紫外/可见光谱技术无损检测苹果的挤压损伤

目的 通过紫外/可见光谱技术结合模式识别算法,建立挤压损伤苹果的Fisher识别模型、K最近邻（KNN）识别模型和偏最小二乘判别分析（PLS-DA）识别模型。方法 以挤压损伤苹果和无损苹果为研究对象,采用光谱仪采集2种苹果的光谱反射率,综合比较不同光谱预处理方法（二阶微分（SD）、标准正态变换（SNV）和多元散射校正（MSC））对各模型识别效果的影响,并利用主成分分析方法（PCA）对预处理后的光谱数据进行降维,并提取能反映损伤苹果的特征光谱。结果 采用主成分分析法选择了累计贡献率超过99%的前7个主成分（P1—P7）作为特征光谱数据,有效地实现了光谱数据的降维;二阶微分对光谱反射率预处理的效果最好;3种判别模型均能满足实际要求,且SD+Fisher和SD+PLS-DA识别模型对校正集和预测集样本的总正确识别率均高达100%。结论 研究结果有助于实现挤压损伤苹果的快速识别。     

基于主成分回归的猕猴桃可溶性固形物无损检测

目的为实现猕猴桃可溶性固形物含量的快速无损检测。方法实验采用反射式光谱采集系统获取不同成熟期“贵长”猕猴桃的反射光谱。比较3种光谱预处理方法(标准正态变换、多元散射校正以及二阶导数)对回归预测模型的影响;采用主成分分析对预处理后的光谱数据进行降维,并基于提取的特征变量,建立猕猴桃可溶性固形物含量的回归预测模型。结果采用主成分分析,从1024个全光谱波段中提取了前16个主成分作为特征变量;基于特征变量建立的回归预测模型具有较好的预测性能,其预测集决定系数R2 P=0.88,剩余预测偏差为2.94。结论基于紫外/可见光谱结合主成分回归可以很好地预测猕猴桃可溶性固形物含量。     

基于高光谱技术的花椰菜农药残留检测

目的 为了快速、无损地检测花椰菜上的农药残留,采用高光谱成像技术分别对花椰菜上是否含有苏云金杆菌、高效氯氰菊酯和虫螨茚虫威等3种农药进行无损检测研究,并且跟踪研究检测效果最好的农药安全间隔期。方法 对含有农药和不含农药的花椰菜样本进行高光谱成像处理,提取感兴趣区域的光谱数据。剔除前后20个波段的原始光谱数据,以降低噪声的影响,针对剩余216个波段(950~1 666 nm)的数据分别采用卷积平滑(S–G)、多元散射校正(MSC)和变量标准化(SNV)等3种算法对光谱数据进行优化。为了提高判别运行速率,采用竞争性自适应重加权算法(CARS)提取3种农药光谱数据的特征波长,并建立偏最小二乘法(PLS)判别模型。结果 基于SNV优化后的PLS模型对花椰菜上3种农药的识别准确率相对最高,其中虫螨茚虫威农药样本的测试效果相对最好,识别率为100%,随后对该农药进行了连续7 d的检测,其结果符合农药的消散规律。结论 将高光谱图像技术应用于花椰菜的农药残留检测具有很好的应用前景。     

基于卷积神经网络的食品塑料包装袋光谱识别

目的实现食品塑料包装袋的快速检测和材质区分。方法研究使用高光谱成像技术在450～950nm波长范围下采集了49组不同食品包装袋样本的光谱数据,利用Savitzky-Golay平滑滤波、数据归一化和主成分分析进行预处理,建立决策树、支持向量机2种传统机器学习模型和卷积神经网络模型,并比较了它们对包装袋材质的识别性能。结果决策树模型与支持向量机模型的验证识别率分别为87.8%和88.9%,卷积神经网络模型的验证识别率高达100%,损失函数值最终下降到0.0171且达到收敛,在分类效果和精度上具有明显的优势。结论高光谱检测方法不破坏检材,重现性好,稳定性强,实现了对食品塑料包装袋的精准识别。卷积神经网络模型对食品包装袋高光谱数据的识别效果最好,为食品包装袋质量检测领域中塑料包装袋的识别鉴定提供依据。     

利用成像光谱仪识别猪肉和牛肉

利用成像光谱仪采集猪肉和牛肉的光谱,应用主成分分析(PCA)对所获得的原始光谱数据进行降维处理,分别利用KNN判别、人工神经网络(ANN)、支持向量机分类(SVM)三种建模方式,建立判别模型,并对猪肉和牛肉各20个预测样品进行识别.结果显示3种分类模型的正确识别率分别为92.5%、97.5%、100%.表明利用成像光谱仪可以实现对猪肉和牛肉的快速、准确、无损分类检测.     

基于多层感知器-Fisher判别分析的车用保险杠红外光谱鉴别

利用光谱检测和数据分析实现不同种类汽车保险杠碎片的快速无损检测和精确识别与分类具有重要的意义,采集8个品牌共计38个汽车保险杠碎片的红外谱图,预处理采用自动基线校正、峰面积归一化、多元散射校正和Savitzky-Golay平滑,通过小波阈值进行去噪处理,借助主成分分析(PCA)提取特征变量,建立基于多层感知器(MLP)的Fisher判别分析(FDA)分类模型。实验结果表明:数据在20维矩阵上特征提取最好,包含的信息量足够大,MLP模型对样本种类的识别准确率为74.70%,在20维特征数据上构建FDA模型,求得Z_1和Z_2判别函数式以及各样本分布散点图,其中35个样本实现了正确的区分和归类,分类准确率为92.1%,相比较单一MLP模型,MLPFDA区分能力更强、精度更高。综上,将红外光谱技术与MLP-FDA模型结合可以实现对车用保险杠碎片的快速无损鉴别,且模型检测精度高,方法具有普适性和借鉴意义。     

基于样本选择的光谱重构研究

目的为了提高使用主成分分析法重构光谱反射率的重构精度。方法利用Matlab进行仿真实验,选择3种不同色卡作为训练样本,使用主成分分析法探究主成分个数和样本间隔对重构结果的影响。结果主成分个数为4时,贡献率均超过99%;样本间隔为10 nm时,RC24色卡重构效果最好,其平均色差2.37ΔE_(ab)~*平均均方根误差为0.0185。结论训练样本的选择会影响光谱重构精度,RC24色卡具有数据量小、重建精度较高的特点,在颜色复制领域可以优先选择。     

基于PCA、FLD和小波包子图融合的人脸识别算法

基于单一理论的识别方法由于各种因素的限制,存在自身固有的局限性。为提高算法的效率和准确率,本文结合主成分分析法(PCA)和Fisher线性判别法(FLD),提出一种基于融合小波包子图的人脸识别方法 FW-PCA-FLD。该方法首先将小波包分解后的人脸子图像根据其能量分布特性进行加权融合,然后利用PCA方法对融合后的小波包图像进行特征提取,最后用Fisher线性判别找到合适的投影空间,通过度量训练样本与测试样本在投影空间上的投影系数进行人脸的分类识别。在CMU PIE人脸库、JAFFE人脸库上的实验结果表明,本文提出的人脸识别算法不仅在正确识别率和识别时间效率上表现较为优越,而且对不同光照、表情、姿态变化下的人脸识别都保持较好的适应性。     

差分拉曼光谱结合化学计量学对白色购物纸袋的检验研究

目的 建立差分拉曼光谱用于无损识别白色购物纸袋的方法。方法 对收集到的60种不同品牌、不同规格的白色购物纸袋进行拉曼光谱测定,对样品的拉曼光谱图进行预处理,根据光谱图对样品进行初步分类,并结合化学计量方法对样品进行分组。应用Fisher判别分析方法对分类结果进行验证。最后应用RBF模型对未知样本进行分类判别。结果 结合样品中所含的碳酸钙、滑石粉、硫酸钡的不同,可初步将白色购物纸袋样品分为五大类,采用K-均值聚类方法继续细分,通过Fisher判别方法对样品分结果进行验证,判别准确率为100%。应用神经网络RBF模型对未知样本进行判别分析,准确率达到89.48%。结论 该方法简便易行,为白色购物纸袋的分类提供了科学的依据,也为公安基层工作的开展提供了便捷的办法。     

基于多元分类模型的颜料红外光谱鉴别

颜料的检验与认定是司法鉴定中一项重要的工作。在传统的分析中,侦查人员往往通过人工逐一比对和分析,其耗时长,误差大,无法满足无损、快速、准确检验现场颜料样本的需求。该文提出一种检验方法,以期实现对物证无损、快速、准确的检验与鉴定。通过采集并分析不同品牌共计48个颜料样本的红外谱图,采用多元散射校正、Savitzky-Golay平滑和峰面积归一化开展预处理工作,建立基于K近邻算法等4种分类模型,从而实现不同颜料间的区分和归类。在区分水粉类颜料和毕加索丙烯画颜料时,相较于K近邻和Fisher判别模型,多层感知器分类模型准确率更高(总体分类准确率为100%),分类结果更好。在经过主成分分析提取特征变量后,分类模型对两类颜料的区分准确率均为100%。应用MLP结合PCA构建的分类模型对颜料样本的区分效果最佳。针对水粉类中的两类即普通水粉类和毕加索水粉类颜料样本,多层感知器分类模型对其的分类准确率为97.2%,针对普通水粉类样本的两个品牌(贝碧欧和晨光),多层感知器分类模型的分类准确率为100%,实验结果理想。利用中红外光谱结合多元分类模型可实现对颜料样本准确的鉴别与区分,其快速无损准确,降低检验鉴定成本,提高检验鉴定效率,可为其他物证的鉴别与分析提供一定的参考。     

苹果可溶性固形物含量的多元线性回归预测

目的可溶性固形物含量是评价苹果品质的重要指标,为开发苹果品质快速检测设备提供理论基础。方法采用高光谱图像采集系统采集"富士"苹果的高光谱图像,并获取感兴趣区域的反射光谱;应用连续投影算法对标准正态变换预处理后的光谱进行降维;基于选取的特征光谱建立预测苹果可溶性固形物含量的多元线性回归模型。结果采用连续投影算法从256个全光谱中提取了12个波长作为特征光谱,明显提升了多元线性回归预测模型的运行效率;基于特征光谱建立的多元线性回归预测模型具有较好的校正性能(R_C=0.804,R_(Cm)=0.665%)和预测性能(R_P=0.859,R_(Pm)=0.413%)。结论研究建立的苹果可溶性固形物含量预测模型性能较稳定,可以满足实际应用需求。     

Independent component analysis in information extraction from visible/near-infrared hyperspectral imaging data of cucumber leaves

Hyperspectral imaging at visible and short near infrared (VIS/SNIR) region has been used to estimate the pigment content of leaves. A complicating feature of measurements with any hyperspectral imaging methodology is the large amount of information generated during the measurement process. In this paper we discuss the identification of the desirable information using independent component analysis (ICA). After hyperspectral image acquisition and pre-processing, the average spectra obtained from the region of interest (ROI) in cucumber leaves were used for model development. Additionally a multi-linear regression model was developed for the prediction of cucumber leaf chlorophyll content. When compared with normal principal component analysis (PCA), the ICA multi-linear regression model provided improved estimates. When the calibration models were applied to an independent validation set, chlorophyll content was reasonably well predicted with a high correlation (r² = 0.774). Depending on the sample, the technique enabled the identification and characterization of the relative content of various chlorophyll types that were distributed within the cucumber leaves. Typically low levels of chlorophyll at leaf margins and higher levels along main vein regions were identified. Our results indicate that hyperspectral imaging exhibits considerable promise for predicting pigments within cucumber leaves and furthermore can be applied non-destructively and in situ to living plant samples.     

Pattern recognition of visible and near-infrared spectroscopy from bayberry juice by use of partial least squares and a backpropagation neural network

Visible and near-infrared reflectance (visible-NIR) spectroscopy is applied to discriminate different varieties of bayberry juices. The discrimination of visible-NIR spectra from samples is a matter of pattern recognition. By partial least squares (PLS), the spectrum is reduced to certain factors, which are then taken as the input of the backpropagation neural network (BPNN). Through training and prediction, three different varieties of bayberry juice are classified based on the output of the BPNN. In addition, a mathematical model is built and the algorithm is optimized. With proper parameters in the training set, 100% accuracy is obtained by the BPNN. Thus it is concluded that the PLS analysis combined with the BPNN is an alternative for pattern recognition based on visible and NIR spectroscopy.     

基于BP网络的苹果硬度高光谱无损检测

目的为了实现基于高光谱成像以及误差反向传播(BP)网络模型的苹果硬度快速无损检测。方法利用高光谱成像采集系统采集采后"富士"苹果的高光谱图像,然后提取整个苹果样本区域的平均反射光谱;利用连续投影算法(SPA)和竞争性自适应重加权算法(CARS)实现对标准正态变换预处理后光谱数据的降维;研究基于全光谱以及特征光谱的预测苹果硬度BP网络模型。结果采用SPA和CARS分别从256个全光谱中提取了18个和16个特征波长,明显提升了预测模型的运行效率,且SPA+BP网络模型具有相对较好的苹果硬度预测能力(rp=0.728,RPm=0.282 kg/cm2)。结论研究表明基于高光谱成像技术和BP网络建立的预测模型可快速无损预测苹果的硬度。     

Removal of analyte-irrelevant variations in near-infrared tissue spectra

This paper describes mathematical techniques to correct for analyte-irrelevant optical variability in tissue spectra by combining multiple preprocessing techniques to address variability in spectral properties of tissue overlying and within the muscle. A mathematical preprocessing method called principal component analysis (PCA) loading correction is discussed for removal of inter-subject, analyte-irrelevant variations in muscle scattering from continuous-wave diffuse reflectance near-infrared (NIR) spectra. The correction is completed by orthogonalizing spectra to a set of loading vectors of the principal components obtained from principal component analysis of spectra with the same analyte value, across different subjects in the calibration set. Once the loading vectors are obtained, no knowledge of analyte values is required for future spectral correction. The method was tested on tissue-like, three-layer phantoms using partial least squares (PLS) regression to predict the absorber concentration in the phantom muscle layer from the NIR spectra. Two other mathematical methods, short-distance correction to remove spectral interference from skin and fat layers and standard normal variate scaling, were also applied and/or combined with the proposed method prior to the PLS analysis. Each of the preprocessing methods improved model prediction and/or reduced model complexity. The combination of the three preprocessing methods provided the most accurate prediction results. We also performed a preliminary validation on in vivo human tissue spectra.     

Prediction of ethylene content in melt-state random and block polypropylene by near-infrared spectroscopy and chemometrics: comparison of a new calibration transfer method with a slope/bias correction method

This paper reports the prediction of the ethylene content (C2 content) in random polypropylene (RPP) and block polypropylene (BPP) in the melt state by near-infrared (NIR) spectroscopy and chemometrics. NIR spectra of RPP and BPP in the melt states were measured by a Fourier transform near-infrared (FT-NIR) on-line monitoring system. The NIR spectra of RPP and BPP were compared. Partial least-squares (PLS) regression calibration models predicting the ethylene (C2) content that were developed by using each RPP or BPP spectra set separately yielded good results (SECV (standard error of cross validation): RPP, 0.16%; BPP, 0.31%; correlation coefficient: RPP, 0.998; BPP, 0.996). We also built a common PLS calibration model by using both the RPP and the BPP spectra set. The results showed that the common calibration model has larger SECV values than the models based on the RPP or the BPP spectra sets individually and is not practical for the prediction of the C2 content. We further investigated whether a calibration model developed by using the BPP spectra set can predict the C2 contents in the RPP sample set. If this is possible, it can save a significant amount of work and cost. The results showed that the use of the BPP model for the RPP sample set is difficult, and vice versa, because there are some differences in the molar absorption coefficients between the RPP and BPP spectra. To solve this problem, a transfer method from one sample spectra (BPP) set to the other spectra (RPP) set was studied. A difference spectrum between an RPP spectrum and a BPP spectrum was used to transfer from the BPP calibration set to the RPP calibration set. The prediction result (SEP (standard error of prediction), 0.23%, correlation coefficient, 0.994) of RPP samples by the transferred calibration set and model showed that it is possible to transfer from the BPP calibration set to the RPP calibration set. We also studied the transfer from the RPP calibration set (the range of C2 content: 0-4.3%) to the BPP calibration set. The prediction result of C2 content (the range of C2 contents: 0-7.7%) in BPP by use of the calibration model based on the transferred BPP spectra from the RPP spectra showed that the transfer method is only effective for the interpolation of the C2 content range by the nonlinear change in the peak intensities with the C2 content.     

基于神经网络的塑料打包带高光谱模式识别

目的 建立一种快速、准确、无损的塑料打包带的检验及分类方法。方法 利用高光谱在波长为350～990 nm的条件下采集52个不同来源的塑料打包带样品的高光谱数据,并对样品进行Savitzky-Golay平滑处理,同时结合主成分分析对样品进行降维。将提取到的主成分进行K-Means聚类,以聚类结果为依据建立径向基函数神经网络（RBFNN）与BP神经网络模型（BPNN）。结果 打包带样品的高光谱谱图在400～500 nm、600～700 nm处有较大区别。实验共提取了5个初始特征值大于1的主成分,可以解释96.633%的原始数据。通过K-means聚类将塑料打包带样品分为6类,Calinski-Harabasz指数为28.76,RBFNN分类准确率为86.7%;BPNN分类准确率为98.1%,BPNN的分类效果更好。结论 研究表明神经网络在高光谱谱图分类处理上具有较高的准确度,同时也验证了高光谱在区分检验塑料打包带类物证的可行性与科学性,为公安机关提供了一种新的检验方法。     

Nondestructive direct determination of heroin in seized illicit street drugs by diffuse reflectance near-infrared spectroscopy

A new method has been developed for the fast and nondestructive direct determination of heroin in seized street illicit drugs using partial least-squares regression analysis of diffuse reflectance near-infrared spectra. Data were obtained from untreated samples placed in standard glass chromatography vials. A heterogeneous population of 31 samples, previously analyzed by a reference method, was employed to build the calibration model and to have a separated validation set. Based on the use of zero-order data for a calibration set of 21 samples, after standard normal variate and quadratic linear removed baseline correction (detrending), in the wavelength range from 1111 to 1647 nm, 8 PLS factors were enough to obtain a root-mean-square error of prediction of 1.3% w/w, with a quality coefficient of 10% for the estimation of the accuracy error in the prediction of heroin concentration in unknown samples and a residual predictive deviation of 5.4.