Research On Typical Part Defect Detection Based On Machine Vision

Machine vision inspection is to use cameras instead of human eyes to complete the detection of products.Compared with the traditional manual detection,machine vision detection has many advantages.In recent years,China has put forward that the manufacturing industry should develop in the direction of intelligence and enhance the competitiveness of Chinese manufacturing.Machine vision plays an important role in intelligent manufacturing.Therefore,many enterprises and universities are studying machine vision.According to the needs of actual production,the machine vision inspection system of self-lubricating gasket is studied in this dissertation.At the beginning of the design,the detection target is analyzed and the whole design of the detection system is carried out.The detection system is divided into a mechanical system,a hardware system,a software system and a control system.After the whole design is determined,the subsystem is relatively independent and complete,and the sub-system is combined into the detection system.After completing the system design,the selection of camera,lens and light source in hardware system is studied.By studying the selection standard of camera and lens,the camera and lens are determined.At the same time,the illumination intensity,illumination mode and illumination color of the light source are studied.According to the detection object in this paper,the light source is selected.After the hardware system was built,the image algorithm is studied.From the acquisition of the image to the final output of the results are studied one by one.There are mainly image recovery processing,the color image is transformed into gray image,the image is located,the position of the target is found,the noise existing in the image is filtered,and then the image is divided into two regions:target and background by binarization.Because the selection of threshold can not separate the background and target well,it is necessary to carry on the morphological processing to the image.The image size information is obtained from the edge of the segmented image,and the image processing algorithm is designed.At the same time,the design of defect detection algorithm is completed.Design the software part of the system.Several different programming software are studied,and LabVIEW is selected as the programming software of this paper.The software part is divided into five parts:user login,image acquisition,image processing,size judgment,data preservation and sorting,and the five parts are designed in detail.Finally,the statistical process control(SPC)of the measured data is carried out to obtain the process capability index of the outer diameter CP=1.04,the process capability index of the inner diameter CP=1.01,and the process capability index of the bolt hole distance is CP=0.58.The process capability of the outer diameter and the inner diameter is better than the distance of the hole and the process of the hole distance needs to be analyzed to improve the process capability.This paper designs a complete detection system for self-lubricating gaskets in industrial production,including user login,user operation interface and detection processing algorithm.which has certain practicability.Compared with traditional manual detection,the system has high stability,rapidity and durability,which is of reference significance for industrial application.