基于L曲线法正则参数优化反演光子相关光谱  

Inversion of photon correlation spectroscopy by means of L-curve optimizing regular parameters

作　　者：李绍新[1] 许晓安[2] 徐军发[1] 王永东[3] 

机构地区：[1]广东医学院基础学院,广东东莞523808 [2]广东技术师范学院教育技术与传播学院,广东广州510665 [3]东莞理工学院化学生物工程系,广东东莞523106

出　　处：《华中科技大学学报（自然科学版）》2010年第6期102-106,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(30972779);东莞市科技计划资助项目(2007108101025)

摘　　要：为解决多分散性光子相关光谱反演运算存在的典型不适定性问题,基于Tikhonov正则化策略采用L曲线法确定正则参数反演光子相关光谱.数值模拟的结果表明,在不同的模拟测量噪声下,该方法能成功反演各种粒径分布.聚苯乙烯乳球实验表明,L曲线法可应用于实际实验数据的反演,确定正则参数反演光子相关光谱,是一种有效的光子相关光谱数据处理方法.Inverse photon correlation spectroscopy （PCS） data with polydispersity is the first Fredholm integral equation problem, which is ill-posed. Inverse PCS data with L-curve based on Tikhonov regular strategy was studied. Numerical experiments show that these methods can invert different distributions with different assumed noise. The latex experiments suggest that L-curve mehods can be applied to inverse real experiments data. L-curve methods choosing regular parameter for inverting PCS data is an effective means of dealing with PCS data.

关 键 词：光子相关光谱 L曲线 正则化 第一类FREDHOLM积分方程 参数反演 

分 类 号：O433.1[机械工程—光学工程]