铁路远程瞭望系统研究与应用

Remote Observation System of Railway and Its Application

本文提出一种铁路远程瞭望系统,该系统通过在铁路沿线安装相机和现场图像处理单元对线路现场图像进行采集和处理,识别是否存在异物等状态。现场图像处理单元将图像和状态发送给远程瞭望服务器进行存储。列车上安装的车载终端或工作人员的手持终端可以自主定位,并根据当前位置向远程瞭望服务器请求列车前方较远距离的线路图像和线路状态并进行显示。该系统可以扩展司机视野,作为现有行车控制系统的补充,保障列车运行安全。同时,本文提出一种基于Markov随机场的异物识别算法,该算法通过Markov随机场建立动态背景模型,可以有效提高异物识别的准确性。现场实验表明该系统可以有效实现限界内线路异物的识别,并能够实现系统预定功能。

In order to improve train operation safety,a remote observation system of railway was put forward.Cameras and field image processing units were installed along the railway line,field image processing units got images from the camera and recognized if there were any foreign objects in the railway line,and after recognition the units sent the result and the image to the remote observation server.Onboard terminals installed on the trains or handheld terminals used by the staff could get access to the images and get the recognition results stored in the server.A terminal could locate its position by itself and request the image and status of the railway line a long distance ahead of it.The remote observation system could extend the driver＇s vision distance,and could be used as a supplement to the current ATC system to improve train operation safety.Meanwhile,this paper proposed a foreign object recognition algorithm based on the Markov random field.The algorithm created a dynamic background model by the Markov random field and could effectively improve the performance of foreign objects recognition.The proposed system was tested in the railway scene.Field experiments show that the system can effectively recognize foreign bodies and achieve intended functions.