<title>Robust three-dimensional vision system by adaptive stereo matching and optical BPEJTC</title>

Abstract

The essential problem in stereo vision system is to search for correspondence points between two different object images projected from the left and right cameras to detect the stereo disparity. So far, several methods were proposed to handle this correspondence problem, but most of them are complicated, need a long calculation time and show the low accuracy of the estimation value because of fixed windowing, etc. The proposed algorithm, firstly, extracted the target object by removing the background noises through the difference image information of the sequential left images and then, controlled the pan/tilt and convergence angle of the stereo camera by using the coordinates of the target position obtained from the optical BPEJTC which was executed between the extracted target image and the input image. Also, we propose a novel 3D vision system to detect and segment the moving object adaptively through controlling the pan/tilt system of stereo cameras by using optical BPEJTC (Binary Phase Extraction joint transform correlator).