Manipulator guide method and device based on 3D machine vision

Abstract

The invention provides a manipulator guide method and device based on 3D machine vision, and is applied to the technical field of manipulator control. The method comprises the steps that point cloud data of an original workpiece are obtained, and the original workpiece is subjected to pose evaluation so that pose data of the original workpiece can be generated; after the position of the original workpiece is changed, point cloud data of the workpiece after change are obtained, and the changed workpiece is subjected to pose evaluation so that pose data of the workpiece after change can be obtained; and the pose variable quantity before and after change is calculated and acts on an original working track planned on the original workpiece to obtain a coarse positioning track, and then, a track after precise positioning is obtained through a preset spatial range and a point position search algorithm for each track point of the coarse positioning track. Thus, the working track of a manipulator can be quickly guided into a system, a quick track switching working mode is included, the process of conducting track guide on the manipulator through machine vision is simplified, and the work amount for switching different operation tracks is decreased.

Description

technical field [0001] The present application relates to the technical field of manipulator control, in particular to a 3D machine vision manipulator guidance method and device. Background technique [0002] In the manufacturing industry, manipulators are commonly used in two scenarios, one for point-to-point operations and the other for continuous trajectory operations; in these two scenarios: [0003] When it comes to the non-negligible contour difference of the processed object itself, machine vision technology is usually used to correct it. In most cases, due to problems such as production capacity, space interference, and equipment budget, it is impossible to pass multiple The contour information of all angles can be obtained by using a single camera or multiple cameras at a time to collect multiple angles of the workpiece, which makes pure visual trajectory planning impossible, and must be obtained by manual teaching or offline programming. job track. [0004] There...

AI Technical Summary

Problems solved by technology

[0007] However, the problem with the above scheme is that the template method is often used to calculate the offset between the relevant position and the standard template in the visual deviation correction part and output it to the manipulator. In this process, the existing manual teaching or offline programming is not fully utilized. The manipulator trajectory needs to manually number the trajectory points and make them correspond to the offset provided by the vision, so the engineering efficiency is low and the degree of intelligence is not high

[0008] When it comes to the required processing trajectory is extremely complex, especially when the posture requirements are relatively high, it is difficult to quickly obtain the required posture and relevant feature points in general offline trajectory programming software and trajectory editing software using machine vision systems. technically difficult issues

Images