-
摘要：视频跟踪是计算机视觉的一个重要的分支,最近几年越来越多的学者和研究人员把目光投向视频跟踪,随着研究的深入和跟踪技术的发展,视频跟踪渐渐成为生产生活中重要的一个组成部分,具有十分广泛的应用前景。它已经在人机交互、语音识别、公共卫生、人工智能、视频监控等诸多领域内得到了广泛的应用。
- 上一篇：车载“动中通”伺服系统的设计与研究
- 下一篇：机载SAR成像条带展宽技术初步研究+Matlab仿真
本文的目的是研究从视频流中提取人体运动，并对提取的运动进行处理，合成满足用户需求的运动的理论、方法和技术，探索利用计算机视觉和图形学再生 人体动画的方法。
本论文主要研究了基于加权颜色的粒子滤波算法,同时讨论了运动目标检测、特征提取和运动目标跟踪，进一步提取运动特征进行模型拟合。在运动目标检测方面本文对差分法进行了讨论，并用帧间差分法提取了运动目标。在运动目标跟踪方面,本文研究了基于卡尔曼滤波的多目标跟踪算法。在运动特征提取方面，本文利用基于形态学的判定算法，识别人体部位并使用OpenGl显示提取的骨架。在模型拟合方面，本文利用人体各个节点间存在的父子关系以及提取的各部位运动向量，在OpenGL平台上开发完成可控人体骨骼模型。关键词 视频检测 运动识别 骨架提取 3DS人体模型 opencv openGl8222
毕业设计说明书（论文）外文摘要
Title Human motion recognition and modeling
Abstract
As an important branch of computer vision technology, the video tracking system has got more and more attention from researchers lately. With the development of tracking technology and deeper research about this, video tracking has day by day become a significant part of manufacture field and has a widely applied prospective future.It has already been used widely in many fields such as Human-Computer Interaction, voice recognition, public health, artificial intelligence, and video supervision.
This dissertation aims to research the theory, approach and technology of how to extract human motion from video streams, and then process and synthesize the extracted motion so as to meet the requirement presented by users. And exploit the approach to apply computer vision and graphics to regenerate human animation.
This paper studied the particle filter algorithm based on weighted color, how to deal with the human motion,and further extraction of the movement characteristics of the model fitting. In moving target detection difference method are discussed.we use inter-frame difference method to extract moving target. This paper studies the multi-objective tracking algorithm based on Kalman filter tracking of moving targets,and the decision algorithm is based on morphology, to identify body parts and use OpenGl to display the extracted skeleton. A controlled human skeleton model is established based on OpenGL platform, Using the parent-child relationship between each node of the body and extract the various parts of the motion vectors.
Keywords VideoDetection MotionRecognition Skeleton extraction
3DS-human model opencv openGl
目录
1 引言 1
1.1 视频人体识别研究现状 1
1.2 主要应用领域 1
1.2.1 运动分析 1
1.2.2 智能安全监控 1
1.2.3 人机交互接口 2
1.2.4 虚拟现实 2
1.3 基础工具及平台介绍 2
1.3.1 Visual C++6.0 2
1.3.2 opencv 2
1.3.3 OpenGl 3
2 目标检测 3
2.1 视屏读入显示及帧预处理 3