Research On Rotating Table Tennis Trajectory Tracking System Based On Three Mesh Stereo Vision

Three eye stereo vision tracking and measuring is a technique which based on binocular vision technology and image processing technique to detect specific moving object for tracking and trajectory reconstruction.The technique combines three cameras into two binocular stereo vision systems.The stereo vision system is widely used in video surveillance,intelligent robot,video encoding,artificial intelligence and military domains.In particular,with the advent of timely playback system,this technology opens a new chapter in the application of sports events.In this paper,the rotating table tennis in the sports competition is taken as the research object,by analyzing the model of rotating table tennis and using the modified Mean-Shift algorithm to track the trajectory of table tennis,the 3D reconstruction of the center of mass and the rotation of the table tennis is realized.The main research contents and results are as follows:(1)Three eyes stereo vision measurement principle and camera calibration.First of all,according to the characteristics of the table tennis game,this paper studies the video capture system of table tennis,and then determines the image acquisition method of three eyes stereo vision;Secondly,the paper builds a camera calibration platform,and according to the shortcomings of the current camera calibration methods and the limitations of the application,this paper proposes an improved method for the calibration of the plane target.The method realizes the calibration of three high speed cameras,which lays the foundation for the reconstruction of the centroid trajectory and rotation trajectory of the table tennis.(2)The centroid trajectory tracking of table tennis.In this paper,the traditional Mean-Shift target tracking algorithm is studied.According to the characteristics of small size and high speed of table tennis,this paper proposed an improved Mean-Shift target tracking algorithm based on motion information and prediction mechanism.Compared with the traditional Mean-Shift algorithm,the motion information and color information are used as the target feature.In order to reduce the effect of background and occlusion,the background and target templates are weighted.Meanwhile,the improved fast Kalman filtering algorithm is introduced to predict the position and the position as a stack position to reduce the tracking error and improve the computing speed.Theoretical analysis and experimental results showthe superiority of the proposed algorithm.(3)The centroid trajectory and rotation trajectory reconstruction of rotating table tennis.First of all,the article compares and analyzes the force situation of the non-rotating ball and eight kinds of rotating table tennis in the flight process,and obtains the tracking model of table tennis.Then,the tracking algorithm is used to track the rotating table tennis,and the least square method is used to obtain the 3D trajectory of the rotating table tennis.The correctness of the trajectory model is verified by experiments.At last,this paper studies the rotating trajectory of the table tennis.Based on the theoretical analysis of the homogeneous coordinate transformation of the rotation of a point on an arbitrary axis,and combining with the centroid trajectory of the table tennis,we can get the rotation of any point on the table tennis.