부산대학교 도서관

Synthetic aperture radar (SAR) acquisitions are particularly useful to produce flood maps thanks to their all-weather and day-night capabilities. However, repetition intervals of radar instruments are in the order of several days for routine operations, reaching daily or higher frequencies only in tasked conditions. Therefore, to follow flood dynamics, images acquired by different sensors at different times may be beneficial. In the present work, multi-temporal SAR intensity, InSAR coherence and optical data are considered to describe a flood event occurred in the Basilicata region (southern Italy) on December 2013. In this case study, optical data have a twofold role: they allow to follow the flood dynamics (because SAR and optical data have been acquired in different dates during the inundation event), and they add information concerning the land cover of the analyzed area. The data fusion approach is based on Bayesian Networks (BNs). It is shown that the synergetic use of different information layers can help detect more precisely the areas affected by the flood, reducing false alarms and missed identifications which may affect algorithms based on data from a single source. The produced flood maps are compared to reference maps, independently obtained; the comparison indicates that the proposed methodology is able to reliably follow the temporal evolution of the phenomenon, assigning high probability to areas most likely to be flooded, reaching accuracies of up to 89%.