usually not a three-dimensional picture. This is due, 

 in the case of large targets, high resolution, and near- 

 bottom scanning, to different rates of parallax 

 change between shadow and target signals [3]. In 

 the case of high-flying stereo-sonar scanning, where 

 the essential imagery is contained in target returns, 

 the three-dimensional illusion is created only if suit- 

 able contrast and resolution are present [ 1 ] . 



Figure 1 shows the geometry of the stereo-sonar 

 system. The equations which take the echo ranges, 

 R, and R,, to the relative target elevation are the 

 following: 



B = (h^ + H^) 



2,1/2 



d = tan"'(h/H) 

 ? = |(R2 - Ri)/2b} - (B/2) 



,„2 f-2,1/2 



V = (Ri - I ) 



y^ = -(I + B)sin0 + tj cos( 



X, = (^ + B)cos0 + 77 sin ( 



(1) 

 (2) 

 (3) 

 (4) 

 (5) 

 (6) 



errors (relative to shore or a fixed seafloor point) 

 below several yards. 



Figure 2 is a schematic diagram of a contour 

 plotting system utilizing the principles of the double- 

 projection, direct-viewing stereoplotter. In the stereo 

 plotting of seafloor contours using side-scan data, the 

 operator would begin somewhere on the crosstrack 

 line passing through the first along-track point of the 

 sonar-stereo scan. Labeling each recognizable target 

 point on this first crosstrack line, the operator would 

 move the viewing screen vertically and horizontally 

 until the no. 1 target point appeared to lie in the 

 plane of and at the center of the screen. His eye-brain 

 system, with the aid of conventional stereo goggles 

 (for example, red-green), would make this possible. 

 Mechanical linkages between viewing screen and the 

 computer would cause the sonar ranges, Rj and R,, 

 to be fed into the computer in feet. (Of course, with 

 optical stereo images in front of the two projectors, 

 no computer is necessary; and the operator simply 

 makes a mark on the map surface directly below the 

 center of the viewing screen.) With stereo-sonar 

 images in front of the projectors, the computer would 

 take Rj and R, and, using Equations 1 through 6, 

 calculate the target position. 



where B = |B| 



B = fish-pair vector 



6 = direction of B, relative to 

 horizontal (positive for 

 fish no. 1 above fish no. 2) 



h = vertical component of B 



H = horizontal component of B 



y^ = vertical distance of target 

 below fish no. 2 (positive 

 for targets below fish no. 2) 



The sixth equation yields, x^, the horizontal distance 

 of the target from fish no. 2. However, unless a 

 shore-based fish navigation system is used, this com- 

 putation is not important. Presumably, it is much 

 easier to keep systematic fish-depth errors below a 

 yard or so than it is to keep the systematic horizontal 



b „ _ Change in parallax (feet) 



Change in range (feet) 



Figure 1. Geometry of stereo-sonar system for 

 computation of target position. 



The problem with the procedure described 

 above is that the operator would have to go through 

 all of the target points on each crosstrack line, and 

 then instruct the computer to arrange the 



