基于单目视觉的工业机器人自动化定位方法研究

Research on Automatic Localization Method for Industrial Robots Based on Monocular Vision

为解决工业机器人移动机械臂抓取作业过程中识别速度慢、定位精度低的问题,首先将相机采集的工件图像进行预处理,然后利用Harris角点检测与基于随机树的特征匹配算法进行了目标工件的识别,利用单目视觉定位模型计算目标工件位姿,最后搭建单目视觉定位系统进行实验。结果表明:与改进的SIFT特征点匹配算法相比,基于随机树的特征匹配算法鲁棒性好、实时性强;基于单目视觉的定位系统能够快速准确识别目标,满足工业自动化生产需求。

In order to solve the problems of slow recognition speed and low positioning accuracy in the grabbing process of industrial robots'mobile robotic arms,the workpiece images collected by the camera were first preprocessed.Then,Harris corner detection and random tree based feature matching algorithm were used to identify the target workpiece.A monocular visual positioning model was used to calculate the pose of the target workpiece.Finally,a monocular visual positioning system was built for experiments.The results show that compared with the improved SIFT feature point matching algorithm,the feature matching algorithm based on random trees has good robustness and strong real-time performance.The positioning system based on monocular vision can quickly and accurately identify targets,meeting the needs of industrial automation production.