基于视频图像处理的车辆检测与跟踪方法研究

发布时间：2018-03-28 18:07

本文选题：视频图像处理　切入点：车辆检测　出处：《长安大学》2014年硕士论文

【摘要】：随着经济的发展和社会的进步，对交通运输的各种需求也迅猛增长，现有交通路网的管理能力与管理设施已经不能满足日益增长的交通流量需求，交通运输问题日益严重，交通拥挤，时常有车辆随意变道的情况发生，引起了许多不必要的交通事故。为了解决这些交通问题，智能交通系统（ITS）的研究被提到了重要位置。运动车辆检测与跟踪系统作为ITS的重要组成部分，成为了目前的研究热点。车辆跟踪技术是实现视频车辆监测的关键技术之一，如何实现基于视频图像处理的车辆检测，以及车辆跟踪的算法研究成为了国内外研究的重要问题。本文将车辆跟踪系统分为车辆检测、车辆跟踪两个主要的技术模块，，并对各模块的主要技术方法进行了介绍。在车辆检测模块，通过对几种常用的检测方法进行分析比较，确定采用基于背景差分法的检测方法进行车辆检测，重点研究了该算法中背景的提取和更新、阈值的选取方法等关键技术。在车辆跟踪模块，重点对车辆跟踪的几种算法进行分析比较，确定使用基于目标车辆质心的跟踪算法进行跟踪标记，最后在跟踪标记的基础上得到追踪目标车辆的运行轨迹。运用MATLAB分析软件及实例对所选择的算法进行验证，确定所选择的算法是否可以达到预期效果。
[Abstract]:With the development of economy and social progress, but also the rapid growth of demand for the various transportation facilities, management ability and management of existing traffic network has been unable to meet the growing traffic demand, traffic problems have become increasingly serious, traffic congestion, vehicles often have occurred change situation, caused a lot of unnecessary traffic accident. In order to solve these traffic problems, intelligent traffic system (ITS) research has been raised to an important position. The moving vehicle detection and tracking system as an important part of ITS, has become a research hotspot at present. Vehicle tracking is one of the key technologies of video vehicle monitoring, how to realize vehicle detection based on video image processing research on vehicle tracking algorithm, and has become an important problem of research at home and abroad.
The vehicle tracking system for vehicle detection, vehicle tracking two technical modules, and the main technical methods of each module are introduced. The vehicle detection module, based on several commonly used detection methods of analysis and comparison, determine the vehicle detection background difference detection method based on the method, focus on background extraction and update the algorithm, the key technology of threshold selection methods. In the vehicle tracking module, focuses on several algorithms of vehicle tracking analysis and comparison, determine the use of the target vehicle centroid tracking algorithm based on tracking markers, finally get the trajectory tracking target vehicle based on tracking markers.
MATLAB analysis software and examples are used to verify the selected algorithms and determine whether the selected algorithm can achieve the desired results.

【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495

【参考文献】