Tracking and Measuring a Moving Object Based on Multi-Sensory Information with Eye-in-Hand Robot System

This paper proposes an algorithm for tracking and measuring a moving object based on multi-sensory information with eye-in-hand robot system. A CCD camera and two laser emittors are mounted on the end-effector of a robot. The tracking control objective is to maintain the end-effector in a desired distance and attitude with the top surface of moving object. This sensory structure and the desired state make possible to acquire simultaneously both gray scale image and range data of two laser spots generated on the top surface, which are used for tracking and measuring. The proposed algorithm utilizes gray scale image for global image processing and range data for local image processing. Gray scale image is used to determine the position of the moving object during the tracking control. Range data provide the information about the relative distance and attitude between the end-effector and the top surface of moving object. Additionally, gray scale image and range data are utilized to measure real geometric dimension of the moving object. In this study, a bootstrap stage is introduced before the succeeding image-based visual feedback control to improve the tracking performance. The experiment results show the effectiveness of the proposed algorithm for tracking and measuring in eye-in-hand robot system.