机构地区: 暨南大学理工学院重大工程灾害与控制教育部重点实验室
出 处: 《光谱实验室》 2010年第2期745-748,共4页
摘 要: 采用FTIR/ATR光谱,偏最小二乘法(PLS)和Savitzky-Golay(SG)平滑方法建立甘蔗糖蜜锤度的快速测定方法和光谱分析的优化模型。基于全谱4500—600cm-1,利用计算机算法平台,把483种SG平滑模式和1—40的PLS因子数任意组合分别建立PLS模型。根据预测效果,最优模型为原谱平滑,2、3次多项式类型,45平滑点数,PLS因子数为6,预测均方根偏差(RMSEP)、预测相关系数(rp)和相对预测均方根偏差(RRMSEP)分别为0.978%、0.902%和1.05%,预测精度很高,并且大幅度优于未做SG平滑处理直接用PLS建模的预测效果。从而表明,FTIR/ATR光谱能够应用于甘蔗糖蜜锤度的快速准确测定,SG平滑模式和PLS因子数的联合大范围筛选能够有效地应用于FTIR/ATR光谱分析的模型优选。 Sugarcane molasses brix was determined fast by FTIR/ATR spectroscopy with partial least squares (PLS) and Savitzky-Golay (SG) smoothing method,and the optimal spectral analysis model was also established. The PLS model was established by any combination of 483 SG smoothing models and 1--40 PLS factors, based on the whole region (4500--600cm^-1) and the computer algorithm platform. According to the prediction effect, the optimal model is the original spectra smoothing, 2 or 3 degree polynomial, 45 smoothing points, the PLS factor of 6,and RMSEP, rp and RRMSEP of 0. 978% ,0. 902 and 1.05% respectively. The prediction accuracy is very high,and it is substantially superior to the prediction effect of the PLS model without SG smoothing. The results show that FTIR/ATR spectroscopy can be used to the rapidly accurate determination of sugarcane molasses brix,and large-scale optimization combining SG smoothing modes and PLS factors can be effectively applied to the model optimization of FTIR/ATR spectral analysis.