A method for detect glue deficiency of workpiece

Abstract

The invention discloses a method for detecting the lack of glue of a workpiece. The method comprises the following steps: S1, calculating an optimal segmentation threshold value to obtain a binary image of a preprocessed image; S2, obtaining a template graph of the binary graph; S3, extracting the object contour information in the template graph; S4, delimiting the target area according to the target contour information; 5, dividing that target region into K equal parts, wherein K is a positive integer not less than 2, and the K is use as a training data set; S6, the image is input to the neural network for classification, and the defect detection result is obtained. The invention can find out the glue-shortage defects in the workpiece, and is quick and simple.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a method for detecting glue shortage of workpieces. Background technique [0002] With the improvement of the quality requirements of industrial products and the popularization of automatic detection, more and more attention is paid to the automatic detection of workpiece defects in the industry, so a reliable, fast and general detection method is needed. [0003] In recent years, surface defect detection technology has been developed rapidly. The existing surface defect detection techniques include statistical method, spectrum method and model method. In statistical methods, the spatial distribution of gray values ​​can be described by features such as gray co-occurrence matrix, autocorrelation coefficient, mathematical morphology, histogram statistical features, and fractals. The premise of the surface defect detection method based on histogram statistical features is...

AI Technical Summary

Problems solved by technology

[0004] The above method is a traditional image processing method. This type of algorithm is suitable for the shape deformation detection of compact workpieces, combined with the feature extraction features of the image itself, and then performs defect detection according to the features. Higher, but due to the complex factory environment, uneven lighting, oil pollution, etc., the quality of the pictures collected on the assembly line cannot meet the requirements of this method, and the realization effect is not ideal

[0005] The reason is that due to the complex factory environment, more and more types of workpieces, and smaller and smaller sizes, there are many factors that affect the quality of the picture

In addition, the factory has many positive samples and insufficient defective samples, which also brings certain difficulties to the lack of glue detection.

Images