Research On Key Technologies Of On-line Detection System Of Cigarette Mark Defects Based On Machine Vision

With the improvement of the quality of modern life,the printing quality and appearance of printed matter have gradually become one of the main factors affecting the purchase of consumers.Tobacco industry attracts many consumers to buy by virtue of rich printing and packaging,but the printing process of tobacco packaging is complicated and there are problems of printing defects.The traditional manual inspection of printing quality is affected by subjective factors and cannot meet the requirements of industrial inspection speed and accuracy.Therefore,it has become a trend to introduce machine vision and image processing technology into high-speed automatic detection of tobacco printed matter instead of manual detection.Therefore,this paper aims to study the online detection system of cigarette labels based on machine vision technology,and install and apply it on the cigarette labels printing production line,and verify the feasibility of the online detection system through experimental tests.In view of the problems existing in the detection of smoke standard,the following research contents are carried out in this topic:(1)The hardware of the online cigarette label defect detection system includes three parts:cigarette label image acquisition,cigarette label image detection processing,defect detection result output and execution.Cigarette image collection requires a high-brightness,high-stability lighting environment,and a bar-shaped LED light source is used for contrast.Considering the practicability of the online detection system,this article uses two linear CCD cameras to capture images,and stitches them through the capture card to ensure stable images.The image processing part of the hardware adopts an industrial control computer with fast computing speed,reliable performance and good stability.The defect detection result output and execution part requires the operation interface on the display to display various detection information to facilitate the operation of the operator.(2)In this paper,a fast image segmentation method for smoke symbol defect image is proposed.According to the obtained smoke beacon image,several common image processing methods are listed,including image gray scale and image denoising.Compared with these algorithms,the efficiency of the proposed algorithm is verified.The classification method of defect features was studied,and a fast identification method of smoke standard defect detection based on Halcon was proposed.Fault-tolerant experiment was carried out,and the detection results were analyzed.(3)The software function of the cigarette label defect online detection system has three main parts: template matching and positioning,defect detection and detection results.The template matching and positioning function is to complete the setting of the shape matching of the cigarette label detection content.The defect detection function realizes the detection,result display and processing of cigarette labels.The inspection result part allows the operator to browse the picture information of the defective cigarette label,and provides support for tracing the cause of the defect.In this paper,the machine vision technology is introduced into the tobacco cigarette packs in the defect detection,in the online print defect detection system based on machine vision,on the basis of the design and implement a set of cigarette label defect detection based on machine vision technology experiment system,the defect testing system developed for smoke,smoke and the developed system is applied to the industrial site marking defect inspection,The system can effectively realize the quality defect detection of tobacco label printed matter,and analyze and sort out the defect data,so as to improve the automation degree of tobacco products and printed matter products,which is of great significance to the development of the whole tobacco industry.There are 77 figures,6 forms and 96 references in this thesis.