Fine hyperspectral classification of rice varieties based on self-attention mechanism

The accurate identification of rice varieties using rapid and nondestructive hyperspectral technology is of practical significance for rice cultivation and agricultural production. This paper proposes a convolutional neural network classification model based on a self-attention mechanism (self-attention-1D-CNN) to improve accuracy in distinguishing between crop species in fields using canopy spectral information. After experimental materials were planted in the research area, portable equipment was used to collect the canopy hyperspectral data for rice during the booting stage. Five preprocessing methods and three extraction methods were used to process the data. A comparison of the classification accuracy of different classification models showed that the self-attention-1D-CNN proposed in this study achieved the best classification with an accuracy of 99.93%. The research demonstrated the feasibility of using hyperspectral technology for the fine classification of rice varieties, and the feasibility of using the CNN model as a potential classification method for near-ground crop monitoring and classification.     

Investigation of human pancreatic cancer tissues by Fourier Transform Infrared Hyperspectral Imaging

Fourier‐transform infrared hyperspectral imaging (FTIR‐HSI) provides hyperspectral images containing both morphological and chemical information. It is widely applied in the biomedical field to detect tumor lesions, even at the early stage, by identifying specific spectral biomarkers. Pancreatic neoplasms present different prognoses and are not always easily classified by conventional analyses. In this study, tissue samples with diagnosis of pancreatic ductal adenocarcinoma and pancreatic neuroendocrine tumor were analyzed by FTIR‐HSI and the spectral data compared with those from healthy and dysplastic samples. Multivariate/univariate approaches were complemented to hyperspectral images, and definite spectral markers of the different lesions identified. The malignant lesions were recognizable both from healthy/dysplastic pancreatic tissues (high values of phospholipids and triglycerides with shorter, more branched and less unsaturated alkyl chains) and between each other (different amounts of total lipids, phosphates and carbohydrates). These findings highlight different metabolic pathways characterizing the different samples, well detectable by FTIR‐HSI.     

基于无人机多源遥感数据的亚热带森林树种分类

树种多样性是生态学研究的重要内容,树木的种类和空间分布信息可有效服务于可持续森林管理。但在复杂林分条件下,获取高精度分类结果的难度大。而无人机遥感可获取局域超精细数据,为树种分类精度的提高提供了可能。基于可见光、高光谱、激光雷达等多源无人机遥感数据,探究其在亚热带林分条件下的树种分类潜力。研究发现:(1)随机森林分类器总体精度和各树种的F1分数最高,适合亚热带多树种的分类制图,其区分13种类别(8乔木,4草本)的总体精度为95.63%,Kappa系数为0.948;(2)多源数据的使用可以显著提高分类精度,全特征模型精度最高,且高光谱和激光雷达数据显著影响全特征模型分类精度,可见光纹理数据作用较小;(3)分类特征重要性从大到小排序为结构信息,植被指数,纹理信息,最小噪声变换分量。     

Hyperspectral imaging system using acousto-optic tunable filter for flow cytometry applications.

A major advantage of flow cytometry is its flexible and open instrument configuration, which is highly suitable for systems integration. This flexibility permits the coupling of auxiliary instrumentation that may offer the measurement of parameters other than those typically measured by this multiparameter measurement technique. On the basis of this advantage, we explore the principle and application of hyperspectral imaging (HSI), which has the potential to be a useful add-on feature to flow cytometry applications. Application of HSI to flow cytometry involves the acquisition of spatial information and rendering it in spectral form. In this work, we describe the development and application of an HSI system which provides both spectral and spatial information. Spectral information was generated by obtaining an entire spectrum of a single sample site within a wavelength region of interest, while spatial information was generated by recording a two-dimensional (2D) image of an area of the sample of interest at one specific wavelength. HSI is a promising additional feature to flow cytometry since it can provide both spatial (image format) and spectral information in addition to the multiparameter information already available from flow cytometry measurements.     

Oil Adulteration Identification by Hyperspectral Imaging Using QHM and ICA

To investigate the feasibility of identification of qualified and adulterated oil product using hyperspectral imaging(HIS) technique, a novel feature set based on quantized histogram matrix (QHM) and feature selection method using improved kernel independent component analysis (iKICA) is proposed for HSI. We use UV and Halogen excitations in this study. Region of interest(ROI) of hyperspectral images of 256 oil samples from four varieties are obtained within the spectral region of 400–720nm. Radiation indexes extracted from each ROI are used as feature vectors. These indexes are individual band radiation index (RI), difference of consecutive spectral band radiation index (DRI), ratio of consecutive spectral band radiation index (RRI) and normalized DRI (NDRI). Another set of features called quantized histogram matrix (QHM) are extracted by applying quantization on the image histogram from these features. Based on these feature sets, improved kernel independent component analysis (iKICA) is used to select significant features. For comparison, algorithms such as plus L reduce R (plusLrR), Fisher, multidimensional scaling (MDS), independent component analysis (ICA), and principle component analysis (PCA) are also used to select the most significant wavelengths or features. Support vector machine (SVM) is used as the classifier. Experimental results show that the proposed methods are able to obtain robust and better classification performance with fewer number of spectral bands and simplify the design of computer vision systems.     

Hyperspectral Image Analysis of Live Cells in Various Cell Cycle Stages

In this study we have explored the use of hyperspectral imaging (HSI) to determine the cell-cycle status of live cells in culture. Live cancer cell lines in culture were either synchronized by release from nocodazole or arrested in various cell-cycle phases with serum starvation (G₁), aphidicolin (S), or nocodazole (G₂/M). The live cells were then stained with the fluorescent DNA binding dyes Heochst 33342 or Dyecycle orange along with propidium iodide or Mitotracker green. Microscopic HSI data was then collected using the PARISS HSI system. Classified spectra were incorporated into spectral libraries; and all spectra acquired from each sample were correlated with library spectra to a user-determined confidence threshold, generating a unique spectral signature for each sample. Examination of these spectral signatures revealed that all cell cycle phases could be objectively differentiated. Ongoing studies employing other viable cell fluorescent dyes, and dyes in combination may provide more robust spectral signatures defining the status and condition of living cells.     

Combining hyperspectral imaging and chemometrics to assess and interpret the effects of environmental stressors on zebrafish eye images at tissue level

Changes on an organism by the exposure to environmental stressors may be characterized by hyperspectral images (HSI), which preserve the morphology of biological samples, and suitable chemometric tools. The approach proposed allows assessing and interpreting the effect of contaminant exposure on heterogeneous biological samples monitored by HSI at specific tissue levels. In this work, the model example used consists of the study of the effect of the exposure of chlorpyrifos‐oxon on zebrafish tissues. To assess this effect, unmixing of the biological sample images followed by tissue‐specific classification models based on the unmixed spectral signatures is proposed. Unmixing and classification are performed by multivariate curve resolution‐alternating least squares (MCR‐ALS) and partial least squares‐discriminant analysis (PLS‐DA), respectively. Crucial aspects of the approach are: (1) the simultaneous MCR‐ALS analysis of all images from 1 population to take into account biological variability and provide reliable tissue spectral signatures, and (2) the use of resolved spectral signatures from control and exposed populations obtained from resampling of pixel subsets analyzed by MCR‐ALS multiset analysis as information for the tissue‐specific PLS‐DA classification models. Classification results diagnose the presence of a significant effect and identify the spectral regions at a tissue level responsible for the biological change.      

Optimized 1D-CNN model for medicinal Psyllium Husk crop mapping with temporal optical satellite data

Deep Learning models are preferred for complex image analysis-based solutions to application-oriented problems. However, the architecture of such models largely influences the results which includes several hyperparameters that need to be tuned. This study aims at developing an optimized 1D-CNN model for medicinal Psyllium Husk crop mapping using open source temporal optical Sentinel-2A/2B satellite data. In this study, a sequential 1D-CNN model architecture was developed by optimizing hyperparameters which includes convolution layers, number of neurons, activation function, and batch size. Psyllium Husk crop fields were mapped in the Jalore district of Rajasthan using Sentinel 2A/ 2B (10 m) optical data. For spectral dimensionality reduction of the data, Modified Soil Adjusted Vegetation Index (MSAVI2) was used to maintain the data dimensionality since temporal data was utilized. The dataset was subsequently refined to include the target crop's specific phenological stages that distinguish it from other closely resembling species. The information corresponding to these specific crop stages was fed to the 1D-CNN model to carry out the classification. A range of training sample sizes were explored to determine the optimal number of training data points. As the output from the model, fractional images are obtained consisting of values proportional to the probability of a pixel lying in the target class. Accuracy assessment was carried out using fuzzy error matrix (FERM) by generating fractional output images from temporal optical PlanetScope data (3m) which was used as a reference. The best overall accuracy among the test cases came out to be 89.85% using conventional MSAVI2 with 1000 training samples.     

Exposure of preterm neonates to toxic metals during their stay in the Neonatal Intensive Care Unit and its impact on neurodevelopment at 2 months of age

BackgroundPremature neonates might be exposed to toxic metals during their stay in the neonatal intensive care unit (NICU), which could adversely affect neurodevelopment; however, limited evidence is available. The present study was therefore designed to assess the exposure to mercury, lead, cadmium, arsenic, and manganese of preterm neonates who received total parenteral nutrition (TPN) and/or red blood cell (RBC) transfusions during their NICU stay and the risk of neurodevelopment delay at the age of 2 months.MethodsWe recruited 33 preterm neonates who required TPN during their NICU admission. Blood samples were collected for metal analysis at two different time points (admission and before discharge). Metals in the daily TPN received by preterm neonates were analyzed. Neurodevelopment was assessed using the Ages and Stages Questionnaire Edition 3 (ASQ-3).ResultsAll samples of TPN had metal contamination: 96% exceeded the critical arsenic limit (0.3 μg/kg body weight/day); daily manganese intake from TPN for preterm neonates exceeded the recommended dose (1 µg/kg body weight) as it was added intentionally to TPN solutions, raising potential safety concerns. All samples of RBC transfusions exceeded the estimated intravenous reference dose for lead (0.19 µg/kg body weight). Levels of mercury, lead and manganese in preterm neonates at discharge decreased 0.867 µg/L (95% CI, 0.76, 0.988), 0.831 (95%CI, 0.779, 0.886) and 0.847 µg/L (95% CI, 0.775, 0.926), respectively. A decrease in ASQ-3-problem solving scores was associated with higher levels of blood lead in preterm neonates taken at admission (ß = −0.405, 95%CI = −0.655, −0.014), and with plasma manganese (ß = −0.562, 95%CI = −0.995, −0.172). We also observed an association between decreased personal social domain scores with higher blood lead levels of preterm neonates before discharge (ß = −0.537, 95%CI = −0.905, −0.045).ConclusionOur findings provide evidence to suggest negative impacts on the neurodevelopment at 2 months of preterm infants exposed to certain metals, possibly related to TPN intake and/or blood transfusions received during their NICU stay. Preterm neonates may be exposed to levels of metals in utero.     

Ground-level spectroscopy analyses and classification of coral reefs using a hyperspectral camera

With the general aim of classification and mapping of coral reefs, remote sensing has traditionally been more difficult to implement in comparison with terrestrial equivalents. Images used for the marine environment suffer from environmental limitation (water absorption, scattering, and glint); sensor-related limitations (spectral and spatial resolution); and habitat limitation (substrate spectral similarity). Presented here is an advanced approach for ground-level surveying of a coral reef using a hyperspectral camera (400–1,000 nm) that is able to address all of these limitations. Used from the surface, the image includes a white reference plate that offers a solution for correcting the water column effect. The imaging system produces millimeter size pixels and 80 relevant bands. The data collected have the advantages of both a field point spectrometer (hyperspectral resolution) and a digital camera (spatial resolution). Finally, the availability of pure pixel imagery significantly improves the potential for substrate recognition in comparison with traditionally used remote sensing mixed pixels. In this study, an image of a coral reef table in the Gulf of Aqaba, Red Sea, was classified, demonstrating the benefits of this technology for the first time. Preprocessing includes testing of two normalization approaches, three spectral resolutions, and two spectral ranges. Trained classification was performed using support vector machine that was manually trained and tested against a digital image that provided empirical verification. For the classification of 5 core classes, the best results were achieved using a combination of a 450–660 nm spectral range, 5 nm wide bands, and the employment of red-band normalization. Overall classification accuracy was improved from 86 % for the original image to 99 % for the normalized image. Spectral resolution and spectral ranges seemed to have a limited effect on the classification accuracy. The proposed methodology and the use of automatic classification procedures can be successfully applied for reef survey and monitoring and even upscaled for a large survey.     

Intraoperative multispectral and hyperspectral label‐free imaging: A systematic review of in vivo clinical studies

Multispectral and hyperspectral imaging (HSI) are emerging optical imaging techniques with the potential to transform the way surgery is performed but it is not clear whether current systems are capable of delivering real‐time tissue characterization and surgical guidance. We conducted a systematic review of surgical in vivo label‐free multispectral and HSI systems that have been assessed intraoperatively in adult patients, published over a 10‐year period to May 2018. We analysed 14 studies including 8 different HSI systems. Current in‐vivo HSI systems generate an intraoperative tissue oxygenation map or enable tumour detection. Intraoperative tissue oxygenation measurements may help to predict those patients at risk of postoperative complications and in‐vivo intraoperative tissue characterization may be performed with high specificity and sensitivity. All systems utilized a line‐scanning or wavelength‐scanning method but the spectral range and number of spectral bands employed varied significantly between studies and according to the system's clinical aim. The time to acquire a hyperspectral cube dataset ranged between 5 and 30 seconds. No safety concerns were reported in any studies. A small number of studies have demonstrated the capabilities of intraoperative in‐vivo label‐free HSI but further work is needed to fully integrate it into the current surgical workflow.      

Benefits of hyperspectral remote sensing for tracking plant invasions

Aim We aim to report what hyperspectral remote sensing can offer for invasion ecologists and review recent progress made in plant invasion research using hyperspectral remote sensing. Location United States. Methods We review the utility of hyperspectral remote sensing for detecting, mapping and predicting the spatial spread of invasive species. We cover a range of topics including the trade‐off between spatial and spectral resolutions and classification accuracy, the benefits of using time series to incorporate phenology in mapping species distribution, the potential of biochemical and physiological properties in hyperspectral spectral reflectance for tracking ecosystem changes caused by invasions, and the capacity of hyperspectral data as a valuable input for quantitative models developed for assessing the future spread of invasive species. Results Hyperspectral remote sensing holds great promise for invasion research. Spectral information provided by hyperspectral sensors can detect invaders at the species level across a range of community and ecosystem types. Furthermore, hyperspectral data can be used to assess habitat suitability and model the future spread of invasive species, thus providing timely information for invasion risk analysis. Main conclusions Our review suggests that hyperspectral remote sensing can effectively provide a baseline of invasive species distributions for future monitoring and control efforts. Furthermore, information on the spatial distribution of invasive species can help land managers to make long‐term constructive conservation plans for protecting and maintaining natural ecosystems.     

Hyperspectral scanning laser optical tomography

In order to study physical relationships within tissue volumes or even organism‐level systems, the spatial distribution of multiple fluorescent markers needs to be resolved efficiently in three dimensions. Here, rather than acquiring discrete spectral images sequentially using multiple emission filters, a hyperspectral scanning laser optical tomography system is developed to obtain hyperspectral volumetric data sets with 2‐nm spectral resolution of optically transparent mesoscopic (millimeter‐centimeter) specimens. This is achieved by acquiring a series of point‐scanning hyperspectral extended depth of field images at different angles and subsequently tomographically reconstructing the 3D intensity distribution for each wavelength. This technique is demonstrated to provide robust measurements via the comparison of spectral and intensity profiles of fluorescent bead phantoms. Due to its enhanced spectral resolving ability, this technique is also demonstrated to resolve largely overlapping fluorophores, as demonstrated by the 3D fluorescence hyperspectral reconstruction of a dual‐labeled mouse thymus gland sample and the ability to distinguish tumorous and normal tissues of an unlabeled mouse intestine sample.      

Classification of desert grassland species based on a local-global feature enhancement network and UAV hyperspectral remote sensing

Grassland ecosystems are an important part of terrestrial ecosystems and are important for building ecological barriers, promoting the pastoral economy, and maintaining social stability. In recent decades, grasslands in northern China have undergone extensive degradation due to the combined effects of global climate change and the anthropogenic overuse of grasslands. An understanding of the spatial distribution of grassland degradation species is helpful for evaluating the process of grassland degradation and formulating appropriate protective measures. This is important for grassland degradation monitoring. To address the limitations of traditional ground surveys and realize intelligent remote sensing grassland degradation monitoring tasks, we use unmanned aerial vehicle (UAV) hyperspectral remote sensing technology to collect data on vegetation species in desert grasslands. In this paper, we propose a local-global feature enhancement network (LGFEN) for the classification of desert grassland species. The method uses the local feature enhancement (LFE) module and global feature enhancement (GFE) module to extract local and spatial information from hyperspectral images (HSIs), respectively. In addition, the introduction of the convolutional block attention module (CBAM) refines the features of HSIs, improving the stability of the classification performance. The results show that the proposed method has superior classification performance compared with existing HSI classification methods. With only 10 training samples per class, the overall accuracy, average accuracy, and kappa coefficient of the proposed method were 98.61%, 97.61%, and 0.9815, respectively. The proposed method provides a new approach for high-precision and high-efficiency dynamic monitoring of grassland ecosystems.     

Hyperspectral discrimination of coral reef benthic communities in the western Caribbean

Determining a subset of wavelengths that best discriminates reef benthic habitats and their associated communities is essential for the development of remote sensing techniques to monitor them. This study measured spectral reflectance from 17 species of western Caribbean reef biota including coral, algae, seagrasses, and sediments, as well as healthy and diseased coral. It sought to extend the spectral library of reef-associated species found in the literature and to test the spectral discrimination of a hierarchy of habitats, community groups, and species. We compared results from hyperspectral reflectance and derivative datasets to those simulated for the three visible multispectral wavebands of the IKONOS sensor. The best discriminating subset of wavelengths was identified by multivariate stepwise selection procedure (discriminant function analysis). Best discrimination at all levels was obtained using the derivative dataset based on 6–15 non-contiguous wavebands depending on the level of the classification, followed by the hyperspectral reflectance dataset which was based on as few as 2–4 non-contiguous wavebands. IKONOS wavebands performed worst. The best discriminating subset of wavelengths in the three classification resolutions, and particularly those of the medium resolution, was in agreement with those identified by Hochberg and Atkinson (2003) and Hochberg et al. (2003) for reef communities worldwide. At all levels of classification, reflectance wavebands selected by the analysis were similar to those reported in recent studies carried out elsewhere, confirming their applicability in different biogeographical regions. However the greater accuracies achieved using the derivative datasets suggests that hyperspectral data is required for the most accurate classification of reef biotic systems.     

Multivariate analysis of Brillouin imaging data by supervised and unsupervised learning

Brillouin imaging relies on the reliable extraction of subtle spectral information from hyperspectral datasets. To date, the mainstream practice has been to use line fitting of spectral features to retrieve the average peak shift and linewidth parameters. Good results, however, depend heavily on sufficient signal-to-noise ratio and may not be applicable in complex samples that consist of spectral mixtures. In this work, we thus propose the use of various multivariate algorithms that can be used to perform supervised or unsupervised analysis of the hyperspectral data, with which we explore advanced image analysis applications, namely unmixing, classification and segmentation in a phantom and live cells. The resulting images are shown to provide more contrast and detail, and obtained on a timescale ∼10² faster than fitting. The estimated spectral parameters are consistent with those calculated from pure fitting.     

Hyperspectral imaging of mucosal surfaces in patients

The aim of this study was to proof applicability of hyperspectral imaging for the analysis and classification of human mucosal surfaces in vivo. The larynx as a prototypical anatomically well‐defined surgical test area was analyzed by microlaryngoscopy with a polychromatic lightsource and a synchronous triggered monochromatic CCD‐camera. Image stacks (5 benign, 7 malignant tumors) were analyzed by established software (principal component analysis PCA, hyperspectral classification, spectral profiles). Hyperspectral image datacubes were analyzed and classified by conventional software. In PCA, images at 590–680 nm loaded most onto the first PC which typically contained 95% of the total information. Hyperspectral classification clustered the data highlighting altered mucosa. The spectral profiles clearly differed between the different groups. Hyperspectral imaging can be applied to mucosal surfaces. This approach opens the way to analyze spectral characteristics of histologically different lesions in order to build up a spectral library and to allow non‐touch optical biopsy. (© 2012 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Parallel morphological/neural processing of hyperspectral images using heterogeneous and homogeneous platforms

The wealth spatial and spectral information available from last-generation Earth observation instruments has introduced extremely high computational requirements in many applications. Most currently available parallel techniques treat remotely sensed data not as images, but as unordered listings of spectral measurements with no spatial arrangement. In thematic classification applications, however, the integration of spatial and spectral information can be greatly beneficial. Although such integrated approaches can be efficiently mapped in homogeneous commodity clusters, low-cost heterogeneous networks of computers (HNOCs) have soon become a standard tool of choice for dealing with the massive amount of image data produced by Earth observation missions. In this paper, we develop a new morphological/neural algorithm for parallel classification of high-dimensional (hyperspectral) remotely sensed image data sets. The algorithm’s accuracy and parallel performance is tested in a variety of homogeneous and heterogeneous computing platforms, using two networks of workstations distributed among different locations, and also a massively parallel Beowulf cluster at NASA’s Goddard Space Flight Center in Maryland.

Hyperspectral imaging combined with machine learning as a tool to obtain high‐throughput plant salt‐stress phenotyping

The rapid selection of salinity‐tolerant crops to increase food production in salinized lands is important for sustainable agriculture. Recently, high‐throughput plant phenotyping technologies have been adopted that use plant morphological and physiological measurements in a non‐destructive manner to accelerate plant breeding processes. Here, a hyperspectral imaging (HSI) technique was implemented to monitor the plant phenotypes of 13 okra (Abelmoschus esculentus L.) genotypes after 2 and 7 days of salt treatment. Physiological and biochemical traits, such as fresh weight, SPAD, elemental contents and photosynthesis‐related parameters, which require laborious, time‐consuming measurements, were also investigated. Traditional laboratory‐based methods indicated the diverse performance levels of different okra genotypes in response to salinity stress. We introduced improved plant and leaf segmentation approaches to RGB images extracted from HSI imaging based on deep learning. The state‐of‐the‐art performance of the deep‐learning approach for segmentation resulted in an intersection over union score of 0.94 for plant segmentation and a symmetric best dice score of 85.4 for leaf segmentation. Moreover, deleterious effects of salinity affected the physiological and biochemical processes of okra, which resulted in substantial changes in the spectral information. Four sample predictions were constructed based on the spectral data, with correlation coefficients of 0.835, 0.704, 0.609 and 0.588 for SPAD, sodium concentration, photosynthetic rate and transpiration rate, respectively. The results confirmed the usefulness of high‐throughput phenotyping for studying plant salinity stress using a combination of HSI and deep‐learning approaches.     

In Vitro Identification of Gold Nanorods through Hyperspectral Imaging

Due to their unique plasmonic and optical properties, gold nanorods (GNR) have shown tremendous potential for nano-based applications extending into a variety of fields including bioimaging, sensor development, electronics, and cancer therapy. These distinctive, shape-specific properties are strongly dependent upon the GNR aspect ratio, thus producing the ability to be targeted for an application by fine-tuning their physical parameters. It is owing to their characteristic spectral signature, which is vastly different from that of a cellular setting, that GNRs are emerging as an ideal candidate for nano-based imaging applications. However, one challenge that has emerged in the field of bioimaging is the need to account for the observed plasmon coupling effect that arises from GNR agglomeration in a physiological environment. In this study, GNRs with aspect ratios of 2.5 and 6.0 were actively identified in an in vitro setting through a hyperspectral imaging (HSI) analysis; which successfully recognized and separated the light scattering pattern of these particles from that of the surrounding cells. Through inclusion of agglomerated GNR spectral patterns in the HSI spectral library, this imaging technique was able to overcome the complication of plasmon coupling, though to varying degrees. These results demonstrate the tremendous potential of GNRs coupled with HSI analysis to advance the field of nano-based sensing and imaging mechanisms.