基于红外光谱特征和傅里叶变换的牛奶成分检测方法

Milk Composition Detection Methods Based on Infrared Spectral Features and Fourier Transforms

目前提出的牛奶成分检测方法对于乳中蛋白质、脂肪以及乳糖的检测能力较差,导致检测相对误差较高。为了解决上述问题,基于红外光谱特征和傅里叶变换研究了一种新的牛奶成分检测方法。通过对液态牛奶成分表征与各种形态变化,进行红外光谱追踪与监测,根据牛奶红外光谱分析结果实现数据预处理,识别处理红外光谱带特征波长,考虑牛奶成分浓度与估计变量分布概率,实现参数估计,根据最小二乘法计算各成分参数估计值后,可计算出牛奶成分预测参照数值,根据成分特征,光谱吸收理论差异,划分观测样本类型,按照牛奶质量等级,不同成分含量浓度高低进行分类,完成特征提取。实验结果表明,基于红外光谱特征和傅里叶变换的牛奶成分检测方法检测相对误差低于0.2%,检测准确率较高。

Currently proposed milk composition detection methods have poor detection ability for protein,fat and lactose in milk,which leads to high relative errors in detection.In order to solve the above problems,a new milk composition detection method based on infrared spectral characterization and Fourier transform was investigated.Through the liquid milk composition characterization and various morphological changes,infrared spectral tracking and monitoring,according to the results of milk infrared spectral analysis to achieve data preprocessing,identification and processing of infrared spectral bands characteristic wavelengths,considering the concentration of the milk components and the estimated distribution probability of the variables,the realization of parameter estimation,according to the method of least squares to calculate the estimated value of the parameters of the components,can be calculated according to the predicted reference value of the milk composition,according to the characteristics of the ingredients According to the difference of spectral absorption theory,the observation sample type is divided,and according to the milk quality grade,the different components content concentration is classified,and the feature extraction is completed.The experimental results show that the relative error of the milk composition detection method based on infrared spectral features and Fourier transform is less than 0.2%,and the detection accuracy is high.