Image Processing Algorithm for Tracking and Picking an Object for an Open Platform Mobile Robot

摘要

Designing a visual tracking system to pick-and-place an object is a complex task because a large amount of video data must be transmitted and processed in real time. In this work, a mobile robot visual control system is proposed which uses a stereo vision to acquire the 3D positions of a target color ball and to control an end-effector to pick the ball up and place it in the specified location. This mobile robotic system is based on the Open Source Computer Vision Library for programming framework with portable components, off-the-shelf commercial hardware, and minimal programming. Two calibrated web cameras are used to track the color ball and a 2-axis end-effector is controlled to pick up the tracked color ball. The visual tracking and servo system as designed in this work is less sensitive to lighting influences and thus performs more efficiently. Experimental results show that the open platform based robot vision system successfully finishes the pick-and-place task and could be scalable to many diverse areas of robotics research.