Research on stereo vision matching algorithm for rescue robot

Abstract

After coal mine gas explosion and other accidents, rapid rescue is an important task. The underground tunnel is generally narrow and the number of people passing through it is limited. Therefore, a real-time detection technique is urgently needed to feedback the mine environment in real time and locate the trapped miners accurately. Under the actual environment, the rescue robot technology is urgently needed in the underground abominable environment. The rescue robot can use the airborne binocular camera to respond the downhole environment to the ground operation platform in real time. Due to the complexity of the environment in the mine, we must study a robust, real-time and accurate matching algorithm to restore the real scenes in the mine. In this paper, the improved census algorithm is used for stereo vision matching, and compared with the traditional SAD algorithm in the matching performance, and the superiority of the improved census algorithm is proved.