在优化粒子滤波跟踪框架下, 设计并实现了一个结合多种图像特征、在多摄像机环境下跟踪人体运动的三

维人体运动跟踪系统1 通过定义三维人体模型、摄像机模型以及观测似然模型, 得到跟踪所需目标函数, 并使用优化

粒子滤波算法进行求解1 实验结果表明, 该系统能够对人体运动进行准确的跟踪和三维重建, 可应用于体育运动分

析和动画制作等领域1-A v ideo-based 3D human body motion t racking system is developed under the optimizat ion part

icle f ilter framework, w hich combines mult iple cues in multiple cameras environment1 By def ining the 3D

body model, camera model and image likelihood model, w e can define the tracking object funct ion and resolve

it using opt imized part icle f ilter1 T he experiment s show that t racking and reconstructing result s made

by our system are accurate and effect ive1 This system can be applied to several areas such as sport ing mot ion

analysis and human animat ion1