Tracking Compensation of a Moving Target for a Biped Robot Based on vision Sensor
In this paper,the visual tracking algorithm for a moving target is proposed for the biped robot of which camera movement is irregular.Hexagonal Matching Algorithm is used to measure the changes of size,location,and rotation angle for a moving object from its image frame.For enhancing the efficiency of the tracking,we can adaptively adjust the starting point and the size of search area from the image information obtained.Finally,by using Affine Transform and Kalman Filter,the position estimation of the moving target is refined against the swing of the camera.Experiments with 20-DOF biped robot using mono vision sensor are implemented to prove the reliability of the proposed method.
Visual Tracking System Biped Robot Hexagonal Matching Algorithm Affine Transform Kalman Filter
Yun-De Shen Dong-Soo Cho Chang-Doo Kee Zhen-Zhe Li
College of Mechanical and Electrical Engineering, Wenzhou University, Wenzhou, China Department of Mechanical Engineering, Chonnam National University, Gwangju, Korea
国际会议
重庆
英文
788-793
2010-12-11（万方平台首次上网日期，不代表论文的发表时间）