A Printing Defect Detection Method Based on Deep Convolutional Generative Networks

Abstract

A printing defect detection method based on deep convolution generation network, the invention relates to the problem of accurate detection of tiny defects in printed matter in defect detection technology. In recent years, deep learning has been widely used in defect detection. Although current methods have made progress on defect detection tasks in simple backgrounds, they still cannot accurately detect subtle defects. To solve this problem, the present invention proposes a printing defect detection method based on a deep convolutional network. First, add an upsampling module to reduce the loss in upsampling. Second, a self-attention mechanism is proposed, so that the network can generate images with more complex structures and more accurate details. Finally, the fitting generator generates the noise distribution of the image, removes the noise, and obtains the defect image. On the basis of ensuring the accuracy rate, the detection precision is improved. The invention is applied to the unsupervised detection of small printing defects.

Description

A printing defect detection method based on deep convolutional generative network technical field The present invention relates to the detection of tiny defects in printed matter Background technique [0002] With the improvement of living standards, people also have higher requirements for the appearance quality of printed matter. China owns In the huge printing market, printing inspection has also become a research hotspot in the academic circle. Traditional manual inspection is not only time-consuming and labor-intensive, Moreover, different quality inspectors have different detection standards, and at the same time, they are also accompanied by false detection and missed detection caused by fatigue. chasing this way The pursuit of quality and efficiency is gradually being neglected and replaced by automated machine vision inspection. Automated detection can Achieving 24-hour inspection at the same standard and speed, without the risk of artificial fatigue, and ...

AI Technical Summary

Problems solved by technology

However, at present, only large-area printing defects can be accurately detected, and accurate detection of small defects cannot be realized.

Images