基于光纤光谱仪对骏枣浸泡梯度的水分定量检测研究  

Study on Quantitative Determination of Water Content in Immersion Gradient of Jujube Based on Optical Fiber Spectrometer

作　　者：张辉 罗华平[1,2] ZHANG Hui;LUO Huaping(College of Mechanical and Electrical Engineering,Tarim University,Alar,843300,Xinjiang,China;Key Laboratory of Colleges&Univers itiesunder the Department of Education of Xinjiang Uygur Autonomous Region)

机构地区：[1]塔里木大学机械电气化工程学院,新疆阿拉尔843300 [2]新疆维吾尔自治区现代农业工程重点实验室

出　　处：《新疆农机化》2022年第2期16-18,共3页Xinjiang Agricultural Mechanization

基　　金：国家自然科学基金项目(11964030,11464039)。

摘　　要：利用光纤光谱技术对50个骏枣进行浸泡梯度的水分分析,通过选用偏最小二乘法(PLS)建立骏枣浸泡梯度水分定量模型。光谱数据预处理方法分别为光程恒定(Con)、多元散射校正(MSC)、标准正态变换(SNV)、S-G平滑。在骏枣浸泡20~60min内,含水率检测精度较低,无法精准预测骏枣的品质信息;在骏枣浸泡120min时含水率检测精度最高,相关系数(R)为0.99462、校正标准偏差(RMSEC)为0.169、预测标准偏差(RMSEP)为1.67。通过试验结果可以看出可见光光谱技术对骏枣浸泡梯度水分定量模型的预测是可行的。In this paper, fiber optic spectroscopy was used to analyze the water content of soak gradient of 50 jujubes. A quantitative model of soak gradient water content of 50 jujubes was established by partial least square method(PLS). Spectral data preprocessing methods are optical path constant(CON), multiple scattering correction(MSC), standard normal transformation(SNV),S-G smoothing. Within 20~60 minutes of soaking in Junjube, the detection accuracy of water content is low, and the quality information of Junjube cannot be accurately predicted. When the jujube was soaked for 120 min, the detection accuracy of water content was the highest, the correlation coefficient(R) was 0.99462, the correction standard deviation(RMSEC) was 0.169, and the prediction standard deviation(RMSEP) was 1.67. The experimental results show that the visible light spectrum technology is feasible to predict the gradient water quantitative model of jujube immersion.

关 键 词：骏枣 水分 偏最小二乘法 可见光光谱 

分 类 号：TS255.7[轻工技术与工程—农产品加工及贮藏工程]