利用衰减全反射附件直接采集猪皮革、牛皮革和羊皮革等3种天然皮革的红外光谱图，分析了皮革氨基酸结构中主要基团的红外吸收峰的特点和位置。对光谱数据预处理后选择合适的光谱区域进行主成分分析，首先构筑主成分得分空间对样品直接进行判别，发现由于牛皮革样品和羊皮革样品的红外光谱图极为相似，它们的主成分得分空间区域大部分重叠在一起。然后以猪皮革、牛皮革和羊皮革等3种天然皮革样品作为3类组分，采用主成分回归方法建立相应的判别模型，3类组分的模型预测相关系数为0.965~0.990，均方差均不大于0.122，预测均方差均不大于0.213，模型具有较好的稳定性。利用检测集样品对模型进行验证，根据各组分的预测值可以快速方便地对3种皮革样品进行识别。

The IR spectra for 3 natural leathers of pig, cow and sheep were collected directly by attenuated total reflectance accessory. The characteristics and locations of IR absorption peaks of main groups in amino acid structure of leather were analyzed. The principal component analysis was preceeded after preprocessing of spectra data and choosing the appropriate areas. Firstly, the samples were identified directly via building principal component score space. It was found that the principal component score space of cow leather sample overlapped with sheep leather sample mostly due to the extremely similar with spectra of the 2 leather samples. Then the discriminant model of principal component regression was established after using 3 natural leather samples of pig, cow and sheep as 3 components. The prediction correlation coefficients of the model for the 3 components were in the range of 0.965-0.990, with RMSEC and RMSEP were no more than 0.122 and 0.213, respectively, which exhibited well stability. The model was verified using the test samples. The results showed that the 3 leather samples could be identified quickly and easily according to the predicted value of each component.