Narrow River Extraction From SAR Images Using Exogenous Information

Monitoring of rivers is major scientific and societal importance due to the crucial resource they provide human activities threats caused by flood events. Rapid revisit synthetic aperture radar (SAR) sensors such as Sentinel-1 or future surface water ocean topography (SWOT) mission are indispensable tools achieve all-weather monitoring bodies at global scale. Unfortunately, spatial resolution these sensors, extraction narrow extremely difficult without resorting exogenous knowledge. This article introduces an innovative river segmentation method from SAR images using a priori databases widths Landsat (GRWL). First, recently proposed linear structure detector used produce map likely line structures. Then, limited number nodes along prior centerline extracted database reconstruct full detection map. Finally, conditional random field approach delineate accurately extent around its centerline. The has been tested on several simulated SWOT data. Both visual qualitative evaluations demonstrate efficiency.