A visual recognition defect detection method based on printed matter characters

Abstract

The invention relates to a visual identification and defect detection technology, in particular to a visual recognition defect detection method based on printed matter characters, This method uses shape matching for registration, The invention adopts OCV and differential shadow method to carry out defect detection, which has high detection accuracy and is suitable for a universal machine vision platform. The method comprises the following steps: firstly, the image of a standard product and the image of a product to be tested are collected by a high-speed high-definition camera, the corresponding ROI region is obtained, and the image character of the ROI region is segmented to obtain the region image of each character; 2, carry out shape matching on that image of the product to be tested toobtain the row and column coordinates, matching scores and correspond template numbers of the searched characters; if the number of matching is not equal to the number of characters, the product is unqualified; otherwise, the searched characters are sequentially detected by an OCV detection system to obtain the printing quality scores of the characters.

Description

technical field [0001] The invention relates to a visual recognition and defect detection technology, in particular to a visual recognition defect detection method based on printed matter characters. Background technique [0002] In our lives, we can see products related to printed characters everywhere. For example, the date on the pharmaceutical packaging, the characters of the product batch number, the characters on the food packaging, the characters on the packaging of consumer goods, etc. Due to a large number of automatic printing, characters appear printing defects, which reduces the qualified rate of products and affects production efficiency. With the development of modern printing industry, people have higher and higher requirements for printing technology, so products must be strictly tested before leaving the factory to control the failure rate. [0003] Traditional detection methods mainly rely on human eyes to detect and reject unqualified products. The adva...

AI Technical Summary

Problems solved by technology

However, in the actual image acquisition process, the illumination may change, and there will be a small amount of deviation in the position of the image. A large number of defective pixels caused by the offset will cover up the existence of point and line defects.

Therefore, the existing surface defect detection system has certain limitations, such as slow detection speed and high detection cost.

Images