George P. Petropoulos

A review of Ts/VI remote sensing based methods for the retrieval of land surface energy fluxes and soil surface moisture

Imagery from remote sensing systems, often combined with ancillary ground information, is able to provide repetitive, synoptic views of key parameters characterizing land surface interactions, including surface energy fluxes and surface soil moisture. Differing methodologies using a wide range of remote sensing data have been developed for this purpose. Approaches vary from purely empirical to more complex ones, including residual methods and those that have their basis in the biophysical properties characterizing a two-dimensional Ts/VI (surface temperature/ vegetation index) scatterplot domain derived from remote sensing observations. The present article aims to offer a comprehensive and systematic review of this latter group of methods, which differ in terms of the complexity and assumptions they entail as well as their requirement for field-based and other ancillary data. Prior to the review, the biophysical meanings and properties encapsulated in the Ts/VI feature space is elucidated, since these represent the building block upon which all the Ts/VI methods described herein are based. The overview of the Ts/VI methods is also very timely, as one such method is being scheduled in the operational retrieval of surface soil moisture content by the National Polar-orbiting Operational Environmental Satellite System (NPOESS), in a series of satellite platforms due to be launched in the next 12 years starting from 2016.

Support vector machines and object-based classification for obtaining land-use/cover cartography from Hyperion hyperspectral imagery

The Hyperion hyperspectral sensor has the highest spectral resolution, acquiring spectral information of Earths surface objects in 242 spectral bands at a spatial resolution of 30m. In this study, we evaluate the performance of the Hyperion sensor in conjunction with the two different classification algorithms for delineating land use/cover in a typical Mediterranean setting. The algorithms include pixel-based support vector machines (SVMs) and the object-based classification algorithm. Validation of the derived land-use/cover maps from the above two algorithms was performed through error matrix statistics using the validation points from the very high resolution QuickBird imagery. Results suggested both classifiers as highly useful in mapping land use/cover in the study region with the object-based approach slightly outperforming the SVMs classification by overall higher classification accuracy and Kappa statistics. Results from the statistical significance testing using McNemars chi-square test confirmed the superiority of the object-oriented approach compared to SVM. The relative strengths and weaknesses of the two classification algorithms for land-use/cover mapping studies are highlighted. Overall, our results underline the potential of hyperspectral remote sensing data together with an object-based classification approach for mapping land use/cover in the Mediterranean regions.

Burnt Area Delineation from a uni-temporal perspective based on Landsat TM imagery classification using Support Vector Machines

Abstract Information on burnt area is of critical importance in many applications as for example in assessing the disturbance of natural ecosystems due to a fire or in proving important information to policy makers on the land cover changes for establishing restoration policies of fire-affected regions. Such information is commonly obtained through remote sensing image thematic classification and a wide range of classifiers have been suggested for this purpose. The objective of the present study has been to investigate the use of Support Vector Machines (SVMs) classifier combined with multispectral Landsat TM image for obtaining burnt area mapping. As a case study a typical Mediterranean landscape in Greece was used, in which occurred one of the most devastating fires during the summer of 2007. Accuracy assessment was based on the classification overall statistical accuracy results and also on comparisons of the derived burnt area estimates versus validated estimates from the Risk-EOS Burnt Scar Mapping service. Results from the implementation of the SVM using diverse kernel functions showed an average overall classification accuracy of 95.87% and a mean kappa coefficient of 0.948, with the burnt area class always clearly separable from all the other classes used in the classification scheme. Total burnt area estimate computed from the SVM was also in close agreement with that from Risk-EOS (mean difference of less than 1%). Analysis also indicated that, at least for the studied here fire, the inclusion of the two middle infrared spectral bands TM5 and TM7 of TM sensor as well as the selection of the kernel function in SVM implementation have a negligible effect in both the overall classification performance and in the delineation of total burnt area. Overall, results exemplified the appropriateness of the spatial and spectral resolution of the Landsat TM imagery combined with the SVM in obtaining rapid and cost-effective post-fire analysis. This is of considerable scientific and practical value, given the present open access to the archived and new observations from this satellite radiometer globally.

A Comparison of Spectral Angle Mapper and Artificial Neural Network Classifiers Combined with Landsat TM Imagery Analysis for Obtaining Burnt Area Mapping

Satellite remote sensing, with its unique synoptic coverage capabilities, can provide accurate and immediately valuable information on fire analysis and post-fire assessment, including estimation of burnt areas. In this study the potential for burnt area mapping of the combined use of Artificial Neural Network (ANN) and Spectral Angle Mapper (SAM) classifiers with Landsat TM satellite imagery was evaluated in a Mediterranean setting. As a case study one of the most catastrophic forest fires, which occurred near the capital of Greece during the summer of 2007, was used. The accuracy of the two algorithms in delineating the burnt area from the Landsat TM imagery, acquired shortly after the fire suppression, was determined by the classification accuracy results of the produced thematic maps. In addition, the derived burnt area estimates from the two classifiers were compared with independent estimates available for the study region, obtained from the analysis of higher spatial resolution satellite data. In terms of the overall classification accuracy, ANN outperformed (overall accuracy 90.29%, Kappa coefficient 0.878) the SAM classifier (overall accuracy 83.82%, Kappa coefficient 0.795). Total burnt area estimates from the two classifiers were found also to be in close agreement with the other available estimates for the study region, with a mean absolute percentage difference of ∼1% for ANN and ∼6.5% for SAM. The study demonstrates the potential of the examined here algorithms in detecting burnt areas in a typical Mediterranean setting.

Discrimination of common Mediterranean plant species using field spectroradiometry

Field spectroradiometry of land surface objects supports remote sensing analysis, facilitates the discrimination of vegetation species, and enhances the mapping efficiency. Especially in the Mediterranean, spectral discrimination of common vegetation types, such as phrygana and maquis species, remains a challenge. Both phrygana and maquis may be used as a direct indicator for grazing management, fire history and severity, and the state of the wider ecosystem equilibrium. This study aims to investigate the capability of field spectroradiometry supporting remote sensing analysis of the land cover of a characteristic Mediterranean area. Five common Mediterranean maquis and phrygana species were examined. Spectra acquisition was performed during an intensive field campaign deployed in spring 2010, supported by a novel platform MUFSPEM@MED (Mobile Unit for Field SPEctral Measurements at the MEDiterranean) for high canopy measurements. Parametric and non-parametric statistical tests have been applied to the continuum-removed reflectance of the species in the visible to shortwave infrared spectral range. Interpretation of the results indicated distinct discrimination between the studied species at specific spectral regions. Statistically significant wavelengths were principally found in both the visible and the near infrared regions of the electromagnetic spectrum. Spectral bands in the shortwave infrared demonstrated significant discrimination features for the examined species adapted to Mediterranean drought. All in all, results confirmed the prospect for a more accurate mapping of the species spatial distribution using remote sensing imagery coupled with in situ spectral information.

Land cover mapping with emphasis to burnt area delineation using co-orbital ALI and Landsat TM imagery

Abstract In this study, the potential of EO-1 Advanced Land Imager (ALI) radiometer for land cover and especially burnt area mapping from a single image analysis is investigated. Co-orbital imagery from the Landsat Thematic Mapper (TM) was also utilised for comparison purposes. Both images were acquired shortly after the suppression of a fire occurred during the summer of 2009 North-East of Athens, the capital of Greece. The Maximum Likelihood (ML), Artificial Neural Networks (ANNs) and Support Vector Machines (SVMs) classifiers were parameterised and subsequently applied to the acquired satellite datasets. Evaluation of the land use/cover mapping accuracy was based on the error matrix statistics. Also, the McNemar test was used to evaluate the statistical significance of the differences between the approaches tested. Derived burnt area estimates were validated against the operationally deployed Services and Applications For Emergency Response (SAFER) Burnt Scar Mapping service. All classifiers applied to either ALI or TM imagery proved flexible enough to map land cover and also to extract the burnt area from other land surface types. The highest total classification accuracy and burnt area detection capability was returned from the application of SVMs to ALI data. This was due to the SVMs ability to identify an optimal separating hyperplane for best classes’ separation that was able to better utilise ALIs advanced technological characteristics in comparison to those of TM sensor. This study is to our knowledge the first of its kind, effectively demonstrating the benefits of the combined application of SVMs to ALI data further implying that ALI technology may prove highly valuable in mapping burnt areas and land use/cover if it is incorporated into the development of Landsat 8 mission, planned to be launched in the coming years.

Flooding extent cartography with Landsat TM imagery and regularized kernel Fisher's discriminant analysis

In this paper the combined use of the regularized kernel Fishers discriminant analysis classifier (kFDA) with Landsat TM multispectral imagery is explored for flooded area cartography purposes. This classifier provides an efficient and regularized solution for the non-linear delineation of pixels corresponding to flooded surface. The flood mapping issue is tackled from both uni- and multi-temporal classification perspectives: the former recasts the problem as a classical image classification procedure – with class water as target; the latter considers the extraction of flooded area as a change detection problem – in which only the non-permanent standing water is considered as flood. As a case study is used a Landsat TM dataset of the James River in South Dakota (USA), a region that experienced a heterogeneous flooding in spring 2011. Findings from our analysis suggest that precisely delineating the exceeding water extent requires a non-linear classifier applied in a multi-temporal setting.

An Overview of the Use of the SimSphere Soil Vegetation Atmosphere Transfer (SVAT) Model for the Study of Land-Atmosphere Interactions

Soil Vegetation Atmosphere Transfer (SVAT) models consist of deterministic mathematical representations of the physical processes involved between the land surface and the atmosphere and of their interactions, at time-steps acceptable for the study of land surface processes. The present article provides a comprehensive and systematic review of one such SVAT model suitable for use in mesoscale or boundary layer studies, originally developed by [1]. This model, which has evolved significantly both architecturally and functionally since its foundation, has been widely applied in over thirty interdisciplinary science investigations, and it is currently used as a learning resource for students in a number of educational institutes globally. The present review is also regarded as very timely, since a variation of a method using this specific SVAT model along with satellite observations is currently being considered in a scheme being developed for the operational retrieval of soil surface moisture by the US National Polar-orbiting Operational Environmental Satellite System (NPOESS), in a series of satellites that are due to be launched from 2016 onwards.

Remote sensing and GIS analysis for mapping spatio-temporal changes of erosion and deposition of two Mediterranean river deltas: The case of the Axios and Aliakmonas rivers, Greece

Abstract Wetlands are among Earths most dynamic, diverse and varied habitats as the balance between land and water surfaces provide shelter to a unique mixture of plant and animal species. This study explores the changes in two Mediterranean wetland delta environments formed by the Axios and Aliakmonas rivers located in Greece, over a 25-year period (1984–2009). Direct photo-interpretation of four Landsat TM images acquired during the study period was performed. Furthermore, a sophisticated, semi-automatic image classification method based on support vector machines (SVMs) was developed to streamline the mapping process. Deposition and erosion magnitudes at different temporal scales during the study period were quantified using both approaches based on coastline surface area changes. Analysis using both methods was conducted in a geographical information systems (GIS) environment. Direct photo-interpretation, which formed our reference dataset, showed noticeable changes in the coastline deltas of both study areas, with erosion occurring mostly in the earlier periods (1990–2003) in both river deltas followed by deposition in more recent years (2003–2009), but at different magnitudes. Spatial patterns of coastline changes predicted from the SVMs showed similar trends. In absolute terms SVMs predictions of sediment erosion and deposition in the studied area were different in the order of 5–20% in comparison to photo-interpretation, evidencing the potential capability of this method in coastline changes monitoring. One of the main contributions of our work lies to the use of the SVMs classifier in coastal mapping of changes, since to our knowledge use of this technique has been under-explored in this application domain. Furthermore, this study provides important contribution to the understanding of Mediterranean river delta dynamics and their behaviours, and corroborates the usefulness of EO technology and GIS as an effective tool in policy decision making and successful landscape management. The latter is of considerable scientific and practical value to the wider community of interested users, given the continued open access to observations from this satellite radiometer globally.

Change detection of surface mining activity and reclamation based on a machine learning approach of multi-temporal Landsat TM imagery

Being able to quantify land cover changes due to mining and reclamation at a watershed scale is of critical importance in managing and assessing their potential impacts to the Earth system. In this study, a remote sensing-based methodology is proposed for quantifying the impact of surface mining activity and reclamation from a watershed to local scale. The method is based on a Support Vector Machines (SVMs) classifier combined with multi-temporal change detection of Landsat TM imagery. The performance of the technique was evaluated at selected open mining sites located in the island of Milos in Greece. Assessment of the mining impact in the studied areas was based on the confusion matrix statistics, supported by co-orbital QuickBird-2 very high spatial resolution imagery. Overall classification accuracy of the thematic land cover maps produced was reported over 90%. Our analysis showed expansion of mining activity throughout the whole 23-year study period, while the transition of mining areas to soil and vegetation was evident in varying rates. Our results evidenced the ability of the method under investigation in deriving highly and accurate land cover change maps, able to identify the mining areas as well as those in which excavation was replaced by natural vegetation. All in all, the proposed technique showed considerable promise towards the support of a sustainable environmental development and prudent resource management.