摘要
根据测量的数据点集,由梯度关系得到采样点和指示函数的积分关系,根据积分关系用划分块的方法获得点集的向量场,计算指示函数梯度场的逼近,构成泊松方程。根据泊松方程使用矩阵迭代求出近似解,采用移动立方体算法提取等值面,对所测数据点集重构出被测物体的模型,泊松方程在边界处的误差为零,因此得到的模型不会存在假的表面框。
An integral relationship between sampling points and indicator function is derived from gradient relationship and based on measured data set of points. According to integral relationship, the points are partitioned into patches to obtain vector field. Then the approximation of gradient field of instructor function is computed to form the Poisson equation. The approximate solution is got by＇matrix iteration based on Poisson equation. Marching cubes algorithm is used to extract the isosurface by selecting isovalue and the model of measured object is reconstructed according to the measured sampling points set. Because the error of Poisson equation is zero in boundary, the obtained model will not have spurious surface frame.
出处
《计算机应用与软件》
CSCD
2009年第4期227-228,231,共3页
Computer Applications and Software
关键词
隐函数
向量场
泊松方程
重构
Implicit function Vector field Poisson equation Reconstruction
