Linear Spectral Unmixing Algorithm Research Articles

Mapping and discrimination of the mineralization potential in granitoids from Banyo area (Adamawa, Cameroon), using Landsat 9 OLI, ASTER images and field observations

The Banyo area is situated in the west Cameroon domain of the Central African Fold Belt. In this study, field data and remote sensing conventional mapping techniques including Red-Green-Blue (RGB) colour combination, principal component analysis (PCA), Crosta-PCA, Band Ratios (BR), Constrained Energy Minimization (CEM), Minimum Noise Fraction (MNF), Spectral Angle Mapper (SAM) classification and Linear Spectral Unmixing (LSU) algorithms were used to process Landsat 9 OLI and ASTER data in order to discriminate various rock units such as biotite-amphibole granites, hornblende-biotite orthogneiss and map associated mineral deposits (e.g., iron-oxide, hydroxyl and corundum minerals). In addition, it was found that RGB combination (743) of OLI bands, PCA of OLI bands (PC123), BR of OLI bands (4/2, 6/7, 6/5), BR of ASTER (4/2, 7/6, 8/9), Crosta-PCA of OLI bands, MNF (123) of OLI bands, SAM, and LSU techniques applied to ASTER bands ascertain potential locations of granitoids and gneisses bearing iron and hydroxyl minerals. On the other hand, this study also revealed that corundum deposits are detected by applying the CEM method to OLI band ratio 6/7 highlighting red tones and processing SAM classification algorithm both using laboratory reference spectra. Automatic (using PCI Geomatica line tool) and manual extractions of lineaments revealed three major trends (N-S, E-W, and WNW-ESE to NW-SE) and an accessory trend (NE-SW to ENE-WSW) that surround hydrothermally altered rocks recognized as containing iron oxide, hydroxyl, and corundum mineral deposits. In the field, these lineaments may group to the tree deformation phases. The D1 phase (NW-SE); the D2 phase (NE-SW to ENE-WSW) and the D3 phase (N-S). Maximum Likelihood Calssification (MLC) and SAM classification methods showed to be more effective in discriminating rocks bearing Iron-oxide, hydroxyl and corundum minerals in the study area.This result permits to confirm that, the combination of remote sensing (Landsat 9 OLI and ASTER data in this case) and field data constitute a useful tool for mineral exploration in the shear zone. However, the accuracy assessment was relied on comparing the different maps produced from Landsat 9 OLI and ASTER data with the ground truth sample points and also, with the earliest geological map of Banyo area. Thus, quantitatively the results were accurate at 99%.

Combining Optical and Radar Satellite Imagery to Investigate the Surface Properties and Evolution of the Lordsburg Playa, New Mexico, USA

Driven by erodible soil, hydrological stresses, land use/land cover (LULC) changes, and meteorological parameters, windblown dust events initiated from Lordsburg Playa, New Mexico, United States, threaten public safety and health through low visibility and exposure to dust emissions. Combining optical and radar satellite imagery products can provide invaluable benefits in characterizing surface properties of desert playas—a potent landform for wind erosion. The optical images provide a long-term data record, while radar images can observe land surface irrespective of clouds, darkness, and precipitation. As a home for optical and radar imagery, powerful algorithms, cloud computing infrastructure, and application programming interface applications, Google Earth Engine (GEE) is an invaluable resource facilitating acquisition, processing, and analysis. In this study, the fractional abundance of soil, vegetation, and water endmembers were determined from pixel mixtures using the linear spectral unmixing model in GEE for Lordsburg Playa. For this approach, Landsat 5 and 8 images at 30 m spatial resolution and Sentinel-2 images at 10–20 m spatial resolution were used. Employing the Interferometric Synthetic Aperture Radar (InSAR) techniques, the playa’s land surface changes and possible sinks for sediment loading from the surrounding catchment area were identified. In this data recipe, a pair of Sentinel-1 images bracketing a monsoon day with high rainfall and a pair of images representing spring (dry, windy) and monsoon seasons were used. The combination of optical and radar images significantly improved the effort to identify long-term changes in the playa and locations within the playa susceptible to hydrological stresses and LULC changes. The linear spectral unmixing algorithm addressed the limitation of Landsat and Sentinel-2 images related to their moderate spatial resolutions. The application of GEE facilitated the study by minimizing the time required for acquisition, processing, and analysis of images, and storage required for the big satellite data.

An Assessment of Water Sources for Heritage Listed Organic Mound Springs in NW Australia Using Airborne Geophysical (Electromagnetics and Magnetics) and Satellite Remote Sensing Methods

Discrete phreatophytic vegetation associated with organic mound springs is present in several places in the semi-arid Walyarta Conservation Park (Park) in northern Western Australia. The mound springs are heritage listed, having significant cultural and environmental significance. Increased industrial (mining and agriculture) development in the region, coupled with a growing demand for groundwater to support these developments, requires an enhanced understanding of how the springs operate and the source of water that sustains their presence. The springs are broadly believed to be situated on geological faults and receive groundwater from artesian sources. However, their association with deeper geological structures and aquifer systems, the focus of this study, is not well understood. This study employed regional- and finer-scale airborne geophysical data, including electromagnetics (AEM) and magnetics, to constrain the sub-basin-scale hydrogeology of the West Canning Basin in Western Australia and to detail tectonic deformation, sedimentological and hydrological processes. The AEM data were inverted using 1- and 2D methods to better define structural discontinuities in the Park, and the results identified the location of faults and other geological structures that were coincident with spring locations. A complementary analysis of spatiotemporal patterns of green vegetation was undertaken using remote sensing data. A model for the extent of green vegetation (in percent), calculated using a constrained linear spectral unmixing algorithm and applied to a select Landsat Thematic Mapper ™ image archive, showed the persistence of green vegetation aligned with interpreted fault systems through extended dry periods. These geophysical and remotely sensed datasets demonstrate that in the Park, the sedimentary aquifers and landscapes are highly compartmentalized and that this constrains aquifer distribution, groundwater quality and the location of wetlands and phreatophytic vegetation. Integrating key information from these datasets allows for the construction of a three-dimensional model that predicts the nature and extent of the critical zone which sustains perennial groundwater discharge within mound springs, drainages and wetlands and provides a framework to assess discharge rates, mixing and, ultimately, sensitivity to changed water availability.

Wetland Mapping Using HJ-1A/B Hyperspectral Images and an Adaptive Sparse Constrained Least Squares Linear Spectral Mixture Model

In this study, we proposed an adaptive sparse constrained least squares linear spectral mixture model (SCLS-LSMM) to map wetlands in a sophisticated scene. It includes three procedures: (1) estimating the abundance based on sparse constrained least squares method with all endmembers in the spectral library, (2) selecting “active” endmember combinations for each pixel based on the estimated abundances and (3) estimating abundances based on the linear spectral unmixing algorithm only with the adaptively selected endmember combinations. The performances of the proposed SCLS-LSMM on wetland vegetation communities mapping were compared with the traditional full constrained least squares linear spectral mixture model (FCLS-LSMM) using HJ-1A/B hyperspectral images. The accuracy assessment results showed that the proposed SCLS-LSMM obtained a significantly better performance with a systematic error (SE) of –0.014 and a root-mean-square error (RMSE) of 0.087 for Reed marsh, and a SE of 0.004 and a RMSE of 0.059 for Weedy meadow, compared with the traditional FCLS-LSMM. The proposed methods improved the unmixing accuracies of wetlands’ vegetation communities and have the potential to understand the process of wetlands’ degradation under the impacts of climate changes and permafrost degradation.

Mapping Oyster Reef Distribution Using Kompsat-2/3 and Linear Spectral Unmixing Algorithm - A Case Study at Hwangdo Tidal Flat

Kim, K.-L. and Ryu, J.-H., 2020. Mapping oyster reef distribution using Kompsat-2/3 and linear spectral unmixing algorithm – A case study at Hwangdo tidal flat. In: Jung, H.-S.; Lee, S.; Ryu, J.-H., and Cui, T. (eds.), Advances in Geospatial Research of Coastal Environments. Journal of Coastal Research, Special Issue No. 102, pp. 246-253. Coconut Creek (Florida), ISSN 0749-0208.Oyster areas are widely used as species of biological indicator, and they are also the most important shellfish in terms of ecosystem economic valuation. Oysters are production on the west coast before 2007 accounted for about 5 % of the nation's total production, but the 2007 Hebei Spirit oil spill caused production to plunge to less than 1 %. Therefore, spatial distribution maps of oyster reefs are required to help local authorities define management strategies. In this study, Kompsat-2/3 was used to map oyster reef distribution and analyze the distribution of oyster reefs based on linear spctral unmixing in Hwangdo tidal flat. A spectral library, collected in situ for various conditions with a field spectroradiometer, was used to conduct linear spectral unmixing classifications on Kompsat-2/3 data. The classification result shows very high accuracy, with an overall accuracy of 95 % or more, and there were misclassifications in some areas. The most causes of misclassification are the similarity of spectral characteristics and the limitations of the spatial and spectral resolutions of satellite images. For this reason, it was difficult to distinguish between oyster reefs and area distributed with many macrobenthos, and small-young oyster reefs were difficult to classify due to very weak reflectivity. In addition, the sand bar, it was is difficult to distinguish between oyster reefs related to dead reefs and sandbars because of the high reflectivity of these areas in the imagery. As a result of analyzing the change in the oyster reef area, it increased in 2019 compared to that in 2015. Especially, the oyster reef area increased in 30-50 % sand content and decreased in 20-30 % and 60-70 %. The changes of sand sediment seem to affect the distribution of oyster reefs. This study may be useful for mapping the distribution of oyster reefs and understanding the spatial variation of their habitat.

The effects of misregistration between hyperspectral and panchromatic images on linear spectral unmixing

ABSTRACT Low probability subpixel target extraction and identification is important for hyperspectral (HS) applications. To better exploit the signatures of targets and extract targets from mixed pixels, we proposed a linear spectral unmixing algorithm combining HS and panchromatic (Pan) images (called SU-Co-Ims). However, misregistration between HS and Pan images is common, which usually has a negative impact on subsequent applications. Our motivation is to attempt to quantify the misregistration error range in which the extracted subpixel target is valid. To determine the maximum acceptable misregistration error (MAME), we focus on analysing the impact of misregistration between images on target extraction, that is, the effects of misregistration between images on linear spectral unmixing. Taking Pan image as reference, HS image and Pan image are intentionally misregistered in the along-track direction. The proposed SU-Co-Ims method is applied to decompose mixed pixels and extract subpixel targets. Spectral angle mapper (SAM) and Euclidean distance (ED) are used to evaluate spectral unmixing error introduced by misregistration between images. Results indicate that spectral unmixing error increases with misregistration error, and the MAME varies from 0.35 to 1.94 pixels for imagers with different spatial resolution. Consequently, accurate image registration remains crucial to unmixing-based subpixel target extraction, but misregistration has a low impact on results when the misregistration error between images is smaller than the MAME.

Estimation of boreal forest floor reflectance from airborne hyperspectral data of coniferous forests

In open forest canopies, such as in boreal forests, forest floor can contribute significantly to the observed top of canopy reflectance. In order to retrieve the biophysical properties of the tree layer, correcting for forest floor is essential. Traditionally, the algorithms for retrieval of forest floor reflectance depend on tree layer information such as leaf area index, canopy cover, and site fertility. To overcome these circular dependencies, we propose an algorithm that can be applied only using airborne remote sensing data. We acquired airborne hyperspectral imagery over the Hyytiälä forest research station (61°50′N, 24°17′E) in central Finland on July 3rd in 2015 using a hyperspectral pushbroom line scanner. The image data had a spectral resolution of 4.6 nm, and the spatial resolution was 0.6 m. We developed a linear spectral unmixing algorithm, which is based on the definition of the reflectance factor, taking into account the variation of incident irradiance inside the canopy. The weights of the mixture can be computed from tree canopy gap fractions, a tree species insensitive leaf albedo, and average tree stand reflectance. Canopy gap fractions were retrieved with empirical methods available in scientific literature. The forest floor reflectance in the near-infrared increased with site fertility in agreement with the forest floor field measurements. Moreover, we found that in near infrared, the reflectance of moderately rich and moist upland forests was significantly different from all other fertility classes. Finally, we tested the reflectance decomposition on the photochemical reflectance index (PRI) known to be heavily affected by understory reflectance and canopy structure, and the forest PRI to be decoupled from the PRI of the over- and understory.

A COMPARATIVE ANALYSIS OF SPECTRAL INFORMATION EXTRACTION ALGORITHMS FOR TARGET DETECTION OF HYDROTHERMAL ALTERATION ZONES USING ASTER SATELLITE IMAGE DATA

Abstract. The focus of this paper is to evaluate the performance of the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data for target detection of hydrothermal alteration zones associated with geothermal (GT) springs as proxy for narrowing areas of interest. The study employed the Per-pixel Spectral Angle Mapper (SAM) and the Sub-pixel Linear Spectral Unmixing (LSU) algorithms for spectral information extraction by using the ASTER satellite image data. In both cases, image endmember spectra specifically for kaolinite, alunite, and illite and calcite zones were selected and extracted by using the Analytical Imaging and Geophysics (AIG)-developed processing methods. The results of the analysis show that both SAM and LSU discriminated targets of interest better when employing image spectra and poorly when using library spectra. However, the Per-pixel SAM is unsuitable for target detection and more suited where the objective of the investigation is to classify whole scene and not particular targets as in this case. The LSU was found to be effective for discriminating alterations associated with the thermal springs especially where image endmember spectra are employed for analysis, thus recommended for prefeasibility mapping of GT related resources.

Identifying high potential zones of gold mineralization in a sub-tropical region using Landsat-8 and ASTER remote sensing data: A case study of the Ngoura-Colomines goldfield, eastern Cameroon

Climatic conditions and vegetation constrain the use of optical satellite imagery as an exploration tool for hydrothermal ore mineralization in tropical and subtropical regions. In this investigation, Landsat-8 and ASTER satellite imagery were used to detect hydrothermal alteration zones associated with gold mineralization in the Ngoura-Colomines region, Eastern Cameroon. The study area contains several gold-bearing quartz veins associated with zones of pyritization, muscovite/sericite, iron oxides, and silicification. Principal Component Analysis (PCA), Independent Component Analysis (ICA), and specialized spectral band ratios were used to extract spectral information related to vegetation, iron oxide/hydroxide minerals, Al–OH, Fe-Mg–OH, carbonate group minerals, and silicification using Landsat-8 data at regional scale. Linear Spectral Unmixing (LSU) algorithm was implemented to ASTER VNIR + SWIR bands for detailed discrimination of hematite, jarosite, kaolinite, muscovite, chlorite and epidote at district scale. The Automated Spectral Hourglass (ASH) technique was employed to extract reference spectra directly from the ASTER bands for producing fraction images of end-members using the LSU. A comprehensive field survey was used to verify the remote sensing results. Petrographic study, X-ray diffraction analysis and reflectance spectroscopy indicated the presence of quartz, goethite and sericite, as well as the absorption features of Fe3+/Fe2+, Al–OH, OH/H2O and SiO2 in the alteration zones. Several hydrothermal alteration zones of iron oxide/hydroxide, clay, carbonate minerals and silicification zones were identified, which are spatially associated with known mining areas and gold occurrences in the study area. High potential prospects were also delineated, including the Ngoura-Colomines prospects and the newly discovered Yangamo-Ndatanga and Taparé-Tapondo prospects in the southwestern and southeastern parts of the study area. Consequently, satellite-based mineral prospectivity maps at regional and district scales were generated for the study area by implementing the fuzzy logic model to the most informative thematic layers derived from image processing results. The satellite-based prospectivity maps are reliable for exploration of new gold prospective zones in the Ngoura-Colomines goldfield.

FPGA Implementation of L1/2 Sparsity Constrained Nonnegative Matrix Factorization Algorithm for Remotely Sensed Hyperspectral Image Analysis

Remotely sensed hyperspectral images provide data of the earth’s surface components. The data provided is collected through airborne devices such as satellites with the capability to collect large amounts of data to be sent to ground stations for processing. The main disadvantage of this scenario is the limited bandwidth connection between the airborne devices and the ground station on Earth which affects the information sending and real time processing. A possible solution is to include an on-board data processor. Field-Programmable Gate Arrays (FPGAs) are excellent target platform that allows the design reconfigurability, powerful computing and high performance levels. One of the most commonly used techniques in hyperspectral data analysis is linear spectral unmixing. In the last decade, L1/2 sparsity constrained Nonnegative Matrix Factorization (NMF), a linear spectral unmixing algorithm, and its extensions have been heavily studied to unmix the hyperspectral images and recover their material spectra. L1/2 regularizer is proven to have much better results in terms of sparsity and accuracy than other regularizers yet, to the best of our knowledge, has not been implemented. In this paper, we present an FPGA design for the L1/2 sparsity constrained NMF (L1/2-NMF) algorithm. The proposed design is tested on both synthetic and real data sets and implemented on Altera Family FPGAs. Implementation results show that the proposed design successfully unmixes the data with maximum frequency of 52.6 MHz and a speedup factor of 3.9 for the synthetic data set and a frequency of 104.32 MHz and a speedup factor of 1.14 for the real data set. The implementation results are compared to the simulation results and ground truth signatures using Spectral Angular Distance (SAD) measure. Calculations show that the implementation results have comparable SAD values to the simulation results.

VALIDATION OF EXTRACTED ENDMEMBERS FROM HYPERSPECTRAL IMAGES

Abstract. An essential step in the characterization of surface materials using hyperspectral image analysis is image classification using endmembers. Spectral unmixing is the best method for hyperspectral image classification. This method assumes that the pixel-topixel variability in scene results from varying proportions of spectral endmembers. Spectral endmembers can be derived from the imagery or measurements in the laboratory or field. The primary objective of this paper was to assess the ability of extracted endmembers against some different solutions for extraction of endmembers for hyperspectral image classification. In this paper, we compared the Pixel Purity Index (PPI) and the Sequential Maximum Angle Convex Cone (SMACC) as two popular methods of endmember extraction with library and field spectral. We used spectral information divergence for detection desirable endmembers from field spectral. For accuracy assessment of spectral mixture analysis and production of endmember abundance images for each of methods, the linear spectral unmixing algorithm is used. After a comparison between the results of these methods, it has been verified that field spectral have a better classification result in comparing with other endmember extraction methods. Also, the PPI has reliable results as an automatic endmember extraction method in comparing.

Landsat-8, Advanced Spaceborne Thermal Emission and Reflection Radiometer, and WorldView-3 Multispectral Satellite Imagery for Prospecting Copper-Gold Mineralization in the Northeastern Inglefield Mobile Belt (IMB), Northwest Greenland

Several regions in the High Arctic still lingered poorly explored for a variety of mineralization types because of harsh climate environments and remoteness. Inglefield Land is an ice-free region in northwest Greenland that contains copper-gold mineralization associated with hydrothermal alteration mineral assemblages. In this study, Landsat-8, Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), and WorldView-3 multispectral remote sensing data were used for hydrothermal alteration mapping and mineral prospecting in the Inglefield Land at regional, local, and district scales. Directed principal components analysis (DPCA) technique was applied to map iron oxide/hydroxide, Al/Fe-OH, Mg-Fe-OH minerals, silicification (Si-OH), and SiO2 mineral groups using specialized band ratios of the multispectral datasets. For extracting reference spectra directly from the Landsat-8, ASTER, and WorldView-3 (WV-3) images to generate fraction images of end-member minerals, the automated spectral hourglass (ASH) approach was implemented. Linear spectral unmixing (LSU) algorithm was thereafter used to produce a mineral map of fractional images. Furthermore, adaptive coherence estimator (ACE) algorithm was applied to visible and near-infrared and shortwave infrared (VINR + SWIR) bands of ASTER using laboratory reflectance spectra extracted from the USGS spectral library for verifying the presence of mineral spectral signatures. Results indicate that the boundaries between the Franklinian sedimentary successions and the Etah metamorphic and meta-igneous complex, the orthogneiss in the northeastern part of the Cu-Au mineralization belt adjacent to Dallas Bugt, and the southern part of the Cu-Au mineralization belt nearby Marshall Bugt show high content of iron oxides/hydroxides and Si-OH/SiO2 mineral groups, which warrant high potential for Cu-Au prospecting. A high spatial distribution of hematite/jarosite, chalcedony/opal, and chlorite/epidote/biotite were identified with the documented Cu-Au occurrences in central and southwestern sectors of the Cu-Au mineralization belt. The calculation of confusion matrix and Kappa Coefficient proved appropriate overall accuracy and good rate of agreement for alteration mineral mapping. This investigation accomplished the application of multispectral/multi-sensor satellite imagery as a valuable and economical tool for reconnaissance stages of systematic mineral exploration projects in remote and inaccessible metallogenic provinces around the world, particularly in the High Arctic regions.

Linear spectral unmixing algorithm for modelling suspended sediment concentration of flash floods, upper Tekeze River, Ethiopia

Flash floods are the highest sediment transporting agent, but are inaccessible for in-situ sampling and have rarely been analyzed by remote sensing technology. Laboratory and field experiments were done to develop linear spectral unmixing (LSU) remote sensing model and evaluate its performance in simulating the suspended sediment concentration (SSC) in flash floods. The models were developed from continuous monitoring in the laboratory and the onsite spectral signature of river bed sediment deposits and flash floods in the Tekeze River and in its tributary, the Tsirare River. The Pearson correlation coefficient was used to determine the variability of correlations between reflectance and SSCs. The coefficient of determination (R2) and root mean square of error (RMSE) were used to evaluate the performance of the generated models. The results found that the Pearson correlation coefficient between SSCs and reflectance varied based on the level of the SSCs, geological colors, and grain sizes. The performance of the LSU model and empirical remote sensing approaches were computed to be R2 = 0.92, and RMSE = ±0.76 g/l in the Tsirare River and R2 = 0.91, and RMSE = ±0.73 g/l in the Tekeze River and R2 = 0.81, RMSE = ±2.65 g/l in the Tsirare river and R2 = 0.76, RMSE = ±10.87 g/l in the Tekeze River, respectively. Hence, the LSU approach of remote sensing was found to be relatively accurate in monitoring and modeling the variability of SSCs that could be applied to the upper Tekeze River basin.

Multicolor re-scan super-resolution imaging of live cells.

Multicolor fluorescence microscopy has proved essential in biological studies. However, the application of conventional multicolor microscopy to imaging subcellular organelles is restricted by its diffraction-limited spatial resolution. Re-scan confocal microscopy (RCM), a novel super-resolution imaging technique, can effectively address this problem. However, previous multicolor RCM imaging methods usually led to spatial mismatch in images due to the sequential scanning of the sample with multiple excitation lasers. We present a new RCM system to achieve multicolor super-resolution imaging. A spectrograph was used as the multicolor detection system, and a linear spectral unmixing algorithm was applied to separate different fluorophores in the spectral image. Moreover, since the image reconstruction process induced an artificial resolution improvement, a gamma correction was introduced to restore the multicolor super-resolution image. By imaging phalloidin-labeled F-actin in breast cancer cells, we found that the lateral resolution of our system is approximately 171 nm, which is a 1.8-fold improvement over that of wide-field imaging. The successful identification of three types of fluorescent beads indicated that our multicolor RCM can resolve different fluorophores whose spectra largely overlap with each other. Finally, we demonstrated that our method is suitable for imaging multicolor-labeled organelles of live cells. Our novel RCM system can acquire multicolor super-resolution images of live cells without spatial mismatch, obvious photobleaching or photodamage. This system may provide a new imaging tool for monitoring dynamic events involving interactions between multiple molecules and organelles in cells.

Palaeo-Channel Bisecting Puri Town, Odisha:Vestige of the Lost River ‘Saradha’?

Puri, an ancient town located on the Odisha (Orissa) coast in eastern India, is associated with the 12th century CE Jagannath Temple. Many ancient and historical texts mention a mythical river named Saradha that flowed across the present Grand Road ( Badadanda ) between the Jagannath and Gundicha temples in Puri, dividing the town into two parts. This study traces the trail of a palaeo-channel beyond and within the heart of Puri town through an integrated study of geology, satellite imagery and ground-penetrating radar (GPR) survey. Various satellite imagery manipulation techniques – band combination, NDVI, MNDWI, linear spectral unmixing algorithm, density slice and spatial profiling – indicate the existence of water components along with vegetation bands, and the presence of a sinusoidal palaeo-channel trace with a V-shaped topographic outline. This is interpreted to represent the remnants of an extinct river valley. GPR survey also suggests that a subsurface river valley, about 128 m wide, existed across parts of the Grand Road. A large, sinusoidal water body near the sea beach at Puri, may represent the last remains of the dried-up channel. Integration of all these features suggests that a river once existed between Jagannath and Gundicha temples in Puri town. The constructed palaeo-channel trail may be that of the lost Saradha river described in ancient texts. The study develops a sequential methodology for identifying palaeo-channels even in urbanized localities like Puri.

Nonlinear Hyperspectral Unmixing Based on Geometric Characteristics of Bilinear Mixture Models

Recently, many nonlinear spectral unmixing algorithms that use various bilinear mixture models (BMMs) have been proposed. However, the high computational complexity and intrinsic collinearity between true endmembers and virtual endmembers considerably decrease these algorithms’ unmixing performances. In this paper, we come up with a novel abundance estimation algorithm based on the BMMs. Motivated by BMMs’ geometric characteristics that are related to collinearity, we conduct a unique nonlinear vertex <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${p}$ </tex-math></inline-formula> to replace all the virtual endmembers. Unlike the virtual endmembers, this vertex <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${p}$ </tex-math></inline-formula> actually works as an additional true endmember that gives affine representations of pixels with other true endmembers. When the pixels’ normalized barycentric coordinates with respect to true endmembers are obtained, they will be directly projected to be their approximate linear mixture components, which removes the collinearity effectively and enables further linear spectral unmixing. After that, based on the analysis of projection bias, two strategies using the projected gradient algorithm and a traditional linear spectral unmixing algorithm, respectively, are provided to correct the bias and estimate more accurate abundances. The experimental results on simulated and real hyperspectral data show that the proposed algorithm performs better compared with both traditional and state-of-the-art spectral unmixing algorithms. Both the unmixing accuracy and speed have been improved.

Spectral mineral mapping for characterization of subtle geothermal prospects using ASTER data

In this study, the performance of ASTER data is evaluated for mapping subtle geothermal prospects in an unexplored tropical region having a number of thermal springs. The study employed a simple Decorrelation stretch with specific absorptions to highlight possible alteration zones of interest related to Geothermal (GT) systems. Hydrothermal alteration minerals are subsequently mapped using Spectral Angle Mapper (SAM) and Linear Spectral Unmixing (LSU) algorithms to target representative minerals such as clays, carbonates and AL-OH minerals as indicators of GT activity. The results were validated through field GPS survey, rock sampling and laboratory analysis using latest smart lab X-Ray Diffractometer technology. The study indicates that ASTER broadband satellite data could be used to map subtle GT prospects with the aid of an in-situ verification. However, it also shows that ASTER could not discriminate within specie minerals especially for clays using SWIR bands. Subsequent studies are aimed at looking at both ASTER and Hyperion hyperspectral data in the same area as this could have significant implications for GT resource detection in unmapped aseismic and inaccessible tropical regions using available spaceborne data.

URBAN MORPHOLOGICAL DYNAMICS IN SANTIAGO (CHILE): PROPOSING SUSTAINABLE INDICATORS FROM REMOTE SENSING

Latin America is one of the world’s most urbanised regions, with more than 80% of inhabitants living in urban areas and over 50 cities with at least 1 million inhabitants. The concept of urban structure types (UST) allows the dynamics of a growing urban environment to be captured in its quantity and quality. They are defined as areas of homogenous appearance in the urban matrix with a recognisable mixture of built-up areas and open spaces. We used the vegetation-impervious-soil (V-I-S) model approach to classify and monitor different types of USTs in Santiago (~800 km2), Chile between 1985 and 2015. The V-I-S model is based on a simplification of the large diversity of urban land cover types in three general categories: vegetation, impervious surfaces and soil. These categories were obtained by processing Landsat-5 TM and Landsat-8 OLI images. First, we applied standard radiometric calibration and co-registration methods to all datasets. Second, using a linear spectral unmixing algorithm we performed a soft classification of urban land cover types (&amp;lt;i&amp;gt;end members&amp;lt;/i&amp;gt;): trees, shrubs, herbaceous plants, soils, buildings, roads and water bodies. All &amp;lt;i&amp;gt;end members&amp;lt;/i&amp;gt; were validated using a combination of photointerpretation on high-resolution images (~1 m) and field data collection (only for 2015). In each pixel we used the resulting probability scores, logically grouped, to obtain final values for each V-I-S component. Third, we used statistical clustering of V-I-S values to create a set of eight pixel groups, which we interpreted as USTs and mapped them for each date. The overall accuracy for V-I-S components in 1985 and 2015 were 78% and 82%, respectively, and errors did not exhibit any spatial correlation. The main sources of differentiation between USTs were the trade-off proportions between vegetation and impervious components, whereas soil proportions remained near 5% across the city in both dates. To analyse the change in UST spatial configuration between dates, we used a set of selected landscape metrics and discussed their use as indicators for sustainable urban development. These indicators relate to the dispersion pattern of urban growth, the connectivity of open green space and the complexity in the composition of the UST types within the different sectors of the city. We were able to identify, using the dynamics exhibited by the USTs, three main zones: (1) city centre, where USTs of high-intensity development predominate, (2) eastern high-income areas whose spatial structure is marked by a relatively high urbanisation intensity with a very large proportion of vegetated spaces, and (3) peripheral areas, with significant changes in composition and configuration of USTs, in recent decades, showing high rates of urbanisation, shifting from low-medium to high densities. We concluded that these patterns and their dynamics are mainly determined by the spatial socio-economic stratification of the population.

URBAN MORPHOLOGICAL DYNAMICS IN SANTIAGO (CHILE): PROPOSING SUSTAINABLE INDICATORS FROM REMOTE SENSING

Abstract. Latin America is one of the world’s most urbanised regions, with more than 80% of inhabitants living in urban areas and over 50 cities with at least 1 million inhabitants. The concept of urban structure types (UST) allows the dynamics of a growing urban environment to be captured in its quantity and quality. They are defined as areas of homogenous appearance in the urban matrix with a recognisable mixture of built-up areas and open spaces. We used the vegetation-impervious-soil (V-I-S) model approach to classify and monitor different types of USTs in Santiago (~800 km2), Chile between 1985 and 2015. The V-I-S model is based on a simplification of the large diversity of urban land cover types in three general categories: vegetation, impervious surfaces and soil. These categories were obtained by processing Landsat-5 TM and Landsat-8 OLI images. First, we applied standard radiometric calibration and co-registration methods to all datasets. Second, using a linear spectral unmixing algorithm we performed a soft classification of urban land cover types (end members): trees, shrubs, herbaceous plants, soils, buildings, roads and water bodies. All end members were validated using a combination of photointerpretation on high-resolution images (~1 m) and field data collection (only for 2015). In each pixel we used the resulting probability scores, logically grouped, to obtain final values for each V-I-S component. Third, we used statistical clustering of V-I-S values to create a set of eight pixel groups, which we interpreted as USTs and mapped them for each date. The overall accuracy for V-I-S components in 1985 and 2015 were 78% and 82%, respectively, and errors did not exhibit any spatial correlation. The main sources of differentiation between USTs were the trade-off proportions between vegetation and impervious components, whereas soil proportions remained near 5% across the city in both dates. To analyse the change in UST spatial configuration between dates, we used a set of selected landscape metrics and discussed their use as indicators for sustainable urban development. These indicators relate to the dispersion pattern of urban growth, the connectivity of open green space and the complexity in the composition of the UST types within the different sectors of the city. We were able to identify, using the dynamics exhibited by the USTs, three main zones: (1) city centre, where USTs of high-intensity development predominate, (2) eastern high-income areas whose spatial structure is marked by a relatively high urbanisation intensity with a very large proportion of vegetated spaces, and (3) peripheral areas, with significant changes in composition and configuration of USTs, in recent decades, showing high rates of urbanisation, shifting from low-medium to high densities. We concluded that these patterns and their dynamics are mainly determined by the spatial socio-economic stratification of the population.

Estimating forest species abundance through linear unmixing of CHRIS/PROBA imagery

The advancing technology of hyperspectral remote sensing offers the opportunity of accurate land cover characterization of complex natural environments. In this study, a linear spectral unmixing algorithm that incorporates a novel hierarchical Bayesian approach (BI-ICE) was applied on two spatially and temporally adjacent CHRIS/PROBA images over a forest in North Pindos National Park (Epirus, Greece). The scope is to investigate the potential of this algorithm to discriminate two different forest species (i.e. beech – Fagus sylvatica, pine – Pinus nigra) and produce accurate species-specific abundance maps. The unmixing results were evaluated in uniformly distributed plots across the test site using measured fractions of each species derived by very high resolution aerial orthophotos. Landsat-8 images were also used to produce a conventional discrete-type classification map of the test site. This map was used to define the exact borders of the test site and compare the thematic information of the two mapping approaches (discrete vs abundance mapping). The required ground truth information, regarding training and validation of the applied mapping methodologies, was collected during a field campaign across the study site. Abundance estimates reached very good overall accuracy (R2=0.98, RMSE=0.06). The most significant source of error in our results was due to the shadowing effects that were very intense in some areas of the test site due to the low solar elevation during CHRIS acquisitions. It is also demonstrated that the two mapping approaches are in accordance across pure and dense forest areas, but the conventional classification map fails to describe the natural spatial gradients of each species and the actual species mixture across the test site. Overall, the BI-ICE algorithm presented increased potential to unmix challenging objects with high spectral similarity, such as different vegetation species, under real and not optimum acquisition conditions. Its full potential remains to be investigated in further and more complex study sites in view of the upcoming satellite hyperspectral missions.