基于可见/近红外光谱和变量优选的南水梨糖度在线检测

利用可见/近红外光谱技术对南水梨糖度进行在线检测研究。南水梨样本以0.3m/s速度传输,并采用USB4000光谱仪在470~1 150nm波段范围内采集南水梨样本的光谱。然后,利用3种变量选择方法对波长变量进行筛选,应用偏最小二乘(PLS)方法分别建立南水梨糖度的在线预测模型,并分析预测模型性能的优劣。结果表明:可见/近红外光谱技术结合变量选择方法在线检测南水梨的糖度是可行的;竞争自适应重加权采样(CARS)方法优于无信息变量消除(UVE)及连续投影算法(SPA);CARS方法可以有效简化预测模型并提高预测模型的性能;南水梨全光谱PLS及CARS—PLS糖度预测模型的预测集相关系数和预测均方根误差(RMSEP)分别为0.940,0.951和0.467%,0.420%。     

基于CARS-SPA的苹果可溶性固形物可见/近红外光谱在线检测

采用CARS(competitive adaptive reweighted sampling)联合连续投影算法(SPA)方法筛选苹果可见/近红外光谱的特征变量,继而联合多种不同建模方法建立苹果可溶性固形物(SSC)预测模型,并对预测模型进行对比研究。研究结果显示,采用CARS-SPA联合筛选出的31个变量,通过采用PLS建立苹果SSC的可见/近红外光谱在线检测模型性能最稳定,其变量数仅为原始光谱的1.69%,预测集的相关系数和均方根误差分别为0.936和0.351%。研究表明采用CARS-SPA能有效提取苹果SSC的光谱特征变量,能有效简化模型并提高模型精度。     

基于近红外光谱结合波长优选检测单颗葡萄的SSC含量

采用无损检测测定单颗葡萄中可溶性固形物(SSC)含量,获得个体和群体信息,以期指导田间管理、葡萄储存条件设置及满足消费者对葡萄口味的不同需求。采用手持式NIR光谱仪在950~1 650nm波长范围采集葡萄的近红外光谱,采用偏最小二乘(PLS)回归建立葡萄SSC预测模型。为了减少冗余无信息变量,增加模型的预测精度和稳定性,采用无信息变量消除法(UVE)、随机蛙算法(RF)筛选出与葡萄SSC含量相关的重要波长变量。结果表明:RF筛选建立的SSC预测模型优于全光谱PLS和UVE筛选建立的模型。RF-PLS模型的校正集、交叉验证及预测集的R2c、R2cv和R~2p分别为0.960 5,0.933 4,0.930 4,校正均方根误差(RMSEC),交叉验证均方根误差(RMSECV)和预测均方根误差(RMSEP)分别为0.638 2,0.829 9,0.868 8。表明通过波长优选后的,基于便携式近红外光谱在预测单颗葡萄SSC含量的应用上完全可行,有较高的预测精度。     

无信息变量消除法在筛选南疆红枣总酸近红外特征波长中的应用

以南疆红枣总酸的快速无损检测为研究对象,利用SPXY(Sample set partitioning based on joint x-y distances)法来划分校正集样本,应用无信息变量消除法(UVE)对南疆红枣总酸近红外光谱(NIRS)特征变量进行筛选,然后用筛选出的变量建立偏最小二乘(PLS)模型,该模型的预测标准偏差(RMSEP)为0.044 7,预测相关系数为Rp为0.877 8,并将UVE筛选的变量建立的PLS模型与全光谱建立PLS模型结果进行比较。结果表明,SPXY法划分的校正集样本合理;UVE优出选全光谱1 557个变量中的420个变量,建立的PLS模型预测效果要好于全光谱建立的PLS模型,UVE能够有效地选取待测成分的特征波长,建立简化的红枣总酸预测模型,降低模型计算量。     

基于特征波长提取的哈密大枣可溶性固形物的高光谱预测

本文利用高光谱图像技术对干制后的哈密大枣可溶性固形物含量(SSC)进行预测研究。使用多种预处理方法对原始光谱进行处理,并对原始光谱和预处理后的光谱分别建立PLS模型,对比分析得出均值中心化(MC)处理效果最佳。对MC处理后的光谱经联合区间偏最小二乘算法(si-PLS)筛选后,再结合遗传算法(GA)和竞争性自适应重加权算法(CARS)提取哈密大枣SSC的特征波长,将提取的波长变量建立哈密大枣SSC的PLS预测模型。结果显示:利用MC-CARS-GA-si-PLS方法提取的16个关键波长变量(仅占全光谱变量的2%)所建立的PLS模型性能优于全光谱PLS模型。该模型的预测集相关系数(Rp)、预测均方根误差(RMSEP)和预测(RPD)分别为0.93、0.48和2.721。该方法提取的波长变量所建立的预测模型,不仅使模型简化,而且增强了模型的预测能力,为高光谱图像技术对水果及其干制品的定量分析研究提供了参考。     

无效变量消除法在油菜籽芥酸近红外无损速测中的应用

探索改善油菜籽芥酸近红外预测模型准确度与精密度的方法,利用无效变量消除法（UVE）,对135个油菜籽样品近红外光谱信号进行筛选,并利用筛选后的光谱对油菜籽芥酸含量进行偏最小二乘法交叉验证。结果表明,UVE法筛选变量后建立的芥酸校正模型对未知样品预测结果的准确度和速度显著优于全波长参与建立的芥酸校正模型。散射校正加一阶导数对光谱预处理,UVE法筛选变量,偏最小二乘法交叉验证建立的校正模型效果最好,其预测值与标准值的相关系数R达到0.92,交叉验证预测均方差为2.2。因此,用UVE进行波长选择后建立的近红外模型,能准确快速地对油菜籽芥酸含量进行定量分析。     

近红外光谱技术结合反向区间偏最小二乘算法-连续投影算法预测哈密瓜可溶性固形物含量

采用近红外光谱技术结合数据降维的方法,建立了哈密瓜可溶性固形物含量的预测模型,对原始光谱进行特征区间选择,共选取了6个子区间,432个光谱变量;将6个联合子区间的光谱数据分别结合特征选择竞争性自适应重加权采样算法、遗传算法、连续投影算法(successive projections algorithm, SPA)提取特征波长;再使用选取的特征波长以及特征区间波长作为模型的输入变量,利用极限学习机和偏最小二乘算法(partial least squares, PLS)建立哈密瓜可溶性固形物含量预测模型。结果显示,反向区间偏最小二乘算法+SPA+PLS建立的预测模型最优,模型的校正集相关系数为0.923 4,预测集相关系数为0.878 8,模型能够准确预测哈密瓜可溶性固形物含量。     

基于变量筛选的温州蜜桔品质的光谱快速检测

目的利用可见/近红外光谱技术结合变量筛选算法建立预测模型。方法采集7个不同批次蜜桔样本的漫透射光谱,预处理优化后,以无信息变量消除法(uninformative variable elimination,UVE)、竞争性自适应重加权法(competitive adaptive reweighting sampling,CARS)及其组合(UVE-CARS)共3种策略来进行光谱有效波段的筛选,建立蜜桔可溶性固形物含量(soluble solid content,SSC)的偏最小二乘预测模型(partial least square,PLS)。结果比较全变量模型和3个特征变量模型的预测性能,UVE-CARS-PLS模型取得了最优的检测效果,相比全变量模型,建模变量数减少了96.5%,其预测集相关系数R_P提升至0.732,预测集均方根误差(root-mean-square error,RMSEP)下降至0.873~0Brix。结论结合多重变量选择算法,可以进一步压缩建模变量数,简化模型,提高模型预测精度,实现区域蜜桔品质的光谱快速检测。     

基于近红外光谱技术预测玉米伏马菌素

为了更好地利用近红外光谱分析技术对玉米伏马菌素含量进行预测,减小因玉米产地间的差异对玉米近红外光谱预测模型的影响,以不同产地的玉米作为研究对象,利用x-y共生距的方法将试验样本划分为校正集与验证集,采用经典的偏最小二乘法分别建立不同产地和混合产地的玉米伏马菌素预测模型,并采用验证集样本分别对模型的预测精度进行验证。为了减小建模及预测过程的运算量,采用连续投影算法(SPA)和竞争性自适应加权算法(CARS)对不同产地玉米的近红外光谱的特征波长进行筛选,筛选出22个特征波长变量作为输入,大大降低了建模及预测过程的运算量,同时预测准确度也有所改善,其预测相关系数达到0.954,为快速、无损地实现对玉米伏马菌素的检测提供了可靠的理论依据。     

基于近红外光谱技术检测苹果气调贮藏期可溶性固形物含量

目的 基于近红外光谱技术结合偏最小二乘（Partial least square, PLS）法和最小二乘支持向量机回归（Least square-support vector regression, LS-SVR）法建立苹果气调贮藏期可溶性固形物（Soluble solids contents, SSC）含量预测模型。方法 在分析了气调贮藏期苹果细胞结构和SSC变化的基础上,采集了可见-近红外（Visible-near infrared, Vis-NIR）波段和长波近红外（Long wave near infrared, LWIR）波段下不同贮藏时间的苹果漫反射光谱信息,利用主成分分析方法（Principal component analysis,PCA）分析不同贮藏期苹果光谱信息分布特征,使用Kennard-Stone（K-S）算法以3:1比例对样本集进行划分,使用多元散射校正（Multiplicative scatter correction, MSC）和SG（Savitzky-Golay）平滑对光谱进行预处理,利用连续投影算法（Successive projections algorithm, SPA）和竞争自适应重加权采样（Competitive adaptive reweighted sampling, CARS）法对光谱进行特征波长提取,并建立SSC预测模型。结果 在LWIR波段下,经MSC预处理和CARS提取特征波长后建立的PLS模型取得了较好的预测精度,模型相关系数为0.900,均方根误差为0.478;经MSC、SG平滑预处理和CARS提取特征波长后建立的LS-SVR模型取得了更好的预测精度,模型相关系数为0.927,均方根误差为0.507。结论 构建的基于可见/近红外光谱无损预测模型可实现对气调贮藏期苹果SSC的准确预测,为高效贮藏技术提供了理论基础。     

Variable Selection in Visible and Near-Infrared Spectral Analysis for Noninvasive Determination of Soluble Solids Content of ‘Ya’ Pear

Informative variable selection or wavelength selection plays an important role in the quantitative analysis of near-infrared (NIR) spectra because the modern spectroscopy instrumentations usually have a high resolution and the obtained spectral data sets may have thousands of variables and hundreds or thousands of samples. In this study, a new combination of Monte Carlo–uninformative variable elimination (MC-UVE) and successive projections algorithm (SPA; MC-UVE-SPA) was proposed to select the most effective variables. MC-UVE was firstly used to eliminate the uninformative variables in the raw spectra data. Then, SPA was applied to determine the variables with the least collinearity. A case study was done based on the NIR spectroscopy for the non-destructive determination of soluble solids content (SSC) in ‘Ya’ pear. A total of 160 samples were prepared for the calibration (n?=?120) and prediction (n?=?40) sets. Three calibration algorithms including linear regressions of partial least square regression (PLS) and multiple linear regression (MLR), and nonlinear regression of least-square support vector machine (LS-SVM) were used for model establishment by using the selected variables by SPA, UVE, MC-UVE, UVE-SPA, and MC-UVE-SPA, respectively. The results indicated that linear models such as PLS and MLR were more effective than nonlinear model such as LS-SVM in the prediction of SSC of ‘Ya’ pear. In terms of linear models, different variable selection methods can obtain a similar result with the RMSEP values range from 0.2437 to 0.2830. However, combination of MC-UVE and SPA was helpful for obtaining a more parsimonious and efficient model for predicting the SSC values in ‘Ya’ pear. Twenty-two effective variables selected by MC-UVE-SPA achieved the optimal linear MC-UVE-SPA-MLR model compared with other all developed models by balancing between model accuracy and model complexity. The coefficients of determination (r ²), root mean square error of prediction, and residual predictive deviation by MC-UVE-SPA-MLR were 0.9271, 0.2522, and 3.7037, respectively.     

Nondestructive Determination of Soluble Solids Content of ‘Fuji’ Apples Produced in Different Areas and Bagged with Different Materials During Ripening

To investigate the feasibility of hyperspectral imaging technique in nondestructive determination of soluble solids content (SSC) of fruits produced in different places and bagged with different materials during ripening, the near infrared hyperspectral reflectance images were acquired on 196 ‘Fuji’ apples picked from four orchards in different areas and bagged with polyethylene film or light-impermeable paper. Mean reflectance spectrum from the regions of interest in the hyperspectral image of each apple was extracted. Standard normal variate (SNV) was used to eliminate the effect of instrument and environment on spectra. The sample set partitioning based on joint x–y distances method was applied to divide the samples into calibration set and prediction set as the ratio of 3:1. Successive projection algorithm (SPA) and uninformative variable elimination (UVE) method were used to select effective wavelengths (EWs) from the full spectra. Partial least squares (PLS), least squares support vector machine (LSSVM), and extreme learning machine (ELM) were used to develop SSC determination models. The results showed that 24 and 122 EWs were selected by SPA and UVE, respectively. The selection of EWs was helpful to SSC determination performance improvement. The optimal SSC prediction model was LSSVM based on selected EWs by SPA, with the correlation coefficient and root-mean-square error of prediction set of 0.878 and 0.908 °Brix, respectively. This study indicates that hyperspectral imaging technique could be used to determine SSC of intact apples produced in different places and bagged with different materials during ripening.     

灵武长枣蔗糖含量的高光谱无损检测

利用高效液相色谱法检测蔗糖含量,同时运用高光谱成像技术结合化学计量方法建立蔗糖预测模型;通过竞争性自适应加权（competitive adaptive reweighted sampling,CARS）算法、连续投影算法（successive projection algorithm,SPA）和无信息消除变量（uninformative variable elimination,UVE）降维处理,建立特征波段和全波段的主成分回归（principal component regression,PCR）、偏最小二乘回归（partial least squares regression,PLSR）和多元线性回归（multivariable linear regression,MLR）模型。结果表明,采用蒙特卡洛方法剔除异常样本后,相关系数由0.611增大到0.846;正交信号校正法预处理效果最佳,RC和RP分别为0.853和0.794;利用SPA、UVE、CARS、CARS＋SPA和CARS＋UVE五种方法提取了5、21、17、10、18 个特征变量,其中CARS-PCR模型最好,校正集、预测集的相关系数为0.861、0.843,校正集、预测集的均方根误差为0.013 mg/g和0.014 mg/g。综上,高光谱成像技术可以实现长枣蔗糖含量的预测,为更深一步探讨枣的内部品质提供参考。     

基于可见/近红外高光谱成像技术的牛肉品种鉴别

利用可见/近红外高光谱成像技术实现荷斯坦奶牛、秦川牛、西门塔尔牛三个品种牛肉的快速无损鉴别。首先,对原始光谱进行预处理并对样本集进行划分;应用竞争性自适应重加权算法(CARS)、连续投影算法(SPA)和无信息变量消除算法(UVE)对预处理后的光谱数据提取特征波长;结合偏最小二乘判别模型(PLS-DA)、K最近邻(KNN)模型及支持向量机(SVM)模型进行全波段及特征波段鉴别分析。结果表明,一阶导数(FD)法为最优预处理方法,利用光谱-理化值共生距离法(SPXY)法划分后的样本模型预测性能最好;利用CARS、SPA和UVE分别选出24、17和19个特征波长;基于CARS法提取的特征波长所建的RBF-SVM模型的校正集与预测集正确率分别为100%、98.82%。由此可见,基于高光谱成像技术能够获得较好的牛肉品种鉴别效果。该研究可为牛肉品种的快速无损鉴别提供参考。     

Variable selection for partial least squares analysis of soluble solids content in watermelon using near-infrared diffuse transmission technique

This work is focused on the variable selection in building the partial least squares (PLS) regression model of soluble solids content (SSC) that is used to evaluate quality grading of watermelon. The spectra were obtained by the near infrared (NIR) spectrometer with the device designed for on-line quality grading of watermelon and the spectra of 680–950 nm were adopted to analysis. The variable selection was based on Monte-Carlo uninformative variable elimination (MC-UVE) and genetic algorithm (GA). In comparison of the performances of the full-spectra (680–950 nm) PLS regression model and the feature wavelengths PLS regression model showed that the MC-UVE–GA–PLS model with baseline offset correction combined multiplicative scatter correction (MSC) pretreatment was much better and 14 variables in total were selected. The correlation coefficients between the predicted and actual SSC were 0.885 and 0.845, the root mean square errors were 0.562 °Brix and 0.574 °Brix for calibration and prediction set, respectively. This work can make a great contribution to the research of on-line quality grading for watermelon nondestructively.     

可见-近红外高光谱成像技术对灵武长枣VC含量的无损检测方法

为探究基于高光谱成像技术预测灵武长枣VC含量的可行性并寻找最佳预测模型。采集100?个长枣样本在波长400～1?000?nm处的高光谱图像,对光谱数据进行预处理;应用遗传算法（genetic algorithm,GA）、连续投影算法（successive projection algorithm,SPA）和竞争性正自适应加权（competitive adaptive reweighted sampling,CARS）算法对原始光谱数据提取特征波长;分别建立基于全光谱和特征波长的偏最小二乘（partial least squares regression,PLS）和最小二乘支持向量机（least squares support vector machine,LSSVM）VC含量预测模型。结果表明,采用标准正态变换预处理算法效果最优,其PLS模型的交叉验证相关系数为0.839?5,交叉验证均方根误差为16.248?2;利用GA、SPA和CARS从全光谱的125?个波长中分别选取出12、5?个和26?个特征波长;基于CARS建立的PLS模型效果最优,其Rc、Rp、校正均方根误差、预测均方根误差分别为0.896?2、0.889?2、10.746?2%、12.145?3%。研究结果表明基于高光谱成像技术对灵武长枣VC含量的无损检测是可行的。