摘要近几年来，基于视觉的目标检测与跟踪技术迅猛发展。它可以运用到智能监控、人机交互、军用装备等很多领域，具有很高的科研和实用价值。

本论文主要对目标检测和跟踪的方法进行了研究，设计并实现了一款可以进行动态目标检测和跟踪的系统。这款系统由目标检测与跟踪程序、计算机、摄像机云台以及其他硬件设备组成。它的主要功能是检测动态目标和进行鲁棒的跟踪，并利用PELCO协议通过串口发送命令来控制摄像机云台的运动。它对目标的遮挡，翻转，形变等不敏感。82526

系统的目标检测算法采用背景减除法，目标跟踪采用Camshift算法，这两种算法运算简单，适用于实时系统。文章对这两种算法有详细的介绍和分析，并用实验结果来验证说明。

毕业论文关键词  目标检测与跟踪  PELCO协议  背景减除法  Camshift

毕业设计说明书外文摘要

Title    Vision-based object detection and tracking       

Abstract In recent years, vision-based target detection and tracking technology are developing rapidly。It can be applied to many areas,such as intelligent monitoring,human-computer interaction,military equipment and so on, with high scientific and practical value。

This paper mainly studies the method of target detection and tracking, and designs and implements a system which can detect and track the moving objects。This system consists of target detection and tracking procedures, computer, PTZ camera and other hardware devices。Its main function is to detect moving targets,carry out robust tracking,and use PELCO protocol  to send commands through the serial port to control the PTZ camera’s movement。It is not sensitive to the target of the occlusion, flip, deformation and so on。

Target detection algorithm of system uses the method of background subtraction  and target tracking algorithm is based on Camshift。These two algorithm are simple and apply to real-time systems。 In the article, there is both a detailed description and analysis of algorithms,and experimental results are used to verify the description。

Keywords  Object detection and tracking  PELCO protocol  Method of background subtraction  Camshift

目次

1  绪论 1

1。1  课题研究的背景 1

1。2  课题研究的目的和意义 1

1。3  国内外研究发展状况 2

1。4  论文主要内容安排 3

2  目标检测与跟踪的常用方法 4

2。1  目标检测 4

2。2  目标跟踪 8

3  目标检测的原理与实现 10

3。1  背景减除 10

3。2  目标分割 11

4  目标跟踪的原理与实现 14

4。1  RGB与HSV颜色空间 14

4。1。1  RGB模型 14

4。1。2  HSV模型 14

4。1。3  RGB与HSV的互相转换 15

4。2  核函数和颜色直方图