Wetland mapping from sparse annotations with satellite image time series and temporal-aware segment anything model

2026-01-16

AI Summary

This study addresses the challenge of accurately mapping wetlands under sparse point annotations, where existing methods struggle to capture strong seasonal and interannual dynamics and fail to effectively leverage temporal remote sensing information. To this end, we propose WetSAM, a novel framework that, for the first time, integrates a temporal-aware mechanism into the Segment Anything Model. WetSAM employs a dual-branch co-optimization architecture featuring hierarchical adapters, dynamic temporal aggregation, and bidirectional consistency regularization to disentangle semantic wetland features from phenological variations. High-quality pseudo-labels are generated through temporally constrained region growing. Evaluated across eight global regions (each approximately 5,000 km²), WetSAM achieves an average F1-score of 85.58%, significantly outperforming current approaches and enabling high-accuracy, scalable wetland mapping with minimal annotation cost.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Semi-Supervised Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate wetland mapping is essential for ecosystem monitoring, yet dense pixel-level annotation is prohibitively expensive and practical applications usually rely on sparse point labels, under which existing deep learning models perform poorly, while strong seasonal and inter-annual wetland dynamics further render single-date imagery inadequate and lead to significant mapping errors; although foundation models such as SAM show promising generalization from point prompts, they are inherently designed for static images and fail to model temporal information, resulting in fragmented masks in heterogeneous wetlands. To overcome these limitations, we propose WetSAM, a SAM-based framework that integrates satellite image time series for wetland mapping from sparse point supervision through a dual-branch design, where a temporally prompted branch extends SAM with hierarchical adapters and dynamic temporal aggregation to disentangle wetland characteristics from phenological variability, and a spatial branch employs a temporally constrained region-growing strategy to generate reliable dense pseudo-labels, while a bidirectional consistency regularization jointly optimizes both branches. Extensive experiments across eight global regions of approximately 5,000 km2 each demonstrate that WetSAM substantially outperforms state-of-the-art methods, achieving an average F1-score of 85.58%, and delivering accurate and structurally consistent wetland segmentation with minimal labeling effort, highlighting its strong generalization capability and potential for scalable, low-cost, high-resolution wetland mapping.

Problem

• wetland mapping
• sparse annotations
• satellite image time series
• temporal dynamics
• segment anything model

Innovation

• temporal-aware SAM
• satellite image time series
• sparse point supervision
• wetland mapping
• pseudo-label generation