虚拟线圈和模板链相结合的车流量视频检测方法研究  

New Method of Traffic Flow Video Detection Based on Combining of Virtual Coil and Templates Chain

作　　者：王慧斌[1] 孟凡[1] 李庆武[1] 吴学文[1] 

机构地区：[1]河海大学计算机与信息学院

出　　处：《微型电脑应用》2010年第11期4-6,2,共3页Microcomputer Applications

基　　金：国家高技术研究发展计划(863计划)(2007AA11Z227);国家自然科学基金(60972101);江苏省社会发展科技项目(BS2007058)

摘　　要：在智能交通系统中,车流量检测对于准确提取交通参数起着重要作用。针对视频检测中虚拟线圈法由于环境和光照等因素影响导致检测不准确的问题,提出一种基于虚拟线圈法和模板链法相结合的车流量检测新方法。首先对采集的交通图像进行车道划分,根据虚拟线圈法设定判别准则,检测路面中的标记区域,同时运用相应的模板对车表面像素灰度变化均匀的区域进行标记,形成模板链,最后通过检测模板链实现车辆的检测。经过实验证明,该方法可以准确地检测出车辆,并具有较好的鲁棒性。In the intelligent transportation systems,traffic flow detection for an accurate traffic parameter extraction plays an important role.For the problem of detection inaccurateness of virtual coil method,a new method of traffic flow video detection based on combining of virtual coil and templates chain is proposed.Firstly,lane partition is done in the traffic image.According to the method of virtual coil,area markers are detected by judgment rule.Pixel gray uniform region of vehicle surface is marked by template.Secondly,the template chain is formed.Finally,the vehicle is detected through the templates chain detection.It is proved that this algorithm can detect vehicle accurately and has the good robustness.

关 键 词：图像处理 车流量检测 虚拟线圈 模板链 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]