Bearing Surface Defect Detection System Based On Multi-angle Light Source Image

Bearing is one of the most common basic parts in modern mechanical equipment.Its basic functions of rotation,support and smoothness are used by various common mechanical products.Therefore,the demand for bearings of different types and models is still high.At present,the world’s bearing production industry has developed into a complete industrial chain including a variety of types,large-scale and new technologies.Since 2005,with the increasing demand for bearings in China,China has become the third largest producer of bearings in the world.At present,the use of bearings has existed in all aspects of mechanical equipment.Defects will inevitably appear in the process of production and use.If the defective bearings are put into use,it will have no small negative impact on the equipment,even cause damage to the equipment,and bring huge losses to users.Therefore,the defect detection technology for bearings is very important.With the rapid development of modern machine vision and automatic control field,bearing detection system has basically got rid of the artificial nature,and automatic detection system based on machine vision and automatic control has become more mature.The detection methods for bearing surface defects are also emerging in an endless stream,but because of the reflective characteristics of the bearing itself and the small structure,the traditional single light source illumination detection method can not achieve perfect detection results for bearing surface defects.In this paper,a bearing image acquisition method based on multi-angle light source illumination is proposed.Firstly,under the condition of coaxial light source illumination,an image with clear boundary and text identification is taken for bearing positioning area division and character recognition.Then,defects on bearing surface are detected by multi-angle light source illumination,because of some minor defects.It can only be detected at a specific angle,so as to propose corresponding defect processing algorithms for four kinds of common defects on bearings,such as notches,rust spots,flat caps and scratches,to find the defect areas on the surface of bearings.Finally,the defects are mapped to the coaxial light source image of bearings.This paper not only realizes the software-related image processing algorithm,but also builds the corresponding hardware automation detection algorithm System.