Research into an intelligent logistics handling robot based on front-end machine vision

Abstract

The rapid development of e-commerce is forcing the logistics industry to become more efficient and intelligent. Conventional handling focuses on simple and repetitive point-to-point handling tasks taught by demonstration, where placement and gripping points are strictly required. In this paper, we take the handling robot as the experimental platform and study the robotic arm control method based on machine vision, which identifies and locates the target object by machine vision, and then controls the robotic arm to accurately grasp the target, so as to achieve the flexible handling operation task. Through image recognition, path planning, object grasping and other functions, it realizes autonomous navigation and cargo handling tasks for robots in complex environments. The application of multi-size fusion based image recognition test experiments showed that the test is close to the real annotation, thus facilitating image recognition and intelligent storage.