The detection methods and devices of abnormal pictures in industrial quality inspection

Abstract

The invention relates to the technical field of industrial quality inspection, and provides a method and device for detecting abnormal pictures in industrial quality inspection. The method includes: acquiring a first picture set in a current cycle; determining a previous cycle adjacent to the current cycle, and acquiring The second picture set in the previous cycle; call the anomaly detection model, and input the first picture set and the second picture set into the anomaly detection model to perform anomaly detection through the anomaly detection model, and obtain the anomaly detection result; when the anomaly detection result is When the first picture set is abnormal, at least one discrete picture is selected from the first picture set through the discrete sample screening model; the abnormal detection result and the discrete picture are displayed on the visual interface; the quality inspection machine is maintained according to the discrete picture. In this way, it is beneficial to avoid the problem that defects cannot be effectively detected due to the destruction of the similarity between the detection data and the sample training set, and can improve the working effect of the quality inspection machine, thereby improving the industrial quality inspection effect.

Description

technical field [0001] The invention relates to the technical field of industrial quality inspection, in particular to a method for detecting abnormal pictures in industrial quality inspection and a device for detecting abnormal pictures in industrial quality inspection. Background technique [0002] The quality inspection machine deployed in the factory performs the work of surface defect detection of workpieces. Its core workflow includes drawing, defect detection model reasoning, and sorting. The distribution of the real-time image data set and the training set of the training defect detection model is similar, which is one of the important conditions for the surface defect detection work to be accurately performed by the quality inspection machine. [0003] When the upstream process or equipment of the factory changes, it will significantly affect the distribution of real-time image data sets, and the conditions of similarity may be destroyed, resulting in the defect det...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problem that the defect detection model cannot effectively detect defects due to the destruction of the similarity between the detection picture data set and the training set, the present invention proposes the following technical solutions

Images