Columnar workpiece classifying and positioning method based on target detection and machine vision

Abstract

The invention discloses a columnar workpiece classifying and high-precision positioning method based on target detection and machine vision. The columnar workpiece classifying and high-precision positioning method comprises the steps of target detection, defect detection and coarse positioning of yolov3 and high-precision positioning of machine vision. Wherein the part of the yolov3 comprises thesteps of making a data set, improving a network structure, adjusting candidate box parameters, carrying out real-time positioning identification and carrying out defect detection. A workpiece image isobtained through an eye-to-hand camera, an image enhancement algorithm is fused, and the candidate box parameters are improved by adopting a vector similarity measurement method. According to the machine vision part, a yolov3 algorithm is used for roughly positioning a position to guide an eye-in-hand camera to obtain an image, feature extraction is carried out on image features, maximum value constraint is adopted, abnormal features are eliminated, finally, workpiece contour features are fitted, and the high-precision position of a target workpiece is obtained.

Description

technical field [0001] The invention relates to industrial robots and machine vision applications, in particular to a high-precision positioning method for columnar workpiece classification based on target detection and machine vision. Background technique [0002] With the development of intelligent manufacturing, industrial robots have the advantages of good versatility and high repeat positioning accuracy. In some industrial automation fields, most of them adopt the method of robot teaching. There is still a long way to go to realize real intelligent manufacturing, and traditional teaching methods cannot meet the needs of intelligent manufacturing. Machine vision technology solves the robot position control requirements very well, but at the same time there is a problem that the recognition flexibility and accuracy are difficult to be compatible. The target detection technology based on deep learning can better meet the flexible requirements of multi-target recognition, ...

AI Technical Summary

Problems solved by technology

[0004] Target detection based on deep learning can be achieved but the accuracy is poor; traditional machine vision recognition and positioning accuracy is high but the detection target is too single

Images