A Single-Hand and Binocular Visual System for EOD Robot

Abstract

Nowadays, the existing EOD robots usually are operated manually; it is very hard to operate. According to the theory of stereo vision, a new explosive ordnance disposal (EOD) robot with a binocular vision system is developed. The vision system captures real-time images, makes camera calibrations with Zhang's method, preprocesses images, matches feature point, then gets the stereo coordinate of the feature point of the suspicious objects, programming with Matlab7.0 and the pattern matching library EasyMatch of machine vision software eVision, it displays the real-time images on the console, and finally, controls EOD robot to grasp the suspicious object automatically and accurately. Experimental results under variable lighting condition, variable shapes and variable distances indicate that the validity in real-time performance and precision of the vision system.