Research And Application Of Tobacco Rod Packaging Inspection System Based On Machine Vision Technology

In the modern tobacco production industry,various automated machinery and equipment are basically deployed.In order to increase production capacity,automated assembly lines are widely used in the production process,and the operation speed of production lines is gradually accelerated.During this rapid production process,defects in the outer packaging such as creases,printing errors,rewinding,de-gluing,skewed labels,broken cartons,etc.are easily produced.In order to solve the defects and flaws caused by cigarette sticks in the process of packing machine production,we developed a machine vision-based inspection system for cigarette sticks,which sends a signal to the rejecting device when the defective sticks are identified,and the rejecting device will reject the sticks after waiting for them to be in place with a certain time delay.The overall design of the detection system,hardware selection,user interaction software design and detection algorithm design are systematically discussed in this paper,mainly in the following aspects:1.the current research status and application background of the tobacco outer package inspection system is thoroughly studied,the overall scheme of the tobacco outer package inspection system is analyzed and initially designed,mainly including designing the composition and mechanical structure of the system,and determining the installation scheme of the light source and CCD camera.2.hardware selection was made for the visual inspection system needs,mainly including light source,CCD camera,lens,etc.,and the type and model were determined through theoretical analysis of hardware characteristics.3.Based on the needs of the production inspection system and the business requirements of the actual production work,a design scheme focusing on user interaction software was proposed,including business logic framework design and module design.In particular,a C#-based graphics engine was developed,whose functions include different types of arithmetic box editing tools in addition to the normal display of captured images.4.An image processing technology based detection algorithm was designed and written for the characteristics of the types of defects in the outer packaging of the cigarette sticks.A new pre-processing method is applied to the images,and image feature statistics,edge detection,template matching,and Hough transform are used to detect defects such as creases,end folds,drawn lines,and drawn ears of the cigarette strips.The existing image processing procedures are improved to enhance the efficiency of the algorithm.5.Experimental testing of the modularity and wholeness of the system.Demonstrates the rationality of the system design and the feasibility of the detection of the algorithm.Finally,the work accomplished in this paper is summarized,and the problems are analyzed for improvement and reflections are made.