U-Shaped Parts Grasping Detection Method Based on Rotated-Yolov5

For improving the production efficiency and reduce labor costs in Industry, robotic grasping method has been concerned. The most common scenario in industry is sorting various parts and components, which is also the most suitable application scenario for robot grasping. For the U-shaped buckle grasping job, this paper proposes a scheme based on objection detection algorithm, fusing of classification angle information. At the same time of meeting real-time requirements, it can achieve more than 0.9 accuracy of the rotation target detection. Then using a mask-based digital image processing method to obtain the grasping points, providing reliable information for the subsequent grasping tasks.