Research And Application Of Assembly Parts Inspection System Based On Computer Vision

With the graduate development of science technology, the use ofcomputer in technical research and development, production and dailylife etc. is more and more and has gained much progress. In modernindustrialized production, the research of intelligent detection system ofindustrial products has been an important direction. The application ofcomputer visual inspecting system could replace manual work to produceproducts and inspect. It could be classified into three types which areinspection, positioning and assembling according to uses and functions.The research and development of this system is a means of highlyintelligent science and technology inspection and an expression ofscientification of modern production way. It could greatly improve theaccuracy of production inspection and satisfy the demands of improvingindustrial production efficiency as well as greatly reduce the productioncost and risk. It is a technical program of highly reducing manpower andmaterials. The system get traditional manufacturing industries out ofcurrent condition of depending on human visual identification. Humanvisual identification couldn’t guarantee to finish the work accurately,quickly and efficiently in the single working condition of high strung andcontinuous.There are incomparably advantages and developmentprospects in building computer visual inspection system. The systemcould identify and inspect spare parts automatically, in no need of human touch. The system has a set of advantages of automation, high efficiencyand low risk. At the same time, it could reduce cost of human productionand inspection, greatly improve production efficiency of manufacturingenterprises and give full play of advantages of high automation.AVI is a non-touch inspection method based on computer visualtechnology and method and comprehensively applying image processing,precision measurement and pattern recognition as well as artificialintelligence technology. Its basic principle is that the target image isanalyzed by computer visual system to get required measurementinformation and then combined with prior information to confirmwhether the products are up to standard or not.The paper is a study and exploration of spare parts inspectionsystem based on computer visual technology. The system mainly usesmethod of manual image segmentation which could avoid defectsresulting from image segmentation and improve the accuracy of imageinspection and guarantee inspection speed. The system uses CCD orCMOS digital camera to get images of spare parts on the assembly lineand could be divided into3stages. First, the system needs to shottemplate images which completely meet qualification requirements ofspare parts, then preprocess them by smoothing filtering, then partitionareas needs to inspect according to the manual set, and then save imagefeatures of every area. Second, conduct image registration of locationbetween target image and standard model image through imageregistration method. Finally, get image features of every area of images tobe detected and compare with standard image to detect the quality of thespare parts and if there are any errors in component assembly so as to make warnings and then to achieve the purpose of intelligent inspection.The paper uses VC++.NET as development platform to build spareparts image inspection experimental system of assembly line. The systemcould be fully used in problems of detecting spare parts in automotiveassembly line, speeding up inspection, improve quality of manufacturingand assembling and strengthen brand competitiveness. It will definitelybring another reform and breakthrough in industrial revolution if thistechnology was greatly promoted and used.Finally, this paper makes a summary of computer visual inspectiontechnology and system and an outlook of following plan, researchdirection and application area related with the topic.