中文摘要：

海洋溢油是主要环境灾害之一,而且近年来其发生频率呈上升趋势。快速地对油品进行种类鉴别、来源评估有利于及时采取应急措施,因此具有重要意义。采用衰减全反射傅里叶变换红外光谱法（ATR-FT-IR）对25种不同来源的油品进行了检测,用不同数据预处理方法对原始光谱进行了预处理,继而用主成分分析（PCA）和系统聚类分析（HCA）方法对光谱进行了分类鉴别。结果表明用多元散射校正（MSC）和连续小波变换（CWT）方法进行数据预处理可以提高分类的准确性,使分类结果与油样的实际来源一致。该方法对正构烷烃差异较大的油品进行了很好的区分,但对差异较小的油品其分辨能力仍有一定局限性。因此提供了一种快速的油品鉴别方法,可用于溢油事件的初步鉴定,从而为油品的进一步鉴定提供有用信息。

英文摘要：

Abstract In the present work, the combination of attenuated total reflectance-Fourier transform infrared spectrometry （ATR- FTIR） and pattern recognition, including principal components analysis （PCA） and hierarchical cluster analysis （HCA）, is used as a fast and convenient analytical tool to classify oil samples. Twenty five samples including crude oils and fuel oils with differ- ent total contents of n-alkanes were analyzed. It was found that multiplicative scatter correction （MSC） and continuous wavelet transform （CWT） as a pretreatment method could improve the classification results of pattern recognition. The classification re- sults were proved to be in agreement with the origin of the oil samples. The oils with high content of n-alkanes and those with low content were classified clearly by this developed method, but it still had some constraint to differentiating oils with little difference. The present work provides a feasible method for quick classification of oils, which can be used for the initial identification of spill oils and afford useful information for the further identification of the oils.