基于高光谱成像技术和连续投影算法检测葡萄果皮花色苷含量

应用高光谱成像技术结合连续投影算法（SPA）实现葡萄果皮中花色苷含量的快速无损检测。采集60 组样本高光谱图像,获取样本光谱曲线,并采用多元散射校正预处理方法提高信噪比。然后采用SPA选择光谱变量,将其作为多元线性回归（MLR）、偏最小二乘（PLS）模型和BP神经网络（BPNN）的输入变量,分别建立SPAMLR、SPA-PLS和SPA-BPNN模型并与全光谱变量PLS模型相比较。结果表明,SPA-MLR、SPA-BPNN和SPA-PLS模型的预测精度均优于全光谱变量PLS模型,其中SPA-PLS模型获得了最佳预测结果,其预测相关系数Rp和预测均方根误差（RMSEP）分别为0.900 0和0.550 6。结果表明,利用近红外高光谱成像技术能够有效检测酿酒葡萄果皮中花色苷含量。     

斜率/截距算法在猕猴桃可溶性固形物含量便携式检测中的应用

基于微型光谱仪搭建一台体积小、易操作的猕猴桃可溶性固形物含量（soluble solids content,SSC）检测装置,并用该装置采集了‘华优’、‘徐香’和‘西选’3?个品种猕猴桃的近红外光谱,利用不同特征波长提取方法从全光谱中提取特征波长,并比较不同方法提取的特征波长及全光谱对3?个品种猕猴桃SSC的偏最小二乘模型预测精度的影响;用斜率/截距算法结合‘华优’猕猴桃大样本模型,预测‘徐香’和‘西选’猕猴桃的SSC。结果表明,连续投影算法对于模型简化效果最好,其对‘华优’、‘徐香’和‘西选’猕猴桃大样本的预测均方根误差（root mean square error of prediction,RMSEP）分别为0.583、0.678?°Brix和0.646?°Brix;用斜率/截距算法对‘华优’猕猴桃SSC模型进行校正时,仅用10?个‘徐香’和50?个‘西选’猕猴桃便能有效地提高对SSC的预测性能,其RMSEP分别为0.966?°Brix和0.875?°Brix。本研究为进一步构建精度更高、更便捷的微型集成式猕猴桃SSC检测仪提供理论依据。     

基于反向传播神经网络和高光谱成像的芒果可溶性固形物含量检测

目的 比较反向传播神经网络(backpropagation algorithm neural network, BPNN)模型与偏最小二乘回归(partial least squares regression, PLSR)模型在预测芒果可溶性固形物含量(soluble solids content, SSC)方面的性能评估。方法 使用高光谱成像仪和全自动折光仪采集芒果的近红外高光谱及SSC数据建立两种预测模型, 通过采用多元散射校正(multiplicative scatter correction, MSC)进行光谱预处理, 利用遗传算法(genetic algorithm, GA)、区间变量迭代空间收缩算法(interval variable iterative space shrinkage algorithms, IVISSA)和变量组合群体分析算法(variable combination population analysis, VCPA)提取特征波长变量, 通过比较不同特征波长提取方法进一步优化对比预测模型。结果 与PLSR模型相比, BPNN模型在预测SSC方面更为有效。而在IVISSA特征波长变量提取后优化的BPNN模型预测能力最佳, 预测集判定系数 、均方根误差(root mean square error of prediction, RMSEP)、残差预测偏差(residual prediction deviation, RPD)分别为0.8641、0.3924和2.7127。结论 该模型可快速、准确地检测芒果的SSC, 并证明可见光-近红外高光谱成像与反向传播神经网络模型相结合有望预测芒果的SSC, 为开发在线芒果SSC无损检测系统奠定基础。     

紫外/可见光谱技术无损检测李子可溶性固形物含量

利用光谱技术结合化学计量学对李子可溶性固形物含量检测进行研究,为李子品质无损检测提供科学方法。通过反射式光谱采集系统获取了"红"李子和"青"李子的平均光谱,并对原始光谱数据进行预处理;应用连续投影算法(SPA)和竞争性自适应重加权算法(CARS)对预处理后的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预测李子可溶性固形物含量的误差反向传播(BP)网络模型。结果表明:利用SPA和CARS算法分别从全光谱的1024个波长中选取出31和104个特征波长;而基于特征波长建立的CARS-BP网络模型效果最优,其相关系数r_c为0.998,r_p为0.887,均方根误差RMSEC为0.026,RMSEP为1.767。这表明光谱技术结合化学计量学进行李子可溶性固形物含量的无损检测具有可行性。     

基于近红外高光谱成像技术的涩柿SSC含量无损检测

对150个涩柿采集900~1 700nm波段的近红外高光谱图像信息,利用蒙特卡罗—无信息变量消除(MC-UVE)和连续投影算法(SPA)对感兴趣区域光谱进行波长优选。通过MC-UVE-SPA优选出924.69,928.05,1 112.72,1 270.91,1 365.3,1 402.42,1 453.06,1 547.69nm 8个特征波长,对应的光谱反射率作为柿子可溶性固性物含量(SSC)检测的偏最小二乘回归(PLSR)检测模型输入,其预测集相关系数rpre=0.942,预测集均方根误差RMSEP=1.009°Brix。结果表明,MC-UVE-SPA可以有效提取与柿子SSC含量相关的特征信息,从而保留较少的波长建立较好的预测模型。     

基于高光谱成像的贵长猕猴桃硬度快速预测

为了提高猕猴桃硬度预测的效率。应用可见/近红外(390～1030 nm)高光谱成像系统获取贵长猕猴桃的高光谱图像信息,并提取每个样品感兴趣区域的反射光谱,采用标准正态变换消除原始反射光谱中的噪声干扰;基于竞争性自适应重加权(competitive adaptive reweighted sampling,CARS)与连续投影算法筛选特征变量,建立误差反向传播神经网络和多元线性回归(multi linear regression,MLR)模型检测贵长猕猴桃硬度。结果表明：应用CARS从256个变量中筛选了35个特征变量,运算效率提升了近11倍(即运算时间从5.84 s降到了0.54 s);构建的CARS-MLR检测模型具有较大的r_c=0.95和r_p=0.92,较小的RMSEC=1.65 kg/cm²和RMSEP=1.99 kg/cm²,且RPD值(2.47)大于2,表明CARS-MLR模型具有非常好的检测性能,利用高光谱成像技术以及化学计量学可以实现贵长猕猴桃硬度的快速无损检测。     

基于VIS/NIR高光谱技术的灵武长枣SSC含量的无损检测研究

利用高光谱技术对灵武长枣可溶性固形物含量(SSC)检测进行研究,为灵武长枣内部品质无损检测提供科学方法。以灵武长枣为对象,对光谱数据进行预处理,应用连续投影算法(SPA)和正自适应加权算法(CARS)进行关键波长的选择,通过偏最小二乘回归(PLSR)和主成分回归(PCR)建立预测模型。实验结果表明:采用去趋势(Detrend)预处理算法效果最优,PLSR模型的交叉验证相关系数(Rcv)为0.809,交叉验证均方根误差(RMSECV)为1.331;通过SPA算法和CARS算法能够有效地对光谱数据进行降维处理,对SPA优选的8个和CARS优选的21个特征变量分别用PLSR和PCR建立预测模型,CARS-PLSR建模效果最优,其相关系数(Rp)为0.864,预测均方根误差(RMSEP)为1.174;研究结果表明基于高光谱成像技术采集的灵武长枣漫反射光谱进行SSC无损检测具有可行性。     

采用二次连续投影法和BP人工神经网络的寒富苹果病害高光谱图像无损检测

为提供苹果病害在线、快速、无损检测的理论依据,采用高光谱成像技术进行了北方大面积种植的寒富苹果病害无损检测研究。寒富苹果的主要病害有炭疽病、苦痘病、黑腐病和褐斑病害。为选择较少的有效波长而利于在线快速检测,首先采集高光谱苹果图像,分割出感兴趣区域并提取光谱信息,然后采用连续投影算法(successive projections algorithm,SPA)从全波长(500~970 nm)中提取了10个特征波长SPA1(502、573、589、655、681、727、867、904、942 nm和967 nm),再对这10个特征波长采用连续投影算法提取3个特征波长SPA2(681、867 nm和942 nm)。最后利用全波长光谱信息、SPA1提取的10个特征波长的光谱信息和SPA2提取的3个特征波长的光谱信息作为输入矢量采用线性判别分析、支持向量机和BP人工神经网络(BP artificial neural network,BPANN)模型进行苹果病害的检测。通过对检测结果分析,最终选择SPA2-BPANN为最佳检测方法,训练集检测率达100%,验证集检测率达100%。结果表明,高光谱成像技术可以有效对苹果病害进行检测,所获得的特征波长可为开发多光谱成像的苹果品质检测和分级系统提供参考。     

高光谱成像技术对灵武长枣果皮强度的无损检测

利用高光谱技术对灵武长枣果皮强度检测进行研究,为灵武长枣外部品质无损检测提供科学方法。采集120个灵武长枣的400~1000 nm的高光谱图像,对光谱数据进行预处理;应用连续投影算法(SPA)、正自适应加权算法(CARS)和无信息变量消除法(UVE)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回归(PLSR)和最小二乘支持向量机(LS-SVM)果皮强度预测模型。结果表明:采用标准正态变换(SNV)预处理算法效果最优,其PLSR模型的交叉验证相关系数(Rcv)为0.8207,交叉验证均方根误差(RMSECV)为9.9630;利用SPA、CARS和UVE法从全光谱的125个波长中分别选取出29个、31个和31个特征波长;而基于全光谱建立的LS-SVM模型效果最优,其预测相关系数(Rp)为0.9555,预测均方根误差(RMSEP)为3.8282;研究结果表明基于高光谱成像技术采集的灵武长枣漫反射光谱进行果皮强度无损检测具有可行性。     

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。     

利用反射和半透射高光谱图像检测水蜜桃早期冷害

研究了应用高光谱图像技术检测‘霞晖5号’水蜜桃早期冷害的方法。通过分析桃果实采后贮藏期间出汁率的变化,确定冷害发生的初始时间。进而针对早期冷害桃果实的400～1 000 nm波段的高光谱反射和半透射图像,应用独立主成分分析方法优选出冷害的特征波长,反射条件下为656、674 nm和704 nm,半透射条件下为640、745 nm和811 nm,并通过提取每个特征波长下的光谱平均值作为Fisher判别方法建模的特征集,所建水蜜桃早期冷害判别模型的预测准确率分别为83.0%和94.0%。结果表明,高光谱技术可以检测桃果实早期冷害,且半透射照射方式的判别结果优于反射方式。     

高光谱图像对白萝卜糠心的无损检测

为实现白萝卜异常品质糠心的无损检测,构建高光谱图像技术检测白萝卜糠心的检测系统。获取了光源透射、反射和半透射模式下白萝卜的高光谱图像信息,结合偏最小二乘分析（partial least squares discriminantanalysis,PLS-DA）、支持向量机（support vector machine,SVM）、人工神经网络（artificial neural network,ANN）3 种算法分别建立白萝卜糠心的识别模型。结果表明：3 种检测模式中,基于透射模式的高光谱图像系统检测准确率最高;3 种预测模型中,ANN模型优于PLS-DA和SVM模型。其中,基于透射模式的ANN模型,高光谱图像对萝卜糠心的检测总体准确率达94.3%,效果最好。因此,采用透射模式的高光谱图像技术对白萝卜糠心的检测是可行的。     

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.     

Identifying CPPU-Treated Kiwifruits Using Near-Infrared Hyperspectral Imaging Technology

The application of forchlorfenuron (N-(2-chloro-4-pyridyl)-N′-phenylurea), short form CPPU, on kiwifruits has become an important factor that influences kiwifruit economic efficiency and the health development of kiwifruit industry. This study aims to investigate the feasibility of using hyperspectral imaging technology to identify kiwifruits treated with CPPU (named as treated kiwifruits) from kiwifruits without CPPU treatment (named as untreated kiwifruits), and to investigate which model, developed for a single variety or for two varieties together, has better identification performance. Two hundred and forty “Xixuan” kiwifruits and 240 “Huayou” kiwifruits (120 treated kiwifruits and 120 untreated kiwifruits for each variety) were used to obtain hyperspectral from 865.11 to 1711.71 nm. The samples were divided into calibration set and prediction set based on Kennard–Stone method as the ratio of 3:1. Standard normal variate transformation was used to preprocess obtained spectra. Successive projections algorithm (SPA) was applied to extract the characteristic wavelengths from full spectra (FS). Support vector machine (SVM) and extreme learning machine (ELM) modeling methods were used to establish identification models of treated kiwifruits based on FS, characteristic wavelengths extracted by SPA, and universal wavelengths (UWs) extracted from characteristic wavelengths selected by SPA for “Xixuan,” “Huayou,” and two varieties together, respectively. The results showed that the number of characteristic wavelengths selected by SPA were 18, 18, and 21 for “Xixuan,” “Huayou,” and two varieties together, respectively. Five UWs were found for the three different samples. The best model was SPA-ELM for both “Xixuan” (99.8 % accuracy rate for predication set) and “Huayou” (100.0 % accuracy rate for predication set). Models of SPA-SVM and SPA-ELM, whose accuracy rate reached 100 % for both calibration and predication sets, had the best performance when the two varieties were used together. The performances of models built for two varieties together were better than that for “Xixuan,” but they were worse than that for “Huayou.”.The study indicates that NIR hyperspectral imaging technique can be used as a noninvasive method for identifying CPPU-treated kiwifruits from untreated ones, and it is potential to develop a model based on multi-varieties together.     

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。     

Application of Long-Wave Near Infrared Hyperspectral Imaging for Measurement of Soluble Solid Content (SSC) in Pear

Soluble solid content (SSC) in fruit is one of the most crucial internal quality factors, which could provide valuable information for commercial decision-making. Near-infrared (NIR) technique has effective potentials for determining the SSC since NIR was sensitive to the concentrations of organic materials. In this study, a novel NIR technique, long-wave near infrared (LWNIR) hyperspectral imaging with a spectral range of 930–2548 nm, was investigated for measuring the SSC in pear, which has never been examined in the past. A new combination of Monte Carlo-uninformative variable elimination (MC-UVE) and successive projections algorithm (SPA) was proposed to select most effective variables from LWNIR hyperspectral data. The selected variables were used as the inputs of partial least square (PLS) to build calibration models for determining the SSC of ‘Ya’ pear. The results indicated that calibration model built using MC-UVE-SPA-PLS on 18 effective variables achieved the optimal performance for prediction of SSC comparing with other developed PLS models (MC-UVE-PLS and SPA-PLS) by comprehensively considering the accuracy, robustness, and complexity of models. The correlation coefficients between the predicted and actual SSC were 0.88 and 0.88 and the root mean square errors were 0.49 and 0.35 °Brix for calibration and prediction set, respectively. The overall results indicated that long-wave near infrared hyperspectral imaging incorporated to MC-UVE-SPA-PLS model could be applied as an alternative, fast, accurate, and nondestructive method for the determination of SSC in pear.