适宜西瓜检测部位提高近红外光谱糖度预测模型精度

为了提高中国厚皮类瓜果的品质质量和出口能力,增强中国水果品质检测装备制造业的技术实力和技术水平。该文以西瓜为对象,对其糖度进行了试验研究。由于西瓜各部位存在差异,因而不同部位采集近红外光谱会对糖度预测模型精度产生影响。采用自主搭建的西瓜内部品质检测系统对不同批次西瓜瓜梗、瓜脐和赤道3个部位采集漫透射光谱信息,分别采用偏最小二乘回归法(partial least squares regression,PLSR)和最小二乘支持向量机法(least squares support vector machines,LS-SVM)2种方法对西瓜糖度建立预测模型,考察西瓜不同检测部位对西瓜糖度预测模型精度的影响。2种预测模型均显示,赤道部位采集光谱所建立的预测模型检测精度较差,而采用瓜脐部位获取光谱信息建立预测模型略好于瓜梗部位,最佳预测相关系数rpre达到0.823,预测均方根误差(root mean square error of prediction,RMSEP)为0.652%。该研究结果表明,不同部位采集光谱信息对最终的检测模型精度有影响,瓜脐部位为该文西瓜内部品质检测装置的较优采集部位。     

近红外光谱检测苹果可溶性固形物

该文目的是通过静态和在线两种方式的对比试验,研究苹果可溶性固形物近红外光谱静态和在线检测的差异。分别在静态（600～950 nm）和在线（600.02～950.92 nm）2种检测方式下,采用间隔偏最小二乘法,寻找苹果可溶性固形物的特征波段,建立了苹果可溶性固形物近红外光谱检测用数学模型,并进行对比分析。试验结果为：与静态检测模型相比,在线检测模型性能稍弱,模型预测相关系数为0.78,预测均方根误差为1.04oBirx。试验结果表明：近红外光谱在线检测苹果可溶性固形物的精度不理想。     

iPLS-SPA变量选择方法在螺旋藻粉无损检测中的应用

该文研究了基于可见-近红外光谱技术的螺旋藻粉类别无损检测方法。采用簇类独立软模式法（SIMCA）建立可见-近红外光谱模型。全波段光谱所建立的模型得到了93.33%的预测集正确率。文章提出了基于间隔偏最小二乘法（iPLS）和连续投影算法（SPA）的组合光谱变量选择方法进行有效波长的选择。该方法从全波段675个变量中选择了5个最优的有效波段,并且得到了96.67%的预测集正确率。和基于全波段光谱、可见光波段光谱和近红外波段光谱进行SPA运算相比,基于iPLS的SPA运算可以有效减少计算时间。研究表明可见-近红外光谱可以用于对螺旋藻粉类别进行无损检测,同时iPLS-SPA是一个有效的光谱变量选择方法。     

连续投影算法在猪肉pH值无损检测中的应用

利用鲜肉的近红外光谱中少量特征波长对其pH值进行预测,可以大幅度降低模型复杂性和计算量,对开发无损检测装置, 实施肉品生产加工过程中pH值监测有重要意义。该文通过连续投影算法（SPA）选择特征波长建立简单多元线性回归模型（SPA-MLR）,并对比了SPA-MLR模型与全波段（5 000～10 440 cm-1）偏最小二乘回归模型（PLSR）及逐步线性回归（SMLR）、遗传算法（GA）选择特征波长所建模型的性能。结果表明经连续投影算法提取37个特征波长建立的模型,所用变量数仅占全波段的2.6%,校正集相关系数0.870,校正集均方根误差为0.094,验证集相关系数0.892,验证集均方根误差为0.085;性能与经多元散射校正预处理的PLSR模型接近,但采用变量数明显减少,优于逐步线性回归和遗传算法选择特征波长建立的模型,表明该方法可较好的选择特征波长,建立简单的预测模型。     

基于多模型证据融合的苹果分类方法

水果内部品质是水果分类的重要依据之一,利用近红外光谱技术对苹果内部品质进行快速无损检测研究有着非常重要的意义。为提高近红外技术分类模型的预测精度,针对单一预测模型适用性差以及硬分割导致分类不确定性等问题,该研究以烟台红富士苹果为研究对象,利用自行研发的水果在线无损检测系统采集苹果近红外光谱以及可溶性固形物含量 (Soluble Solids Content, SSC),分别采用偏最小二乘（Partial Least Squares, PLS）和极限学习机（Extreme Learning Machine, ELM）法建立苹果预测分类模型,根据SSC的预测值与分类边界的距离提出三角形质量函数生成方法,通过证据理论的Dempster组合规则融合质量函数从而实现2种模型的融合,并探讨基于三角形质量函数的证据理论融合模型对预测精度的影响。研究结果表明：PLS分类模型的准确率为92.25%,ELM分类模型的准确率为 93.80%,而提出多模型融合方法的分类准确率达到了95.35%。而且,该研究提出的三角形质量函数生成法与硬分割生成的质量函数相比方法更符合实际,通过PLS、ELM模型和DS融合模型的混淆矩阵可以看出,融合模型实现了苹果SSC处于分类边界值时的准确分类,三类苹果被错误分类的个数均有减小。该研究提出的多模型证据融合方法不仅提高了模型的预测精度,而且更好地表达了关于类标预测的不确定性,为苹果的在线无损检测分类提供研究基础。     

连续投影算法在砂糖橘总酸无损检测中的应用

酸度是评价砂糖橘品质的重要指标之一,为了消除光谱变量间的共线性影响、减少建模变量以提高校正速度,该文应用连续投影算法(SPA)对砂糖橘总酸近红外光谱无损检测模型进行优化。利用连接点修正方法修正近红外光谱,结合学生化残差图和模型回归图剔除异常样本,利用SPXY(sample set partitioning based on joint x-y distances)方法划分样本集,最后利用SPA进行变量选择,比较SPA选择的变量建模和全光谱变量PLS模型的预测效果,并分析橘皮对总酸模型的预测精度的影响程度。     

蛋壳品质的近红外光谱检测分析

蛋壳品质对蛋品孵化、贮存和运输均有重要影响。为了探索近红外光谱技术快速检测蛋壳品质的方法,该文在鸡蛋蛋壳品质指标相关性分析的基础上进行了蛋壳品质的近红外光谱检测分析,研究比较了不同建模方法、不同光谱预处理方法和不同波段范围对预测结果的影响。结果表明：在5段特征波长范围内建立的经过多元散射校正的偏最小二乘回归（PLSR,partial least squares regression）模型对蛋壳强度的预测结果最好,相关系数r为0.86,校正、预测均方根误差分别为4.42、7.53 N;同时蛋壳百分比（蛋壳质量/蛋质量）的PLSR模型的相关系数r为0.92,校正、预测误差分别为0.313%、0.529%;蛋壳厚度的PLSR模型的相关系数r为0.81,校正、预测误差分别为0.0176、0.0234 mm。研究结果表明应用近红外光谱技术预测蛋壳品质是可行的,为蛋壳品质的快速无损检测提供了一种新的方法。     

基于音频和近红外光谱融合技术的西瓜成熟度判别

为了满足西瓜成熟度的快速无损检测需求,该研究主要利用声学技术、近红外光谱技术结合K最近邻法(k-nearest neighbor,KNN)、线性判别分析(linear discriminant analysis,LDA)和反向传播人工神经网络(back propagation artificial neural network,BP-ANN)3种化学计量学方法对不同成熟度的西瓜进行定性判别;同时采用联合区间偏最小二乘筛选法(synergy interval partial least squares,Si-PLS)分别建立声学技术、近红外光谱技术、融合技术的西瓜可溶性固形物预测模型。结果表明融合技术处理结果均优于单一信号,其LDA模型数据的西瓜成熟度模型识别率较佳,校正集和预测集的识别率分别为100.00%和91.67%。同时,基于融合技术所建立的西瓜可溶性固形物预测模型效果较佳,其校正集的均方差根误差(root mean squared error of the calibration set,RMSECV)为0.601%,预测集的均方差误差(root mean squared error of the prediction set,RMSEP)为0.725%,相比的单独音频信号其均方根误差分别降低了0.081、0.068个百分点。研究结果可为高精度的西瓜品质快速鉴别提供参考。     

基于变量选择的蚕茧茧层量可见-近红外光谱无损检测

以蚕茧茧层量为研究对象,研究了基于可见-近红外光谱技术的蚕茧茧层量无损检测方法。采用最小二乘支持向量机（least square-support vector machine,LS-SVM）建立可见-近红外光谱模型。采用无信息变量消除算法（uninformative variable elimination, UVE）与连续投影算法（successive projections algorithm, SPA）相结合选取光谱有效波长。结果表明,基于UVE-SPA法进行变量选择,最终将原始光谱的600个光谱变量减少到了8个（673,937,963,982,989,992,995和1 008 nm）。基于此8个变量建立的LS-SVM模型得到了预测集的确定系数（Rp2）为0.5354,误差均方根（RMSEP）为0.0373的预测结果。表明可见-近红外光谱可以用于对蚕茧的茧层量进行无损检测,同时UVE-SPA是一种有效的光谱变量选择方法。     

基于可见-近红外光谱的滩羊肉多品质指标无损检测

针对近红外光谱技术在生鲜肉品质检测中预测模型适用范围窄、检测指标单一、模型稳定性差、难以有效应用于生产检测等问题,本研究采集不同月龄宁夏滩羊宰后3个时期4个部位肉的可见-近红外光谱信息,测定色泽、pH值、蒸煮损失、剪切力以及蛋白质、粗脂肪和水分含量,利用2个波段(370~1 050 nm、900~1 700 nm)的光谱数据分别构建各个指标的偏最小二乘回归(PLSR)预测模型以实现滩羊肉多品质指标同步无损检测。结果表明,两波段中各品质指标的PLSR预测模型相关系数(R)均大于0.80,第二波段中水分含量PLSR模型预测集R可达0.941;两波段中各品质指标预测模型的性能较好,其中370~1 050 nm波段的光谱数据对样品色泽参数预测效果更好。综上所述,可见-近红外光谱技术可实现滩羊肉7个品质指标的快速无损检测。本研究结果为滩羊肉品质控制和滩羊屠宰加工企业优质特色产品的生产提供了技术支撑。     

基于介电特性及ANN的油桃糖度无损检测方法

为了探索利用果品的介电特性无损预测内部品质的可能性,该文采用矢量网络分析仪测量了10 d贮藏期间,300个99-1油桃在20～4500 MHz频率下的相对介电常数和介电损耗因子,以糖度作为内部品质指标,基于x-y共生距离的样本划分法确定了含243个样本的校正集和57个样本的预测集;建立了预测油桃糖度的偏最小二乘、支持向量机及极限学习机模型,并综合比较了采用全频谱以及利用无信息变量消除法和连续投影算法分别提取的特征变量作为各模型输入变量时,对各模型拟合效果的影响。结果表明：连续投影算法结合极限学习机预测效果最好（预测相关系数为0.887,预测均方根误差为0.782）;与全频谱和无信息变量消除法相比,连续投影算法在简化模型及提高模型稳定性方面性能良好。该研究结果表明,基于油桃介电特性无损检测糖度是可行的,可为应用介电特性无损检测果品的内部品质指标提供了一种新方法。     

金银花贮藏过程中绿原酸含量的高光谱无损检测模型研究

绿原酸(chlorogenicacid,CGA)是评价金银花品质的重要指标。为了实现金银花贮藏期间CGA含量变化的快速有效检测,该文采集了500个不同贮藏时间(0～20d)的金银花高光谱图像,构建CGA含量的高光谱检测模型。为了提高模型性能,采用savizky-golay卷积平滑(SG),移动窗口平滑(moving average),标准正态变量(standard normal variable,SNV),基线校正(baseline correction,BC),多元散射校正(multiplicative scatter correction,MSC),正交信号校正(orthogonal signal correction,OSC)6种预处理方法并建立偏最小二乘回归(partial least squares regression,PLSR)模型,确定SNV方法为最佳预处理方法,其预测集的R2为0.976 6,RMSE为0.271 1%。为了简化校准模型,利用无信息变量消除(uninformative variable elimination,UVE),连续投影算法(successive projections algorithm,SPA),竞争性自适应加权算法(competitive adaptive reweighted sampling,CARS)以及UVE-CARS、UVE-SPA等方法对SNV预处理后的光谱提取特征波长。然后,分别基于全光谱数据和所选特征变量数据,建立线性偏最小二乘回归(PLSR)和非线性BP神经网络模型。结果表明:UVE-CARS算法可以有效地减少提取变量个数(共提取26个,仅占全光谱范围的3.2%),PLSR和BP模型的预测集R2分别为0.974 6和0.978 4,RMSE分别为0.286 3%和0.250 3%。非线性BP模型预测结果整体优于线性PLSR模型,在BP模型中,UVE-CARS-BP预测精度最高,预测集的R2和RMSE的值分别为0.978 4, 0.250 3%。综上,基于高光谱成像技术建立的SNV-UVE-CARS-BP模型,可以实现金银花贮藏过程中CGA含量变化的快速无损预测。     

Quantitative analysis of soil nitrogen and carbon at a farm scale using visible and near infrared spectroscopy coupled with wavelength reduction

Reducing large spectral datasets to parsimonious representations of wavelengths is of value for efficient storage and easing analysis, in addition to the potential to use a simpler and cheaper spectrophotometer. This study evaluated the potential of calibrating visible and near infrared (vis‐NIR) spectra to total nitrogen (N), total carbon (C), organic C and inorganic C in soil on a 15‐ha farm, with the aim of comparing several wavelength reduction algorithms and rates in terms of model prediction accuracy. We explored the uninformative variables elimination (UVE), UVE coupled with successive projections algorithm (SPA) and two uniform‐interval wavelength reduction approaches (UWR‐I and UWR‐II) with successive wavelength reduction rates (WRRs) of 2, 5, 10, 20, 50, 100, 200, 500 and 1000. The standard normal variate (SNV)‐transformed absorbance spectra of soil samples recorded from 400 to 2499 nm at 1‐nm intervals were used. The calibration sets were subjected to a partial least squares regression (PLSR) with leave‐one‐out cross‐validation. Prediction results showed that UVE can reduce wavelength variables significantly while retaining good model prediction accuracy. The UVE‐SPA produced only three or four wavelengths, with which PLSR models achieved competitive prediction performance, compared with those based on all 2100 wavelengths, with coefficient of determination (R²) of 0.91, 0.89, 0.91 and 0.53 and residual prediction deviation (RPD) of 3.53, 2.95, 3.27 and 1.53 for soil total N, total C, organic C and inorganic C, respectively. The UWR tests showed that PLSR models responded insensitively to various WRRs from 2 to 100. The models calibrated for the 100‐nm interval spectra (21 remaining wavelengths) performed almost as well as those for the 1‐nm interval spectra. Although these findings might be valid only at the farm scale, it is recommended that the proposed wavelength reduction algorithms for more soil types and soils originated from larger areas should be examined.     

Nondestructive determination of total soluble solid content and firmness in plums using near-infrared reflectance spectroscopy

The fruit industry requires rapid, economical, and nondestructive methods for classifying fruit by internal quality, which can be built into the processing line. Total soluble solid content and firmness are the two indicators of plum internal quality that most affect consumer acceptance. These parameters are routinely evaluated using methods which involve destruction of the fruit; as a result, only control batches can be analyzed. The development of nondestructive analytical methods would enable the quality control of individual fruits. Near-IR spectroscopy (NIRS) was used to assess total soluble solid content (SSC, degrees Brix) and firmness (N) in intact plums. A total of 720 plums (Prunus salicina L. cv. 'African Pride', 'Black Diamond', 'Fortune', 'Laetitia', 'Larry Anne', 'Late Royal', 'Prime Time', 'Sapphire', and 'Songold') were used to obtain calibration models based on reference data and near-IR spectral data. Standard errors of cross-validation (SECV) and coefficients of determination for cross-validation (r(2)) were (0.77 degrees Brix; 0.83) for total soluble solids content and (2.54 N; 0.52) for firmness. Results suggest that NIRS technology enables fruit to be classified in terms of total soluble solid content and firmness, thus allowing increased sampling of each production batch and ensuring a given quality with greater precision and accuracy.     

红茶感官品质及成分近红外光谱快速检测模型建立

以在发酵过程中小叶种工夫红茶为研究对象,分别建立了基于近红外光谱检测技术的感官品质评分和理化品质指标(茶黄素、茶红素、茶褐素、儿茶素和酚氨比)的定量分析模型。在模型建立过程中,探讨了特征变量优选方法对预测模型的影响。首先,对获取的近红外光谱数据进行标准正态变量变换法(standard normal Z transformation,SNV)预处理,进而采用联合区间偏最小二乘回归(synergy interval PLS,Si-PLS)、随机蛙跳算法(shuffled frog leaping algorithm,SFLA)、竞争性自适应权重取样法(competitiveadaptivereweightedsampling,CARS)和连续投影(successive projections algorithm,SPA),筛选出各品质指标的最优特征波长变量;最后基于优选波长分别建立各发酵品质指标的偏最小二乘法(partial least squares regression,PLS)线性预测模型和支持向量机(support vector regression,SVR)非线性预测模型。模型结果比较表明,Si、CARS、SFLA和SPA等变量筛选方法可有效压缩变量,以及进一步提高模型精度。非线性模型的预测均方根误差值(root-mean-square error of prediction,RMSEP)均明显小于PLS模型,相关性系数(correlation coefficient,R)和相对分析误差(relative percent deviation,RPD)均高于PLS模型。对于红茶发酵品质的检测上,非线性模型性能优于线性模型。感官品质、茶褐素和儿茶素的最优变量SVR预测模型的RPD值分别为3.923、3.234和5.462,酚氨比和茶红素模型的RPD值分别为2.815和2.223。除茶黄素的评价模型外(RPD为1.77),基于最优特征波长的各品质指标SVR模型的RPD值均大于2,表明模型具有极好的预测性能。研究结果为实现工夫红茶发酵品质的近红外光谱快速检测的实际应用奠定理论基础。     

Ash and Mineral Contents in Leaves of Woody Species: Analysis by Near‐Infrared Reflectance Spectroscopy

Abstract

Near‐infrared reflectance spectroscopy (NIRS) was evaluated for its effectiveness to determine ash and mineral concentrations [potassium (K), magnesium (Mg), copper (Cu), iron (Fe), and zinc (Zn)], in a total of 182 leaf samples of 17 woody species located in the central‐western region of the Iberian Peninsula. Chemical analysis revealed great variability in all leaf mineral elements. This variability was mainly related to differences in leaf habit (deciduous versus evergreen) and to differences in mean leaf longevity and among leaf age classes within evergreen species. A set of samples including all 17 species and leaf age classes was used to develop the calibration equations using multiple linear regression (MLR) and partial‐least squares regression (PLSR). The set of samples that did not enter in the calibration was used for external validation. In general, the most satisfactory results were obtained using PLSR and derivative transformations. Despite the strong heterogeneity of the samples included in the study, the results showed that NIRS can be employed as an effective tool, alternative to the more time‐consuming standard methods. The best predictive model was obtained for ash content. Models with acceptable accuracy were obtained in the prediction of K and Mg contents. However, their applicability for the determination of trace elements was more limited.