基于Canny边缘检测和聚合接续法的路轨边缘提取方法

在铁路视频路障检测报警系统设计中,关键技术之一是提取路轨边缘以确定障碍物检测的有效范围,对范围内障碍物报警。本文提出一种基于Canny边缘检测技术和聚合接续法的路轨边缘提取方案。通过直方图均衡化,实现对原始图像的空间增强,突出图像细节,经过Canny边缘检测提取图像的边缘组合,采用本文提出的聚合接续法提取出符合条件的边缘组合中最长的两条线,并对其实施接续,可达到路轨边缘提取的目的。本算法经过MATLAB编程实现,并移植到TI的DM642视频开发板上进行硬件测试,能够有效地实现路轨边缘的提取,为铁路路障检测报警系统设计提供参考。

Rail-framework Abstraction Based on Canny Edge Detection and Poly-connection

In the design of the rail-block detection and alarming system,the Abstraction of rail framework is one of the key techniques,which is used to confirm the valid detection range and alarm the existence of obstacles.The method of rail-framework Abstraction based on the Canny edge detection and poly-connected method was brought forward in this paper.By histogram equalization,the original image was enhanced to stress some details,and the image’s edges group was Abstracted through Canny edge detection.Next,the two longest lines were gotten and poly-connected from the edges group,which stand for the rail-framework by the rules brought forward later in this paper.This arithmetic is realized by MATLAB,and tested in the TI DM642 video DevKit.The result shows that it can Abstract the rail-framework effectively and be referred to in design of the rail-block detection and alarming system.