The work reports some experiments in classifying and mapping the areas which were affected by wildfires in centre-eastern Sicily in June 2007. The study exploited moderate resolution images acquired by ASTER sensor (Advanced Spaceborne Thermal Emission and Reflection Radiometer) and compared four types of classification, all of which based on spectral indices specially developed for this kind of application: BAI (Burned Area Index), NBR (Normalized Burned Ratio) and MIRBI (Mid-Infrared Bispectral Index). After assessing each index by means of direct interpretation and set-up of appropriate threshold values, the method was optimized by their multiple thresholding. This approach was used for both single images and multi-temporal series. For each index some confusion errors with different kinds of surface (e.g. farmed areas, urban areas, water bodies) were reported. The simultaneous use of the indices made the classification more accurate with fewer commission errors.
Burned areas mapping by multispectral imagery: a case study in Sicily, Summer 2007 / Conte P.; Bitelli G.. - STAMPA. - (2009), pp. 277-284. (Intervento presentato al convegno 7th International Workshop of the EARSeL Special Interest Group (SIG) on Forest Fires tenutosi a Matera nel 2-5 September 2009).
Burned areas mapping by multispectral imagery: a case study in Sicily, Summer 2007
CONTE, PAOLO;BITELLI, GABRIELE
2009
Abstract
The work reports some experiments in classifying and mapping the areas which were affected by wildfires in centre-eastern Sicily in June 2007. The study exploited moderate resolution images acquired by ASTER sensor (Advanced Spaceborne Thermal Emission and Reflection Radiometer) and compared four types of classification, all of which based on spectral indices specially developed for this kind of application: BAI (Burned Area Index), NBR (Normalized Burned Ratio) and MIRBI (Mid-Infrared Bispectral Index). After assessing each index by means of direct interpretation and set-up of appropriate threshold values, the method was optimized by their multiple thresholding. This approach was used for both single images and multi-temporal series. For each index some confusion errors with different kinds of surface (e.g. farmed areas, urban areas, water bodies) were reported. The simultaneous use of the indices made the classification more accurate with fewer commission errors.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
