A Novel Target Tracking System for the Amphibious Robot based on Improved Camshift Algorithm

Abstract

In the past, there are some problems in target tracking, such as poor effect and easy loss. This paper propose a target recognition system based on improved Camshift algorithm, which is mainly used in marine rescue, marine garbage search, aquaculture surface target tracking and other fields. To the flexible application of amphibious robot in sea, land and other fields, RGB-D camera was equipped on the amphibious robot for target recognition and tracking. We have done the research of target recognition in the early stage. In this paper, CAMSHIFT and Kalman filter were used together. Kalman filter predicted the moving position of the target in the next frame and fed it back to the previous frame. Because the judgment of the edge node can greatly shorten the communication time with the terminal, so as to improve the efficiency of target tracking. Experiments show that the amphibious robot can accurately identify and track the target, and the delay is controlled within one second