一种适应城市交通道口的改进型自适应背景提取算法  

An Algorithm of Background Extraction Based on Statistics of Histogram Combining with Multi-frame Average

作　　者：李晓飞[1] 梅中辉[1] 

机构地区：[1]南京邮电大学通信与信息工程学院,江苏南京210003

出　　处：《系统仿真技术》2008年第4期242-246,共5页System Simulation Technology

摘　　要：介绍了背景的提取与自适应更新的算法,提出了1种基于直方图统计与多帧平均混合的算法。这种算法克服单纯的多帧图像在车流量多、车速低时出现拖尾和模糊的现象,且相对于混合高斯模型,其计算复杂度并不大。该算法能够利用现有的TIDM642DSP硬件平台实现,实验图像表明其结果与前2种算法相比,提取的背景更符合实际场景。The paper introduces background ward an algorithm of background extraction extractions with self-adaptive update algorithm and puts forbased on statistics of histogram combining with multi-frame average. It avoids the image trail-blur phenomena using pure multi-frame method on traffic jam, and has relative low computation complexity comparing with the mix Gauss model. It can run on the TI DM642 DSP hardware platform, the experimental image shows the better result and gets better background image than two methods before.

关 键 词：智能交通系统 视频检测 背景的提取 自适应更新算法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]