A Passive Optical Motion Capture Method towards Occlusion Conditions Based on Multi-vision System

When using passive optical motion capture technology to measure the 6 degree of freedom(6-DoF) motion of mobile robots, the target attached to it may be obscured by the obstacles in the environment. The occlusion will lead to the marker points missing and further affect the accuracy of the pose estimation. In this paper, a passive optical motion capture method based on feature reconstruction and state estimation is proposed. The feature reconstruction is conducted by the means of monocular EPnP algorithm to make use of the priori information and keep the dimensions of the observations constant. Then an adaptive unscented Kalman filter(UKF) is adopted to accomplish state estimation. Finally the information from different binocular systems is fused in a multi-vision system to further improve the robustness of the motion capture system. The measurement results of our system show that the average absolute position error is 7.892mm and the average absolute attitude error is 1.395° during the measurement of the 6-DoF motion under some severe occlusion conditions. The results demonstrate the effectiveness and augmentability of the proposed method.