부산대학교 도서관

This paper develops a vision system to predict ball trajectory for the ping-pong robot. The stereo vision system with two cameras is used to measure ball center. A multi threshold segmentation algorithm is applied to precisely locate the center of ball in image and the triangulation algorithm is applied to compute the 3D position in the world coordinates. Six consecutive coordinates are collected and the least squares method is used to determine the initial states of the ball. From these initial states, the following flight trajectory of the ball is predicted using two physical models: aerodynamics model and rebound model. Then the states of the ball at striking point are calculated from the predicted trajectory. Experimental results show that the developed system can achieve a good predicting precision in real-time.