毕业论文移动版

摘要公共安全问题日益突出，越来越受到大家的关注与重视。公共场所内异常事件的及时发现将无疑有助相关部门对事件的处理，能大大降低各种损失。近年来，在智能监控的高速发展下，异常事件的检测已经成为图像处理、模式识别等领域的重点研究热点。异常事件的检测有两个基本的问题，一是正常基本事件的表示，二是异常事件检测模型的建立。检测方法有两个环节，先是视频数据的特征提取，然后对事件行为模式建模分析与识别。最后将测试特征与预先标定特征进行匹配，从而达到异常事件检测的目的。本文提出一种基于低秩分解的异常事件检测方法，将从训练视频提取到正常特征进行低秩分解获取字典，以测试特征关于字典稀疏表示的重建误差作为判断异常事件的标准，重建误差较大的，既是异常事件。82584

毕业论文关键词  智能监控  异常事件检测  低秩分解  稀疏表示

毕业设计说明书外文摘要

Title       Abnormal event detection method based on         low-rank decomposition               

Abstract Public security problem increasingly prominent, more and more concern and attention by all of us。Public places within the exception events in time will certainly help to related department's handling of the event, can greatly reduce various losses。In recent years, under the rapid development of intelligent monitoring, abnormal event detection has become the focus of the fields such as image processing, pattern recognition research hot spot。Abnormal event detection, there are two basic questions, one is the normal basic event said, the second is the establishment of the abnormal event detection model。Detection method has two part, first the feature extraction of video data, and then modeling analysis and recognition of events behavior patterns。The last match the test features and calibration in advance, so as to achieve the purpose of the abnormal event detection。This paper puts forward a kind of abnormal event detection method based on low-rank decomposition, extracted from the training video to the normal characteristics of low-rank decomposition to get the dictionary, to test the characteristics of the reconstruction error of sparse representation dictionary as a standard of judging abnormal events, the reconstruction error is bigger, is both an abnormal event。

Keywords  Intelligent monitoring, Abnormal event detection, Low-rank decomposition, Sparse representation

目   次

1  绪论 1

1。1  研究背景与目的 1

1。2  国内外研究现状 2

1。2。1  基本事件表示 3

1。2。2  异常事件检测模型建立 4

1。3  论文结构 6

2  相关理论基础 7

2。1  预备知识 7

2。2  低秩简述 7

2。3  低秩矩阵恢复 8

2。3。1  经典主成分分析法PCA 8

2。3。2  鲁棒主成分分析RPCA 8

2。3。3  鲁棒主成分分析RPCA的求解 9

2。3  低秩矩阵表示 12

2。4  低秩子空间分解算法 (责任编辑：qin)