Study On Defect Testing System Of Bearing Roller Surface Based On Machine Vision

As China's industry continues to improve product quality,stricter testing of bearing components is a key step in improving product quality.Due to the low efficiency of manual inspection and the strong dependence on experience,machine vision inspection has been widely used in industrial inspection.Bearings are the core components of industrial production,so the use of visual technology to detect bearing components can improve production efficiency while ensuring production safety.This paper analyzes the shortcomings of the current use of machine vision to detect bearing roller defect detection methods,combined with the actual inspection needs of the factory,and proposes the use of a single camera combined with bright and dark field machine vision detection method to design the bearing roller defect detection system.Finally,a set of bearing roller defect detection system with fast speed and high detection rate was designed to solve the problem of low detection efficiency at present.Experiments have proved that the system can quickly and accurately detect surface defects of rollers,and has certain practical value for improving production efficiency.Firstly,the roller cylinder spreading device based on double rollers is designed according to the shape characteristics and motion mode of the rollers,which can ensure the stable operation of the rollers under high-speed movement and will not interfere with the image acquisition.By cooperating with sensors and controllers,it is ensured that only valid information is collected,the processing of useless information by the system is reduced,and the processing speed of the system is accelerated.Secondly,this article discusses the characteristics of common light sources in the actual industrial environment,analyzes the light characteristics of the roller surface,customizes the special arc light source and selects and designs a set of efficient visual acquisition system.According to the characteristics of roller reflection,a light source system with multi-light source and multi-angle illumination is designed,which is used as a visual illumination system for image collection of roller surface defects.Different light sources are used to construct the bright field and the dark field to ensure that the defect characteristics in the image are obvious,so that the bright and dark fields exist in the same field of view at the same time,and a single area array camera is used for acquisition.While reducing the cost of the acquisition hardware,it ensures that the roller can still acquire high-quality images of the surface for image analysis under high-speed motion.Then,for the defects such as scratches,scratches and rust on the rolling surface of the roller,a set of efficient defect detection algorithms is designed,and a defect detection algorithm based on dynamic threshold segmentation and blob analysis of average gray scale is proposed.The verification in the experimental environment shows that the system can realize the division and detection of the appearance defects of the tapered roller,and the detection accuracy rate reaches more than 95%,which meets the needs of the production environment.Finally,the Halcon software is used to develop image processing algorithms,and the interactive interface is designed based on QT,and a set of automatic defect detection software for defects such as nicks,scratches and rust has been developed.It can realize the fast detection of the defective images and display and statistics of the detection results.The experimental results show that: the hardware device designed in this paper,combined with the inspection software,can quickly and accurately detect roller defects.Compared with the same type of software and algorithms,it has great advantages in real-time and accuracy.