基于Kalman滤波器的车辆检测与跟踪系统的实现  

Implementation of vehicle detection and tracking based on Kalman filter

作　　者：赵莉[1] 陈泉林[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《电子测量技术》2007年第2期165-168,共4页Electronic Measurement Technology

基　　金：上海市教委高等学校科学技术发展基金(217302)资助项目

摘　　要：智能交通系统是解决城市交通拥挤最有效的方式,其中交通信息采集设备是交通系统管理的基础与前提,而基于视频图像处理的交通信息检测器较其他类型检测器,具有信息量丰富,安装和维护成本低廉的特点。本文用基于Kalman滤波器的方法实现了交通信息采集设备中的车辆检测与跟踪。它采用了一种自适应背景更新算法,通过分割、二值化、腐蚀膨胀得出前景图像,以包含前景图像的矩形框的中心作为Kalman滤波器的跟踪特征,对运动车辆进行跟踪估计得出车辆的运动轨迹和速度。一系列的视频实验表明,该方法简单可行而且对天气、光照变化、阴影有很强的适应能力。Intelligent transportation system（ITS） is the most effective solution to the growing urban traffic jam. Traffic information collection system is the fundament and precondition for traffic management system in ITS. Compared to other kinds of traffic information detectors, video-based traffic information detector has the advantages of abundant information , low installation cost and low maintenance cost. An approach to detect and track moving vehicles based on Kalman filter is presented in this paper. An adaptive background update is used. Foreground objects can be gained through segmenting foreground objects and corrupting and expanding to the binary images. Kalman filtering is used in vehicles tracking and the tracking parameters are the center of the rectangles which include the foreground objects. The track and the velocity of the vehicles can be attained by Kalman filtering tracking vehicles. Experiments on video streams show that the method is simple and feasible and detection and tracking performance has its strong adaptability to weathers, lighting changes, shadows.

关 键 词：自适应背景更新 车辆检测与跟踪 KALMAN滤波 

分 类 号：TN713[电子电信—电路与系统]