基于NIRS与PLS-DA的红提产地鉴别研究

Identification of red grape origin based on NIRS and PLS-DA

红提产区分布广、生态类型复杂,不同产地红提果实品质和市场效应具有较大差异。研究基于近红外光谱技术(NIRS)与偏最小二乘判别(PLS-DA)对不同产地的红提果品进行识别和鉴伪。试验购置了美国红提、秘鲁红提、国产红提,每种45个,共135个样品。在4000~12000 cm^-1范围内采集所有样品的近红外漫反射光谱。对光谱数据进行主成分(PCA)分析,结果发现PCA方法无法完全实现3种产地红提的聚类鉴别。最后,建立PLS-DA模型,对3种产地的红提进行判别,识别率均为100%。

The red grape producing areas are widely distributed and the ecological types are complex.The fruit quality and market effects of red grapes in different producing areas are quite different.Based on near-infrared spectroscopy(NIRS)and partial least squares discrimination(PLS-DA),the red grape fruits from different regions were identified and authenticated.American red grape,Peruvian red grape and domestic red grape were purchased from supermarkets,each containing 45 samples,totaling 135 samples.Near-infrared diffuse reflectance spectra of all samples were collected in the range of 4000~12000 cm^-1.Principal component(PCA)analysis of the spectral data showed that the PCA method could not fully realize the cluster identification of the three types of red grape producing areas.Finally,the PLS-DA model was established to discriminate the three red grape producing areas,and the recognition rate was 100%.