Product defect online classification method in industrial visual inspection

Abstract

The invention discloses a product defect online classification method in industrial visual inspection. The method comprises the following process steps: S1, establishing a classification target function based on manifold regularization: establishing the classification target function by adopting an one vs the rest strategy; S2, acquiring a defect area from the industrial camera image: acquiring the defect area from the industrial camera image by adopting Blob analysis; S3, extracting defect area features: extracting the defect area features by adopting Blob analysis; S4, putting the samples into an online classifier for learning: putting the defect area features as the samples into the online classifier for learning; S5, returning a defect classification result: classifying the samples through an online classifier to obtain a defect classification return result. According to the method, the flaw area characteristics of the defective products are obtained through the Blob analysis, andthe data information of the defective products can be counted in real time; defected products are classified through an online learning method based on manifold regularization, so that the algorithm complexity can be reduced, the algorithm error rate can be reduced, and the classification efficiency can be effectively improved.

Description

technical field [0001] The invention relates to the field of industrial visual inspection, in particular to a product defect classification method in industrial visual inspection. Background technique [0002] Visual inspection is to use machines instead of human eyes for measurement and judgment. Visual inspection refers to converting the captured target into an image signal through machine vision products (that is, image capture device, divided into CMOS and CCD), and sending it to a dedicated image processing system. Digitized signals; the image system performs various operations on these signals to extract the characteristics of the target, and then controls the on-site equipment actions according to the results of the discrimination. Is a valuable mechanism for production, assembly or packaging. It is invaluable in its ability to detect defects and prevent defective products from being shipped to consumers. Visual inspection generally uses machine learning. [0003]...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is that the existing product defect classification method in industrial visual inspection has low real-time performance, low classification efficiency and insufficient classification accuracy. Aiming at the above-mentioned deficiencies in the prior art, a method for obtaining defective products through Blob analysis is provided Defect area features can count defective product data information in real time; classify defective products through online learning methods based on manifold regularization, which can reduce algorithm complexity, reduce algorithm error rate, and effectively improve classification efficiency. Product defects in industrial visual inspection online taxonomy

Images