Modeling Research On Trajectory Of Ping-pong Based On Binocular Vision

The Ping-pong robot, as a real-time, intelligent robot integrated research platform, has become a hot topic of research in recent years. In this paper, modeling research on trajectory of ping-pong based on binocular vision is investigated. The recognition of the high-speed flying ping-pong, trajectory modeling and drop point forecasting are researched. The main work includes:The ping-pong recognition and tracking system with binocular vision is developed by using OpenCV. The system can accurately identify high-speed flying ping-pong, calibrate camera, realize the 3D positioning of ping-pong combining the camera model. What's more, the trajectory tracking and drop point forecasting are realized according to the result of recognition and positioning.In this paper, motion blur of flying ping-pong, as well as the demands by an adversarial ping-pong robot, are analyzed, and after that we propose an integrated method called CS-BS-EF, which stands for a comprehensive integration and optimization of in being methods of color-segmentation, background-subtraction and ellipse-fitting and, overcomes the drawbacks of each single. Now, CS-BS-EF is used for detecting smear scope and calculating and tracking flying ping-pong.The analysis of the resent years' ping-pong movement trajectory model has been researched and compared over software and hardware cost, real-time nature, calculate accuracy. Based on the model of Chinese Academy of Sciences, a ping-pong movement trajectory model is built.