基于单目视觉的工业机器人拆垛系统设计与实验

Design and Experiment of Industrial Robot Stacking System Based on Monocular Vision

基于单目视觉提出了工业机器人拆垛系统构想。为实现工业机器人视觉拆垛系统中机器人精确运动控制与工件位姿识别,运用D-H位移矩阵法建立了机器人运动学模型,得到机器人末端相对机器人坐标系的位姿信息。基于形状模板匹配法提出目标图像识别算法,得到各个目标图像在相机坐标系下的位姿信息;将三维视觉模型与机器人运动学模型进行信息融合,建立机器人视觉拆垛控制系统数学模型。基于CCD工业相机、4-DOF工业机器人搭建视觉定位抓取实验系统。通过对空间目标进行抓取的实验,验证基于单目视觉的工业机器人拆垛系统的正确性、精确性、鲁棒性。

Base on monocular vision,the concept of industrial robot stacking system is proposed. In order to realize robotics motion control and recognition of the workpiece pose accurately in robotic visual stacking system,the kinematics model of robot was established by using D-H displacement matrix method,and the position posture information of the robot terminal actuator basing on robot coordinate system could be obtained. The algorithms of image recognition was proposed by using the shape matching approach,and the image coordinate information of the targets basing on camera coordinate system could be obtained. The three dimensional( 3 D) vision model and the robot kinematics model were fused to build a mathematical model of robot stacking system. Set up an experimental system for visual positioning and grabbing by using the charge coupled device( CCD) industrial camera and the four degee of freedom( 4 DOF)industrial robot. Verify the correctness,accuracy and robustness of the robot stacking systemby the grasping experiment of objects in space.