【摘要】 提出了一种三维数据场形状特征的可视化方法．首先讨论数据场中特征的定义、特征的属性以及有关特征可视化方法研究中存在的问题，提出对多时间片序列数据场进行特征跟踪的方法．其次给出三维数据场特征抽取的方法，并提出一种三维边界特征的表示方法，所抽取的特征用三维曲面来绘制，用主分量分析法估计特征点处的法向，通过构造一个规则的距离场绘制三维特征边界曲面．最后讨论算法的实现与存在的问题．更多还原

【Abstract】 This paper presents a visualization method of 3D shape based on feature extraction in time dependent three dimensional data field.A feature tracking method is provided to track the evolution of the features.A method to represent 3D feature by using three dimensional surface is given.The normal at a given point on the surface is estimated by principal component analysis method.The surface is rendered by constructing the iso surface of a regular data field that is formed by a signed distance function.更多还原