Table 2 Computing times for determining 
a conjugate point by HITAC M-280H 
Case 1 Conventional Method 
Case 1-1 B/H=0.5 18.3 m.sec. 
Case 1-2 B/H=1.0 19.2 m.sec. 
Case 2 Triplet Method 
Case 2-1 Independent 69.0 m.sec.* 
Case 2-2 Simultaneous 21.8 m.sec. 
* Error check path was not well designed. 
Figure 3 shows a given contour image of digital terrain 
model of the test area. The contour interval is 25 meters. 
(original is color) 
Figure 4 (a),(b),(c) and (d) shows computer generated images 
for the cases of 1-1, 1-2, 2-1 and 2-2 respectively. 
Figure 5 (a),(b),(c) and (d) shows the error distribution 
images for the four corresponding cases. 
CONCLUSIONS 
(1) The triplet method with three rays showed the better 
accuracy and reliability than the conventional stereo 
method with two rays, though the computing time was more. 
(2) The simultaneous check method was more reliable than the 
independent method because maximum error was much less. 
The computing time for the simultaneous check method 
increased only 10-20 percent as compared with the 
conventional method. 
(3) The triplet method would be able to solve the error 
detection problem which has been involved in the 
conventional image correlation with two rays. 
References 
(1) Shunji Murai and Ryosuke Shibasaki; Geometric Correction 
of Linear Array Sensor Data; Com.I Symposium, Australia, 
1982 
(2) Shunji Murai and Ryosuke Shibasaki; A Study on Three 
Dimensional Measurement with Use of Stereo Linear Array 
Sensors; Com.IV, 15th ISPRS Congress, Brazil, 1984 
230