视频考桩系统设计与实现

发布时间：2018-03-26 16:06

本文选题：视频监控　切入点：角点法　出处：《电子科技大学》2014年硕士论文

【摘要】：视频监控技术既是一门多学科的交叉与结合的科学性学科,也是当前被广泛采用的一种技术手段。视频监控是当前的研究热点,随着计算机技术的发展与图像处理技术的提高,视频监控技术正越来越深入地渗透到娱乐、教育、医疗、安防、运动等领域,总体来说,视频监控技术呈现出数字化、网络化、集成化、智能化的发展趋势。本文通过对传统桩考系统的分析研究,结合最新的视频监控技术,深入研究角点法与粒子群优化算法等理论方法,实现自动跟踪所监控运动的车辆目标,在此基础上设计一套完整的视频桩考系统,在任意天气条件下均能准确、无误地工作。本文研究的主要内容包括:在分析了传统桩考系统的需求与现状的基础上,首先提出了基于角点检测算法的车辆轮廓的提取算法,针对角点法中常用的直线投影法和SUSAN检测法两种算法进行了优劣分析与论证,并对SUSAN算法进行了自适应阈值的改进;然后描述了当前桩考系统视频监控用到的三种算法,即光流场法、背景差分法、帧间差分法,及对各个算法优缺点进行评测的基础上,提出了基于粒子群优化的车辆跟踪算法;最后,在此基础上,使用Microsoft Visual Studio 2008开发工具在windows环境下设计实现了软件算法与系统整体功能结构。经实验表明,经过改进的SUSAN检测算法和粒子群优化算法能很好地符合视频监控中的车辆轮廓识别和车辆跟踪的需求,基于该算法设计的本视频桩考系统对移动目标的提取识别稳定性高,速度快,效果好,适应任意天气条件下任意车辆的视频识别与检测,准确率高,提高了考试系统的公平、公正、准确性,达到预期目标。
[Abstract]:Video surveillance technology is not only a interdisciplinary and scientific subject, but also a widely used technical means. Video surveillance is a hot research topic at present, with the development of computer technology and the improvement of image processing technology. Video surveillance technology is more and more deeply penetrated into entertainment, education, medical care, security, sports and other fields. In general, video surveillance technology presents digitization, networking, integration, etc. Through the analysis and research of the traditional pile test system and the latest video surveillance technology, this paper deeply studies the angle point method and particle swarm optimization algorithm to realize the automatic tracking of the moving vehicle target. On this basis, a complete video pile test system is designed, which can work accurately and unmistakably under any weather conditions. The main contents of this paper are as follows: based on the analysis of the requirements and present situation of the traditional pile test system, Firstly, a vehicle contour extraction algorithm based on corner detection algorithm is proposed. The advantages and disadvantages of linear projection method and SUSAN detection algorithm, which are commonly used in corner method, are analyzed and proved, and the adaptive threshold of SUSAN algorithm is improved. Then, the paper describes three algorithms used in video surveillance of piling test system, namely optical flow field method, background difference method, inter-frame difference method, and evaluates the advantages and disadvantages of each algorithm, and proposes a vehicle tracking algorithm based on particle swarm optimization. Finally, on this basis, the software algorithm and the whole function structure of the system are designed and implemented under the windows environment by using the Microsoft Visual Studio 2008 development tool. The improved SUSAN detection algorithm and particle swarm optimization algorithm can well meet the requirements of vehicle contour recognition and vehicle tracking in video surveillance. The video pile test system designed based on this algorithm has high stability for the extraction and recognition of moving targets. The speed is high, the effect is good, and the video recognition and detection of any vehicle under any weather condition is suitable, the accuracy is high, and the fairness, justice and accuracy of the examination system are improved, and the expected goal is achieved.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN948.6;TP18

【相似文献】