光谱数据预处理策略选择及应用  

Selection and Application of Spectral Data Preprocessing Strategy

作　　者：孙嘉豪 张伟[1] 施鉴芩 李艳坤[1] SUN Jia-hao;ZHANG Wei;SHI Jian-qin;LI Yan-kun(Department of Environmental Science and Engineering,North China Electric Power University,Baoding,Hebei 071003,China)

机构地区：[1]华北电力大学环境科学与工程系,河北保定071003

出　　处：《计量学报》2023年第8期1284-1292,共9页Acta Metrologica Sinica

基　　金：国家级大学生创新创业训练计划(S202210079067);中央高校基本科研业务费(2017MS135)。

摘　　要：光谱预处理是建立光谱计量学模型中的重要环节,目前,已出现诸多光谱预处理方法。根据预处理目的与效果的不同,分类综述了散射校正、尺度缩放、平滑处理、基线校正与滤噪中常用光谱预处理方法的提出、发展、原理、特点和近年来的应用实例;根据目前预处理方法使用时存在的问题,结合应用实例探讨了预处理方法集成策略的原理、发展和特点,并对其研究趋势和发展前景进行了分析,以期为光谱预处理方法策略的选择和使用提供依据和参考。Spectral pre-processing is an important link in the establishment of spectral measurement metrology model,and at present,many spectral preprocessing methods have appeared.Based on the purpose and effect of the pretreatment,the origin,developments,theories,characterizes and actual application examples of common spectral preprocessing methods in scatter correction,scaling,smoothing,baseline correction and noise filtering areas were reviewed.According to the problems existing in the application of preprocessing methods,the principle,developments and characteristics of the ensemble strategy of preprocessing methods were discussed,and its research trend and development prospect were analyzed,for providing basis and references for the selection of spectral preprocessing methods and strategies.

关 键 词：计量学 光谱预处理 集成策略 数据分析 

分 类 号：TB99[一般工业技术—计量学] TB96[机械工程—测试计量技术及仪器]