Research on Robotic Arms Based on Target Recognition and Positioning Technology

Abstract

This article proposes a multi-objective detection system based on the Yolov5 algorithm for object grasping using robotic arms in real-life scenarios. Compared to earlier versions of the Yolo algo-rithm and R-CNN algorithm, this algorithm has the characteristics of low computational complexity and high accuracy. Monocular cameras are often used in the market to endow the manipulator with a two-dimensional vision module. This paper, based on the monocular camera, compared with the binocular camera, which has a high cost in acquiring three-dimensional data