基于机器视觉的Baxter机器人抓取算法研究

Research on Baxter Robot Grasping Algorithm Based on Machine Vision

针对传统机器人结构化示教抓取无法满足智能制造生产过程中实际抓取需求的问题,对机器人在非结构化场景下的多目标灵活抓取进行了研究,在分析了传统机器人抓取检测方法的基础上以智能协作式机器人Baxter作为操作对象,融合机器视觉技术提出了一种基于改进的Canny边缘检测和Hough变换的机器人抓取检测分类算法,并最终实现了机器人的目标抓取。通过实验证明,该算法能够有效提升Baxter机器人在完成抓取任务过程中的精准性和鲁棒性,该研究对智能制造环境下的机器人灵活抓取任务实现有着重要的案例参考意义。

Aiming at the problem that the traditional robot structured teaching grasping cannot meet the actual grasping requirements in the intelligent manufacturing production process,the robot’s multi-target flexible grasping in unstructured scenes is studied,and the traditional robot grasping detection method is analyzed.On the basis of this,the intelligent collaborative robot Baxter is used as the operating object,and machine vision technology is combined to propose a robot grasping detection classification algorithm based on improved Canny edge detection and Hough transform,and finally achieves the robot’s target grasping.Experiments have shown that the algorithm can effectively improve the accuracy and robustness of Baxter robots in the process of completing grasping tasks.The research in this article has important case reference significance for the realization of flexible robot grasping tasks in intelligent manufacturing environments.