高速公路隧道车辆检测方法研究

发布时间：2018-06-28 04:49

本文选题：监控图像 + 隧道　；参考：《大连海事大学》2017年硕士论文

【摘要】：随着我国高速公路的快速发展,高速公路隧道数量和里程也在迅速增加,与之伴随而来的是高速公路隧道照明中存在着巨大的电能浪费现象,尤其是在一些车流量较小的隧道,照明灯具常亮造成了巨大的能源损失。为了解决这种现象,本课题组设计一种"车进灯亮,车走灯暗"的隧道节能控制系统。该系统主要由环境信息感知子系统、车辆行为感知子系统、照度实时监测子系统、隧道照明及调光子系统、隧道照明通信子系统、服务器隧道照明监控软件组成。其中车辆行为感知子系统由以下两部分组成:放置在隧道前一段距离处的红外传感器和地感线圈,负责检测是否有车辆驶进隧道,若有车驶进,此时调亮隧道灯光;安装在隧道内墙壁上可以覆盖整个隧道区域的监控摄像机,负责采集隧道内部各段区域的监控图像,通过处理隧道内每个摄像机获取的监控图像判断隧道内是否有车辆经过,若判断隧道内无车,且隧道外部也没有新的车辆即将驶进隧道时,则调暗整个隧道。如果隧道内车辆不能被准确检测,则会影响到隧道内的行车安全,所以对于隧道内车辆有无的及时准确的检测具有重要意义。基于这种目的,本文采取了一种基于监控图像的隧道内车辆有无检测方法:首先利用中值法得到不同调光等级下的隧道背景图像,将其保存在服务器中;监控摄像机获得隧道内部各段区域的实时图像,截取图像中路面部分作为感兴趣区域,将其与无车背景图像中相应的感兴趣区域进行差分运算;对差分后的图像进行基于边缘检测的二值化处理,目的是提取目标轮廓信息,去除目标以外的杂点;再对二值化图像进行分块处理,统计特定块区域内非零点的个数;最终通过非零点阈值来判断隧道内车辆的有无;阈值的获得来源于在吉林省通化市快大茂镇"赤柏隧道"大量实验统计结果的经验值;每当隧道内照明亮度发生变化时,重新获取该亮度下的背景图像存入数据库。该车辆检测方法在赤柏隧道实验超过半年以上,实际测试及使用结果表明:该方法实现简单,且当隧道有车时,检测为无车的漏检率为0,当隧道无车时,检测为有车的误检率小于5%。保证赤柏隧道实验阶段在达到隧道照明规范的情况下,取得了良好的节能效果。
[Abstract]:With the rapid development of expressway in our country, the number and mileage of expressway tunnels are increasing rapidly, which is accompanied by the huge waste of electric energy in the lighting of highway tunnels, especially in some tunnels with small traffic flow. Lighting lights often cause a huge loss of energy. In order to solve this problem, a tunnel energy saving control system is designed. The system consists of environmental information perception subsystem, vehicle behavior perception subsystem, illumination real-time monitoring subsystem, tunnel lighting and dimming subsystem, tunnel lighting communication subsystem, server tunnel lighting monitoring software. The vehicle behavior sensing subsystem consists of the following two parts: infrared sensor and ground sensing coil placed in front of the tunnel are responsible for detecting whether any vehicle is driving into the tunnel. If there is a vehicle coming in, the tunnel lights are turned on at this time; A surveillance camera installed on the wall of the tunnel that can cover the entire tunnel area, is responsible for collecting the monitoring images of each section of the tunnel, and judging whether there are vehicles passing through the tunnel by processing the monitoring images obtained by each camera in the tunnel. If it is judged that there is no vehicle in the tunnel and there is no new vehicle coming into the tunnel, the whole tunnel will be darkened. If the vehicle in the tunnel can not be detected accurately, it will affect the traffic safety in the tunnel, so it is of great significance for the timely and accurate detection of the vehicle in the tunnel. For this purpose, this paper adopts a method to detect whether there is a vehicle in the tunnel based on the monitoring image. Firstly, the median method is used to obtain the tunnel background image with different dimming levels, which is stored in the server. The real-time image of each region in the tunnel is obtained by the surveillance camera, and the road surface part of the image is taken as the region of interest, and the difference operation is carried out with the corresponding region of interest in the background image without vehicle. In order to extract the contour information of the target and remove the clutter points outside the target, the binary image is processed in blocks to count the number of non-zero points in the specific block area. Finally, the non-zero threshold is used to determine whether there is a vehicle in the tunnel. The threshold is obtained from the experience of a large number of experimental statistical results of the "Chipao Tunnel" in Quidomao Town, Tonghua City, Jilin Province. Whenever the lighting brightness changes in the tunnel, Re-retrieve the background image under this brightness and store it in the database. The vehicle detection method has been tested in Chipao Tunnel for more than half a year. The practical test and application results show that the method is simple, and when there is a vehicle in the tunnel, the rate of missing detection is zero, and when there is no vehicle in the tunnel, the detection rate is zero. The false detection rate of the vehicle detected was less than 5%. In the experiment stage of Chipao tunnel, good energy-saving effect is achieved under the condition of tunnel lighting specification.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495;TP391.41

【相似文献】