The Dimension Measurement And Appearance Defects Detection Of The Mobile Phone Battery

At present, in industrial field of dimension measurement and apparen ce defects detection on mobile phone battery, precision demand is higher and higher. At present, size measurement and surface defect detection is artificial in the manufacture of mobile phone batteries, it is time-consuming and labor, hard to meet the demand of different standardized criterions. As machine vision becomes more and more mature, automatic testing system used in mobile phone battery checking develops rapidly as well, and detecting technology based on machine vision is an advanced technology in field of size measurement and apparent defect s detection on mobile phone battery.This thesis adopts machine vision to perform a detailed study of dimension measurement and apparence defects detection on mobile phone battery. In respect of size measurement on mobile phone battery, at first, select camera, light source and lens according to demand of measure precision up to 0.02 millimeter, to build appropriate hardware testing platform, and then perform calibration with method of Zhengyou Zhang. Secondly, do a series pretreatment in collected images, including median filtering to remove the image noise and detecting edge of cell phone image. Finally, put forward a improved matching algorithm based on correlation coefficient, realizing image matching in the process of measurement. After that, do Hough transform to fit a straight line, obtaining coordinate information of the line as results of the measurement.In the respect of apparence defects detection on mobile phone battery, putting forward image registration method based on affine transformation. And then, difference method was employed to realize the image to be detected defect judgment, followed by image binarization, morphology processing, Blob analysis, connectivity analysis with detected defect image. At last, fit minimum circumscribed rectangle of detected defects, to obtain information such as siz e, number and location of defects, and classify flaws to be found.At the end of this thesis, algorithm put forward above is carried out to measure the size of the mobile phone battery, obtaining multiple sets of data, whose standard deviation is less than 0.01 mm, angle measurement of the standard deviation is less than 0.05 °, meeting requirements of this article according to analysis of test result. The algorithm has been successfully applied to industrial production owing to its high precision and in accord with the requirement of industrial production. A series of phone battery images are used to test the algorithm, verifying effectiveness of this defect detection algorithm according to test result, whose defect detection accuracy is 97.33%. Known from test result, this defect detection algorithm can effectively detect the surface defects including surface stains, surface white dots and pattern missing of mobile phone batteries with high accuracy.