摘要
使用自行设计的近红外透射光谱仪系统,针对牛奶的蛋白质、脂肪和乳糖建立定量分析模型进行检测,利用Fisher分类方法和BP神经网络分类方法建立了针对酸奶品牌的分类模型。结果表明,上述两种方法都可以用于酸奶品牌和种类的分类检测,其中BP神经网络方法具有更高的准确性。
A special-designed near-infrared transmission spectroscopy system is used for the quantitative detection of protein, fat and lactose in milk. Fisher classification and BP neural network are used to establish a classification model for yogurt brands. Results show that both methods are suitable for the classification of the brand and type of yogurt. BP neural network is considered to have higher accuracy.
出处
《中国奶牛》
2015年第22期30-34,共5页
China Dairy Cattle
基金
国家科技支撑计划(2012BAD12B08-2)
关键词
近红外透射检测
光谱仪
定量分析
Fisher分类
BP神经网络
Near infrared transmittance detection
Spectrometer
Quantitative analysis
Fisher classification
BP neural network
