毕业论文5.30.doc

基于视频的人体骨架步态协同研究 摘要 基于视频的人体运动分析是计算机视觉研究领域的重要课题之一，其研究在人体动画、游戏、人机交互、视频监控、体育运动分析等邻域均有着广阔的应用前景。本文从本课题基于图像处理、计算机视觉、计算机动画、骨运动学等学科，结合计算机动画领域的发展方向，对基于视频的人体骨架提取以及步态协同方式进行了研究。 首先，通过对彩色视频图像进行序列化处理、灰度图像转化、正则滤波降噪、帧差法确定背景模型、背景差分法定义运动图像区域、形态学处理、边缘检测等步骤，清晰、准确地定义了目标人体步态轮廓。其次，通过分析目前人体骨架提取的主要算法，结合人体骨架在步态协同中的关键性，提出了“带膝关节的星形人体骨架模型”。 然后，结合骨运动学的理论对步态协同方式进行程序上的分解，即：关节点运动、骨骼动作、运动姿态。并利用四元数，对带膝关节的星形人体骨架的特征点进行位置角度信息的提取。最后，在MATLAB平台上建立基于群组关系的三维虚拟动画模型，将Simulink作为接口、特征点位置角度信息作为输入信号对虚拟人物进行控制，使之按照预定的姿态进行运动，实现步态协同。 本文的研究对人体骨架模型的描述方式以及计算机动画制作方式的发展具有一定的意义。 关键词：带膝关节的星形人体骨架，步态协同，特征点，四元数，虚拟现实  HUMAN BODY SKELETON EXTRACTION AND SYNERGY KINEMATICS BASED ON VIDEO ABSTRACT Video-based human motion analysis is one of the important topics in the field of computer vision research,the research on human animation、games、human-computer interaction、video surveillance、sports analysis both have broad application prospects . This article from the issue based on image processing、computer vision、computer animation、kinematics bones and other disciplines , combined with the development direction of the field of computer animation , based on human skeleton extraction and synergistic way video gait were studied . First, a color video image sequence processing, grayscale image conversion, regular noise filtering, frame difference method to determine the background model, background subtraction method to define the motion picture area, morphological processing, edge detection step, clearly and accurately defines the target human gait silhouette. Secondly, through the analysis of the main current of human skeleton extraction algorithm, combined with human skeleton in gait coordination of critical, proposed the Star of the human skeleton model with the knee. Then, combining the theory of gait kinematics bone collaborative manner on the decomposition process, namely: movement joints, bones action, athletic stance. And using quaternions, the feature points of the star with knee human skeleton extraction position angle informati