针对蒙特卡罗法运算量大的问题,通过引入区域分解和逐级光线分裂技术,减少杂散光线与表面交点求解时所涉及的表面数及跟踪光线数,降低计算量.区域分解时,考虑子区域与表面数的匹配关系;并基于杂散光传播过程的模拟信息,逐级确定光线分裂倍数.以大衰减比抑制结构和多传感器系统为例,进行杂散光分析.结果表明,对复杂系统杂散光传播的蒙特卡罗法模拟,应用区域分解和逐级光线分裂技术,可以在保持计算精度的前提下,有效地提高计算效率.
A technique combined domain decomposition and staged ray splitting is presented to decrease computational complexity by reducing surfaces and rays for intersections solution.Domain decomposition is based on relation between sub-zone and number of surfaces.Splitting ratio is determined with propagation information of stray light.A suppression structure and a multi-sensor system are analyzed.It shows that the technique can improve computational efficiency while maintaining acceptable accuracy.