Research On The Automatic Detection System Of Roller Bearings Surface Defects

In this paper,we discuss the application of machine vision and LabVIEW software platform in the roller bearing surface defect detection. Introducing machine vision and virtual instrument technology into the traditional detection system, the system solves commendably the problem of low efficiency and standard disunity which appear in traditional roller bearing detection methods.The automatic detection system of roller bearing surface defect according to the different functions are divided into three parts: motion control part,image acquisition part, image processing and analysis part.The motion control part is mainly responsible for transmitting, positioning and sorting rollers,the whole part controlled by computer. Image acquisition part is mainly responsible for image collection by CCD after receiving the position signal that given by motion control part,after that,sending the picture to computer by connecting with IPC,and giving a feedback signal to motion control part,a start signal to image processing part. Image processing and analysis part is responsible for image pre-processing, segmentation,edge determination after receiving a start signal from image acquisition part,giving the results of the analysis and feedback to the position control of the elements of motion control signal for the future of the sub-roller sorting.The part center on edge determination and application of image processing software. Through comparing several edge detection operators,finally selected Canny edge detection operator for the system operator. In this paper, image acquisition and image processing part are based on the virtual instrument development platform LabVIEW and image processing software IMAQ Vision.Using control function that provide by IMAQ Vision,collecting image > image pre-processing n edge determination and template matching defect identification.Through simulated tests,the automatic detection system of roller bearings surface defect based on machine vision and virtual instrument technology can meet the testing requirements.The system has a good accuracy and efficiency of detection, high reliability and good stability, and a good application prospects.