Research On Image Transmission And Processing In Tobacco Cleaning

In the process of tobacco production,due to the impact of the production process and the production environment some debris will be mixed.These debris will greatly reduce the quality of tobacco,so the impurity removal of tobacco has become an important part of tobacco production.In view of the low efficiency and high cost of manual sorting,tobacco impurity removal machine based on machine vision are widely used in tobacco impurity removal.Tobacco impurity removal machine mainly includes tobacco delivery system,image acquisition system,image transmission system,image processing system and foreign material rejection system.This paper first introduces the development of tobacco impurity removal system.general introduces the working process of tobacco impurity removal machine and the realization of related technologies,and then mainly analyzes the status quo of image transmission system and image processing system.The defects of image transmission and image processing in the existing tobacco impurity removal machine are improved,designed and implemented.As for the image transmission system of the tobacco impurity removal machine,image transmission system includes sending and receiving images and a variety of instructions.In view of the diversity of transmission information and the demand of network transmission system versatility,this paper uses the UDP transmission protocol as the basis to realize the reliable transmission of the network information.In the application layer,respectively,using sliding window mechanism and maximize the transmission method to ensure the reliability and real-time transmission,and the two methods are implemented,tested and compared.both methods can meet the application standards.As for the image processing system of the tobacco impurity removal,due to both the shape and texture of tobacco leaf and debris are very complicated,color is used by the existing tobacco impurity removal machines to identify the tobacco leaves and debris.In this paper,a method of applying BP neural network to color classifier to remove tobacco impurity is designed,and a complete set of training and testing software system is written,the color values of tobacco leaves and sundries are trained by BP neural network,the results show that the method can meet the practical application requirements.In order to solve the problem that it is difficult to distinguish the tobacco leaves and sundries with similar color in the impurity removal of tobacco leaves using the BP neural network color classifier,a method of applying the convolution neural network to the tobacco impurity removal is designed.And a set of image processing system based on convolutional neural network is written,including sample collection production software,training software,test application software.through the test,the tobacco impurity removal machine based on convolution neural network further improves the impurity removal accuracy of tobacco.