Hyperspectral Data Analysis Research Articles

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Alternating proximal gradient method for sparse nonnegative Tucker decomposition

Multi-way data arises in many applications such as electroencephalography (EEG) classification, face recognition, text mining and hyperspectral data analysis. Tensor decomposition has been commonly used to find the hidden factors and elicit the intrinsic structures of the multi-way data. This paper considers sparse nonnegative Tucker decomposition (NTD), which is to decompose a given tensor into the product of a core tensor and several factor matrices with sparsity and nonnegativity constraints. An alternating proximal gradient method (APG) is applied to solve the problem. The algorithm is then modified to sparse NTD with missing values. Per-iteration cost of the algorithm is estimated scalable about the data size, and global convergence is established under fairly loose conditions. Numerical experiments on both synthetic and real world data demonstrate its superiority over a few state-of-the-art methods for (sparse) NTD from partial and/or full observations. The MATLAB code along with demos are accessible from the author's homepage.

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Reliable Cell Segmentation Based on Spectral Phasor Analysis of Hyperspectral Stimulated Raman Scattering Imaging Data

Hyperspectral stimulated Raman scattering (SRS) imaging has rapidly become an emerging tool for high content analyses of cell and tissue systems. The label-free nature of SRS imaging combined with its chemical specificity allows in situ and in vivo biochemical quantification at submicrometer resolution without sectioning and staining. Current hyperspectral SRS data analysis methods are based on either linear unmixing or multivariate analysis, which are not sensitive to small spectral variations and often provide obscure information on the cell composition. Here, we demonstrate a spectral phasor analysis method that allows fast and reliable cellular organelle segmentation of mammalian cells, without any a priori knowledge of their composition or basis spectra. We further show that, in combination with a branch-bound algorithm for optimal selection of a few wavenumbers, spectral phasor analysis provides a robust solution to label-free single cell analysis.

PerTurbo Manifold Learning Algorithm for Weakly Labeled Hyperspectral Image Classification

Hyperspectral data analysis has been given a growing attention due to the scientific challenges it raises and the wide set of applications that can benefit from it. Classification of hyperspectral images has been identified as one of the hottest topics in this context, and has been mainly addressed by discriminative methods such as SVM. In this paper, we argue that generative methods, and especially those based on manifold representation of classes in the hyperspectral space, are relevant alternatives to SVM. To illustrate our point, we focus on the recently published PerTurbo algorithm and benchmark against SVM this generative manifold learning algorithm in the context of hyperspectral image classification. This choice is motivated by the fact that PerTurbo is fitted with numerous interesting properties, such as low sensitivity to dimensionality curse, high accuracy in weakly labelled images classification context (few training samples), straightforward extension to on-line setting, and interpretability for the practitioner. The promising results call for an up-to-date interest toward generative algorithms for hyperspectral image classification.

Special issue on Advances in Hyperspectral Data Processing and Analysis

Special issue on Advances in Hyperspectral Data Processing and Analysis

Detection of Callosobruchus maculatus (F.) infestation in soybean using soft X-ray and NIR hyperspectral imaging techniques

Soybean (Glycine max L.) is a major oilseed crop grown throughout the world and, total post-harvest losses of soybean are approximately 10%, and 3% of produced soybean is lost during storage. Cowpea weevil (Callosobruchus maculatus (F.)) is the major storage pest which causes extensive storage losses of legumes. Detection of early stages of cowpea weevil infestation could assist farmers and storage facility managers in implementing suitable control practices for insect disinfestations. Soft X-ray and near-infrared (NIR) hyperspectral imaging techniques were used to acquire images of soybeans infested by egg, larval, and pupal stages of C. maculatus along with uninfested and completely damaged (hollowed-out after emergence of adults) soybeans. From soft X-ray images, totally, 33 features (12 histogram and 21 textural features) were extracted and from hyperspectral data 48 features were extracted (30 histogram and 18 spectral features) for analysis. Linear and quadratic discriminant analysis (LDA and QDA) models were developed using these extracted features to classify different stages of infestation. The LDA classifier for soft X-ray images correctly identified more than 86% of uninfested soybeans and 83% of soybeans infested with all developmental stages of C. maculatus except the egg stage. Pair-wise LDA classification models developed from NIR hyperspectral data yielded more than 86 and 87% classification accuracy for uninfested and infested seeds, respectively. The QDA pair-wise classifiers positively differentiated more than 79% uninfested seeds from infested seeds. The principal component analysis of NIR hyperspectral data identified the wavelengths of 960 nm, 1030 nm and 1440 nm being responsible for more than 99% of spectral variability. Combining soft X-ray features with hyperspectral features increased the classification accuracies for egg and larvae compared to either imaging system used alone.

Hyperspectral imaging of bruises

Introduction Hyperspectral imaging combines a camera with a tunable optical filter to record reflectance spectrophotometry data over a range of wavelengths for each image point. The data can be mathematically processed. Bruises comprise extravasated blood (haemoglobin) in the skin that is broken down to haemosiderin and bilirubin, the presence of which can be demonstrated by analysis of hyperspectral data. Aims To develop hyperspectral image data processing to demonstrate the spatial distribution of bruises and their temporal changes. Method Bruise-type injuries were created by injection of blood into skin of rats (with Animal Ethics Approval 1 ). Hyperspectral images were captured using a Chemimage HSI examiner 100 QD at a range of time points. Hyperspectral data were processed to analyse the first derivative to quantify the accumulation of degradation products of haemoglobin; tissue was also examined histologically. Results Hyperspectral images were successfully obtained; however, initial processing did not provide a defined correlation with time or histological changes. Discussion Initial results using simple analysis of first derivative values did not produce a clear correlation with accumulation of haemoglobin degradation products over time. Advanced processing techniques will be tested and the outcome of these will be presented.

Hyperspectral Data Analysis for Detecting Lead Pollution in Rice

This paper presents a new method for detecting lead pollution in rice by analyzing hyperspectral data. First, preprocessing method is used to remove the outliers which deviate so much from other hyperspectral data. Then, dimensionality-reduced data are made by using discrete wavelet transform. Finally, linear discriminant analysis is utilized to extract the feature which characterizes polluted and unpolluted rice. The experimental result based on the proposed method shows the good performance in detecting lead pollution in rice.

A Fast Smoothing Algorithm for Post-Processing of Surface Reflectance Spectra Retrieved from Airborne Imaging Spectrometer Data

Surface reflectance spectra retrieved from remotely sensed hyperspectral imaging data using radiative transfer models often contain residual atmospheric absorption and scattering effects. The reflectance spectra may also contain minor artifacts due to errors in radiometric and spectral calibrations. We have developed a fast smoothing technique for post-processing of retrieved surface reflectance spectra. In the present spectral smoothing technique, model-derived reflectance spectra are first fit using moving filters derived with a cubic spline smoothing algorithm. A common gain curve, which contains minor artifacts in the model-derived reflectance spectra, is then derived. This gain curve is finally applied to all of the reflectance spectra in a scene to obtain the spectrally smoothed surface reflectance spectra. Results from analysis of hyperspectral imaging data collected with the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) data are given. Comparisons between the smoothed spectra and those derived with the empirical line method are also presented.

State-of-the-art Microanalysis at the nm-scale and Smaller: Going from Pretty Pictures to Quantitative Analysis of Hyperspectral Data

Extended abstract of a paper presented at Microscopy and Microanalysis 2013 in Indianapolis, Indiana, USA, August 4 – August 8, 2013.

Application of the wavelet transform for feature extraction in the analysis of hyperspectral laser-induced fluorescence data

Remote-sensing approaches for environmental protection and exploration have evolved rapidly in the last decade. Among the new operational tools, hyperspectral Fluorescent LiDAR System (FLS®) lidar has demonstrated a high sensitivity and the ability to function in complex environments for real-time, robust oil-spill monitoring on airborne or ship-borne analytical platforms. The capabilities of such analytical platforms include real-time analysis of laser-induced fluorescence (LIF) data. Although numerous examples of the application of signal theory to the analysis of hyperspectral data appear in the remote-sensing literature, the conventional data analysis strategies are not well adapted to the practical issues of the LIF applications. The aim of this article is to provide a new approach for LIF lidar analytical platforms, which is focused on the specifics of hyperspectral LIF data. The approach is based on structural data analysis and interpretation, through which more detailed spectral matching is performed. This article is based on a simulated experiment in which the spectra of actual seawater and well-known types of petroleum products were combined to demonstrate the wavelet-transform-based analysis of LIF data. The final part of the article demonstrates the application of the wavelet transform to the structural analysis of LIF data from field experiments for the detection and identification of oil products in difficult environmental conditions.

PlanetServer: Innovative approaches for the online analysis of hyperspectral satellite data from Mars

PlanetServer is a WebGIS system, currently under development, enabling the online analysis of Compact Reconnaissance Imaging Spectrometer (CRISM) hyperspectral data from Mars. It is part of the EarthServer project which builds infrastructure for online access and analysis of huge Earth Science datasets. Core functionality consists of the rasdaman Array Database Management System (DBMS) for storage, and the Open Geospatial Consortium (OGC) Web Coverage Processing Service (WCPS) for data querying. Various WCPS queries have been designed to access spatial and spectral subsets of the CRISM data. The client WebGIS, consisting mainly of the OpenLayers javascript library, uses these queries to enable online spatial and spectral analysis. Currently the PlanetServer demonstration consists of two CRISM Full Resolution Target (FRT) observations, surrounding the NASA Curiosity rover landing site. A detailed analysis of one of these observations is performed in the Case Study section. The current PlanetServer functionality is described step by step, and is tested by focusing on detecting mineralogical evidence described in earlier Gale crater studies. Both the PlanetServer methodology and its possible use for mineralogical studies will be further discussed. Future work includes batch ingestion of CRISM data and further development of the WebGIS and analysis tools.

On the CFAR Property of the RX Algorithm in the Presence of Signal-Dependent Noise in Hyperspectral Images

In this paper, we investigate the constant false-alarm rate (CFAR) property of the RX anomaly detector which is widely used for the analysis of hyperspectral data. The RX detector relies on an adaptive scheme where the mean vector and the covariance matrix of the background are locally estimated from the image pixels themselves. First, demeaning is accomplished by removing the estimated local background mean value, and then, the covariance matrix is estimated in a homogeneous neighborhood of each pixel. In principle, if the local mean is perfectly removed and the covariance matrix is estimated from background pixels sharing the same covariance matrix, the RX algorithm has the CFAR property, which is highly desirable in practical applications. The CFAR behavior of the algorithm also requires the spatial stationarity of the random noise affecting the hyperspectral image. In data collected by new-generation sensors, such an assumption is not valid because photon noise contribution, which depends on the spatially varying signal level, is not negligible. This has motivated us to analyze the behavior of the RX algorithm with respect to the CFAR property in data affected by signal-dependent (SD) noise. In this paper, we show both theoretically and experimentally that the SD noise is one of the causes of the non-CFAR behavior of the RX detector that we have experienced in many practical situations. We propose a strategy to enhance the robustness of the anomaly detection scheme with respect to the CFAR property based on an adaptive nonlinear transform aimed at reducing the dependence of the noise on the signal level. Experiments on simulated data and real data collected by a new hyperspectral camera are also presented and discussed.

Hyperspectral Remote Sensing Data Analysis and Future Challenges

Hyperspectral remote sensing technology has advanced significantly in the past two decades. Current sensors onboard airborne and spaceborne platforms cover large areas of the Earth surface with unprecedented spectral, spatial, and temporal resolutions. These characteristics enable a myriad of applications requiring fine identification of materials or estimation of physical parameters. Very often, these applications rely on sophisticated and complex data analysis methods. The sources of difficulties are, namely, the high dimensionality and size of the hyperspectral data, the spectral mixing (linear and nonlinear), and the degradation mechanisms associated to the measurement process such as noise and atmospheric effects. This paper presents a tutorial/overview cross section of some relevant hyperspectral data analysis methods and algorithms, organized in six main topics: data fusion, unmixing, classification, target detection, physical parameter retrieval, and fast computing. In all topics, we describe the state-of-the-art, provide illustrative examples, and point to future challenges and research directions.

Randomized Approximation Methods for the Efficient Compression and Analysis of Hyperspectral Data

Hyperspectral imaging techniques such as matrix-assisted laser desorption ionization (MALDI) mass spectrometry imaging produce large, information-rich datasets that are frequently too large to be analyzed as a whole. In addition, the "curse of dimensionality" adds fundamental limits to what can be done with such data, regardless of the resources available. We propose and evaluate random matrix-based methods for the analysis of such data, in this case, a MALDI mass spectrometry image from a section of rat brain. By constructing a randomized orthornormal basis for the data, we are able to achieve reductions in dimensionality and data size of over 100 times. Furthermore, this compression is reversible to within noise limits. This allows more-conventional multivariate analysis techniques such as principal component analysis (PCA) and clustering methods to be directly applied to the compressed data such that the results can easily be back-projected and interpreted in the original measurement space. PCA on the compressed data is shown to be nearly identical to the same analysis on the original data but the run time was reduced from over an hour to 8 seconds. We also demonstrate the generality of the method to other data sets, namely, a hyperspectral optical image of leaves, and a Raman spectroscopy image of an artificial ligament. In order to allow for the full evaluation of these methods on a wide range of data, we have made all software and sample data freely available.

Mapping Regolith and Gossan for Mineral Exploration in the Eastern Kumaon Himalaya, India using hyperion data and object oriented image classification

Crystalline in the Kumaon Himalaya, India near Askot area is a prominent site of the base metal mineralization and gossanised surface. This area is hosted by the sulphides and sulphates of Cu, Pb, Zn and Au and Ag mineralization with the altered rocks like sericite chlorite schist, gneiss etc. Due to the deep weathering this area is also a good illustration site of the gossanised outcrop.Regolith mapping through the multispectral remotely sensed data using different parametric and nonparametric classification algorithm has been used for many years. In recent years, object oriented classification for classification of object rather than pixel has gained a good success within the geospatial community. On the other hand, space borne hyperspectral remote sensing has gained a great success in identification of the minerals from space. The narrow contiguous bands of this hyperspectral remote sensing data can provide more information of the chemical content of the different minerals.In this study EO1 (Earth Observation) Hyperion hyperspectral sensor data has been evaluated for regolith and gossan mapping using the object oriented image classification technique. The efficacy of the hyperion data is evaluated in and around the Askot base metal mineral deposits for hydrothermal alteration and base metal minerals. Three sites were selected by regolith mapping using object oriented image classification method and were evaluated by the aid of hyperspectral data for alteration minerals. During the field verification it was found that the mineralization within these sites was associated with the dolomite, gneiss and schists. Strong hydrothermal activity and shearing were dominated at these sites. Minerals like carbonates, sulphates and sulphides of copper, lead, zinc and silver, magnetite along with the altered minerals like, mica, chlorite, talc etc. were found at the target sites. eCognition™ and ENVI® software’s were used for the object based classification and for the hyperspectral data processing, respectively. It was concluded from this study that object oriented classification is a suitable classification algorithm for regolith and gossan mapping as it uses the spatial, spectral and textural information of the remote sensing dataset. Target areas suggested by the object oriented classification method were examined by the hyperspectral data analysis. Results of the hyperspectral analysis were in good confirmation of the minerals found in the field as well as in the lab. Though the signal to noise ratio of the scene was low but still hyperion data was able to highlight the mineralogy of the Askot and nearby areas. It was also concluded by this study that the spectral analysis was very helpful for identification of the gossan and limonite over the conventional petrography analysis.