一种基于图像处理的铁轨自动检测方法

Automatic Railroad Detection Approach Based on Image Processing

提出了一套局部线型检测算子(LLPD)用于货运编组站铁轨的自动检测。利用该算子提取铁轨边缘,去除检测结果中的干扰信息,结合Hough变换的思想,实现了断线相连。该检测结果与其他边缘提取方法相比较,对复杂背景上铁轨的检测更有效,该方法能定位铁轨上的障碍物,对减少铁路调车事故有重要的意义。

This paper proposes a set of local line pattern detector（LLPD） for automatic rail detection in freight railway stations.It uses LLPD to extract the features of rail edges,and discards the fallacious patterns from the result of the previous step.It re-connects the broken lines using Hough transform.Compared with other edge detection methods,the new technique is efficient for rail detection with complex surrounding.The proposed method can locate the obstructions on the railroads and thus of great importance for reducing railroad accidents.