基于视频的复杂交通场景中车辆实时检测系统的设计与实现  

Design and Implementation of Vehicle Real-time Detection System for Complex Traffic Scene Based on the Video Technique

作　　者：李冬芬[1] 杨丽娟[1] 余国伟[1] 施键兰[1] 

机构地区：[1]福建农林大学东方学院,福州350017

出　　处：《软件》2012年第6期29-31,共3页Software

摘　　要：在基于视频的交通监控系统中,车辆检测是关键技术。为了提高车辆检测的实时性,本文建立的车辆检测系统通过设置检测区域,使系统处理的数据量大大减少,并结合了帧间差分和背景相减的方法检测出车辆及其阴影,同时利用颜色不变定律消除阴影,只检测出车辆。该车辆检测系统具有简单易用,实时性较强的特点。In the video traffic monitoring system,the vehicle detection is the key technology.In order to improve the real-time performance,detect-area was set up in established vehicle detection system,making the amount of data processing greatly reduced.In this system,vehicle and its shadow were detected by the frame difference and background subtraction.At the same time,the vehicle was only detected by applying the color constancy to remove shadow.The system was simple and valid,fitting for real-time detection.

关 键 词：车辆检测 帧间差分 背景相减 阴影消除 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]