基于OpenCV的铁路导轨自动识别研究  

Research on railway automatic recognition by OpenCV

作　　者：黄凯[1] 余学才[1] 丁建平[1] 李亚南[1] 唐飞[1] 

机构地区：[1]电子科技大学光电信息学院,四川成都610054

出　　处：《计算机工程与设计》2013年第2期574-578,600,共6页Computer Engineering and Design

摘　　要：在铁路交通安全检测中,铁路导轨的准确识别是判断导轨线上有无路障的前提,利用红外热像仪采集的视频图像,提出了一种在OpenCV开源平台下,实现基于数字图像处理的铁路导轨自动识别方案。通过获取图像序列帧,对其首先进行预处理,边缘检测;然后通过链码跟踪算法,实现线段的提取和判别;最后提出了统计链码值出现次数,用其最大值从方向上判别线段,并实现了断线相连。实验结果表明,在复杂的外界环境下能准确地获得直道导轨和弯道导轨的信息,具有很强的实时性和可靠性。In the railway traffic safety testing, the precise recognition of the rails is the premise of judging whether there is road block in the guide rail line or not. Using the video image collected by the thermal infrared imager, a new scheme based on OpenCV open source platform is proposed, and the rails automatic identification that based on digital image processing is realized. By getting image sequence frame, pretreatment and edge detection are finished, and then through the chain code tracking algorithm, the line and the extraction of discrimination is realized, finally, The maximum number of the chain code size that appeared to remove the interference lines of some wrong directions , and the disconnection linked is realized. All experimental re stilts show that, the proposed method can accurately obtain the information of the straight guide and the corner guide in the com plex outside environment, and has the very strong real time and reliability.

关 键 词：OPENCV 红外图像 视频检测 导轨识别 边缘检测 链码跟踪 线段连接 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]