基于自适应Tikonhov正则化约束的红外光谱重构及在药品辨识中的应用（英文）  

Infrared spectra reconstruction with adaptive tikhonov regularization and its applications in drug identification

作　　者：刘秋霞[1] 肖莎[1] 郭振辉[1] 刘海[1] 

机构地区：[1]广州军区广州总医院老年重症医学科,广东省老年感染与器官功能支持重点实验室,广州市老年感染与脏器功能支持重点实验室,广东广州510010

出　　处：《计算机与应用化学》2017年第3期234-238,共5页Computers and Applied Chemistry

基　　金：Supported by Guangzhou science and technology plan project health care coordination and innovation of major projects(201508020253);Science research project of Guangzhou science and technology project(2014J4100033);Guangdong Provincial Key Laboratory of Geriatric Infection and Organ Function Support(2012A061400010);Guangzhou Provincial Key Laboratory of Geriatric Infection and Organ Function Support(7411675081103)~~

摘　　要：针对红外光谱常常遭受到谱带重叠和随机噪声的问题,本文提出了一种基于自适应Tikhonov正则化约束的红外光谱信号重构方法。该方法能有效的抑制平坦区域的噪声和保存谱带的结构信息,不同光谱区域的结构信息可以通过自适应项来区分。本文所建立的数学模型能有效通过交替最小值最优方法求解。对比实验结果表明,该方法能有效的分裂重叠的谱带并抑制Poisson噪声。重构的红外光谱易于提取光谱特征并解释未知的药品成分。Infrared spectrum often suffers from common problems of band overlap and random noise. In this paper, a blind reconstruction algorithm with adaptive Tikhonov regularization is introduced. The structure information in different spectral regions is incorporated into regularization by using the adaptive term to distinguish spectral structure from noise region. The proposed algorithm can effectively suppress the noise in fiat regions as well as preserved the spectral structure and detailed information. Moreover, we show that the resulting minimization problems can be efficiently solved by the alternation minimization method. The comparative experimental results including quantitative and qualitative analysis manifest that the proposed method can effectively split the overlap bands, suppress Poisson noise and reduce artifacts. The deconvolution infrared spectrum is more convenient for extracting the spectral feature and interpreting the unknown drugs.

关 键 词：光谱数据处理 红外光谱学 光谱分析 药品识别 

分 类 号：Q71[生物学—分子生物学] TQ015.9[化学工程]