A dual polarimetric method for detecting burnt areas: the May 2022 Stromboli Island (Italy) wildfire

In this study, multi-polarimetric Synthetic Aperture Radar (SAR) features extracted from Sentinel-1 SAR data are used to detect burnt area (BA) related to the wildfire that hit Stromboli Island, Italy, in May 2022. First, a new dual polarimetric (DP) change detector, based on the normalized difference of the covariance matrices collected at two different times, is used to detect the changes that occurred over the scene after the wildfire. Then, the BA map is extracted using a clustering algorithm based on the k-means method. Finally, to assess the performance of the proposed approach, the DP detector is compared with single-polarimetric ones. In contrast, optical images from Sentinel-2 are exploited as reference data for evaluating the accuracy of the DP method. Results obtained by processing a pair of C-band SAR images collected before and after the 2022 Stromboli wildfire confirm the soundness of the proposed approach in detecting the BA.