高光谱参数和逐步判别的苎麻品种识别

为了探讨基于高光谱的苎麻品种识别和分类的方法,在大田栽培条件下,采集了4个不同基因型苎麻品种共927个叶片高光谱数据。根据苎麻叶片高光谱反射曲线,选择了2组特征参数： 基于高光谱波形峰谷反射率和位置参数(V1组)、基于偏度和峰度参数(V2组)。运用逐步判别的方法,通过设置不同F值筛选不同个数的变量,分别建立基于2组特征参数的多个Fisher线性判别函数,并从计算量、正确率和稳定性三方面对所建立的判别函数进行分析比较。结论： (1)所有组合的判别函数总体平均正确率为91.1%,标准差总体均值为1.2%;(2)综合权衡,在所有组合中,V2组且14≥变量个数n≥8判别效果最好--计算量中等,正确率和稳定性均高于平均值,其中,13个变量的Fisher判定函数平均正确率最高有94.2%,标准差最低为0%;(3)若优先考虑正确率,V1组且22≥变量个数≥15正确率最高,平均正确率最大有95.5%,但计算量比较大,稳定性中等,标准差最低为0.9%。研究表明,利用高光谱参数结合逐步判别方法识别苎麻品种是可行的。     

基于高光谱图像的玉米种子特征提取与识别

玉米种子的形态特征是玉米品种识别的重要因素之一.采用高光谱成像系统获取9个品种共432粒玉米种子的高光谱反射图像,对图像进行校正和预处理,提取每个样本在563.6~911.4 nm共55个波段范围内的形状特征.分别利用单波段、多波段和全波段下的玉米种子形状特征结合偏最小二乘判别法进行模型分类.结果显示,全波段范围内训练集和测试集的平均正确识别率达到98.31%和93.98%,均优于多波段和单波段的正确识别率.研究表明,该方法能充分利用高光谱图像中可见光和近红外区域的有效特征信息,较准确地鉴别玉米品种,为玉米品种的自动识别领域提供了一种新方法.     

基于高光谱图像的玉米种子特征提取与识别

玉米种子的形态特征是玉米品种识别的重要因素之一.采用高光谱成像系统获取9个品种共432粒玉米种子的高光谱反射图像,对图像进行校正和预处理,提取每个样本在563.6～911.4nm共55个波段范围内的形状特征.分别利用单波段、多波段和全波段下的玉米种子形状特征结合偏最小二乘判别法进行模型分类.结果显示,全波段范围内训练集和测试集的平均正确识别率达到98.31%和93.98%,均优于多波段和单波段的正确识别率.研究表明,该方法能充分利用高光谱图像中可见光和近红外区域的有效特征信息,较准确地鉴别玉米品种,为玉米品种的自动识别领域提供了一种新方法.     

基于红外光谱的石斛品种判别分析

测试了30个品种206株石斛茎的红外光谱,光谱显示,石斛的主要物质成分为纤维素等多糖物质。用石斛的傅里叶变换红外光谱结合Wilks’Lambda逐步判别分析法对石斛品种进行识别研究,比较分析了不同的光谱范围和不同的训练样本数对识别结果的影响。对1 800~1 250cm-1光谱范围的287个变量进行品种判别分析,每个品种的训练样本分别为2,3,4,5,6个时,对训练样本的回判正确率都是100%;以余下的样本作为测试样本进行品种预测的正确率分别为79.4%,91.3%,93.0%,98.2%,100%。同时对1 800~1 500,1 500~1 250,1 250~600,1 250~950,950~600cm-1等不同光谱范围,按五种不同的训练样本情况,相同的判别分析方法进行判别分析比较,1 800~1 250,1 800~1 500和950~600cm-1光谱范围的变量更适宜进行石斛品种的判别分析;每个品种的训练样本达到3个及以上时,建立的模型判别分析结果较好。结果表明,傅里叶变换红外光谱结合逐步判别分析法对不同品种的石斛进行鉴别,能够较好的识别石斛的品种,为石斛品种的鉴别提供了一种简便、易行的新方法。     

基于高光谱成像技术的不同产地小米判别分析

高光谱成像技术被广泛应用于农产品的检测。基于高光谱成像技术结合机器学习算法无损鉴别不同地区的小米样本。将来源7个省份共计23份样品的小米样本根据地理区域划分为东北地区、河北、陕西、山东和山西共5大类,其中东北地区共6份样品,山西地区5份样品,河北、陕西和山东各4份样品。将每份样品均分为10等份并利用高光谱成像仪采集900～1 700 nm波段内小米的高光谱数据。为了减少光照不均匀和暗电流对实验的影响,对采集到的高光谱数据进行黑白校正。利用ENVI软件选取小米高光谱图像的感兴趣区域(ROI),每份小米样品选取9个ROI。计算ROI内的平均光谱值,以此平均值作为该样本的一条光谱记录,最后共收集到2 070条光谱曲线,其中东北类540条,山西类450条,其他河北类、山东类、陕西类各360条。为了减少样品表面的不平整性引起的散射现象,进而影响小米的真实光谱信息,对收集到的原始光谱进行多元散射校正预处理(MSC)。采用随机划分法对校正过后的光谱数据划分训练集和测试集,测试集占的比例为0.3。利用线性判别分析(LDA)对不同产地小米的光谱数据进行可视化分析,将测试集代入训练好的LDA模型,做出预测结果的混淆矩阵(Confusion Matrix),结果表明LDA对于陕西和山西类的预测准确率为0.84和0.99,对于东北、河北和山东的预测准确率仅为0.68,0.68和0.40。进而采用递归特征消除(RFE)对小米的光谱信息进行特征选择,去除冗余的信息,提高模型的预测准确率。将RFE分别与支持向量机(SVM)和逻辑回归(LR)结合,对不同产地小米的判别进行对比分析。将小米光谱数据的训练集分别代入SVM-RFE和LR-RFE模型并结合3折交叉验证技术,以模型F值的微平均(Micro-averaging)最优选择出相应的特征子集。结果表明,LR-RFE选择的波长数为74个,其模型的Micro_F为0.59;SVM-RFE选择的波长数为220,其模型的Micro_F为0.66。将选择后的特征子集应用到测试集并将测试集分别代入SVM和LR模型,采用模型预测结果的混淆矩阵和模型的受试者工作特征曲线(ROC)作为评价方法。结果表明SVM-RFE对东北地区、河北、陕西、山东和山西的预测准确率分别为1,0.37,0.72,0和1,其ROC曲线下面积(AUC)分别为0.82,0.92,0.93,0.70和0.99。LR-RFE的预测准确率分别为0.92,0,0.97,0和0.80,其AUC分别为0.72,0.74,0.94,0.66和0.88。从预测结果可以看出SVM-RFE模型的综合分类性能优于LR-RFE,而对陕西类的判别LR-RFE要优于SVM-RFE,对于河北类和山东类两个模型都不能有效判别。这两个模型的预测准确率相比LDA有了一定的提升。     

基于高光谱图像多光谱参数的草莓成熟度识别

为了建立多光谱参数用于草莓成熟度的自动识别,采用高光谱图像技术,通过提取草莓样本ROI的平均光谱,计算已有的八个成熟度参数Ind1,Ind2,Ind3,IAD,I1,I2,I3,I4的参数值,并结合Fisher线性判别法判断八个参数对于三种成熟度(成熟、接近成熟、未成熟)草莓样本的分类识别效果,发现基于I4参数的线性判别分析模型的识别效果最佳,建模集和预测集识别准确率分别为90%和91.67%;基于草莓样本的光谱特征,提取与草莓成熟度相关的三个波长535,675和980 nm,并基于这三个波长和已有的参数形式,构建了四个用于草莓成熟度检测的新参数：i1,i2,i3,i4,通过Fisher线性判别法判断四个参数的分类识别效果,发现基于参数i1,i2和i4的线性判别分析模型的识别效果均比参数I4好,建模集和预测集识别准确率为95.83%,95.83%,95.83%和95%,95%,96.67%。结果表明新建立的多光谱参数i1,i2和i4可以用于草莓成熟度的自动分类识别,为草莓成熟度的在线检测提供了理论依据。     

基于高光谱成像和判别分析的黄瓜病害识别

利用光谱成像技术(400～720 nm)识别黄瓜白粉病、角斑病、霜霉病、褐斑病和无病区域。构建高光谱图像采集系统进行样本图像的采集,预处理和光谱信息的提取。由于获得的原始光谱数据量很大,为了减少后续运算量,提高准确率,采用逐步判别分析和典型判别分析两种方法进行降维。逐步判别从55个波段中选择12个波段,典型判别从55个波段中提取2个典型变量。利用选择的光谱特征参数建立病害识别模型。逐步判别构建的模型对训练样本和测试样本的判别准确率分别为100%和94%,典型判别构建的模型对训练样本和测试样本的判别准确率均为100%。说明利用高光谱成像技术可以进行黄瓜病害的快速、准确识别,并为实现可见光谱范围内黄瓜病害的田间实时在线检测提供了可能。     

高光谱技术的羊肉品种多波段识别研究

利用可见/近红外(400～1 000 nm)及近红外(900～1 700 nm)高光谱成像技术对宁夏地区滩寒杂交、盐池滩羊、小尾寒羊三个品种羊肉进行识别研究。针对不同波段光谱特点,分别优选出Baseline及SG卷积平滑光谱预处理方法,并运用连续投影算法(SPA)提取特征波长,结合线性判别分析(LDA)及径向基核函数支持向量机(RBFSVM)模型进行全波段及特征波长识别分析。结果表明不同波段高光谱对羊肉品种识别均获得较好效果,其中400～1 000 nm波段采用Baseline-Fullwave-RBFSVM及12个特征波长下准确率为100%与98.75%,900～1 700 nm波段采用Baseline-Fullwave-RBFSVM及7个特征波长下准确率为96.25%与87.80%;RBFSVM非线性分类准确率高于LDA线性判别结果,400～1 000 nm波段识别准确率优于900～1 700 nm波段,说明三种羊肉在色泽纹理上差异比成分含量显著,利用高光谱成像技术结合RBFSVM方法能够获得较优的羊肉品种识别效果。     

基于高光谱数据的高寒草地营养状况的研究

高寒草地牧草营养状况的遥感监测是草地合理利用的基础,是草地动态监测的难点。通过对高寒草地牧草营养成分和对应时间高光谱数据的分析,探讨草地营养状况与植被高光谱数据间的关系。结果表明,两波段比高光谱指数模型与草地营养biomass,air-DM,P,CF,CP指标间显著相关;MAXR多元回归模型,当选取波段数达到五个,模型与各营养指标均显著相关。总体上,MAXR回归模型要优于两波段比模型,MAXR模型对air-DM,P,CF进行反演,可获得较高的精度。     

高光谱技术融合图像信息的牛肉品种识别方法研究

高光谱图像包含了大量的光谱信息和图像信息,采用高光谱成像技术对牛肉品种进行识别。获取可见-近红外(400~1000 nm)光谱范围内的安格斯牛、利木赞牛、秦川牛、西门塔尔牛、荷斯坦奶牛五个品种共252个牛肉样本的高光谱图像。在ENVI软件中对高光谱图像进行阈值分割并构建掩膜图像,获取样本的感兴趣区域(ROI),并结合伪彩色图对牛肉样本的反射率指数进行可视化表达;采用Kennard-Stone(KS)法对样本集进行划分以提高模型的预测性能;对原始光谱采用卷积平滑(SG)、区域归一化(Area normalize)、基线校正(Baseline)、一阶导数(FD)、标准正态变量变换(SNV)及多元散射校正(MSC)等6种方法进行预处理;采用竞争性自适应重加权算法(CARS)提取特征波长。然后利用颜色矩对不同牛肉样本的颜色特征进行提取;对原始光谱图像进行主成分分析,结合灰度共生矩阵(GLCM)算法,提取主要纹理特征。最后结合偏最小二乘判别(PLS-DA)算法建立牛肉样本基于特征波长、颜色特征以及纹理特征的识别模型。KS法将牛肉样本划分为校正集190个,预测集62个;将未经预处理的光谱数据与经过6种不用预处理的光谱数据进行建模分析,结果发现经FD法处理后的光谱数据所建模型的识别率最高;结合CARS法对经FD法预处理后的光谱数据进行特征波长提取,共提取出22个波长;利用颜色矩和GLCM算法分别提取出每个牛肉样本的9个颜色特征、48个纹理特征。将特征波长数据与颜色、纹理特征信息进行融合建模,结果表明,基于特征光谱+纹理特征的模型识别效果最佳,其校正集与预测集识别率分别为98.42%和93.55%,均高于特征光谱数据模型识别率,说明融合纹理特征后使样本分类信息的表达更加全面;融合颜色特征后模型的校正集识别率均有所增加,但预测集识别率稍逊,颜色特征虽携带了部分有效信息,但这些信息与牛肉样本的相关性不大。因此,寻找与牛肉样本相关性更大的颜色特征是提高模型识别率的重要途径之一。该研究结果为牛肉品种的快速无损识别提供了一定的参考。     

Coronary disease identification using the discriminant analysis

Summary This paper presents a method for identification of coronary disease using the discriminant analysis. The set of discriminant variables used is composed by the general characteristics of the patient, the lipemic values, collected by a blood analysis, and some results of a cardiac catheterism. By using all the variables a total level of classification of about 89.9% is reached, separating the healthy patients from the sick patients affected by a coronary disease. The discriminant analysis is also performed using the general characteristics of the patient and the lipemic values only (i.e. noninvasive medical examinations), and general characteristics and some results of a cardiac catheterism only. In conclusion, the diagnosis of coronary artery disease, obtained by discriminant analysis of noninvasive data, seems to be good enough to discriminate sick from healthy people and to assign patients to a further protocol.     

Simultaneous species and sex identification of silkworm pupae using hyperspectral imaging technology

To obtain high-quality raw silk and improve the economic values of sericulture industry, sex needs to be discriminated first before cross-breeding. Much work has been reported about sex identification. However, to realize automatic separation of silkworm pupae, the species also needs to be classified, which no research has ever explored. Hence, this paper studied the feasibility of visible and near-infrared hyperspectral imaging technology to identify the species and sex of silkworm pupae. 288 hyperspectral images of silkworm pupae were collected and the average spectra were extracted from the region of interest, around the tail region of silkworm pupae. Successive projection algorithm was served as a variable selection method to choose the optimal wavelengths from the full spectra. At the same time, principal component analysis was used to choose the characteristic images. Then, the gray-level co-occurrence matrix was implemented on the first three principal component images (accounted for 99.05% of the total variances) to extract 48 textural features. Partial least squares discriminant analysis and support vector machine models were built, respectively, based on the spectral data, textural data and fusion data that included spectral and textural data, in which the support vector machine model based on the fusion data, gave the best species and sex identification result with an accuracy of 95.83%. It demonstrated that the hyperspectral imaging technology could be a new and nondestructive method to replace the manual work.     

新型干涉高光谱成像系统的光束剪切特性分析

提出了一种基于新型双折射横向剪切分束器的高光谱成像方法,采用的横向剪切分束器主要由Wollaston棱镜和角锥反射体组成.在分析双折射分束器的偏光结构和分光机理的基础上,利用光线追迹方法分析了光束在Wollaston棱镜中的传播特性,通过计算光束在双折射分束器中的传播方向及出射位置,推导出调制光程差的理论表达公式.根据理论推导结果,分别仿真分析了系统在不同扫描模式下光程差与入射光视场角以及角锥顶点偏移量的调制关系.基于理论分析结果搭建了实验装置,对光程差分析结果进行验证,实验结果与理论分析结果匹配较好.所提方法可以提高剪切光束的平行性,保证干涉条纹的高调制度,降低了复原光谱准确度对光学装调精度和元件加工精度的依赖性,具有结构稳定、复杂度低的显著特点.     

Blood oxygenation monitoring using hyperspectral imaging after flap surgery

In this paper, we present a new method for the assessment of skin flap perfusion based on the non-invasive monitoring of skin blood oxygenation using hyperspectral imaging. This method consists of generation of oxyhemoglobin and deoxyhemoglobin maps in skin flap from hyperspectral image using an analytic model based on Beer–Lambert law. The results obtained on a fasciocutaneous sural flap have revealed that the oxyhemoglobin and deoxyhemoglobin maps facilitate the visualization of some differences between different areas of the flap, an aspect that clinical examination fails to do. In conclusion, the proposed method could provide a new option for a more accurate assessment of flap survival.     

High speed measurement of corn seed viability using hyperspectral imaging

Corn is one of the most cultivated crops all over world as food for humans as well as animals. Optimized agronomic practices and improved technological interventions during planting, harvesting and post-harvest handling are critical to improving the quantity and quality of corn production. Seed germination and vigor are the primary determinants of high yield notwithstanding any other factors that may play during the growth period. Seed viability may be lost during storage due to unfavorable conditions e.g. moisture content and temperatures, or physical damage during mechanical processing e.g. shelling, or over heating during drying. It is therefore vital for seed companies and farmers to test and ascertain seed viability to avoid losses of any kind. This study aimed at investigating the possibility of using hyperspectral imaging (HSI) technique to discriminate viable and nonviable corn seeds. A group of corn samples were heat treated by using microwave process while a group of seeds were kept as control group (untreated). The hyperspectral images of corn seeds of both groups were captured between 400 and 2500 nm wave range. Partial least squares discriminant analysis (PLS-DA) was built for the classification of aged (heat treated) and normal (untreated) corn seeds. The model showed highest classification accuracy of 97.6% (calibration) and 95.6% (prediction) in the SWIR region of the HSI. Furthermore, the PLS-DA and binary images were capable to provide the visual information of treated and untreated corn seeds. The overall results suggest that HSI technique is accurate for classification of viable and non-viable seeds with non-destructive manner.     

利用光谱和质谱成像技术实现指纹痕量检测

近年来,依赖于先进光源的化学成像技术迅速发展,极大提高了痕量检测的准确性,在公共安全、环境、食品、医药、考古等领域具有重要的实用价值.在痕量检测中,通过将成像技术与光谱测量技术、质谱技术等相结合,能够同时获取检验对象的物质组成和二维图像信息,不仅可以揭示材料表面的痕量物质成分及其分布,还可以在提高检验灵敏度的情况下,减少甚至避免传统检测手段所需要的特殊显现剂,因此与其他检验方法具有良好的兼容性.本文以指纹检验这一典型的痕量检测问题为例,阐述基于光谱和质谱成像技术的化学成像方法在痕量检测领域中的应用,从定向针对特定组分的化学成像和非定向的直接化学成像两个方面,综述了在指纹显现或显现增强中获得应用的主要成像手段,包括可见-近红外成像、红外成像、拉曼成像、质谱成像等.     

Excitation-resolved hyperspectral fluorescence lifetime imaging using a UV-extended supercontinuum source

We present a time-gated, optically sectioned, hyperspectral fluorescence lifetime imaging (FLIM) microscope incorporating a tunable supercontinuum excitation source extending into the UV. The system is capable of resolving the excitation spectrum, emission spectrum, and fluorescence decays in an optically sectioned image.     

Detection of ochratoxin A contamination in stored wheat using near-infrared hyperspectral imaging

Near-infrared (NIR) hyperspectral imaging system was used to detect five concentration levels of ochratoxin A (OTA) in contaminated wheat kernels. The wheat kernels artificially inoculated with two different OTA producing Penicillium verrucosum strains, two different non-toxigenic P. verrucosum strains, and sterile control wheat kernels were subjected to NIR hyperspectral imaging. The acquired three-dimensional data were reshaped into readable two-dimensional data. Principal Component Analysis (PCA) was applied to the two dimensional data to identify the key wavelengths which had greater significance in detecting OTA contamination in wheat. Statistical and histogram features extracted at the key wavelengths were used in the linear, quadratic and Mahalanobis statistical discriminant models to differentiate between sterile control, five concentration levels of OTA contamination in wheat kernels, and five infection levels of non-OTA producing P. verrucosum inoculated wheat kernels. The classification models differentiated sterile control samples from OTA contaminated wheat kernels and non-OTA producing P. verrucosum inoculated wheat kernels with a 100% accuracy. The classification models also differentiated between five concentration levels of OTA contaminated wheat kernels and between five infection levels of non-OTA producing P. verrucosum inoculated wheat kernels with a correct classification of more than 98%. The non-OTA producing P. verrucosum inoculated wheat kernels and OTA contaminated wheat kernels subjected to hyperspectral imaging provided different spectral patterns.     

半监督判别分析的跨库语音情感识别

针对训练样本与测试样本来自不同语音情感数据库造成特征向量空间分布不匹配的问题,采用半监督判别分析减小二者的差异。首先寻找有标签的训练样本和来自另一个库的部分无标签训练样本之间的最优投影方向。基于一致性假设即相近的点更有可能具有相同的类别,利用p近邻图对无标签训练样本相近点之间的关系进行建模,从而获得无标签样本的分布信息。在保证无标签样本间流形结构的同时,使所有训练样本类间散度和类内散度的比值达到最大,从而得到最优的投影方向。采用两组实验进行验证,第1组用eNTERFACE库训练去测试Berlin库,识别率为51.41%,第2组用Berlin库训练测试eNTERFACE库,识别率为45.76%,相比未采用半监督判别分析的识别结果分别有了13.72%和22.81%的提高,说明该算法的有效性。通过实验前后数据的可视化分析,说明利用半监督判别分析确实减小了不同库之间特征向量空间分布的不匹配问题,从而提高跨库语音情感识别率。     

半监督判别分析的跨库语音情感识别

针对训练样本与测试样本来自不同语音情感数据库造成特征向量空间分布不匹配的问题,采用半监督判别分析减小二者的差异。首先寻找有标签的训练样本和来自另一个库的部分无标签训练样本之间的最优投影方向。基于一致性假设即相近的点更有可能具有相同的类别,利用p近邻图对无标签训练样本相近点之间的关系进行建模,从而获得无标签样本的分布信息。在保证无标签样本间流形结构的同时,使所有训练样本类间散度和类内散度的比值达到最大,从而得到最优的投影方向。采用两组实验进行验证,第1组用eNTERFACE库训练去测试Berlin库,识别率为51.41%,第2组用Berlin库训练测试eNTERFACE库,识别率为45.76%,相比未采用半监督判别分析的识别结果分别有了13.72%和22.81%的提高,说明该算法的有效性。通过实验前后数据的可视化分析,说明利用半监督判别分析确实减小了不同库之间特征向量空间分布的不匹配问题,从而提高跨库语音情感识别率。