Detectarea și localizarea incendiilor de pădure și urmărirea evoluției acestora pe baza imaginilor satelitare de înaltă rezoluție temporală.

Early detection of vegetation fires and forest fires in particular is of great importance and helps in fire risk management so that alerting and intervention can take place quickly and the necessary measures can be taken in time to protect the population and sensitive infrastructure. Remote sensing is successfully used in this respect, especially with the launch into orbit of Earth observation sensors with high temporal resolution such as MODIS or VIIRS. The present research was carried out in a forested area within the administrative borders of Jiana and Pătulele communes in Mehedinți County (South-West Romania). The year 2021 was unusual in terms of vegetation fires (agricultural and forest fires), these occurring with a remarkable frequency, following a period of drought combined with local practices of burning stubble to clear the agricultural parcels. Thus, in the period 01.07-30.09.2021, 852 fires in the county started from stubble and dry vegetation on uncultivated land, of which 35 fires occurred in the forest, 24 surface fires in mature stands and 11 fires burned forest plantations. To detect, locate the fires and track their evolution, Sentinel 2 satellite images (which have spatial resolutions of 10 and 20 m and a revisit time of 5 days) were used, freely available on the Copernicus platform, as well as two images recorded by the Doves satellite constellation. It was confirmed that fire hotspots derived from high temporal resolution satellite data (MODIS and VIIRS) and provided by EFFIS correctly captured the forest fires analyzed in the study area. At the same time, by combining high spatial resolution satellite data (Sentinel 2 and Doves), the almost daily evolution of fires can be faithfully and precisely reconstructed, on the condition that the sky is not covered by clouds at the time of image recording. [ABSTRACT FROM AUTHOR]