Design And Research Of Visual Inspection System For Tobacco Conveying State

With the development of the cigarette factories and fine of cigarette production increase,according to conventional smoke cigarettes factory original wind system has not fully meet the requirements of the processing of thin cigarette,cigarette factories need according to the realtime transmission of wind pipe of wire feed system of tobacco state with air conveying system parameters such as wind speed,summed up a set of new wind parameters standard wire feeding system.Therefore,it is an urgent technical problem to detect the transmission state of tobacco in wind wire feeding system.At present,most of the several kinds of tobacco conveying status detection equipment that can be found in the research have disadvantages such as large amount of pipeline reconstruction,high equipment cost and complex operation,so they have not been put into actual production.In view of the above problem,this thesis will research in recent years the hot machine vision technology is applied to the design of the state of tobacco conveying visual detecting system,the system can effectively solve the pipeline modification quantity big,the equipment is expensive and complicated problem,and under the efficient detection algorithm of blessing,improved the detection accuracy and speed.The main contents of this thesis are as follows:（1）The development environment of visual inspection system is set up.According to traditional visual system structure,summarizes and state of the preliminary design of tobacco conveying visual detection system hardware structure,and according to the testing requirements and the work scene of each part of the hardware selection,and then set up in the computer visual inspection system for software development VS2019,Qt and Open CV platform,made full preparations for follow-up study.（2）The methods of image acquisition and preprocessing are studied.The basic image knowledge is briefly introduced,the image acquisition is studied,and the calibration process of the industrial camera image acquisition is studied.The image preprocessing technology required before the special algorithm research of the collected image was compared experimentally,and the recognition scheme based on color space was determined to extract the tobacco information in the image and the Canny algorithm was used to detect the tobacco edge,which laid a foundation for the subsequent tobacco contour extraction.（3）Two graphic processing algorithms for detecting the transport state of tobacco were proposed.After sorting out and analyzing the existing image processing algorithms,detailed algorithm research and experimental verification are made for the two detection directions of tobacco spatial distribution state monitoring and tobacco speed detection,and the flow of tobacco transport state detection algorithm is finally determined.A tobacco extraction algorithm based on HSV color feature and a tobacco motion vector detection algorithm based on contour extraction and SIFT feature point matching were proposed.（4）The overall structure and interactive interface of the visual inspection system are designed.After completing the preliminary verification of tobacco conveying status detection in the experimental environment,the overall structure and functional modules of the visual detection system were studied,and the human-computer interaction interface of the visual detection system was designed by using Qt and VS2019 software development platforms.