MagicBathyNet: A Multimodal Remote Sensing Dataset for Bathymetry Prediction and Pixel-based Classification in Shallow Waters

TL;DR

MagicBathyNet introduces a comprehensive open dataset combining multimodal remote sensing images, bathymetry, and seabed annotations to advance deep learning methods for shallow water mapping.

Contribution

It provides the first open benchmark dataset for deep learning-based bathymetry and seabed classification in shallow waters.

Findings

Benchmark results for state-of-the-art methods on the dataset.

Open access to data, models, and code for reproducibility.

Facilitates further research in shallow water remote sensing.

Abstract

Accurate, detailed, and high-frequent bathymetry, coupled with complex semantic content, is crucial for the undermapped shallow seabed areas facing intense climatological and anthropogenic pressures. Current methods exploiting remote sensing images to derive bathymetry or seabed classes mainly exploit non-open data. This lack of openly accessible benchmark archives prevents the wider use of deep learning methods in such applications. To address this issue, in this paper we present the MagicBathyNet, which is a benchmark dataset made up of image patches of Sentinel2, SPOT-6 and aerial imagery, bathymetry in raster format and annotations of seabed classes. MagicBathyNet is then exploited to benchmark state-of-the-art methods in learning-based bathymetry and pixel-based classification. Dataset, pre-trained weights, and code are publicly available at www.magicbathy.eu/magicbathynet.html.