A Study of Image Processing Based Object Depth Estimation

Abstract

Issue of object's depth estimation from binocular vision technique is studied in this paper. A binocular vision platform is set up to detect and estimate the object's depth. Mathematical derivation of object's depth from the two cameras of the proposed platform is re-visited to reveal the linkage between system parameters such as focal length and the estimation error. Both of simulation and experimental results are obtained for parametric analysis of estimation error. A comparison with Kinect is also given to demonstrate the superiority of the proposed design.