本文研究了利用近红外光谱技术结合Adulterant Screen算法建立的咖啡快速鉴别方法。采用傅里叶变换近红外光谱仪采集咖啡样品光谱，建立标样（材料光谱）光谱数据库，采集掺假物光谱数据，建立掺假物光谱数据库，建立了咖啡分类模型及掺假物模型，运用Adulterant Screen算法技术对模型的主成分及掺假成分进行计算分析，建立了咖啡掺假快速鉴别模型。本咖啡快速鉴别方法对含量在2%、5%、10%、15%、20%、30%及40%巴西莓果粉和5%、10%、15%、20%、30%及40%大麦掺假咖啡可以实现有效的掺假鉴别，最低识别咖啡中巴西莓果粉及大麦掺假含量分别为2%及5%。利用近红外光谱技术结合Adulterant Screen算法建立咖啡快速鉴别模型可以鉴别咖啡中巴西莓果粉和大麦等掺假物，为咖啡样品掺假鉴别提供了一种快速、可靠、无损的检测方法，能有效的运用于咖啡样品掺假鉴别的日常检测工作中。

The near infrared spectroscopy combined with adulterant screen algorithm was used to establish coffee fast identification method in this word. The near infrared spectrometer was used to collect coffee spectral data, the spectral database of standard samples (material spectrum), the spectral data of adulterants, and then to establish the spectral database of adulterants. The classification model and adulterant model of coffee were established. The Adulterant Screen was used to analyze the principal components of the model and composition of adulteration, and then to establish a coffee adulteration fast identification model. The method can be used to identify adulterated coffee containing 2, 5, 10, 15, 20, 30 and 40 percent of Brazil berry fruit power. It can be also used to identify adulterated coffee containing 5, 10, 15, 20, 30 and 40 percent of barley. The lowest identified concentrations of adulterated acai and barley in coffee were 2% and 5%, respectively. The near infrared spectroscopy combined with Adulterant Screen algorithm based fast identification model can identify coffee in adulteration, such as brazil berry fruit powder and barley. It provides a fast, reliable and nondestructive testing method for the adulteration identification of coffee samples. It can be used in routine testing of coffee sample adulteration.

质检总局科技计划项目(2016IK054)；广东检验检疫局科技计划项目(2016GDK17)；“十三五”国家重点研发计划项目（2017YFF0211302）