摘要：视频图像的运动目标分析技术可以对视频图像中感兴趣的运动目标进行检测、分类、跟踪、识别、场景理解等。运动目标检测、分类是将运动目标从序列图像背景中分离出来，是目标跟踪、行为识别、场景理解等后期处理的基础。62262
当今社会中，车辆作为主要交通工具之一，处处可见其的影子，为了能更方便更系统地对车辆进行管理，视频图像处理技术在智能监控系统中经常被使用，如怎么检测视频中运动的车辆，怎么跟踪指定的车辆，怎么检测车辆的型号等一系列的功能，形成了车辆检测系统，系统地分析交通车辆视频，从中得到需要的信息。
本文主要对车辆视频中运动的车辆的检测进行了研究，先对基础的图像处理技术进行了介绍，之后介绍了几种常见的，被广泛运用的经典算法，着重研究了其中的高斯混合模型和支持向量机两种检测算法，背景减除法中的关键是背景模型的建立和更新，在支持向量机中，主要是通过建立检测正负样本，结合HOG特征，从而检测出目标，之后用MATLAB工具进行相关实验，实现了两种方法对车辆视频中的运动车辆进行检测，检测结果比较理想，体现了两种方法的各自的特点。
毕业论文关键词： 高斯混合模型；支持向量机；HOG特征；运动目标检测
Study on Moving Target Detection Algorithm for Vehicle Detection System
Abstract: Moving target video image analysis techniques can analysis moving target in detection, classification, tracking, identification, scene understanding and etc. Moving object detection, classification is a moving target sequences isolated from the background image, is the basis for target tracking, behavior recognition, scene understanding and other post-processing.
Today's society, as one of the main transport vehicle, cars can be seen everywhere. In order to be more convenient and more systematic management of vehicles, video image processing technology in the intelligent monitoring system is often used. Such as how to detect vehicle motion video, and how to track specific vehicles, how to detect a vehicle model and a series of functions, which formed a vehicle detection system. It can do systematic analysis of video transport vehicles, and derive the required information.
This paper focuses on the study, the video motion detection of vehicle. Firstly, the basis of image processing techniques were introduced. Then, some of classical algorithms which are widely used and common were introduced. This paper focus on the two algorithms: Gaussian mixture model and Support Vector Machine. The key of the background subtraction is the establishment and the update of the background model. In support vector machine, we can detect the target through the establishment of the positive and negative testing samples combined with HOG features. Then, do some experiments by using MATLAB. The detection of moving vehicles with these two algorithms were implemented by making experiments. The experiments reflect the characteristics of two algorithms.
Keywords: GMM; SVM; Hog; moving target detection
目录
摘要 i
Abstract i
目录 iii
1 绪论 1
1.1 研究目的和意义 1
1.3 本文工作 3
2 HOG车辆检测系统中的运动目标检查算法研究:http://www.751com.cn/jisuanji/lunwen_68422.html