Cigarette appearance defect detection method based on deep learning

Abstract

The invention relates to a cigarette appearance defect detection method based on deep learning. The cigarette appearance defect detection method comprises the following steps that , a front image anda back image of a cigarette are collected through a camera assembly; histogram equalization processing is carried out on the collected images; performing median filtering processing on the processed image; cutting the processed image in proportion, establishing a gray template based on an image pyramid, and correcting the position of the image according to the position and direction of the image;combining the corrected upper image and lower image into one image; cutting the processed image into three images; generating masks in background areas of the cigarette holder section image, the LOGOsection image and the cigarette body section image respectively, and taking maskless areas as areas of interest 1-3 respectively; and establishing a model data set by using the region-of-interest sample set: establishing and optimizing a deep learning algorithm model through the model data set, and performing defect detection on the cigarette image by using a deep learning model. The deep learningmodel used in the invention can accurately classify defects, and can provide information for manufacturers to eliminate mechanical faults and improve production schemes.

Description

technical field [0001] The invention relates to cigarette detection, in particular to a method for detecting appearance defects of cigarettes based on deep learning. Background technique [0002] 1. The existing on-line inspection systems for cigarette appearance based on machine vision technology all use area array CCD cameras, and the collected cigarette end face images are two-dimensional plane information, which cannot realize the problem of end face cavity depth (three-dimensional information) defect detection, and The existing detection system can only realize the defect detection of the circumferential surface or the end surface, and has the disadvantage of single function. [0003] 2. The existing on-line inspection system for cigarette appearance based on machine vision technology can only reduce the droplet interference, but cannot identify the droplet and eliminate the interference. The existing method of eliminating droplet interference is to install two cameras...

AI Technical Summary

Problems solved by technology

[0002] 1. The existing on-line inspection systems for cigarette appearance based on machine vision technology all use area array CCD cameras, and the collected cigarette end face images are two-dimensional plane information, which cannot realize the problem of end face cavity depth (three-dimensional information) defect detection, and The existing detection system can only realize the defect detection of the circumferential surface or the end surface, and has the disadvantage of single function

[0003] 2. The existing on-line inspection system for cigarette appearance based on machine vision technology can only reduce the droplet interference, but cannot identify the droplet and eliminate the interference.

However, when the droplets are in the overlapping position of the field of view of the two cameras, the droplets will appear on the two images at the same time, causing misjudgment

[0004] 3. The existing online cigarette appearance inspection system based on machine vision technology uses image comparison algorithms to detect defects. Since each cigarette is at a different angle on the roller, the position of the collected image with logos or patterns will also change. , using a comparison algorithm to detect patterned cigarettes is likely to cause misjudgment, so the existing comparison algorithm cannot realize defect detection in the area of ​​patterned cigarettes

Images