Combining ASTER multispectral imagery analysis and support vector machines for rapid and cost-effective post-fire assessment: a case study from the Greek wildland fires of 2007

Abstract

Abstract. Remote sensing is increasingly being used as a cost-effective and practical solution for the rapid evaluation of impacts from wildland fires. The present study investigates the use of the support vector machine (SVM) classification method with multispectral data from the Advanced Spectral Emission and Reflection Radiometer (ASTER) for obtaining a rapid and cost effective post-fire assessment in a Mediterranean setting. A further objective is to perform a detailed intercomparison of available burnt area datasets for one of the most catastrophic forest fire events that occurred near the Greek capital during the summer of 2007. For this purpose, two ASTER scenes were acquired, one before and one closely after the fire episode. Cartography of the burnt area was obtained by classifying each multiband ASTER image into a number of discrete classes using the SVM classifier supported by land use/cover information from the CORINE 2000 land nomenclature. Overall verification of the derived thematic maps based on the classification statistics yielded results with a mean overall accuracy of 94.6% and a mean Kappa coefficient of 0.93. In addition, the burnt area estimate derived from the post-fire ASTER image was found to have an average difference of 9.63% from those reported by other operationally-offered burnt area datasets available for the test region.

Assessing the influence of atmospheric and topographic correction and inclusion of SWIR bands in burned scars detection from high-resolution EO imagery: a case study using ASTER

“…Furthermore, in a study over a site in Greece, Koutsias et al (2012) reported an increase in the reflectance of burned surfaces observed in the middle-infrared region (2.08-2.35 lm or band 7) of the Landsat TM sensor (which corresponds to ASTER bands 5, 6, 7 and 8) because of changes in water content. The improving incidence of ASTER SWIR band inclusion to the classification process was also reported in another study by (Petropoulos et al 2010a), in which the authors concluded that the benefit from advanced land imager (ALI) sensor use in burned area extraction appears to be partially due to its higher number of spectral SWIR bands. Findings presented herein are also in close agreement to those reported by others who have explored the use of either the ML or the SVMs classifier in burn scar extraction from multispectral imagery.…”

“…Findings presented herein are also in close agreement to those reported by others who have explored the use of either the ML or the SVMs classifier in burn scar extraction from multispectral imagery. For example, Petropoulos et al (2010a) combined ASTER and Landsat TM data, respectively, with the SVMs in mapping the extent of the burned area of the same fire event. The results presented herein outperformed the performance of the artificial neural network (ANN) and spectral angle mapper (SAM) classifiers when combined with the Landsat imagery reported to that study by the authors.…”

İran Tahran şehri güney bölgesinde kent alanlarının sınıflandırılmasında SVM yöntemi ile hiperspektral görüntü ve tekstür bilgilerinin birlikte kullanılması

“…The reason behind this selection is its better performance in previous studies (Colgan et al, 2012;Huang et al, 2002;Petropoulos et al, 2010). Moreover, this kernel needs less parameter for tuning compared to the other ones (2 parameters (cost and sigma) compared to 4 parameters in polynomial and 3 in sigmoid kernels).…”

An intercomparison of burnt area estimates derived from key operational products: the Greek wildland fires of 2005–2007

“…Different types of remote sensing data have been exploited for more than 20 yr in performing various fire analysis investigations, including mapping the extent of burnt areas (e.g. Trigg and Roy, 2007;Roy and Boschetti, 2009;Petropoulos et al, 2010).…”