鲜枣可溶性固形物最小二乘支持向量机动态检测研究

用可见/近红外光谱动态检测鲜枣的可溶性固形物含量。试验时样品以0.1m/s的速度运动,采集其可见/近红外漫反射光谱(350～2500nm)。用平均平滑法对120个赞皇枣样品、118个郎枣样品的光谱进行消噪处理,采用连续投影算法提取其特征波长,并建立相应的最小二乘支持向量机预测模型SPA/LS-SVM;同时将赞皇枣在500～1100nm范围的可见/短波近红外平滑光谱数据,郎枣在700～1500nm范围的平滑光谱数据用最小二乘支持向量机建立Smooth/LS-SVM预测模型,并对各自预测集样品(30个)的可溶性固形物含量进行预测和对比分析。结果表明:SPA/LS-SVM模型预测相关系数(赞皇枣0.833,郎枣0.847)与Smooth/LS-SVM模型预测相关系数(赞皇枣0.848,郎枣0.857)相差不大,且前者更精简,预测速度快,预测时间短,可以作为鲜枣可溶性固形物含量的一种动态检测方法,但模型的精度和稳定性需进一步提高。     

基于多源光谱分析技术的鱼油品牌判别方法研究

多源光谱分析技术被用于鱼油品牌快速无损鉴别。采用可见光谱分析技术、短波近红外光谱分析技术、长波近红外光谱分析技术、中红外光谱分析技术和核磁共振光谱分析技术采集了7种不同品牌的鱼油的光谱特征,并应用偏最小二乘判别分析法(partial least squares discrimination analysis,PLS-DA)和最小二乘支持向量机(least-squares support vector machine,LS-SVM)建立判别模型并比较判别结果。基于长波近红外光谱的PLS-DA模型和LS-SVM模型取得了最高识别正确率,建模集和预测集识别正确率均达到100%。采用中红外光谱和核磁共振谱分别建立的LS-SVM模型,也可以获得100%的判别正确率。而可见光谱和短波近红外光谱则判别准确率较差。且LS-SVM算法较PLS-DA更加适合用于建立光谱数据和鱼油品牌之间的判别模型。研究结果表面长波近红外光谱技术能够有效判别不同鱼油的品牌,为将来鱼油品质鉴定便携式仪器的开发提供了技术支持和理论依据。     

壶瓶枣轻微损伤可见/近红外光谱动态判别模型研究

为了研究快速识别轻微损伤壶瓶枣与完好壶瓶枣的有效方法,本文以轻微损伤壶瓶枣和完好壶瓶枣为研究对象,动态采集轻微损伤壶瓶枣和完好壶瓶枣的近红外光谱数据。采用S-G平滑与多元散射校正(MSC)相结合的方法预处理光谱数据,分别以预处理后的全光谱(FS)数据和采用主成分分析(PCA)法提取主成分、采用连续投影算法(SPA)提取特征波长作为输入变量,建立偏最小二乘判别分析(PLS-DA)和最小二乘支持向量机(LS-SVM)模型,比较4种损伤壶瓶枣及完好壶瓶枣的判别准确性。结果表明:采用PCA提取主成分有较明显的优势,对4种损伤壶瓶枣的判别准确性均能满足实际要求,且采用PCA-LS-SVM模型对4种轻微损伤壶瓶枣和完好壶瓶枣的正确判别率最佳,分别达到100%、86%、100%、100%和100%,总的正确判别率为97.2%。该研究为轻微损伤壶瓶枣的动态判别提供了新的理论基础。     

基于微型近红外光谱仪的油菜籽含油率模型参数优化研究

为实现油菜籽含油率快速无损检测,采用微型近红外光谱仪,结合竞争性自适应重加权(CARS)、遗传算法(GA)、连续投影算法(SPA)、无信息变量消除法(UVE)、向后区间偏最小二乘法(BIPLS)、联合区间偏最小二乘法(SIPLS)等方法优选油菜籽含油率近红外光谱特征波长,建立偏最小二乘回归(PLSR)和最小二乘支持向量机(LS-SVM)定量分析模型,同时对LS-SVM模型参数进行优化。研究表明,对PLSR模型,BIPLS+GA优选的26个特征波长建模效果最好,其预测相关系数(Rp)和预测均方根误差(RMSEP)分别为0.9330和0.0075,对LS-SVM模型,SIPLS+GA优选的13个特征波长建模效果最好,预测相关系数(Rp)和预测均方根误差(RMSEP)分别0.9192和0.0055。证明了波长优选和参数优化可有效简化油菜籽含油率近红外光谱定量分析模型,提高模型预测精度和稳定性,为进一步拓展微型近红外光谱仪的应用提供技术参考。     

基于NIR高光谱成像技术的滩羊肉内部品质无损检测

利用近红外高光谱成像技术对滩羊肉蛋白质和脂肪含量、pH值进行无损检测研究。通过高光谱系统(900~1700 nm)采集69个羊肉样本信息,先对全波段下的原始光谱和预处理后光谱建立偏最小二乘回归(PLSR)模型,对比优选出最佳预处理算法,后采用PLSR的加权β系数法提取特征波长,建立特征波长下各品质参数的PLSR模型,分析预测效果。结果表明:羊肉蛋白质、脂肪含量、pH值最佳预处理方法为基线校准(Baseline)、多元散射校正与S-G卷积平滑结合算法(MSC+SG)和原始光谱;利用特征波长建立预测模型,决定系数(RP2)分别为0.83、0.86和0.72,预测均方根误差(RMSEP)为0.57、0.09和0.12,可替代全波段建模。利用近红外高光谱成像技术对羊肉内部品质进行快速无损检测是可行的。     

盐池滩羊肉脂肪含量的高光谱预测模型构建

本文利用可见-近红外高光谱成像技术预测冷鲜滩羊肉脂肪含量,优选最佳预测模型。测定90个滩羊背最长肌的脂肪含量并采集其光谱图像,对原始光谱进行不同种预处理后,构建了全波段下的偏最小二乘回归(PLSR)和主成分回归(PCR)的光谱预测模型。为减少模型运算次数,在预处理效果最优的全波段模型上采用连续投影算法(SPA)、应用竞争性自适应重加权(CARS)、变量组合集群分析(VCPA)和波长空间迭代收缩(IVISSA)方法提取特征波长,构建脂肪含量的光谱预测模型。结果表明:采用归一化(Normlize)预处理后光谱构建的PLSR全波段模型效果最好,校正集模型相关系数(Rc)达到0.921;采用多元散射校正(MSC)预处理后光谱构建的PCR全波段模型效果最好,其校正集模型相关系数(Rc)达到0.850;在4种提取特征波长过程中,IVISSA算法所提取特征波长的交互验证均方根误差(RMSECV)最低,为0.0072;Normlize-IVISSA-PLSR模型较其他3种算法所构建的PLSR模型效果最优,其校正集相关系数(Rc)和预测集相关系数(Rp)值分别为0.931和0.754,表明利用高光谱技术对盐池滩羊肉脂肪含量进行预测是可行的。研究成果为冷鲜滩羊肉品质在线光谱快速无损检测系统开发提供理论依据。     

NIR高光谱成像技术联用SPA算法 快速检测五花肉的过氧化值

利用近红外(NIR)高光谱成像技术结合连续投影算法(SPA)快速、无损检测五花肉的过氧化值。通过高光谱成像系统采集样品的光谱图像,提取其反射光谱信息,经过基线校正(BC)、高斯滤波平滑(GFS)、中值滤波平滑(MFS)、卷积平滑(SGS)、移动平均值平滑(MAS)、标准正态变量变换(SNV)、多元散射校正(MSC)七种预处理后,利用偏最小二乘(PLS)建立预测模型。使用SPA筛选最优波长,重新预算,构建优化的PLS模型和多元线性回归(MLR)模型。结果显示,经过BC预处理(R_P=0.960,RMSEP=5.15×10-4g/100 g)和原始数据RAW(R_P=0.960,RMSEP=4.89×10-4g/100 g)的全波段PLS模型(F-PLS)预测过氧化值效果较好。优化结果显示,RAW的MLR模型(R_P=0.968,RMSEP=4.12×10-4 g/100 g)预测效果更好。研究表明,NIR高光谱成像技术联用SPA算法可潜在实现对五花肉过氧化值的快速无损检测。     

油茶果自然霉变程度的可见/近红外与中短波近红外光谱检测

分析利用可见/近红外光谱（400～1 000 nm）与中短波近红外光谱（900～1 700 nm）对不同自然霉变程度油茶果检测判别的可行性,实验同时采集不同霉变程度油茶果赤道阴面、阳面和接合面三点的两波段光谱,样品平均光谱的主成分分析（principal component analysis,PCA）发现不同霉变程度样品同组内具有一定聚类效果且PC1和PC2对于判别不同组间样品有效,全光谱偏最小二乘判别分析模型结果显示原始光谱已具有足够信息,建立的模型性能比预处理后全光谱更优。进一步进行特征波长选取,发现相比于PC载荷,连续投影法在两光谱范围选取波长建立的简化模型均为最优,预测集判别准确率与Kappa系数均为84.4%与0.766 7。结合预测集混淆矩阵发现,两光谱范围最优简化模型预测不同霉变组样品特异度相当,均在0.84以上,但900～1 700 nm中短波近红外光谱对于中等霉变程度的判别灵敏度（0.72）略高。本研究表明近红外光谱技术可用于油茶果的自然霉变程度检测,可见/近红外与中短波近红外光谱能力相当,考虑到仪器成本问题,可见/近红外光谱具有更好的实时检测应用前景。     

基于可见/近红外光谱技术的橘小实蝇侵染柑橘检测研究

目的 使用可见/近红外光谱技术实施橘小实蝇侵染柑橘不同时期的无损检测。方法 研究选取人工制备的不同侵染时期的柑橘样本作为研究对象,利用搭建的可见/近红外光谱系统测量的光谱信息结合人工标定的侵染时期,对原始光谱进行了5种预处理,采用竞争性自适应重加权算法(competitive adaptive reweighted sampling, CARS)、连续投影算法(successive projections algorithm, SPA)两类方法提取反映侵染柑橘时期变化的光谱特征波长,应用偏最小二乘判别分析(partial least squares discriminant analysis,PLS-DA)建立基于特征波长光谱的柑橘侵染时期分类模型,对比分析不同光谱预处理方法的模型分类效果。结果 原始光谱经多元散射校正(multiplicative scatter correction, MSC)预处理的模型分类效果最佳,分别经CARS方法和SPA方法优选出了34和16个光谱特征波长。采用MSC-CARS-PLS-DA方法构建的模型分类效果最好,总准确率、假阳率分别为96.8%和0.0...     

基于近红外高光谱技术快速检测豌豆蛋白掺假牛肉

利用近红外高光谱成像技术结合偏最小二乘模型(PLSR)对牛肉掺入豌豆蛋白进行快速无损检测。将豌豆蛋白按照1%～30%(w/w)的浓度梯度掺入牛肉糜中(掺入浓度间隔1%)，共获得93个样品以采集其光谱信息。经移动平均值平滑(Moving average smoothing,MAS)、高斯滤波平滑(Gaussian filter smoothing,GFS)、基线校正(Baseline correction,BC)、S-G卷积平滑(Savitzky Golay convolution smoothing,SGCS)、标准正态变量校正(Standard normal variable correction,SNV)等5种方法预处理光谱信息后，利用PLSR算法构建预测模型。然后采用回归系数法(Regression Coefficient,RC)、逐步回归法(Stepwise)和连续投影算法(Successive Projections Algorithm,SPA)筛选最优波长进行模型优化。结果显示，经过GFS预处理所构建的全波段PLSR模型预测效果更好(R_P^2<...     

Nondestructive evaluation of jujube quality by visible and near-infrared spectroscopy

This study compared prediction ability of interactance, transmission measurements of visible and near-infrared (Vis/NIR) spectroscopy in detecting the soluble solids content (SSC) of jujubes. Calibration models relating Vis/NIR spectra to SSC were developed based on partial least squares regression (PLSR) with respect to the logarithms of the reciprocal absorbance (log (1/R)), its first and second derivatives (D₁log (1/R), D₂log (1/R)). The PLSR models for prediction samples resulted correlation coefficients (r_p) of 0.74-0.91 and root mean square error of prediction (RMSEP) of 2.018-3.200 °Brix for interactance; r_p of 0.63-0.73 and RMSEP of 3.517-3.863 °Brix for transmission, respectively. The results indicate that interactance displays an obvious advantage over transmission measurement.The reflectance measurement was used to access the discrimination potential in sorting external insect-infested jujubes from intact class. Stepwise discriminant analysis (SDA) was performed to identify the effective wavelengths that best discriminated the insect-infested jujubes from intact jujubes and to derive a discriminant function in classifying the jujubes showing external infestation and those that were free of infestation. The results showed that log (1/R) had better correct classification rate than D₁log (1/R), and D₂log (1/R) for classifying intact, insect-infested and stem-end classes.     

Assessing the temperature influence on the soluble solids content of watermelon juice as measured by visible and near-infrared spectroscopy and chemometrics

The visible/near-infrared (Vis/NIR) spectra may be affected by environmental variations, such as temperature fluctuation. In the current study, this type of influence was investigated. The spectra of watermelon juice samples were collected and analyzed under nine temperatures, varying from 0 °C to 40 °C at intervals of 5 °C. Discriminant analysis (DA) and particle least square regression (PLSR) statistical methods were used to establish calibration models. The regression models were compared under nine different temperatures. Local and global compensation models were also established. The results indicated that there was an influence of temperature on the Vis/NIR spectra. The average absorbance shifted with the change of temperature. The PLSR model established at 20 °C performed better than at other temperatures. The results also indicated that the local models were sensitive to changes in temperature. Additionally, the global model showed good prediction ability, which made temperature fluctuation a negligible interference.     

Comparison of different modes of visible and near-infrared spectroscopy for detecting internal insect infestation in jujubes

This study compared the abilities of the interactance, reflectance, and transmission modes of visible and near-infrared (Vis/NIR) spectroscopy in detecting internal insect-infested jujubes. Statistical analysis was performed to identify the effective wavelengths that best discriminated the insect-infested jujubes from intact jujubes and to derive a discriminant function in classifying the jujubes showing internal insect infestation and those that were free of infestation. The highest correct classification rates obtained from the above modes were 100%, 90.0%, and 97.0%, respectively. The interactance mode in the long-wave NIR (LWNIR) range is preferable to the transmission mode in the visible and short-wave near-infrared (VSWNIR) ranges. Furthermore, the transmission mode in the VSWNIR range displayed an obvious advantage over the reflectance mode in every range. The results indicate that it is possible to use both the interactance and transmission modes to develop a system in detecting the internal qualities of jujubes.     

Quantification of Nitrogen Status in Rice by Least Squares Support Vector Machines and Reflectance Spectroscopy

The estimation of nitrogen status non-destructively in rice was performed using canopy spectral reflectance with visible and near-infrared reflectance (Vis/NIR) spectroscopy. The canopy spectral reflectance of rice grown with different levels of nitrogen inputs was determined at several important growth stages. This study was conducted at the experiment farm of Zhejiang University, Hangzhou, China. The soil plant analysis development (SPAD) value was used as a reference data that indirectly reflects nitrogen status in rice. A total of 64 rice samples were used for Vis/NIR spectroscopy at 325–1075 nm using a field spectroradiometer, and chemometrics of partial least square (PLS) was used for regression. The correlation coefficient (r), root mean square error of prediction, and bias in prediction set by PLS were, respectively, 0.8545, 0.7628, and 0.0521 for SPAD value prediction in tillering stage, 0.9082, 0.4452, and −0.0109 in booting stage, and 0.8632, 0.7469, and 0.0324 in heading stage. Least squares support vector machine (LS-SVM) model was compared with PLS and back propagation neural network methods. The results showed that LS-SVM was superior to the conventional linear and non-linear methods in predicting SPAD values of rice. Independent component analysis was executed to select several sensitive wavelengths (SWs) based on loading weights; the optimal LS-SVM model was achieved with SWs of 560, 575–580, 700, 730, and 740 nm for SPAD value prediction in booting stage. It is concluded that Vis/NIR spectroscopy combined with LS-SVM regression method is a promising technique to monitor nitrogen status in rice.     

Discrimination of Producing Areas of Auricularia auricula Using Visible/Near Infrared Spectroscopy

Visible and near infrared (Vis/NIR) spectroscopy combined with chemometric methods was applied for the discrimination of producing areas of Auricularia auricula. Four major varieties of commercial A. auricula were prepared for spectral acquisition. Some pretreatments were performed, such as Savitzky–Golay smoothing, standard normal variate, and the first and second Savitzky–Golay derivative. The scores of the top four latent variables, extracted by partial least squares, were considered as the inputs of back propagation neural network (BPNN) and least squares-support vector machine (LS-SVM). The performance was validated by 60 validation samples. The excellent recognition ratio was 98.3% by BPNN and 96.7% by LS-SVM model with the threshold prediction error ±0.1. The results indicated that Vis/NIR spectroscopy could be used as a rapid and high-precision method for the discrimination of different producing areas of A. auricula by both BPNN and LS-SVM methods.     

Non-destructive discrimination of paddy seeds of different storage age based on Vis/NIR spectroscopy

The potential of visible/near infrared reflectance (Vis/NIR) spectroscopy for non-destructive discrimination of paddy seeds of different storage age was examined based on Vis/NIR spectroscopy coupled with chemometrics. Data from 210 samples of paddy seed were collected from 325 to 1075 nm using a field spectroradiometer. The spectral data were processed and analyzed by chemometrics, which integrated the methods of wavelet transform (WT), principal component analysis (PCA) and artificial neural networks (ANN) modelling. The noise of spectral data was filtered and diagnostic information was extracted by the WT method. Then, diagnostic information from WT was visualized in principal components space, in which the structures with the storage period were discovered. Finally, the first eight principal components, which accounted for 99.94% of the raw spectral variables, were used as the input for the ANN model. A promising model was achieved with a high discrimination accuracy rate of 97.5%. Thus, an effective and non-destructive way to discriminate paddy seeds of different storage periods was put forward.     

Moisture content detection of maize seed based on visible/near-infrared and near-infrared hyperspectral imaging technology

To realise accurate and nondestructive detection on moisture content of maize seed based on visible/near-infrared (Vis/NIR) and near-infrared (NIR) hyperspectral imaging technology, the hyperspectral images on two sides (embryo and endosperm sides) of each maize seed of four varieties were collected. The effects of average spectra extraction regions, that is centroid region and whole seed region, and different spectral preprocessing methods, were investigated. Uninformative variable elimination (UVE) was used to extract the feature wavelengths, and the partial least squares regression (PLSR) prediction models were established. The results showed that extracting the average spectra from the centroid region did better than from the whole seed region, and S-G smoothing was prior to other preprocessing methods. The PLSR models established with NIR spectra had better performance than that with Vis/NIR spectra. The model developed for a single variety was superior to that for all varieties together.     

A comparative study for the quantitative determination of soluble solids content,pH and firmness of pears by Vis/NIR spectroscopy

Visible and near infrared (Vis/NIR) spectroscopy was investigated to determine the soluble solids content (SSC), pH and firmness of different varieties of pears. Two-hundred forty samples (80 for each variety) were selected as sample set. Two-hundred ten pear samples (70 for each variety) were selected randomly for the calibration set, and the remaining 30 samples (10 for each variety) for the validation set. Partial least squares (PLS) and least squares-support vector machine (LS-SVM) with different spectral preprocessing techniques were implemented for calibration models. Different wavelength regions including Vis, NIR and Vis/NIR were compared. It indicated that Vis/NIR (400–1800 nm) was optimal for PLS and LS-SVM models. Then, LS-SVM models were developed with a grid search technique and RBF kernel function. All LS-SVM models outperformed PLS models. Next, effective wavelengths (EWs) were selected according to regression coefficients. The EW-LS-SVM models were developed and a good prediction precision and stability was achieved compared with PLS and LV-LS-SVM models. The correlation coefficient of prediction (r_p), root mean square error of prediction (RMSEP) and bias for the best prediction by EW-LS-SVM were 0.9164, 0.2506 and −0.0476 for SSC, 0.8809, 0.0579 and −0.0025 for pH, whereas 0.8912, 0.6247 and −0.2713 for firmness, respectively. The overall results indicated that the regression coefficient was an effective way for the selection of effective wavelengths. LS-SVM was superior to the conventional linear PLS method in predicting SSC, pH and firmness in pears. Therefore, non-linear models may be a better alternative to monitor internal quality of fruits. And the EW-LS-SVM could be very helpful for development of portable instrument or real-time monitoring of the quality of pears.