Feasibility study of assessing cotton fiber maturity from near infrared hyperspectral imaging technique

Background Fiber maturity is a key cotton quality property,and its variability in a sample impacts fiber processing and dyeing performance.Currently,the maturity is determined by using established protocols in laboratories under a controlled environment.There is an increasing need to measure fiber maturity using low-cost(in general less than $20000)and small portable systems.In this study,a laboratory feasibility was performed to assess the ability of the shortwave infrared hyperspectral imaging(SWIR HSI)technique for determining the conditioned fiber maturity,and as a comparison,a bench-top commercial and expensive(in general greater than $60000)near infrared(NIR)instrument was used.Results Although SWIR HSI and NIR represent different measurement technologies,consistent spectral characteristics were observed between the two instruments when they were used to measure the maturity of the locule fiber samples in seed cotton and of the well-defined fiber samples,respectively.Partial least squares(PLS)models were established using different spectral preprocessing parameters to predict fiber maturity.The high prediction precision was observed by a lower root mean square error of prediction(RMSEP)(<0.046),higher R_(p)^(2)(>0.518),and greater percentage(97.0%)of samples within the 95% agreement range in the entire NIR region(1000-2500 nm)without the moisture band at 1940 nm.Conclusion SWIR HSI has a good potential for assessing cotton fiber maturity in a laboratory environment.

Evaluation of growth characteristics of Aspergillus parasiticus inoculated in different culture media by shortwave infrared(SWIR) hyperspectral imaging

The growth characteristics of Aspergillus parasitic us incubated on two culture media were ex-amined using shortwave infrared(SWIR,1000-2500 nm)hyperspectral imaging(HSI)in this work.HSI images of the A.parasiticus colonies growing on rose bengal medium(RBM)and maize agar medium(MAM)were recorded daily for 6 days.The growth phases of A.parasiticus were indicated through the pixel number and average spectra of colonies.On score plot of the first principal component(PC1)and PC2,four growth zones with varying mycelium densities were identified.Eight characteristic wavelengths(1095,1145,1195,1279,1442,1655,1834 and 1929 nm)were selected from PC1 loading,average spectra of each colony as well as each growth zone.F urthermore,support vector machine(S VM)classifier based on the eight wavelengths was built,and the classification accuracies for the four zones(from outer to inner zones)on the colonies on RBM were 99.77%,9935%,99.75%and 99.60%and 99.77%,9939%,99.31%and 98.22%for colonies on MAM.In addition,a new score plot of PC2 and PC3 was used to differ-entiate the colonies incubated on RBM and MAM for 6 days.Then characteristic wavelengths of 1067,1195,1279,1369,1459,1694,1834 and 1929 nm were selected from the loading of PC2 and PCg.Based on them,a new SVM model was developed to diferentiate colonies on RBM and MAM with accuracy of 100.00%and 9999%,respectively.In conclusion,SWIR hyperspectral image is a powerful tool for evaluation of growth characteristics of A.parasiticus incubated in diferent culture media.

Near-Infrared Imaging Using a High-Speed Monitoring Near Infrared Hyperspectral Camera(Compovision)

This review paper reports near-infrared(NIR)imaging studies using a newly-developed NIR camera,Compovision.Compovision can measure a significantly wide area of 150mm×250mm at high speed of between 2and 5s.It enables a wide spectral region measurement in the 1 000~2 350nm range at 6nm intervals.We investigated the potential of Compovision in the applications to industrial problems such as the evaluation of pharmaceutical tablets and polymers.Our studies have demonstrated that NIR imaging based on Compovision can solve several issues such as long acquisition times and relatively low sensitivity of detection.NIR imaging with Compovision is strongly expected to be applied not only to pharmaceutical tablet monitoring and polymer characterization but also to various applications such as those to food products,biomedical substances and organic and inorganic materials.

Sunglint Impact on Atmospheric Soundings from Hyperspectral Resolution Infrared Radiances

The mid-wave infrared band （3-5 #rn） has been widely used for atmospheric soundings. The sunglint impact on the atmospheric parameter retrieval using this band has been neglected because the reflected radiances in this band are significantly less than those in the visible band. In this study, an investigation of sunglint impact on the atmospheric soundings was conducted with Atmospheric InfraRed Sounder ob- servation data from 1 July to 7 July 2007 over the Atlantic Ocean. The impact of sunglint can lead to a brightness temperature increase of 1.0 K for the surface sensitive sounding channels near 4.58 #m. This contamination can indirectly cause a positive bias of 4 g kg-1 in the water vapor retrieval near the ocean surface, and it can be corrected by simply excluding those contaminated channels.

Using Longwave Infrared Hyperspectral Imaging for a Quantitative Atmospheric Tracer Monitoring in Outdoor Environments

Quantitative real-time retrieval of concentration-lengths (CL) through gaseous plumes is an important tool for environmental monitoring, enabling remote monitoring of emissions from industrial facilities and risk assessment in scenarios of toxic gas releases. The adoption of LWIR (Long Wave InfraRed) hyperspectral imaging as a leading technique for remote gas plume detection paved the way for an introduction of a precise CL estimation and two-dimensional (2D) mapping. A novel methodology for evaluating and characterizing the performance of a retrieval algorithm is presented. The algorithm utilizes state-of-the-art retrieved hyperspectral 2D mapping and is applied on a series of localized atmospheric tracer gas (SF6) releases in monitored environmental conditions. The retrieved CL distributions are compared to a numeric atmospheric T&D (Transport and Diffusion) model. Satisfactory agreement between retrieved and simulated CL prediction is manifested, and the uncertainty involved is quantified. Possible sources for the remained discrepancies between retrieved and simulated CL values are characterized, and methods to minimize them are discussed.

Impacts on Initial Condition Modification from Hyperspectral Infrared Sounding Data Assimilation: Comparisons between Full-Spectrum and Channel-Selection Scheme Based on Two-Month Experiments Using CrIS and IASI Observation

This paper discusses the performance difference between full-spectrum and channel-selection assimilation scheme of hyperspectral infrared observation, e.g. CrIS and IASI, on improving the accuracy of initial condition in numerical weather prediction. To accomplish this, we develop a 3D-Variational data assimilation system whose observation operator is a principal-component based fast radiative transfer model, which equips the direct assimilation of full-channel radiance from hyperspectral infrared sounders with high computational efficiency. This project’s primary goal is to demonstrate that assimilation of infrared observation in a full-channel mode could improve the accuracy of initial condition compared to selected-channel assimilation. Results show that full-channel assimilation performs better than selected-channel assimilation in modifying low and middle troposphere (1000 - 700 hPa, 700 - 400 hPa) temperature and water vapor field, while marginal improvements from temperature and water vapor field could be found over upper troposphere (400 - 100 hPa). This research also proves the feasibility of an alternative path to data assimilation for the full usage of hyperspectral infrared sounding observation in numerical weather prediction.

Observation of Atherosclerotic Plaque Phantoms through Saline or Blood Layers by Near-Infrared Hyperspectral Imaging

We observed atherosclerotic plaque phantoms using a novel near-infrared (NIR) hyperspectral imaging (HSI) technique. Data were obtained through saline or blood layers to simulate an angioscopic environment for the phantom. For the study, we developed a NIR-HSI system with an NIR supercontinuum light source and mercury-cadmium-telluride camera. Apparent spectral absorbance was obtained at wavelengths of 1150 - 2400 nm. Hyperspectral images of lipid were constructed using a spectral angle mapper algorithm. Bovine fat covered with saline or blood was observed using hyperspectral images at a wavelength around 1200 nm. Our results show that NIR-HSI is a promising angioscopic technique with the potential to identify lipid-rich plaques without clamping and saline injection.

近红外光谱及成像在果品无损检测中的应用

近红外光谱及成像技术作为新兴传感无损检测技术,具有绿色、无污染、非破坏性等优势,为果品品质领域的快速、无损检测提供新的方法和手段。文章综述近红外光谱及成像技术在果品品质检测中的研究进展,介绍近红外光谱和成像技术的基本原理,分析近红外光谱技术在果品品质检测中的化学组成、成熟度评估和小型化设备开发的应用现状,探析近红外光谱成像技术在果品外观和质量、病虫害诊断中的应用现状,阐述其在果品组分分布和缺陷识别等方面的技术优势。同时,探讨近红外光谱及成像技术与人工智能、物联网结合应用的巨大潜力,以及对提高果品检测效率、精度和改善供应链智能管控等方面的应用前景做出展望,以期为果品无损检测的应用研究提供参考。

大幅宽多谱段高光谱红外探测器研究

报道了基于分子束外延(Molecular Beam Epitaxy,MBE)碲镉汞(Mercury Cadmium Telluride,MCT)技术的大幅宽、多谱段、大像元高光谱红外探测器的最新研究进展。采用MBE技术制备出高质量MCT材料;采用成熟的n-on-p技术路线制备探测芯片,并针对特殊形状的大像元进行了优化;高光谱专用读出电路设计针对短波、窄谱段、小信号等典型特征进行了优化,并针对光谱应用加入行增益可调等功能设计。测试结果表明,组件基本性能良好,有效像元率大于99.5%,平均量子效率优于70%。

婴幼儿奶粉中多种掺假物近红外高光谱图像检测方法

奶粉市场是食品掺假行为频发领域,其中婴幼儿配方奶粉价格高,其质量是消费者、生产企业和执法部门关注的重点。近红外高光谱成像(Near infrared-hyperspectral imaging,NIR-HSI)技术结合化学计量学和机器学习算法可以检测奶粉中单一掺假物含量。基于NIR-HSI技术研究了不同品牌婴幼儿奶粉中多掺假物(三聚氰胺、香兰素和淀粉)的定量预测。对基于像素点预处理后的高光谱图像划分感兴趣区域(Region of interest,ROI),提取ROI平均光谱。基于经典的过滤式特征选择算法拉普拉斯分数(Laplacian score)(无监督)和ReliefF(有监督)挑选建模关键变量,建立偏最小二乘回归模型(Partial least squares,PLS)。开发包含自定义选择层的一维卷积神经网络模型(One-dimensional convolutional neural networks,1DCNN)。自定义层根据权重系数绝对值,可确定重要波长变量。Laplacian score-PLS模型对预测集中奶粉、三聚氰胺、香兰素和淀粉质量分数预测结果均方根误差分别为0.1110%、0.0570%、0.0349%和0.3481%。ReliefF-PLS模型对预测集中奶粉、三聚氰胺、香兰素和淀粉预测结果均方根误差分别为0.1998%、0.0540%、0.0455%和0.1823%。1DCNN模型对预测集中奶粉、三聚氰胺、香兰素和淀粉质量分数预测结果均方根误差分别为0.8561%、0.0911%、0.0644%和0.2942%。对Laplacian score、ReliefF和自定义选择层挑选出的前15个重要波长进行对比分析,不同特征选择方法挑选的特征波长子集有所区别,但都选择1210、1474、1524、1680 nm等附近波长。基于ReliefF-PLS模型的可视化结果表明了其良好的预测能力。

鄂东南矿集区典型矽卡岩-斑岩矿床蚀变矿物短波红外(SWIR)光谱研究与勘查应用

鄂东南矿集区是我国著名的矽卡岩-斑岩型铜铁金矿床产区,研究程度较高,近年来矿产勘查进入了深部找矿的新阶段,需要较新的勘查方法与思路。本文介绍了利用短红外光谱(SWIR)技术,对鄂东南典型的铜绿山矽卡岩型Fe-Cu-Au矿床、鸡冠嘴矽卡岩型Cu-Au矿床、铜山口矽卡岩-斑岩型Cu-Mo-W矿床进行的蚀变矿物综合研究最新成果。在确定蚀变矿物组合类型、形成期次、蚀变分带的基础上,揭示了主要蚀变矿物SWIR特征值的指示规律,并进一步提取各个矿床的蚀变矿物SWIR勘查标志,如铜绿山矿床富Fe绿泥石(Pos2250>2253nm)、高结晶度高岭石族(Pos2170>2170nm,Dep2170>0.18)、白云母族-蒙脱石异常Pos2200值(>2212nm或<2202nm)、高岭石、迪开石及皂石的大量出现,可以作为铜绿山铜铁金矿床有效的蚀变矿物勘查标志;鸡冠嘴矿床白云母族-蒙脱石Pos2200特征值的高值(>2209nm)区域对矿体位置具有较好的指示性;铜山口矿床绿泥石的高Fe-OH吸收峰位值(Pos2250>2249nm)和高Mg-OH吸收峰位值(Pos2335>2333nm)的高频出现,可以作为铜山口铜钼钨矿床的有效勘查标志。这些研究成果表明蚀变矿物可以为鄂东南矿集区提供有效的勘查标志体系,同时,这些最新的勘查标志也在铜绿山铜铁金矿床得到了初步的应用和验证。

基于可见光-近红外高光谱信息与数据融合的木质化鸡胸肉的判别模型构建

木质化鸡胸肉(wooden breast,WB)制约肉鸡行业发展,传统触诊检测方法耗时且效率低,为提升高光谱成像(hyperspectral imaging,HSI)技术检测鸡胸肉木质化程度的效果,本论文以白羽鸡鸡胸肉为研究对象,将其划分4个木质化等级,采集其在400~1000和1000~2000 nm内的HSI信息,通过不同光谱预处理算法和特征波段筛选方法,建立基于全波段、特征波段和HSI数据融合的偏最小二乘判别分析(Partial least squares-discriminant analysis,PLS-DA)模型和支持向量机(Support vector machine,SVM)模型。结果显示,SVM模型比PLSDA模型更适于判别鸡胸肉木质化程度,基于1000~2000 nm内全波段和特征波段的最佳模型预测集总体正确率均高于400~1000 nm内的模型,基于两波段HSI数据融合的木质化判别模型优于基于单一波段(包括全波段和特征波段)的模型,最佳模型预测集总体正确率为96.7%,能较好地区分出4个木质化等级,且对4个等级的判别准确率均可达90%以上。研究结果为HSI实现木质化鸡胸肉的准确无损检测提供技术支持。

褐土全氮含量Vis/NIRS组合预测模型的构建

准确掌握农田土壤全氮含量对于评估土壤肥力,合理施用氮肥具有重要意义。为综合利用各个单预测模型的优势,提升整体预测性能、降低模型方差,提高鲁棒性,以农田褐土土壤为研究对象,基于近红外-可见光高光谱数据,提出了一种基于标准差的预测有效度组合预测模型(CPM),用于预测土壤全氮含量。对原高光谱数据采用Savitzky-Golay平滑和一阶微分变换,采用树模型进行特征波段提取,利用决策树回归(DTR)(模型1)、高斯核回归(GKR)(模型2)、随机森林回归(RF)(模型3)、LASSO回归(模型4)、多层感知器回归(MLP)(模型5)5个单预测模型,通过单预测模型的线性组合建立组合预测模型。结果表明:(1)通过广义简约梯度优化算法求得组合预测模型中5个单预测模型的权重分别为ω_(1)*=0.407,ω_(2)*=0.378,ω_(30*=0.215,ω_(4)*=0,ω_(5)*=0;(2)对于所有数据,基于5种单预测模型和组合预测模型对土壤全氮含量预测的预测有效度分别为M_(1)=0.855,M_(2)=0.856,M_(3)=0.847,M_(4)=0.785,M_(5)=0.796,M_(CPM)=0.880;与单模型预测有效度最大值相比,组合预测模型预测有效度提高了2.924%;(3)对于所有数据,基于5种单预测模型和组合预测模型对土壤全氮含量预测的预测精度和标准差分别为E(A_(1))=0.924,σ(A_(1))=0.075,E(A_(2))=0.928,σ(A_(2))=0.077,E(A_(3))=0.923,σ(A_(3))=0.082,E(A_(4))=0.882,σ(A_(4))=0.109,E(A_(5))=0.889,σ(A_(5))=0.104,E(A_(CPM))=0.937,σ(A_(CPM))=0.066,与单模型预测精度最大值相比,组合预测模型预测精度提高了0.970%,模型稳定性提高了12.000%,且为优性组合预测。组合预测模型可用于可见光-近红外光谱数据的农田褐土土壤全氮含量的有效估测,可为农田土壤全氮含量的快速监测提供依据和参考。

界面聚合法制备耐热型山梨醇水溶液微胶囊

为获得具有植物叶片近红外精细反射光谱特征的伪装材料,采用界面聚合法制备山梨醇水溶液微胶囊。选择聚脲为壁材,设计了两步升温的包覆工艺,通过反应温度的优化得到了水分含量达50%以上的山梨醇水溶液微胶囊。扫描电子显微镜下观察到微胶囊呈表面较为光滑的球形;紫外/可见/近红外分光光度计结果表明,微胶囊的反射光谱在1450 nm和1930 nm处呈现与绿色植被相似的水分吸收谷。微胶囊具有优良的水分保持能力,在25℃、相对湿度50%环境中放置96 h后水分含量变化率不超过3%;热重分析和差示扫描量热结果表明,微胶囊水分的流失起始温度超过200℃。该微胶囊轻质、水分含量高、耐热性强,制备工艺简单,与植物叶片近红外反射光谱匹配度高,在对抗高光谱探测技术领域具有良好应用前景。

地面目标的中波红外高光谱成像特性测量

利用Telops公司设计的Hyper-Cam中波高光谱成像仪对水泥地上某坦克模型进行3~5μm波段的红外高光谱成像测量。对实验数据进行了逐像素点的大气校正。基于矫正后的实验数据,分析了仪器误差、随机误差、大气校正传递误差等对测试不确定度的影响。分析了该坦克模型在不同谱段上的亮度分布和模型上不同部位的光谱分布特性。结果表明:在波数为2000~3000 cm-1波段,其相对不确定度一直稳定在10%以内,但是在大于3000 cm-1的波段,其误差迅速上升,主要因为常温物体的中波辐射低,在该波段目标辐射与大气路程辐射接近,而引起的噪声增大,造成的测试信噪比降低。该坦克模型光谱辐射亮度测试数据的平均相对不确定度在20%以内,整体误差较低。光谱分布上,坦克各部分特征点和水泥地表的光谱辐射亮度在波长为4.2~5μm波段时的差异性比波长为3.0~4.2μm波段时的差异性更大,且探测器接收到的4.2~5μm波段的光谱辐射亮度要大于3.0~4.2μm的光谱辐射亮度。由于4.3μm波段处于大气吸收带,大气透射率几乎为0,以至于无法获得目标的光谱辐射亮度的真实测量值。空间分布上,模型的辐射亮度主要集中在侧边涂层上,炮口处和履带处的辐射亮度较小。目标各部位辐射亮度分布与周边水泥地表的辐射亮度有着明显差异性。在3~5μm的辐射亮度中,4.4~4.8μm波段的辐射亮度占比较高,该波段成像效果最好。3.2~3.8μm波段的辐射亮度占比较低,该波段成像效果较差。处于大气吸收带的4.2~4.4μm波段,由于大气透射率几乎为0,无法获得其准确的空间辐射亮度分布。研究表明该拓展中波高光谱成像仪在地面目标的中波红外高光谱成像特性研究上具有目标区分精确、误差小、“图像光谱合一”的特点,测试数据可以应用于目标的红外高光谱隐身设计、目标的分类与识别等领域。

基于近红外高光谱图像的花生内部霉变快速判别方法研究

目的针对外观正常但内部存在不同程度霉变的花生,探索采用近红外高光谱成像技术结合机器学习方法构建花生内部霉变快速无损判别模型的可行性。方法采集100粒内部霉变和100粒健康花生的近红外高光谱图像构成数据集,将多种经典光谱预处理方法与支持向量机(support vector machine,SVM)组合建立花生内部霉变判别模型,并采用蒙特卡洛-无信息变量消除法(Monte Carlo-uninformative variable elimination,MC-UVE)找出霉变判别中有效的光谱特征波长。结果将Savitzky-Golay卷积平滑方法和二阶求导光谱预处理方法与SVM组合,对内部霉变严重样本判别的总体识别准确率可达95%,对不同程度内部霉变样本的平均识别准确率为88%;基于MC-UVE筛选得到10、5、3个特征波长构建的模型总体识别准确率为90%、85%和82%。结论实验结果表明高光谱技术结合机器学习可为花生内部霉变的快速、无损判别提供可行的解决方案,同时特征波长筛选为基于光电原理的霉变花生色选机系统开发提供了参考。

中红外时间拉伸频率上转换高光谱成像系统研究

高光谱成像技术是一种分析物质化学成分、物理性质和形态等信息的非接触式、无损检测方法。然而,中红外波段分子指纹谱测量,受限于探测器响应速度和固有噪声,传统的高光谱成像技术难以同时实现高速和高信噪比测量。基于时间拉伸频率上转换的光谱测量技术具有高测量速度、高光谱分辨率、宽光谱覆盖等优势,与高光谱成像技术结合后,可为快速获取样品种类与形态信息提供可靠手段。搭建了中红外时间拉伸频率上转换高光谱成像系统,同源产生1047 nm泵浦光与1550 nm信号光平均功率分别为2 W和100 mW,结合同步泵浦技术,可在两块周期性极化铌酸锂晶体中分别实现中红外脉冲的产生与频率上转换,完成将中红外波段分子指纹谱信息传递至近红外波段,解决了难以高速、低噪声探测中红外光谱信息的难题。通过对晶体温度与工作通道的调谐,中红外波段探测范围可覆盖2700~3900 nm,具有多样品检测能力。对频率上转换的近红外脉冲进行时间拉伸并与高光谱成像相结合,通过逐点扫描对比色皿中苯溶液吸收光谱及空间分布信息进行实测,光谱数据与傅里叶变换红外光谱仪测量结果高度吻合,能在8 s内完成600μm×1200μm空间区域的高光谱成像,单个像素点采集时间12.9 ns,光谱测量速度77.6 MSpectra·s^(-1),光谱分辨率5.8 cm-1,证实了系统对光谱覆盖范围内液相分子的高速光谱测量及空间分辨能力。解决了传统高光谱测量方式在中红外波段存在响应速度慢、积分时间长、信噪比低的难点,能以10~7帧率的光谱刷新率完成多组分样品检测及其形态测量,为材料与生物成像分析等领域提供了新途径。

基于近红外高光谱技术的杧果可溶性固形物含量无损检测

【目的】近红外高光谱成像技术(NIR-HSI)在水果内部品质的无损检测方面具有快速、准确和无损的特点。旨在利用NIR-HSI技术分析不同品种杧果的可溶性固形物含量,并探讨400~1000nm波段范围内的光谱差异和可溶性固形物含量的响应。【方法】选择贵妃杧果和台农1号杧果作为研究对象,使用NIR-HSI技术获取杧果样本的光谱数据。采用CARS-PLS模型分析可溶性固形物含量与各波段光谱反射率的相关系数。为了验证模型的性能,计算了建模R^(2)、斜率Slope、截距和RMSE等指标。【结果】得到CARS-PLS模型的性能指标:建模R^(2)为0.8806,斜率为0.8515,截距为12.208,RMSE为0.6366。这些指标表明该模型具有较高的建模拟合度和预测精度。【结论】应用NIR-HSI技术对杧果可溶性固形物含量进行检测具有可行性。为进一步研究不同水果可溶性固形物含量的高精度模型奠定了基础。通过NIR-HSI技术的应用,可以提供一种非破坏性且高效准确的方法,用于水果品质评估和检测。这对农产品质量控制和市场营销具有重要的意义。

二维相关红外光谱和高光谱成像技术检测黄精干燥过程中多糖的变化

[目的]研究黄精“一蒸一制”炮制工艺中干燥过程多糖的变化。[方法]采用常压隔水蒸对黄精切片进行蒸制,然后在60℃条件下进行热泵干燥,利用近红外光谱(NIR)、高光谱成像(HSI)和中远红外光谱技术对黄精干燥过程中的特征光谱进行解析,并结合二维相关红外光谱(2D-IR)分析方法研究黄精干燥过程中的多糖变化。[结果]蒸汽加热11 h后黄精多糖含量为3.39%,干燥11 h后多糖含量为6.67%,整体来看,黄精在干燥过程中多糖含量呈逐渐增多的趋势。通过二维相关分析可知,多糖的特征官能团在不断发生变化,在1 016 cm^(-1)处,其官能团变化顺序均为—CH2的C—H伸展→O—H基团在水中拉伸和变形→N—H基团的伸缩振动→O—H拉伸和C—O拉伸组合→C—H拉伸和C—C拉伸组合→吡喃糖苷;HSI技术结合化学计量学,利用PLSR建立多糖的光谱预测模型,模型的R2p为0.903。[结论]红外光谱、HSI技术结合2D-IR技术能很好地监测黄精在干燥过程中的多糖变化规律。

基于遗传算法和深度神经网络的近红外高光谱检测猪肉新鲜度

为系统评估基于深度学习的智能辅助高光谱成像系统在猪肉新鲜度指标检测中的效果,采集了猪肉在4℃冷藏12 d中挥发性盐基氮(volatile basic nitrogen,TVB-N)、菌落总数(total viable count,TVC)以及900~2500 nm近红外光谱数据。基于Python的TensorFlow和Keras平台,对高光谱数据进行处理,建立了深度神经网络的定量检测模型。并利用遗传算法(GA)选择与猪肉新鲜度相关的特征光谱波段。结果表明,遗传算法选取波段对光谱模型的性能有明显提升。当光谱波段数达到35和50时,GA+ANN模型预测精度高于全波段的线性回归模型。TVC为预测指标的预测性能优于TVB-N,TVC测试集最佳R_(p)^(2)为0.877,RMSEP为0.575;预测TVB-N的最佳R_(p)^(2)为0.826,RMSEP为1.01。此外,通过研究还发现,遗传算法优选的近红外光谱波段与肉品的O-H,N-H,C=O等分子振动吸收带有较高的吻合度。本研究为处理近红外和高光谱数据提供了新的方法,也为猪肉及其他肉品新鲜度快速无损检测提供了技术参考。