摘要
近红外光谱定性分析在纺织品检测领域内已有大量研究,这些研究有着各种不同的定性模型方法。本文介绍了常用的模式识别方法,如聚类分析、距离判别法、主成分分析法、SIMCA(簇类软独立法)、PLS-DA(偏最小二乘-判别分析法)、人工神经网络法、SVM(支持向量机法)、合格性测试及相似度匹配等,并分析了每种方法的适用情况。总结了近红外光谱法的应用存在的问题及现状并进行了展望。
Near-infrared spectroscopy qualitative analysis has been extensively studied in the field of textile testing,with a variety of different qualitative modeling methods being developed.Each method has its own advantages and disadvantages.This paper introduces commonly used pattern recognition methods,such as cluster analysis,distance discriminant method,principal component analysis(PCA),SIMCA(Soft Independent Modeling of Class Analogy),PLS-DA(Partial Least Squares-Discriminant Analysis),artificial neural network method,SVM(Support Vector Machine),qualification testing,and similarity matching,and analyzes the applicability of each method.The paper summarizes the existing issues and current status of near-infrared spectroscopy applications and provides an outlook for its future development.
作者
贾凯凯
周兆懿
黄怡婧
JIA Kaikai;ZHOU Zhaoyi;HUANG Yijing(Shanghai Institute of Quality Inspection and Technical Research,Shanghai 200040,China)
出处
《中国纤检》
2025年第5期50-55,共6页
China Fiber Inspection
关键词
近红外光谱
定性分析
鉴别
纺织品检测
主成分分析
PLS-DA
near-infrared spectroscopy
qualitative analysis
identification
textile testing
principal component analysis
PLS-DA
