System and method for sorting surface defects of aluminum profiles in real time

Abstract

The invention provides a system and a method for sorting surface defects of aluminum profiles in real time. The system comprises a transmission module, an image acquisition module, an image processing module and a mechanical arm motion control module, wherein the transmission module conveys a to-be-detected product to a detection area, the image acquisition module completes real-time acquisition of an aluminum profile plate image in the detection area and transmits acquired images to the image processing module, the image processing module comprises a defect detection model, detection results of the detection model are output to the mechanical arm motion control module through processing of the defect detection model, the mechanical arm motion control module selects a motion path according to the detection result information, and a defective plate is grabbed and conveyed to a target area by the mechanical arm motion control module. The system and the method for sorting surface defects of the aluminum profiles in real time can improve the detection precision of the model and realize automation of the quality inspection process of the aluminum profiles.

Description

technical field [0001] The invention relates to the technical field of aluminum profile surface defect detection, in particular to a real-time sorting system and method for aluminum profile surface defects. Background technique [0002] As one of the important materials in modern industry, aluminum profiles are widely used in transportation, construction, industrial manufacturing and other fields. Before aluminum profiles are put into use, the quality of aluminum profiles needs to be tested. Defects are formed for various reasons, and different defects need to be reprocessed through different processing procedures to prevent quality problems of the ex-factory aluminum profiles. [0003] The real-time sorting of aluminum profile surface defects is mainly based on manual inspection and traditional machine vision inspection. Among them, the manual detection method has the disadvantages of low accuracy, poor real-time performance, and being greatly affected by subjective factor...

AI Technical Summary

Problems solved by technology

However, deep learning networks need to collect large-scale training samples, and when applied to the field of defect detection, defect samples are often difficult to obtain

Vision-based defect detection is not stable enough, affected by ambient light, dust, etc., if there is no large-scale labeled training samples, the generalization of the deep learning model is not enough, the detection accuracy needs to be improved, and as the network depth deepens, the model responds The speed will be reduced, affecting the detection efficiency

Images