Defect detection method based on image grayscale features

Abstract

The invention discloses a defect detection method based on image grayscale features, comprising: 1) performing the following steps on the measured feature region image to obtain: ① grayscale processing of the image to obtain a grayscale image; ② in the grayscale image Create a selection area in the current selection area, and each pixel in the current selection area generates eigenvalues; concatenate the eigenvalues ​​in the selection area by row / column to form the feature vector of the current pre-selected pixel point; ③ mark a new pre-selected pixel point, and repeat step ② until Obtain the eigenvector of the last pre-selected pixel point; ④sort all eigenvectors and remove some eigenvectors; weight the retained eigenvectors; concatenate all weighted eigenvectors to form a global eigenvector; 2) Calculating the cosine similarity 3) Marking the feature image type corresponding to the larger similarity value as the shape of the measured object; this method can have low requirements on imaging quality, and is suitable for quality inspection of products at industrial sites.

Description

technical field [0001] The invention relates to the field of visual detection, in particular to a defect detection method based on image grayscale features. Background technique [0002] Controlling product quality has always been an enduring topic in the field of processing and manufacturing. There are various methods of quality inspection. With the development of computer technology, more and more vision-based product inspection methods have replaced traditional inspection methods. The existing The visual inspection method has the following problems: high requirements on imaging quality, when the gray level difference between the object to be tested and the workpiece body is small, the key feature point information cannot be accurately extracted, resulting in false detection; for this kind of problem, on the one hand, it is necessary to use expensive On the other hand, a stable test environment is required, and the requirements for the external environment are high, which ...

AI Technical Summary

Problems solved by technology

[0002] Controlling product quality has always been an enduring topic in the field of processing and manufacturing. There are various methods of quality inspection. With the development of computer technology, more and more vision-based product inspection methods have replaced traditional inspection methods. The existing The visual inspection method has the following problems: high requirements on imaging quality, when the gray level difference between the object to be tested and the workpiece body is small, the key feature point information cannot be accurately extracted, resulting in false detection; for this kind of problem, on the one hand, it is necessary to use expensive camera, the cost increases; on the other hand, a stable test environment is required, and the requirements for the external environment are high, which cannot meet the use in the complex production environment of the industrial site

Images