, the 
volume measurement using Digital Still Camera (NIKON D80 f 
= 28mm, lOMPixel) with PI-3000 was performed. Also 
comparison of measuring time was also performed. 
Figure 13 shows the result of volume measurement in this 
experiment. Table 3 is a comparison result of volume 
measurement from Digital Video Camera and Digital Still 
Camera with PI-3000. 
Figure 10. Result of Exterior Orientation Procedure 
for Model Gravel Mound 
Figure 11. Result of TIN Modelling from 
the 3D Coordinate of Natural Feature 
Table 2 shows the result of volume measurement. The volume 
error of this experiment was 1.97% for standard volume of 
gravel mound. This accuracy is enough if considering that error 
of volume is permitted to 10% in mining field. 
Standard Volume of 
Gravel 
7630[cm3] 
Calculated Volume of 
Gravel 
7480[cm3] 
Error 
1.97[%] 
Table 2. Result of Volume Measurement for 
Model Gravel Mound 
3.2.2 Application for Real Gravel Mound: The application 
experiment for real gravel mound was performed as the accu 
racy assessment for real field measurement. Figure 12 shows 
the test site of this experiment. The ground control point of this 
field was measured by Total Station TOPCON GPT-9000A. 
Figure 12. Test site of Real Gravel Mound 
a: Result of Digital Still Camera (standard volume) 
b: Result of Digital Video Camera 
Figure 13. Result of Volume Measurement 
for Real Gravel Mound 
Standard Volume of 
Gravel 
8.69[m3] 
Calculated Volume of 
Gravel 
8.82[m3] 
Error 
1.50% 
Table 3. Result of Volume Measurement for 
Real Gravel Mound 
The volume error of Digital Video Camera result was 1.5% 
from standard volume which obtained from Digital Still Camera 
with PI-3000. In addition, the measuring time of Digital Video 
Camera was about 1/10 of Digital Camera measurement. If 
considering these results, volume measurement using Digital 
Video Camera is useful for the mining field. 
4. CONCLUSITON 
The authors describe the application of robust regression for 
exterior orientation for video image sequences in this paper. 
The robust tracking and robust bundle adjustment process 
developed in this study use the LMedS method for detection of 
outlier in corresponding points. In addition, in order to increase 
the robustness of bundle adjustment process and for solution of 
the limitation for camera motion, authors used LMedS in the