Overcoming spectral and tidal challenges in mangrove mapping with a multi-modal deep learning network fusing phenological and texture features

Key Points

• To develop a multi-modal deep learning framework for accurate mangrove mapping, addressing spectral and tidal challenges.
• Introduced a novel multi-modal deep learning framework integrating phenological and texture features.
• Utilized attention mechanisms for adaptive feature fusion.
• Tested the framework across three different coastal sites in China.
• Achieved superior mapping accuracy across the tested sites.
• Demonstrated strong performance even with limited training samples.