Research On The Online Detection Technology For The Surface Defect Of Tapered Roller

Bearing is a kind of precision mechanical components to reduce friction between theaxle and the axle bed, which consists of “The big four”, the outer ting, the roller, theholder, the inner ring. The roller is the most vulnerable parts of rolling bearings, and itsquality determines the life and stability of bearings. Roller surface defects will increasethe bearing vibration, decrease the rotating precision and bearings’ life, therefore, thedetection of bearing surface defects is meaningful. Nowadays, Roller surface defectsdetetion is mainly depended on manual work. In order to improve the efficiency anddetection reliability, a tapered roller on-line detection system was designed.The defects on3different surfaces was Summarized, Rust, scratches, spots, lack ofgrinding, lack of punching. Based on machine vision detection, using the method ofbinaryzation to distinguish the surface defects on the flank and bigger end surface, methodof feature extraction to judge the size, method of edge detection to detect the lack ofpunching on the smaller end surface.The physical construction of the roller surface defect detection system was designedin this paper, which consisted of loading, conveying, unfolding, laying-off mechanism, torealiz the online detection of roller. A control system of roller detection was designed inThis paper. The detection system is composed of illuminant, camera and image processor.Control system software is written on labview. Detection system software is written in theimage processor.The Pros and cons of various methods of image processing based on experiments wascompared, finding out the optimum method of roller detection. The precision, mistake rateand miss rate of the system was tested. The system can detect defects ofrust, scratches, spots, lack of grinding, lack of punching. The minimum detectable defectdiameter is0.2mm. The mistake rate is less than2%and the miss rate is less than5%.