摘要
提出了一种散射介质光学参量的干涉测量方法.在蒙特卡罗数值模拟的基础上引入聚焦高斯光源模型和基于比例缩放的压缩算法,实现了对光子后向散射分布和干涉特性的快速模拟.利用计算数据训练了前向人工神经网络进行了介质光学参量的反向求解.实验中以脂肪乳注射剂和生物染色剂印度墨水组成的悬混液为样本,通过光学干涉系统测量得到了后向散射光依赖于深度的干涉光强分布,并对其中散射和吸收系数进行了反向计算.所得结果能够正确反映散射和吸收系数与散射和吸收物质浓度之间的线性关系,验证了该方法的可行性.
A method of coherent measuring optical parameters of scattering medium is presented based on Monte Carlo simulation.The condensed scaling method is introduced to accelerating process of calculation.A tri-layer structure BP-artificial neural network is developed for reverse calculation of scattering and absorption coefficients of medium.Fat emulsion and stain mixture of IntralipidTM and India ink are taken as scattering and absorbing sample for coherent measurement by fiber Michelson interferometer.Results of experiments show the linear relation between scattering or absorption coefficients and concentration of IntralipidTM or India ink.The method is helpful to nondestructive testing components of turbid material.
出处
《光子学报》
EI
CAS
CSCD
北大核心
2012年第7期781-785,共5页
Acta Photonica Sinica
基金
国家自然科学基金(No.61008057)
广东省自然科学基金(No.9151065005000008)资助
关键词
蒙特卡罗
干涉测量
人工神经网络
高斯光束
Monte Carlo
Coherent measurement
Artificial neural network
Gaussian beam
