A Deep Learning-Based Approach for Mangrove Monitoring

TL;DR

The paper tackles the global mangrove monitoring problem by introducing MagSet-2, a new open dataset that fuses Global Mangrove Watch annotations with Sentinel-2 imagery and multi-zone stratification for robust evaluation. It systematically benchmarks six deep learning architectures across CNN, Transformer, and Mamba families, using a two-phase protocol to ensure fair comparison and scalability, with a consistent parameter budget and training regime. The Swin-UMamba Mamba model consistently delivers the strongest mangrove segmentation performance on both sampled and complete MagSet-2 data, outperforming CNN and Transformer baselines and highlighting convergence considerations for larger datasets. The work provides a valuable benchmark, methodological rigor, and practical implications for improved, globally scalable mangrove monitoring from optical satellite data.

Abstract

Mangroves are dynamic coastal ecosystems that are crucial to environmental health, economic stability, and climate resilience. The monitoring and preservation of mangroves are of global importance, with remote sensing technologies playing a pivotal role in these efforts. The integration of cutting-edge artificial intelligence with satellite data opens new avenues for ecological monitoring, potentially revolutionizing conservation strategies at a time when the protection of natural resources is more crucial than ever. The objective of this work is to provide a comprehensive evaluation of recent deep-learning models on the task of mangrove segmentation. We first introduce and make available a novel open-source dataset, MagSet-2, incorporating mangrove annotations from the Global Mangrove Watch and satellite images from Sentinel-2, from mangrove positions all over the world. We then benchmark three architectural groups, namely convolutional, transformer, and mamba models, using the created dataset. The experimental outcomes further validate the deep learning community's interest in the Mamba model, which surpasses other architectures in all metrics.

Figures (5)

• Figure 1: Mangrove Position (neon blue) and the different Mangrove Zones (green) Dataset based on the Global Mangrove Watch (GMW) v3.2020
• Figure 2: Sentinel-2 Spectral Display and Vegetation Analysis: Starting from the top left with the RGB bands, followed by the NIR band, Vegetation NIR, and SWIR band in sequence. On the bottom row, from left to right, we have the estimated NDVI, NDWI, NDMI indices, and the targeted Mangrove locations for predictive modeling.
• Figure 3: Comparative performance on sampled test of MagSet-2, using Training Set Loss (left), Test Set F1 Score (center), and Test Set Intersection over Union (IoU) (right). Each line represents a model: U-Net (neon blue), PAN (red), MANet (black), BEiT (green), SegFormer (yellow), and Swin-UMamba (dark blue) trained over 100 epochs. Lower loss values, higher F1 and IoU values indicate better performance. Swin-UMamba consistently shows superior performance over all metrics.
• Figure 4: Comparative visual segmentation results of mangrove areas. The first column shows the original satellite images, the second column depicts the ground truth segmentation, and the subsequent columns display the segmentation results from U-Net, PAN, MANet, BEiT, SegFormer, and Swin-UMamba models, resp.
• Figure 5: Comparative performance on complete MagSet-2, using Training Set Loss (left), Test Set F1 Score (center), and Test Set Intersection over Union (IoU) (right). Each line represents a model: MANet (black), SegFormer (yellow), and Swin-UMamba (dark blue) trained over 100 epochs. Lower loss values, higher F1 and IoU values indicate better performance. Again, Swin-UMamba consistently shows superior performance over all metrics.