基于近红外光谱技术建立沉香含油量预测模型  

Prediction Models of Oil Content of Agarwood Based on Near Infrared Spectroscopy

作　　者：林艳 何紫迪 毛积鹏 蒋开彬 刘天颐 黄少伟 

机构地区：[1]华南农业大学林学与风景园林学院广东省森林植物种质资源创新与利用重点实验室,广东广州510642

出　　处：《热带作物学报》2018年第1期182-188,共7页Chinese Journal of Tropical Crops

基　　金：国家林业公益性行业科研专项"黎蒴等华南重要乡土树种良种选育"(No.201204303)

摘　　要：为建立沉香(Aquilaria sp.)含油量的近红外光谱预测模型,在950~1 650 nm的光谱范围内,使用DA7200 NIRS分析仪收集了64个沉香样本的光谱数据,采用偏最小二乘法(PLS)建立回归模型,并选择最佳预处理方法和最佳主成分数,建立沉香含油量近红外光谱模型。结果表明,采用卷积平滑法(S-G)对光谱进行预处理且当最佳主成分数为7时,可达到最优模型,其校正集相关系数(RC)和校正集均方根误差(RMSEC)分别为0.980 9和0.958 9,交互验证集相关系数(RV)和交互验证集均方根误差(RMSEV)分别为0.697 4、1.029 0。说明预测值与测量值具有显著的相关性,该模型的预测准确度较高,可以满足对沉香结香品质进行快速预测的要求。The spectral data of 64 agarwood samples between the spectrum of 950 nm to 1 650 nm were collected using DA7200 NIRS analyzer to estabilish a prediction model of near infrared spectroscopy of agarwood oil content. A regression model was established using the partial least squares（ PLS） method, and selecting the best pretreatment method and the optimal number of principal components to set up a model of the near infrared spectra of the oil content. Results showed that the smoothing（ S-G） method was best for spectral preprocessing,and when the best optimum principal component number was 7 can achieve the optimal mode. The related coefficient of calibration（RC） and root mean square error of calibration（RMSEC） was 0.980 9, 0.958 9; the related coefficient of validation（ RV） and root mean square error of validation（ RMSEV） was 0.697 4, 1.029 0. The prediction value has a significant correlation with the measured value, and the prediction accuracy of the model is high, which can meet the requirement of rapid prediction of agarwood quality.

关 键 词：近红外 沉香 含油量 预测模型 

分 类 号：S722.5[农业科学—林木遗传育种]