Method for identifying dyed orange based on near infrared spectroscopy

The invention discloses a method for identifying a dyed orange based on near infrared spectroscopy. The method comprises the following steps: (1) collecting data and establishing a qualitative model:collecting spectral data of collection points on the surface of an orange sample, dividing the spectral data into a prediction set and calibration set, and establishing the qualitative model; (2) carrying out major component analysis: performing spectral correction, determining the major component according to the cluster of the qualitative model in a corrected image, and carrying out data dimension reduction on independent variables having a correlation; and (3) identifying the dyed orange through the major component analysis. The method has the advantages of simplicity in operation and low cost. 本发明公开了种基于近红外光谱鉴别染色橙的方法，包括如下步骤：(1)采集数据并建立定性模型：采集橙子样品表面采集点的光谱数据，将所述的光谱数据分为预测集和校正集，建立定性模型；(2)主成分分析：进行光谱校正，根据校正图像中定性模型的聚类来确定主成分，将具有相关性的自变量进行数据降维；(3)通过主成分分析鉴别染色橙。本发明方法具有操作简便、成本低廉的优点。