长波二阶导数近红外光谱-偏最小二乘法测定白酒中乙醇含量

Long Wave Near Infrared Spectroscopy-partial least Square Detect Alcohol Content in Distilled Spirits

根据近红外光谱的振动吸收强度与有机分子官能团含量的线性关系,用偏最小二乘法对白酒的近红外光谱和其中的乙醇浓度建立相关模型。为了优化模型,我们对白酒的近红外光谱进行一阶导数处理,选择最佳的波长范围和最适的因子数。然后用所建的模型对预测集和白酒样品进行预测,所得结果令人满意。有望成为白酒快捷而准确的检测方法。

Base on the linear relationship between the absorption intensity of long wave near infrared spectroscopy and the content of the organic molecule＇s functional group, we have applied partial least square to establish a model for determination of alcohol content in Distilled Spirits. The first derivative spectra have been applied to develop this model. This model has been optimized by selecting the best wavelength range and the most suitable factors. Using this model for predicting alcohol content in the validation and Distilled Spirits purchased from the market, a satisfactory result is obtained. It is a simple and accurate method for determination of alcohol content in Distilled Spirits.