Design And Implementation Of Flaw Detection For Filtered Cigarette Based On Image Processing

During the process of producing cigarettes,the quality of raw materials may be poor and the process of machining may be subject to deviations.As a result,some of the cigarettes produced during processing may be defective.Common defective products include wrong teeth,local wrinkles,dark spots of filters,spots of smoke sticks,and punctures of cigarette sticks.Therefore,it is necessary to carry out the inspection of the cigarettes produced by the processing to ensure the quality of the listed cigarettes and to prevent the presence of defective cigarettes.At present,many cigarette factories use artificial eyes to determine whether each cigarette is qualified or not to achieve the detection of cigarette smoke.The detection method is more likely to be misjudged,and the efficiency is low and labor intensity is high.Therefore,researching and designing a cigarette branch inspection method to realize detection automation is an important way for cigarette companies to ensure brand quality.The article takes the image of cigarettes collected by Hongta Group,aiming at the wrong teeth,local wrinkles,dark spots of filter rods,spots of tobacco rods,and punctures of cigarette rods in the production process of cigarette packs,and designs a method for the detection of cigarette smoke.The main research work done in this paper is as follows:(1)After analyzing the grey scale image samples of Hongta Group cigarette smoke,and comparing the commonly used threshold segmentation methods,the Manual threshold method was used to binarize the gray image of cigarettes;the bwareaopen function was also used.The result of binary image is processed for reprocessing to eliminate a portion of the interference region.(2)Analyze and study the binarization method for the gray image of cigarette.The morphological preprocessing of the extracted target cigarettes was performed.By comparing with the commonly used edge detection operators,the Canny operator was finally selected for edge extraction to obtain the edge parts of the tobacco rod and the filter connected to the target cigarette.The segmentation algorithm locates the target edge,divides the cigarette stick and the filter tip of the target cigarette branch,and pretreats the cigarette stick and the filter tip obtained by dividing the target cigarette branch.(3)For different types of cigarettes such as wrong teeth,partial wrinkles,dark spots of filters,spots of smoke sticks,and punctures of cigarette sticks,the different extraction methods for cigarette sticks and filter tips are described separately.The quality of the defects was evaluated,and the results of cigarette smoke detection were finally obtained,which laid the foundation for the development of the cigarette flaw detection system.