基于机器视觉的铁路货车滚动轴承表面缺陷检测  

Detection of railway fright car rooler bearing surface defects base on computer vision

作　　者：涂宏斌[1] 周新建[1] 

机构地区：[1]华东交通大学CAD/CAM研究室,南昌330013

出　　处：《电子测量技术》2006年第6期202-204,共3页Electronic Measurement Technology

摘　　要：针对目前轴承缺陷检测操作复杂、速度慢、精确度低等缺点,利用数字图像处理技术,设计出基于CCD成像技术的轴承缺陷非接触在线检测系统,讨论了系统中灯光、摄像机的安装位置对成像的影响,提出利用图像特征的麻点的检测方法,该系统能对麻点缺陷进行识别,对缺陷的区域进行测定。现场测试表明,系统操作简单、速度快、精确度高、为轴承缺陷检测提供了一种新方法。Based on CCD image measuring and processing techniques, an online non-touch measuring system for measurement of bearing pitting is developed. The effects of system illumination and configuration of cameras are emphasized, and a novel pitting measuring algorithm are brought forward based on image processing. Experiments indicate that the system has a strong ability of anti-disturbance and high detection accuracy.

关 键 词：图像分割 表面缺陷 滚动轴承 机器视觉 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]