Research On Key Technologies Of Cigarette Identification System

In the field of tobacco sorting and distribution,as customer orders become more complex and diversified,in order to improve the accuracy of sorting and the efficiency of distribution,it is necessary to develop a cigarette identification and error correction system that matches the sorting line.According to the characteristics and technical requirements of the tobacco sorting line,this paper conducts research on the accurate and efficient cigarette identification system.The research content is as follows:First,after analyzing the composition of the cigarette sorting line and the layout of the recognition equipment,a cigarette recognition system based on the fusion of multi-angle code scanning and front image recognition was proposed,and a semi-closed tunnel shooting device and a smoke posture correction device were designed.,And multi-angle scanning code layout,improve the image quality of cigarettes and the reading rate of cigarette barcodes.Secondly,the image preprocessing is studied,including image filtering,image morphology processing,geometric transformation,image segmentation methods,and through comparative experiments,it is determined that the Otsu threshold segmentation method can distinguish the target area from the background area to the greatest extent.Then the related methods of target region extraction are studied,the principles and methods of common template matching are introduced and analyzed,and a two-time template matching method of coarse matching followed by fine matching is proposed for target region extraction.Experiments show that this method is feasible and effective.Finally,the research on the identification method of cigarettes,the image recognition algorithm of cigarettes,and the fusion of the two recognition technologies;in the barcode recognition of cigarettes,according to the characteristics of the EAN-13 code,the determination of the number of modules,the number of empty modules and the number matching are proposed.The two-step bar code recognition method of verification and threshold adjustment can accurately and quickly identify bar codes;in the aspect of cigarette image recognition,the RBM(Limited Boltzmann Machine)model and DBN(Deep Confidence Network)are introduced,and then a An optimized algorithm of the Deep Belief Network(DBN)based on the Elite Ant System(EAS)is used as the cigarette image recognition method in this paper.The experiment shows that the EAS-DBN algorithm meets the system requirements;in terms of the integration of the two recognition technologies,comprehensive consideration After the respective characteristics of the two identification technologies,the identification logic method based on "or logic" is determined;and the field experiment shows that the cigarette identification system in this paper has a high recognition rate and a low false alarm rate,which effectively improves the cigarette sorting Line sorting efficiency.