视频动态目标压缩感知跟踪技术及实现
摘要视频动态目标跟踪在视频监控，军事目标探测，智能交通等领域具有广泛的应用前景。本文围绕视频动态目标跟踪方法展开研究。
本文首先综述了视频动态目标的国内外发展，在此基础上本文主要工作主要分为以下三个方面：
（1）研究和分析了经典的目标跟踪算法，包括背景背景差分法，帧间差分法，光流法，TLD跟踪算法，MeanShift跟踪算法，STC跟踪算法。
（2）重点研究了基于压缩感知的跟踪算法，并且对比性的给出了TLD跟踪算法，MeanShift跟踪算法，STC跟踪算法的实现流程和实验评估。
（3）设计了压缩感知跟踪算法的视频演示系统，分析了压缩感知跟踪算法的性能，实验表明压缩感知跟踪算法性能较好。49131
关键词：目标跟踪 压缩感知 特征提取 特征降维 朴素贝叶斯分类器 在线学习 对比实验
毕业论文设计说明书外文摘要
Title Video dynamic target compressive tracking technology and its implementation
Abstract
Video dynamic target tracking has a wide application prospect in many fields including video surveillance，military target detection,intelligent transportation and so on.In this paper,we mainly focus on video dynamic target tracking algorithm.
In this paper,firstly,we summarizes the domestic and foreign development of video dynamic target tracking methods.Then based on this,the main works are listed as follows:
(1)The investigation and analysis on classical object tracking algorithms are made in this paper such as background difference method,the inter frame difference method,the optical flow method,TLD tracking method,MeanShift Tracking method and STC tracking method.
(2)We mainly focus on the compressive tracking algorithm in this paper and give the comparative implementation process and experiment evaluation of TLD tracking method,MeanShift tracking method and STC tracking method.
(3) A demo system of compressive tacking algorithm was designed in this paper. We analyzed the performance of compressive tracking method.The experimented results show that the compressive tracking method has a good performance.
Keywords Object tracking Feature extraction Feature dimension reduction Naive Bayes Classifier Online study Comparative experiment
1 引言 1
1.1 研究背景 1
1.2 国内外现状分析 1
1.3 存在的问题 3
1.4 论文的主要工作及内容安排 3
2 运动目标检测与跟踪 5
2.1 引言 5
2.2 运动目标检测的常见方法 5
2.2.1 背景差分法 5
2.2.2 帧间差分法 5
2.2.3 光流法 6
2.3 目标跟踪算法 7
2.3.1 目标跟踪算法的基本流程 7
2.3.2 目标跟踪算法的分类 7
2.4 一些经典的目标跟踪算法 9
2.4.1 TLD跟踪算法 9
2.4.2 Mean-Shift跟踪算法 10
2.4.3 STC跟踪算法 13
2.5 本章小结 15
3 压缩感知跟踪算法