Research On Target Detection And Tracking Based On Binocular Vision Of Wheeled Robot

ABSTRACT: With the continuous development of science and technology, visual robot in the production of human life has got more and more applications. Binocular vision mobile robot is one of the most active branches, and it has obtained more and more attention. Mobile robot studied in this paper is equipped with a platform as binocular vision, binocular mobile robot moving target detection and passive target tracking. The details of the robot with the control system design are as follows.First, this paper reviews the development in the world of binocular stereo vision, problems in stereo vision, development trends, the contents of this research, as well as key issues. Designed according to the actual situation of this subject, the general scheme of binocular vision robot system and design the hardware system of the binocular vision are introduced. The binocular vision positioning for the track is in the next section.Secondly, the algorithm of mobile robot based on binocular vision to track moving object is studied. In Simulink environment, the typical movement target tracking algorithms are developed and realized. Based on the background of difference between three algorithms, this paper gives a special algorithm of detecting specific shape of dynamic target—background differencing based on shape characteristics algorithm. The feasibility and superiority of the algorithm is verified.Afterwards, the algorithm of mobile robot based on binocular vision to track and locate static target is studied. An algorithm based on the YCbCr Gaussian space of location and tracking specific color target has been adopted. Through the experimental comparison, the superiority of the proposed algorithm is demostrated compared with the simple threshold segmentation method. Through combining the shape characteristics of method with the YCbCr color space method, an algorithm of identify and positioning a specific shape and color of the static target is proposed—named as the static target location and tracking algorithm, which based on the shape and color limited. The algorithm could be used to locate and track a specific target in multi-target environment. In addition, the smallest difference algorithm of location and tracking static targets is studied.Finally, structure and electromechanical control system is designed for the mobile robot. The control strategy is developed for the mobile robot to locate and track target object independently.