Research On The Detection System Of Brake Disc Surface Defects Based On Machine Vision

With the development of computer technology,machine vision has reached a period of rapid development,so that more and more applications based on machine vision have made great breakthroughs.Based on machine vision technology,this paper realizes the detection and size measurement of automobile brake disc surface defects.Brake disc is one of the key parts in the automobile braking system,which is assembled together with the wheel hub.Through dry friction with the brake lining,great friction is generated to reduce the wheel speed and thus slow down the braking.When driving at high speed,the automobile will make an emergency stop and the brake disc would suffer considerable wear and tear.So,the quality of brake disc can directly affects the performance of the brake and vehicle safety.When the amount of brake wear increases,it will lead to a shorter service life.Because the carbon content of the existing brake disc is too high,there will be a lot of sand holes,pores and other defects in the casting process.These surface defects of the brake disc are usually inspected manually.Due to the difference of manual discriminant standards,the inspection efficiency is low and the quality standards are not unified,which wastes manpower and material resources.Therefore,this paper studies the detection of brake disc surface defects based on machine vision,and mainly completes the following tasks:(1)The camera,lens and other hardware were selected and the light source and lighting method suitable for detecting the surface of the brake disc were designed according to the technical requirements of the brake disc acceptance standards.On this basis,the imaging model was established to build the imaging platform,and the camera was calibrated.(2)The sharpness of the image has increased,cause we carried out the image enhancement based on histogram equalization.This paper studies the shortcomings of common image denoising algorithms and designs an image denoising algorithm based on Contourlet transform maximum posterior estimation.This algorithm can precisely decompose the surface image of brake disc in multi-scale and multi-direction more,which can effectively remove the image noise and keep the edge and detail information of the image well.(3)Two types of holes and lines defects were obtained,according to the classification of surface defects generated by the casting process.Two sets of different image processing schemes are adopted in the paper,because the texture and structure details of these defects are different.Firstly,SIFT algorithm is used to accurately locate the defects,and then pixel calibration is carried out based on shape matching to determine whether the defect areas of all the holes are defects.Feature description based on image gradient vector is used for line defect detection,and then FPLE algorithm of momentum search is used to obtain defect skeleton.(4)The dimension measurement is realized because the contour of brake disc is fitted by using the algorithm of least square fitting circle.And then the industrial surface defect detection system V1.0 based on machine vision was used to detect the brake disc and complete the statistics of the results.