James Olaleye 
T LT 51 Ti5 
= (9) 
but 7, and T, are direction vectors, so that (9) may be written as: 
1 m- TT 5 1 T, b 
— T = (10) 
=r | -T,b 
The determinant of the normal equation (10) is given by 
| FT — y 
| pE| IL. ^is 1-5. (11) 
igure 5 shouli II. 1 
ver, in reality, E 
“mathematica 
. vector P is of 
Fig. 4). Also, PED : 
which is non- Zero unless the two rays are parallel. Then the solution to (10) is 
om the vector 
$, I| rr TIT. : 
= — DL = m (12) 
$ 2 D 1,7 1 s 
ngth of vector 
d values fors The values of s;, s» obtained from 12 provide the minimum length of the want of intersection of the conjugate rays. In 
practice, the two rays are assumed equally weighted so that we may locate the actual point of their intersection as being 
midway along vector p. Hence, the vector corresponding to this point of intersection may be written as: 
1 
T -T2sT 4—P 
- 
or by substituting (7) for p and simplifying, we obtain 
b sTi tel: 
T= BUE (13) 
Applying the ARDOVS operational rules 1&2 (and noting that the axes of the image R- space are now relabelled for 
an object C-space), the object space location represented by the resulting vector T may be computed by the following 
vector product equation (See Olaleye 1992): 
R$ 
P=P+s RT (14) 
R'T 
(8) 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000.