三角平面Halton点采样策略及其性能分析

Halton Points Sampling Strategy and Performance Analysis for Triangle Plane

针对一般随机采样的局限性,提出了基于Halton点采样的原理与方法.给出了Halton点的定义,分析了其算法复杂度,在此基础上给出了三角平面的采样策略,包括几何区域的划分、三角平面与矩形平面的坐标映射、采样点坐标计算原理;并将该方法与Jittered采样法、改进的LHS采样法进行了比较.实验结果表明,Halton点采样策略比一般的随机采样策略具有更好的采样点分布.

Aiming at the limi algorithm with Halton points are tation of common random sampling strategy, the sampling principle and adopted. The definition of Halton points is introduced and the algorithm complexity is analyzed, then the sampling strategy of triangle plane is introduced, including geometry region partitioning, coordinate mapping between triangle plane and rectangle plane, and computing principle to the coordinates of sampling point. The experiment results compared with Jittered sampling and improved LHS sampling are given. The experiment show that the sampling strategy using Halton points is better than the random sampling strategy at sampling points distribution.