Stereo vision VLSI processor based on a recursive computation algorithm

Abstract

Stereo vision is a well-known method for three-dimensional instrumentation. A major issue of stereo vision is to establish reliable correspondence between images. This paper presents a high-speed corresponding matching algorithm based on recursive SAD (sum of absolute differences) computation. Since the recursive SAD computation can be performed in parallel and have regular data dependencies, the proposed algorithm is well suitable for VLSI implementation. A VLSI processor architecture based on area-time minimization is also discussed to achieve highly-parallel processing and a simple interconnection network among processing elements.