Research On Size Measurement Of Buake Wheel Cover Based On Machine Vision

The workpiece plays an important role in daily and industrial production,and the size of the workpiece is an important basis for the inspection of qualified size.Size measurement is an important method to obtain data,check the size is the traditional manual detection,this contact detection is time-consuming,the accuracy is not high,has been gradually adapted to the detection of modern industrial scene,especially in the production of industrial automation high precision products.With the development of technology and production,based on machine vision detection method has been applied to industrial automation detection,machine vision detection with high accuracy,high speed,high efficiency,non contact point,it has the broad application prospect.This project is from the design of detection of Zhejiang Wenzhou a large enterprise project,based on the principle of machine vision detection and digital image processing technology,mainly on software system design and build detection system,image preprocessing,edge processing,camera calibration of sub-pixel edge detection,segmentation,error analysis and other aspects of a lot of research work.To achieve the two micron level measurement indicators and the development of XG software testing system.The main research work is as follows:The first analysis of the development status at home and abroad are compared on machine vision research,discusses the machine vision of small size brake wheel cover research based on the outlook;and then focus on measurement and control system hardware design,including the light source,camera,lens,image acquisition card selection,detection system platform,improved camera calibration Zhang Zhengyou calibration method based on the inside and outside parameters of the camera,to prepare for the following image processing and detection accuracy.Then the camera detection of the relevant image pre-processing,filtering,smoothing,noise reduction and other related work.In view of the two image segmentation,the OSTU threshold method is used to segment the image.The software design is the first in the image processing edge detection operator according to the merits of several classic comparison,using Matlab image processing interface related image analysis,finally decided to use the improved Canny operator is used to find the pixel edge tracing method,to suppress noise,eliminate the false edge.Edge detection algorithm of low precision,sub-pixel edge detection algorithm to make the edge segmentation and positioning more accuracy,is proposed based on the first five times spline interpolation on the edge of the image edge segmentation amplification,and then combined with the Gauss surface fitting to the edge fitting after subdivision,improve the pixel subdivision accuracy.The improved algorithm based on the VC software development platform,the elliptic contour segmentation of piecewise fitting standard circle,parameter estimation,finally detect the size parameters,has certain self adaptability,to achieve two times the development of XG detection software,basically reached the experimental accuracy index.Finally,the error analysis of the experimental data is carried out,and the factors that may influence the measurement results are analyzed in detail.In this paper,according to the requirements of production testing requirements of the design of the visual brake wheel cover detection system,in the follow-up to the actual experiment can be verified to achieve the basic parameters of the relevant data detection.This detection method can greatly improve the detection speed and efficiency,and re-examination rate greatly reduced.It has been proved that this improved design algorithm is a practical and practical method,which can reduce the labor cost and translate the experimental theory into practical results.