FusionAttNet framework for hierarchical attention driven Sentinel 1 and Sentinel 2 fusion for semi arid land cover classification in Far North Cameroon

Accurate land cover mapping in semi-arid regions remains challenging due to spectral homogeneity of sparse vegetation, seasonal variability, and persistent cloud cover. In this study, we propose a FusionAttNet, a novel deep learning framework integrating Sentinel-1 SAR and Sentinel-2 optical data through modality-aware hierarchical attention for land cover classification case study of Far North Cameroon. Our approach employs parallel sensor-specific processing streams (extracting cross-modal feature synergies dynamically via deep attention layers) and a three-tiered attention mechanism (pixel-patch-landscape) to resolve spatial ambiguities in semi-arid landscapes. Enhanced by modality-aware augmentation and focal loss with label smoothing, FusionAttNet achieved 96.97% overall accuracy and a 0.9681 weighted F1-score on a single held-out spatial test split (98.29% ± 2.10% mean accuracy across 5-fold repeated spatial cross-validation) on a multi-seasonal dataset over a 12-month temporal window spanning the year 2021, outperforming a naive feature-stacking baseline (63.64%) and single-modality optical-only (81.82%) and SAR-only (93.94%) variants under an identical, spatially block-partitioned, leakage-controlled evaluation protocol. Key innovations include: (1) landscape-level attention capturing phenological transitions in Sahelian ecotones (2), cross-modal indices mitigating cloud-induced optical data gaps, and (3) SAR-optimized augmentation preserving backscatter textures. Results demonstrate a 33.3%-point absolute accuracy gain of the full hierarchical attention architecture over the naive feature-stacking baseline, supporting the value of multi-scale attention for semi-arid land cover monitoring.