一种基于行车区域改进的高速公路车辆检测法  

An Improved Method Based on Trafficked Area in Highway Vehicle Detection

作　　者：王健[1] 王命延[1] 汪泉[1] 施志强[1] 

机构地区：[1]南昌大学信息工程学院,江西南昌330031

出　　处：《计算机与现代化》2008年第10期34-36,共3页Computer and Modernization

摘　　要：在高速车辆速度的检测中,视频检测因其本身的优势得到了越来越多的人的青睐,而视频测速中的运动目标提取通常会采用背景差法来实现,但在实际应用中这种方法很容易受到周边环境的影响,特别是绿化带和其他车道车辆的影响显著。本文针对高速公路中车辆行驶在车道线中的特点,先提取车道线,根据车道线求出行车区域,再对行车区域进行背景差提取运动目标,消除了周边环境对运动目标的影响。实验证明,该方法很好地克服了周边环境的影响,消除了误检目标,可以准确地检测出所需的运动车辆。There are more and more people pay attention to video vehicle speed detection in high way vehicle detection. And people usually take baekgrotmd subtraction to detect moving object, but when we take it into practice, this method can be easily influenced by circumstances, especially influenced by the vehicle from other ways and trees aside highway. This article pays attention to the lane feature, it detects the lane firstly, then detects the ROI and uses the ROI to take background subtraction, reaching the aim to decrease influence of circumstances. Our experiment justifies that this method can effectively decrease the influence of circumstances, and can detect vehicle in ROI exactly.

关 键 词：行车区域 车道线 运动车辆检测 感兴趣窗口 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]