基于机器视觉的机械臂双目测距系统研究

Research on Manipulator Binocular Distance Measurement System Based on Machine Vision

针对工业机器人在非接触式的工作环境下获取目标三维信息的潜在问题,设计出了一套基于机器视觉的机械臂双目测距系统。该系统采用Matlab和OpenCV对CMOS双目工业相机进行双目标定,并根据标定得到的相机内外参数进行立体校正,然后结合双目视觉的SGBM立体匹配算法计算出相应环境下的深度信息,最后利用二维图像和深度信息辅助机械臂控制中心获取目标工件的三维信息并作出相应命令操作。根据实验分析,该测距系统测量时间短、鲁棒性好、测量精度高。

In order to enable industrial robots to acquire target 3D information in a non-contact working environment,a manipulator binocular distance measurement system based on machine vision was designed.The system uses Matlab and OpenCV to calibrate the CMOS binocular industrial camera and perform stereo calibration according to the internal and external parameters of the camera,and then calculate the depth information in the corresponding environment by combining the SGBM stereo matching algorithm of binocular vision.Finally,the 2D image and depth information are used to assist the robot arm control center in obtaining the 3D information of the target workpiece and making the corresponding command operation.The experimental analysis indicates that the ranging system has short measurement time,good robustness and high measurement accuracy.