【摘要】 本文给出一种基于特定对象边缘抽取与绘制的特征可视化方法．该方法以在三维规则数据场中构造的一组Ｈａｒａｌｉｃｋ正交多项式为基，在以各网格点为中心的子区域内求灰度分布的拟合函数，通过对拟合函数二阶方向导数零交叉点的判断，抽取特定对象的边缘点，在此基础上，考虑在１８－连通意义下，三维边缘点的三角面片和多边形的表示与拼接。本文介绍这种特征可视化方法的实现，同时给出实验结果．更多还原

【Abstract】 This paper presents an approach for visualizing features of special objects contained in 3D rectilinear datafield. By constructing a group of 3D Haralick orthogonal polynomials, the distribution of value is fit in an area around each point in datafield. After that, the edge of a special object is detected by judging if there isa zero crossing occurred in the area, and the cells are divided into two groups: border cell and no-border cell. The border cell is rendered, together with some no-border cell, in polygons that show 18-adjacency. The realization, as well as some experiment results, is given in the paper.更多还原