Synergistic use of Sentinel-1 and Sentinel-2 for improved LULC mapping with special reference to bad land class: a case study for Yamuna River floodplain, India

High accuracy land use/land cover (LULC) mapping of Yamuna Chambal ravines for reclamation and conservation of these degraded/badlands is indispensable. Integration of freely available SAR datasets along with medium to high resolution optical data is one of the best approach for high accuracy LULC mapping. The objective of the presented study is to evaluate the fusion technique for Sentinel-1 SAR data and Sentinel-2 optical data for high accuracy LULC mapping in order to assess the area occupied by these negative landforms i.e., ravines. The VH-polarization fused image with Sentinel-2 optical data gives the best accuracy of 85% followed by VV-polarization fused image with same datasets of 84% accuracy whereas Sentinel-1 and Sentinel-2 provides the accuracy of 60 and 80%, respectively. The prepared LULC maps shown that bad land (Ravine class) occupied an area in the range of 600–700 km2 using combinations of different datasets as the wastelands in the area required immediate reclamation and conservation measures to be adopted. However, asymptotic performance of fusion technique for SAR and optical data further elucidate its successful implementation and dominancy over other datasets for improved LULC mapping.