Chemometric Analysis of Multisensor Hyperspectral Images of Precipitated Atmospheric Particulate Matter
Chemometric Analysis of Multisensor Hyperspectral Images of Precipitated Atmospheric Particulate Matter
复制标题
DOI：
10.1021/acs.analchem.5b02272
复制
发表时间：
2015-09-15
影响因子：
7.4
通讯作者：
Lohninger, Hans
中科院分区：
文献类型：
--
作者：
Ofner, Johannes;Kamilli, Katharina A.;Lohninger, Hans
多传感器高光谱数据的化学计量学分析允许对沉淀的大气颗粒物进行全面的基于图像的分析。大气颗粒物沉淀在铝箔和分析拉曼显微光谱，随后通过电子显微镜和能量色散X射线光谱。所有获得的图像都是面积为100 × 100 μ m（2）的同一斑点。这两个高光谱数据集和高分辨率扫描电子显微镜图像被融合成一个多传感器高光谱数据集。这个多传感器数据立方体进行了分析，使用主成分分析，层次聚类分析，k-均值聚类，顶点分量分析。通过对多传感器数据进行详细的化学计量分析，可以对沉淀颗粒进行广泛的化学解释，其结构和组成有助于全面了解大气颗粒物。
The chemometric analysis of multisensor hyperspectral data allows a comprehensive image-based analysis of precipitated atmospheric particles. Atmospheric particulate matter was precipitated on aluminum foils and analyzed by Raman microspectroscopy and subsequently by electron microscopy and energy dispersive X-ray spectroscopy. All obtained images were of the same spot of an area of 100 X 100 mu m(2). The two hyperspectral data sets and the high-resolution scanning electron microscope images were fused into a combined multisensor hyperspectral data set. This multisensor data cube was analyzed using principal component analysis, hierarchical cluster analysis, k-means clustering, and vertex component analysis. The detailed chemometric analysis of the multisensor data allowed an extensive chemical interpretation of the precipitated particles, and their structure and composition led to a comprehensive understanding of atmospheric particulate matter.