利用改进的背景模型实现车辆检测  

Vehicle detection using improved background estimation

作　　者：郭莉琼[1] 王小鹏[1] 

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《微型机与应用》2010年第16期50-53,共4页Microcomputer & Its Applications

摘　　要：为解决基于背景差分的车辆检测办法易受交通状况影响的问题,首先建立基于区间分布的自适应背景模型,然后利用改进的背景更新算法对建立的背景模型选择性更新,最后结合阈值分割和形态学处理实现运动车辆检测。实验结果表明,该算法在交通堵塞或临时停车等复杂交通环境中有很好的背景提取和更新效果。与经典的算法相比,该车辆检测算法在实时性和准确性方面都有所提高。The application of computer vision technology to traffic detecting is an important subject in ITS, and moving vehicle detecting is the basic part. To solve background difference in the context of practical application problems, firstly, a quickly adaptive background mode/ based on the interval distribution is mentioned. Secondly, a selective updating of the background model at the pixel level is proposed. At last, it combines threshold segmentation and morphological operate to improve the accuracy of the detection. The experimental result shows that： the proposed algorithm is compared with some previous method, showing its effectiveness and excellent real-time performance in traffic scenes. Especially, to prevent the background model from incorporating objects that are slow moving or stopped for a time gap.

关 键 词：车辆检测 自适应背景模型 选择性更新 阈值分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]