Research On Rubber Ring Dimensions And Defects Image Detection System

Rubber sealing ring is an important industrial product.In the domestic,rubber seal ring defect detection methods are mainly manual inspection.This detection method has the following disadvantages,big labor intensity,easy to fatigue,introducing the artificial error and data stability is poor.At present,machine vision has been widely used in industrial inspection.In this paper,aiming at the specific structure of rubber ring,the detection system based on machine vision was studied.The contents of the study are as follows:(1)Analyzed the existing problems in the development of domestic rubber seal testing technology in domestic and introduced the background and significance of the research in this paper.(2)According to the geometric structure of the rubber sealing ring,defecting characteristics and oil cooling radiator manufacturer's requirements,the technical index of the testing system was formulated.Design detection system of rubber seal ring based on machine vision.According to the system design requirements,selected the appropriate surface light source,CCD camera,lens.By the TracePro optical simulation software and orthogonal experiment method,the uniform tunable LED light source system based on hos t computer was designed.The mechanical transmission system was designed for driving the rubber sealing ring do linear displacement motion and rotation motion in horizontal plane,and collecting seal edge and surface images..(3)The basic image processing algorithms suitable for rubber seal detection was studied,including filtering and image segmentation.A new weighted fast median filtering algorithm was proposed.Compared with the mean filter,Gauss filter and median filter,this algorithm not only has good filtering effect,but also has the fastest detection speed.The detection rate of this kind of filtering algorithm is nearly twice as fast as the median filter.(4)Based on analyzing the traditional method,according to the size of defect detection of rubber sealing ring,put forward a new algorithm based on coordinate transformation.The experimental data showed that the mean value of this method is 35.01 mm and theme an square deviation was 0.01 mm.The mean value of the traditional visual method is 35.05 mm and the mean square deviation was 0.05 mm.So the mean and mean square deviation of the method are lower than the traditional detection methods.The detection method in this paper has high accuracy and good stability,and it is in accordance with the technical index of the detection system.(5)In order to detect the surface quality defects of rubber sealing ring,a method based on sequential marking method was proposed.The experimental data showed that the least accurate detection rate of this method is 98%,higher than the detection system technical indicators.(6)Written a microcontroller program,LED light source control software and rubber sealing ring detection software...