PCA-LVQ法及其在RS-FTIR大气环境监测数据处理中的应用

将主成分分析(PCA)用于遥感傅里叶变换红外光谱(Remote Sensing Fourier Transform Infrared:RS-FTIR)的特征提取,结合学习矢量量化(LVQ)神经网络,实现了PCA-LVQ对大气中的8组分混合体系进行快速定性分析的建模方法。并与单纯的LVQ神经网络、反向传播人工神经网络(BP-ANN)得到的结果进行了比较。PCA-LVQ显示出较好的处理数据的能力,它不仅提高了运算速度,而且提高了模型的预测准确度,分类精度达到91.7%。PCA-LVQ的这一预测精度及运算速度,足以满足遥感傅里叶变换红外光谱对大气中有毒气体的实时、在线监测的需要。

Application of PCA-LVQ in Processing of RS-FTIR Data for Atmospheric Environment Monitoring

Principle component analysis(PCA),in combination with learning vector quantization(LVQ) neural network,was applied to extract information features from remote sensing Fourier transform infrared spectroscopic(RS-FTIR) data,and rapid qualitative analysis of eight components in atmosphere was demonstrated using this PCA-LVQ approach.Compared with LVQ neural network and BP-ANN,PCA-LVQ can process data more efficiently,with a much faster computation speed and a higher prediction accuracy.The classification precision was as high as 91.7%.The high prediction accuracy and fast computation speed of PCA-LVQ can meet the needs of real-time and on-line monitoring of atmospheric toxic species via FTIR.