利用傅里叶近红外光谱仪(FT-NIR)对金沙蓝钻酒样品进行了光谱采集,采用一阶求导与Savitzky-Golay卷积平滑相结合的预处理方法结合仪器自动优化谱区(4119.20~9881.46 cm-1)建立了金沙蓝钻酒的相似度匹配分析模型,同时采用该模型对已知样品进行相似度计算。结果显示:蓝钻酒样品的相似度均达到95%以上,而其他酒样的相似度均在60%以下,其中赖茅相似度最高,达到了59.24%;在所有乙醇水溶液样品中,体积分数为50%的乙醇水溶液相似度最高,为18.23%。实验中,验证样品全部判别准确,准确率达到100%,表明经过优化的模型预测精度和可靠性较高,可用于实际检测。 The sapphire blue wine samples were collected by Fourier transform infrared spectroscopy (FT-NIR) spectroscopy. The preprocessing method combined with Savitzky-Golay convolution smoothing was used to optimize the spectral region (4119.20 ~ 9881.46 cm-1), a similarity matching analysis model was established for sapphire blue wine, and the similarity of known samples was calculated using this model. The results showed that the similarity of blue wine sample reached more than 95%, while the similarity of other wine samples were below 60%, and the similarity of Lai Mao was the highest, reaching 59.24%; in all aqueous ethanol samples, the volume fraction The highest similarity of 50% ethanol aqueous solution was 18.23%. In the experiment, all the samples were verified accurately with the accuracy of 100%, indicating that the optimized model has higher prediction accuracy and reliability and can be used for actual testing.