基于计算机视觉的铁轨表面缺陷检测系统  

Rail Surface Defects Detection System Based on Computer Vision Technology

作　　者：王海涛[1,2] 甄理[1] 杨春霞[2] 陈如冰[3] 丁克勤[4] 李冬[1] 王平[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016 [2]中国商飞上海飞机设计研究院,上海200235 [3]苏州市计量测试研究所,苏州215128 [4]中国特种设备检测研究院,北京100013

出　　处：《无损检测》2011年第11期38-41,共4页Nondestructive Testing

基　　金：国家自然基金资助项目(50907032);国家科技支撑计划资助项目(2009BAK58B02);国家质检总局科技计划资助项目(2009QK162);江苏省科技厅资助项目(BK2010503;BZ200951;BE2009162);苏州市科技局资助项目(SYG201024)

摘　　要：计算机视觉技术是计算机科学的一个重要分支,它涉及到图像处理、模式识别、人工智能、光学等多个领域。基于计算机视觉技术的基本理论,设计了铁轨表面缺陷检测的软硬件平台,并运用该方案对铁路现场采集到的铁轨缺陷图像进行处理。试验结果表明,该方案能够正确识别出铁轨表面缺陷的大小和位置,具有一定的实用价值。Computer vision technology is an important branch of computer science, which involves image processing, pattern recognition, artificial intelligence, optics and other fields. Based on the basic theory of computer vision technology, a hardware and software platform was designed for detecting the rail surface defects. The result data from the experiment, which was obtained with defect images collected on the railway, showed that this method was capable of identifying the size and location of surface defects and had stated applicability.

关 键 词：计算机视觉 缺陷检测 特征提取 

分 类 号：TG115.28[金属学及工艺—物理冶金]