HyMap Image Research Articles

Retrieving key benthic cover types and bathymetry from hyperspectral imagery

Bathymetry and the spatial distribution of benthic cover in coastal waters are of key importance in managing and monitoring our coastal water environments. Currently very little of the Western Australian shallow coastal water habitats are mapped, and for those maps that do exist, the spatial resolution generally is poor and the information is dated. Aircraft and space-borne hyperspectral sensors have been shown to be useful in imaging substrate features in shallow coastal waters. This paper describes a method for quantitatively estimating both bathymetry and benthic cover in shallow waters from hyperspectral imagery. The method incorporates a shallow water reflectance model, which accounts for the water column absorption and backscattering, water depth and substrate reflectance. The model was tested against simulated reflectance data, demonstrating the models' ability to retrieve appropriate fractional coverage of sediment, sea grass and brown algae for depths ranging from 1 - 12 m. The model was applied to a HyMap image encompassing a portion of the Jurien Bay Marine Park off the coast of Western Australia. The retrieved benthic cover products were compared to underwater video observations sampled within the image. The comparison shows the method's great potential for characterizing key aspects of marine ecosystems from remotely sensed hyperspectral data.

Estimating and mapping grass phosphorus concentration in an African savanna using hyperspectral image data

We tested the utility of imaging spectroscopy and neural networks to map phosphorus concentration in savanna grass using airborne HyMAP image data. We also sought to ascertain the key wavelengths for phosphorus prediction using hyperspectral remote sensing. The remote sensing of foliar phosphorus has received very little attention as compared to nitrogen, yet it plays an equally important role in explaining the distribution and feeding patterns of herbivores. Band depths from two continuum‐removed absorption features as well as the red edge position (REP) were input into a backpropagation neural network. Following a series of experiments to ascertain the optimum wavelengths, the best trained neural network was used to predict and ultimately to map grass phosphorus concentration in the Kruger National Park. The results indicate that the best trained neural network could predict phosphorus distribution with a coefficient of determination of 0.63 and a root mean square error (RMSE) of 0.07 (28% of the mean observed phosphorus concentration) on an independent test data set. Our results also show that the absorption feature located in the shortwave infrared (R 2015–2199) contains more information on phosphorus distribution, a region that has hardly been explored before in most spectroscopic experiments for phosphorus as compared to the visible bands. Overall, the study demonstrates the potential of imaging spectroscopy in mapping grass phosphorus concentration in savanna rangelands.

River Floodplain Vegetation Scenario Development Using Imaging Spectroscopy Derived Products as Input Variables in a Dynamic Vegetation Model

River floodplains are becoming increasingly subject to multifunctional land-use. In this contribution, we are linking imaging spectrometer derived products with a dynamic vegetation model to improve the simulation and evaluation of scenarios for a river floodplain in the Netherlands. In particular, we are using airborne HyMap imaging spectrometer data to derive Leaf Area Index (LAI), spatial distribution of Plant Functional Types (PFT), and model dominant species abundances as input for the ecological model. We use the dynamic vegetation model (DVM) SMART2-SUMO to simulate vegetation succession under scenarios of changing abiotic conditions and management regimes. SMART2 is a soil chemical model whereas SUMO describes plant competition and resulting vegetation succession. We validate all remote sensing derived products and the DVM calibration independently using extensive field sampling. We conclude that the dynamic vegetation models can be successfully initialized using imaging spectrometer data at currently unprecedented accuracy. However, all efforts undertaken for validation in this contribution may significantly exceed capacities for national or continental scale application of the proposed method.

The Importance of Scale in Object-based Mapping of Vegetation Parameters with Hyperspectral Imagery

In recent years, object-oriented image analysis has been widely adopted by the remote sensing community. Much attention has been given to its application, while the fundamental issue of scale, here characterized by spatial object-definition, seems largely neglected. In the case of vegetation parameters like aboveground biomass and leaf area index (LAI), fundamental objects are individual trees or shrubs, each of which has a specific value. Their spatial extent, however, does not match pixels in size and shape, nor does it fit the requirements of regional studies. Estimation of vegetation parameters consequently demands larger observation units, like vegetation patches, which are better represented by variably shaped objects than by square pixels. This study aims to investigate optimal object definition for biomass and LAI. We have data from 243 field plots in our test site in southern France. They cover a vegetation range from landes to garrigue to maquis, which is considered to be the climax vegetation in the area. A HyMap image covers the area. The image is subjected to a Minimum Noise Fraction (MNF) transformation, after which it is segmented with ten different heterogeneities. The result is ten object sets, each having a different mean object size. These object sets are combined with the original image with the mean band values serving as object attributes. Field observations are linked to the corresponding objects for each object set. Using Ridge regression, relations between field observations and spectral values are identified. The prediction error is determined for each object set by cross validation. The overall lowest prediction error indicates the optimal heterogeneity for segmentation. Results show that the scale of prediction affects prediction accuracy, that increasing the object size yields an optimum in prediction accuracy, and that aboveground biomass and LAI can be associated with different optimal object sizes. Furthermore, it is shown that the accuracy of parameter estimation is higher for object-oriented analysis than for per-pixel analysis.

Mapping geology associated with manganese mineralisation using spectral sensing techniques at Woodie Woodie, East Pilbara

Mineralogical and geological maps are now possible through a new generation of airborne and satellite systems including the 126-band airborne HyMap and 14-band satellite-borne ASTER imaging sensors. Data from both these imaging technologies were tested for their ability to map geological host rock and mineralogies associated with the Woodie Woodie manganese (Mn) mineral deposits. A detailed study was conducted over the deposits using processed ASTER and HyMap data in conjunction with field and laboratory data. In addition, processed ASTER products were also generated for other prospective mineralised terrains over a larger area within the surrounding Bangemall Basin.Seamless maps were generated in this study from ASTER and HyMap imagery to represent surface abundances of either mineral groups or specific mineral identities. The ASTER imagery, at 15 to 90 m resolution, enabled the generation of maps designed to represent the abundance of broad mineral groups, including MgOH/carbonate, quartz/silica, ferric, and ferrous iron (within silicate or carbonate). However, the accuracy of these products was observed to be limited, mainly by the broad spectral resolution of the ASTER bands. Narrower spectral bands are generally required to discriminate and map specific minerals, and separate the effects of vegetation and the atmosphere. In contrast, 5 metre resolution HyMap data demonstrated accurate mapping of abundances of minerals such as ferric iron and dolomite/calcite, as well as dry and green vegetation. Image products representing ferrous-rich carbonate units and opaques (including manganese) were also generated.The results of this study showed that the best products for mapping the Carawine Dolomite and Pinjian Chert Breccia units hosting the Mn mineralisation are obtained by using HyMap data and included: carbonate abundance; ferrous iron content within carbonate units; and ferric iron abundance. The ASTER quartz map product also proved useful to identify silicified dolomite. The geological interpretation of these mineral maps was aided by the comparison and integration with digital elevation model (DEM) and Thorium information from detailed airborne radiometric surveying.

Atmospheric correction of hyperspectral data over dark surfaces via simulated annealing

A method [atmospheric correction via simulated annealing (ACSA)] is proposed that enhances the atmospheric correction of hyperspectral images over dark surfaces. It is based on the minimization of a smoothness criterion to avoid the assumption of linear variations of the reflectance within gas absorption bands. We first show that this commonly used approach generally fails over dark surfaces when the signal to noise ratio strongly declines. In this case, important residual features highly correlated with the shape of gas absorption bands are observed in the estimated surface reflectance. We add a geometrical constraint to deal with this correlation. A simulated annealing approach is used to solve this constrained optimization problem. The parameters involved in the implementation of the algorithm (initial temperature, number of iterations, cooling schedule, and correlation threshold) are automatically determined by using a standard simulated annealing theory, reflectance databases, and sensor characteristics. Applied to a HyMap image with available ground truths, we verify that ACSA adequately recovers ground reflectance over clear land surfaces, and that the added spectral shape constraint does not introduce any spurious feature in the spectrum. The analysis of an AVIRIS image of Central Switzerland clearly shows the ability of the method to perform enhanced water vapor estimations over dark surfaces. Over a lake (reflectance equal to 0.02, low signal to noise ratio equal to about 6), ACSA retrieves unbiased water vapor amounts (2.86 cm/spl plusmn/0.36 cm) in agreement with in situ measurements (2.97 cm/spl plusmn/0.30 cm). This corresponds to a reduction of the standard deviation by a factor 3 in comparison with standard unconstrained procedures (1.95 cm/spl plusmn/1.08 cm). Similar results are obtained using a Hyperion image of DoE ARM SGP test site containing a very dark area of the land surface.

Correcting brightness gradients in hyperspectral data from urban areas

The analysis of airborne hyperspectral data is often affected by brightness gradients that are caused by directional surface reflectance. For line scanners these gradients occur in across-track direction and depend on the sensor's view-angle. They are greatest whenever the flight path is perpendicular to the sun–target–observer plane. A common way to correct these gradients is to normalize the reflectance factors to nadir view. This is especially complicated for data from spatially and spectrally heterogeneous urban areas and requires surface type specific models. This paper presents a class-wise empirical approach that is adapted to meet the needs of such images. Within this class-wise approach, empirical models are fit to the brightness gradients of spectrally pure pixels from classes after a spectral angle mapping (SAM). Compensation factors resulting from these models are then assigned to all pixels of the image, both in a discrete manner according the SAM and in a weighted manner based on information from the SAM rule images. The latter scheme is designed in consideration of the great number of mixed pixels. The method is tested on data from the Hyperspectral Mapper (HyMap) that was acquired over Berlin, Germany. It proves superior to a common global approach based on a thorough assessment using a second HyMap image as reference. The weighted assignment of compensation factors is adequate for the correction of areas that are characterized by mixed pixels. A remainder of the original brightness gradient cannot be found in the corrected image, which can then be used for any subsequent qualitative and quantitative analyses. Thus, the proposed method enables the comparison and composition of airborne data sets with similar recording conditions and does not require additional field or laboratory measurements.

Estimating nitrogen in mixed forests from HyMap data using band-depth analysis and branch-and-bound algorithm

Spatial information of nitrogen concentration (Nc) is of great interest because of its role in photosynthesis, ecosystem productivity and thus influences global cycling of carbon and oxygen. Imaging spectroscopy offers a means to assess this compound. Nc was estimated in mixed forests in Switzerland from airborne HyMap data using band-depth analysis. Instead of stepwise regression, an exhaustive search algorithm has been applied to select significant wavebands in order to build relationships between transformed reflectance and fieldmeasured Nc. This study confirms that partitioning data into vegetation types yielded in higher R2. R2 was largest for the homogeneous coniferous sample. A preclassification of the HyMap images is therefore recommended. The tested branch-and-bound algorithm performed well in selecting important known nitrogen absorption wavebands. A comparison with other subset selection methods is planned.

Lithologic mapping of the Mordor, NT, Australia ultramafic complex by using the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER)

Spectral measurements made in the Mordor Pound, NT, Australia study area using the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), in the laboratory and in situ show dominantly Al-OH and ferric-iron VNIR–SWIR absorption features in felsic rock spectra and ferrous-iron and Fe,Mg-OH features in the mafic–ultramafic rock spectra. ASTER ratio images, matched-filter, and spectral-angle mapper processing (SAM) were evaluated for mapping the lithologies. Matched-filter processing in which VNIR + SWIR image spectra were used for reference resulted in 4 felsic classes and 4 mafic–ultramafic classes based on Al-OH or Fe,Mg-OH absorption features and, in some, subtle reflectance differences related to differential weathering and vegetation. These results were similar to those obtained by match-filter analysis of HyMap data from a previous study, but the units were more clearly demarcated in the HyMap image. ASTER TIR spectral emittance data and laboratory emissivity measurements document a wide wavelength range of Si-O spectral features, which reflect the lithological diversity of the Mordor ultramafic complex and adjacent rocks. SAM processing of the spectral emittance data distinguished 2 classes representing the mafic–ultramafic rocks and 4 classes comprising the quartzose to intermediate composition rocks. Utilization of the complementary attributes of the spectral reflectance and spectral emittance data resulted in discrimination of 4 mafic–ultramafic categories; 3 categories of alluvial–colluvial deposits; and a significantly more completely mapped quartzite unit than could be accomplished by using either data set alone.

Spectrodirectional remote sensing for the improved estimation of biophysical and -chemical variables: two case studies

Over the past few years, significant advancements are made in the acquisition, processing, analysis and interpretation of quantitative directional and high spectral resolution data. In particular, the broader availability of air- and spaceborne directional imaging spectrometer data supports the estimation of biophysical and -chemical variables with unprecedented accuracy and in calibrated physical units. We describe in this paper two experiments that we carried out to demonstrate regional performance of spectral and directional-based retrieval approaches in vegetated areas. In the first case study, we focus on a mountain forest located in South-Eastern Switzerland representing a boreal forest like ecosystem. DAIS7915 imaging spectrometer data have been acquired with simultaneous ground measurements. We describe the soil–vegetation–atmosphere radiative transfer using a combination of the PROSPECT, GeoSAIL, and ATCOR models. In the second case study, we acquired spectrodirectional data on ground using a field goniometer in parallel with several HyMap imaging spectrometer overflights. Both cases demonstrate conditions for the estimation of biophysical and -chemical canopy properties with reduced uncertainties by respecting the full spectral coverage and directionality of the data. We conclude that the derived canopy variables represent the actual spatial distribution of properties as they occur in the landscape.

Comparison between field measurements and airborne visible and infrared mapping spectrometry (AVIRIS and HyMap) of the Ronda peridotite massif (south-west Spain)

The high spectral resolution of Airborne Visible Infrared Imaging Spectrometer (AVIRIS; 224 channels from 400 to 2455 nm) and HyMap (127 channels between 437 and 2485 nm) images is necessary to conduct geological analysis with remote petrological determinations of rock types or soils, or to determine vegetation groups. When airborne images and field spectra are well adjusted between each other, and when the vegetation does not interfere in the analysis, the spectral shape analysis (SSA) method represents an easy treatment to reveal a large amount of geological information. The method presented in the paper takes into account both wells and peaks of spectra resulting from a combination of absorption features and continuum shapes. It was conducted on the Ronda peridotite, in the south-west of Spain, which was imaged by AVIRIS in 1991 and by HyMap in 2000, and which was partially sampled in the field using a GER 3700 spectrometer in 1997, 2000 and 2001. In this study, the AVIRIS and HyMap images are processed to infer geological features using first photo-interpretation of colour composite images and then using the SSA method. This allows us to distinguish easily the peridotite massif from its surrounding rocks, to identify petrological variations inside the peridotite, including different varieties of serpentines, but also inside the surrounding rocks, particularly the carbonate-bearing ones which comprise marbles and dolomites.

Hyperspectral indices for characterizing upland peat composition

The erosion of blanket peat is a major environmental issue in the UK. Maps of erosion extent and peat composition, especially humification and moisture content, would aid our understanding of the erosion process and provide information for management decisions. HyMap images, acquired as part of the SAR and Hyperspectral Airborne Campaign (SHAC), were used to test candidate indices of peat composition for eroded blanket peat in the southern Pennines. Peat physical properties, including moisture content and degree of humification (measured as transmission), were derived in the laboratory and related to the remotely sensed data. Strong correlations were found between HyMap SWIR reflectance and transmission, but other peat physical properties were not significantly correlated. Spectral indices were calculated to express the depth of cellulose, lignin and water absorption features. Strong positive correlations were found between transmission and an adjusted cellulose absorption index (CAI), r 0.71, and the gradient of its shoulders between 2020 and 2200 nm, r 0.89. Other indices also performed well. Normalized indices performed better because they allowed for differences in brightness. Higher moisture content in poorly humified peats may have reinforced the effect of deeper ligno-celluloic absorptions, but further sampling is required to test this. The results suggest the potential for hyperspectral remote sensing to provide information on surface peat composition across large areas.