Accurate thermal anomalies identification
Thermal anomalies on ground, as seen by satellite platforms, can be generated by surface warming like Volcanoes eruptions, or by the presence of fires. The latter is the case analyzed in the framework of the KLAUS Project, where the Italian State Forestry Corps (Corpo Forestale dello Stato, CFS) was involved as end-user of a fully automatic satellite-based thermal anomalies detection system.
Using the SOMAFID system, two web services have been developed to permit the end-user to run:
- a fully automatic subscription service (Fire Detection and one-day Vegetation Change subscription Service) that provides, every day, an updated map of identified fires
- an on-demand service (Fire Detection from MODIS Catalogue Image) that permits the user to select a satellite image starting from 2000 (global coverage) on which the SOMAFID system is applied to retrieve thermal anomalies maps.
The SOMAFID system has been calibrated and validated on data provided by CFS over the Calabria Region, Italy, on a dataset of all the fires identified from July to September 2009. Fires have been divided by their burned areas in three main classes: between 10ha to 50ha, between 50ha and 100ha, larger than 100ha.
The validation results showed that SOMAFID is able to detect 9 fires out of 10 detectable, over different types of background and fire types.