Machine vision-based method for conducting color quantization of colored bottle cap image

Abstract

The invention discloses a machine vision-based method for conducting the color quantization of a colored bottle cap image. The method comprises the steps of (1) collecting the image of a colored bottle cap; (2) according to the type of the bottle cap, selecting the number K of colors required for the image after being processed; (3) according to the selected number K of colors, subjecting the collected image of the colored bottle cap to color clustering quantization; (4) according to a judgment condition for ending the clustering, judging whether the clustering is completed or not, and obtaining K color types after the completion of the clustering (clustering results are different for different kinds of bottle caps); (5) for the images of the colored bottle cap obtained later, assigning colors obtained during the previous clustering step to a new image according to the principle of shortest Euclidean distance. The method is applied to the pretreatment process for the defect detection of bottle caps. Based on the method, the basic color features of processed color images are retained, and the influence of light sources, lens and other environmental factors on the imaging quality is reduced.

Description

Technical field [0001] The invention relates to a method for quantifying the color of a color bottle cap image based on machine vision. Background technique [0002] In industrial automated production lines, especially in the production of metal bottle caps, due to trademark printing, cutting, stamping and other processes, the surface of the cap is prone to defects and needs to be tested. Existing automatic inspection systems based on machine vision are all based on the gray-scale image of bottle caps. As the patterns of bottle caps become more and more, the detection accuracy and stability of the detection system based solely on gray-scale images are reduced. . A pre-processing method is needed to process color bottle cap images. Summary of the invention [0003] The technical problem to be solved by the present invention is to overcome the shortcomings of the existing bottle cap defect detection system, which only detects the gray image of the bottle cap without detecting the ...

AI Technical Summary

Problems solved by technology

Existing automatic inspection systems based on machine vision are all based on grayscale images of bottle caps. With more and more patterns on bottle caps, the detection accuracy and stability of detection systems based solely on grayscale images have decreased.

Images