Research On Automatic Detection Technology Of Shaft Part’s Multiparameter Based On Image Processing Technology

Quality test is the most important part of machinery production. Quality Parts isone of the important factors that determine the level of quality of mechanical products.The traditional detection methods basically use contact measurement,but the contactmearsurement always cost too much time and occupy too much source.According tothe request of company,this article has developed a device that can automaticallydetect shaft part’s multi-parameter based on image processing technology.This devicecan improve the efficiency and quality of detection of shaft part.The general detection devices can only do the single parameter detection.Thedevice developed by this article can detect the both parameters.So how to detect theboth parameters in the same time is the biggest difficulty.According to the difficulty of this article,this article give the solution fromstructural design and algorithm.The mearsurement of dimensional accuracy use therelative detection method,the mearsurement of surface quality use the thresholddetection method. In the structural design this paper refered to the common partdetection equipment,choose the right image sensors and light sources.This paperdesigned the lighting system that combine frontal lighting system with back lightingsystem,achieved the requirement that a device must detect the surface quality and thedimensions of shaft part at the same time.In terms of detection software, this paperuse sub-pixel detection algorithm to improve the detection accuracy of surface qualityand dimension.This paper firstly introduced the traditional edge detection algorithmsuch as Robert,Sobel,Canny,Otsu，ect.This can let we know the outline of edgedetection.This paper haved compare the traditional edge algorithm detection resultsby detecting the shaft parts.Through the comparison this paper find the fitablealgorithm in the traditional algorithm.In next step,This paper haved gived the detailedintronduction about the curve fitting,interpolation,spatial ment and Zenike ment.Bycomparing the results of thes four subpixel algorithms,This paper prove the spatial ment algorithm is more suitable for the shaft parts than the other algorithms.In thispaper also design a set of experiments to test the stability of these fouralgorithms.According to the result of the eiperiment,this paper prove the stability ofspatial ment algorithm is the best,Zenike algorithm rank the second,interpolation isthe third, curve fitting is the worst.Finally,this paper proposed the algorithm that suitfor the detection of shaft part based on the research of traditional algorithms andsub-pixel algorithms.At the end of the article,this paper did the experiment based on the previousresearch. Through the experimental data and processing results,this paper verify thereasonableness of the designed algorithm and device.