摘要
收集代表性绿茶样品158个,直接对新鲜样品进行近红外光谱扫描,运用改进偏最小二乘法(MPLS)在4种不同的光谱数据预处理方式下进行水分含量建模,并用验证集对最优模型进行检验。结果显示,光谱数据在散射处理方式SNV+Detrend下经过一阶导数处理后的预测结果最优,其定标标准差(SEC)为0.32%,样品预测值和实测值之间的决定系数(RSQ)为0.861,预测标准差(SEP)为0.5%,偏差(Bias)为-0.1%。说明应用近红外光谱分析技术实现绿茶中水分含量的无损检测是可行的,并可得到较为满意的预测效果。
158 representative samples of green tea were collected. The fresh samples were scanned directly by NIRS. The moisture content model was built up with MPLS at four different spectral data preprocessing methods, and the optimal model was tested by the test set. The experimental results indicated that the forecast results of spectral data was optimal at scatter- ing approach SNV＋Detrend by one rank differential coefficient derivative disposal, its calibration standard deviation （SEC） was 0.32%, the coefficient of determination （RSQ） between forecast and actual values of sample was 0.861. Forecast standard deviation （SEP） was 0.5% and deviation （Bias）to-0.1%. The results showed that using NIRS technology to achieve nondestructive testing of moisture content in green tea was feasible. The forecast results were satisfactory.
出处
《湖北农业科学》
北大核心
2007年第6期956-959,共4页
Hubei Agricultural Sciences
基金
湖北省自然科学基金资助项目(2007ABA351)
关键词
绿茶
水分
无损检测
green tea
moisture content
nondestructive testing
