基于OpenCV的运动目标检测与跟踪
毕业论文关键词 背景差分 目标检测 目标跟踪 CamShift 混合高斯
Title Moving Object Detection and Tracking Based on OpenCV
Abstract
Detection and tracking of moving target in video stream is the core of many visual
applications. The OpenCV provides a variety of mathematical functions and image processing
methods to meet each need, replacing the shackles of the underlying basic codes, improving the
social production efficiency. In general, the method of background subtraction is usually used
for the static video surveillance, and the target is highlighted from the video and background. In
this paper, the general process of the background difference detection object, all the typical
background modeling algorithms and their advantages are described, then the background model
of the GMM is described
The Gaussian mixture background model of the background difference method for target
detection, using CAMSHIFT algorithm is applied to the tracking of targets. The experimental
results show that for similar to the simple static background, the recognition rate by this program
is quite high, and tracking are rarely seen out with satisfactory results.
Keywords background subtraction,target tracking and detection,CamShift,GMM
目次
1绪论 1
1.1 引言 . 1
1.2 国内外研究与应用现状 . 1
1.3 OpenCV简介 2
1.4 开发环境搭建 . 2
1.5 本章小结 . 3
2图像预处理 4
2.1 图像噪声的概念与分类 . 4
（1）按产生的原因分类 . 4
（2）按噪声频谱分类 . 5
（3）按信号关系分类 . 5
（4）按概率密度函数分类 . 5
2.2 直方图均衡 6
2.3本文中所采取的滤波 . 8
2.4 数学形态学 9
2.4.1膨胀与腐蚀 10
2.4.2 开运算与闭运算 . 12
2.5 本章小结 12
3运动目标检测 13
3.1 运动目标检测的基本方法 13
3.1.1 帧间差分法 . 13
3.1.2 光流法 . 13
3.1.3背景差分法 15
3.2 本文的具体实现 18
4运动目标跟踪 20
4.1 CamShift算法 . 20
4.2 程序的运行及演示 23
4.3 本章小结 24
结 论 . 25
致 谢 . 26
参考文献 . 27
1 绪论
1.1 引言
在过去十年中许多用以识别实时运动的方法被提出。用于类似视频监控，交通监测，车
辆追踪，自主导航，计算机视觉等生产生活中。对于视觉信息的捕获、辨别与分析成为了研
究领域的重要课题。智能化的视频监控就是从最基础的实现对目标的检测识别到运用模型等
方法的跟踪技术，最终并轨人工智能实现对目标行为的理解判别。识别算法采用一系列图片
下一篇：基于自由能的遥感图像质量评价