基于近红外光谱特性分析的花生含水率检测  

Moisture Detection of Peanuts Based on Near-infrared Spectral Characteristics

作　　者：陈天华[1] 雷春宁[1] 李月[1] 

机构地区：[1]北京工商大学计算机与信息工程学院,北京100048

出　　处：《食品科学技术学报》2013年第5期50-54,共5页Journal of Food Science and Technology

摘　　要：以花生检测国家标准GB/T 5497为基础,采用近红外光谱检测技术对花生含水率是否达标进行检测.实验配制了30个不同含水率的花生样本,其中18个样本含水率达到国家标准,12个未达标,将样本分为训练集和测试集,通过近红外实验获取不同含水率的花生对不同波长光的吸收情况,将采集的数据作为BP神经网络的输入参数,在训练集对神经网络进行学习和训练,然后采用该模型,对测试集花生含水率是否达标进行测试.实验表明,基于近红外光谱技术和神经网络的识别方法可全部正确识别测试集样本.Based on the national standard of peanuts （ GB/T 5497） , the near infrared ray spectroscopy （NIRS） technology was used to test the moisture content of peanuts. Thirty kinds of peanuts with different moisture contents were prepared and the moisture contents of 18 samples reached the national standard whereas other 12 did not. To get the different wavelengths＇ absorbing states of different samples, NIRS technology was applied and samples were separated into the training set and test set. The data was used as the input parameters of the BP neural network. The neural network was tested and recorded in the training set and then used to test the moisture content of peanuts in the test set. The results showed that the identification method based on NIRS technology and neural network can correctly identify the testing samples.

关 键 词：近红外检测 人工神经网络 含水率检测 花生品质 

分 类 号：TS210.7[轻工技术与工程—粮食、油脂及植物蛋白工程] TS207.3[轻工技术与工程—食品科学与工程]