Fusion and atmospheric correction techniques on multitemporal and multisensor satellite data for the detection of burnt areas in Western Peloponnese, Greece

Abstract

The objective of this study was to process multitemporal and multisensor satellite data for the detection of burnt areas in Western Peloponnese, Greece. The broader region combines at the same time the characteristics of an urban, a coastal and a rural area. In 1986, 1998 and in 2000 three big fires have burnt more than 500.000.000 m² of forest and rural land accordingly to the local authorities. In order to detect the vegetation changes for the period 1984-2001 we used the following multitemporal and multisensor satellite images in which we applied different vegetation indexes. A Landsat 5 TM cloud free subscene, acquired on July 27 1984 and on September 18 1986, Two KFA-1000 images of September 1986, A Landsat 7 ETM cloud free subscene, acquired on July 28 1999, Four Terra Aster cloud free scenes acquired on August 31 2000 and on August 18 2001. As the images have been acquired from different sensors and at different dates we used absolute atmospheric correction algorithms in order to reduce the phenomena of atmospheric attenuation. We fused multispectral ETM data with panchromatic data as well as TM multispectral data of 1984 & 1986 with high-resolution data of the Russian camera KFA-1000. All the fused images have been resampled in 15 meters resolution in order to compare them with Aster Vnir data (that have 15m resolution). The local authorities have mapped the burnt areas using traditional methods. We used the produced maps in order to check the results of the use of Vegetation Indexes with the above satellite data for burnt areas detection. All the indexes gave good results in the detection of burnt areas. SAVI and NDVI gave the most precise results. We produced thematic maps of the burnt areas. The general conclusion is that we can use multitemporal and multisensor satellite data with the vegetation indexes for the mapping of burnt areas and the vegetation monitoring. Atmospheric correction and data fusion techniques should be used in order to make the multisensor and multitemporal satellite data comparable.