Research on Autonomous Grasping of Target Based on Machine Vision

Abstract

Abstract In order to improve the work efficiency and performance of the robotic arm in the product sorting process, a robot autonomous grasping control method based on visual recognition feedback control is proposed. Aiming at the limitations of traditional robot teaching control or grasping object recognition based on a single feature, an improved template matching target identification method based on contour Hu moment feature for contour matching combined with color and shape features is proposed. Through carrying out autonomous grasping experiments and analysis, the experimental results show that the visual recognition feedback control method effectively improves the accuracy of robot grasping.