Answers are grounded in this paper and cite their sources.
This article presents the multitemporal analysis of a Synthetic Aperture Radar (SAR) image series of an area affected by several fires during the years 2000 and 2001 in Central Portugal. An initial study was carried out to determine the best conditions to acquire optimal SAR imagery. Burnt areas were classified using neural network techniques. The neural network classification presented an overall accuracy of 92.11% using an entire European Remote Sensing (ERS) SAR time series, whereas an accuracy of 89.90% was achieved when using a subset of scenes selected through principal components analysis (PCA). Finally, the burnt area maps obtained were compared to estimates from the Portuguese Forest Services (DGF, Direc@@o Geral das Florestas) and the European Forest Fires Damage Assessment System (EFFDAS). This comparison showed that the SAR-based methodology provided higher accuracy than the other systems. The sensitivity of the SAR data to determine burnt severity allowed the discrimination of partially burnt areas and isles within the perimeter of the fire. These results show that the analysis of the temporal variation in the ERS-SAR backscatter coefficient permits the extraction of accurate and reliable information on the position and extent of burnt areas in Mediterranean forest environments.
No takes yet. Share an insight, caveat, or question.
Gimeno et al. (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: