MARIDA: Marine Debris Archive for Sentinel-2 Semantic Segmentation

Provider
Kikaki, Kakogeorgiou, Mikeli, Raitsos and Karantzalos (National Technical University of Athens)
License
CC-BY-4.0 (source)
Format
JSON (in sampled files)
Stored at
Azure Blob Storage, https://radiantearth.blob.core.windows.net/mlhub/marida
Access
Free and public. No account with the data provider is needed.
DOI
10.1371/journal.pone.0262247
Category
Earth Observation & Geospatial
Source
https://doi.org/10.1371/journal.pone.0262247

About MARIDA: Marine Debris Archive for Sentinel-2 Semantic Segmentation

A benchmark for detecting floating marine debris and plastic litter in Sentinel-2 multispectral imagery, from the National Technical University of Athens. Annotators drew polygons on Sentinel-2 scenes over coastal waters at sites with reported marine debris events in several countries. Beyond marine debris the 15 classes cover dense and sparse Sargassum macroalgae, natural organic material, ships, clouds, marine water, sediment-laden and turbid water, shallow water, foam, waves, cloud shadows, wakes and mixed water, with a confidence level per annotation, so the task is weakly supervised pixel-level semantic segmentation. Measured 2026-11-15: 4,536 files (4.7 GB): 256 x 256 pixel 11-band GeoTIFF patches with class and confidence masks under patches/ /, the source polygons as shapefiles and GeoJSON, a labels_mapping.txt and train/val/test splits. Load with rasterio; the authors provide Random Forest and U-Net baselines. Released under CC BY 4.0; cite Kikaki et al. (2022), PLOS ONE 17(1): e0262247.

How to access MARIDA: Marine Debris Archive for Sentinel-2 Semantic Segmentation in Python

import adlfs
fs = adlfs.AzureBlobFileSystem(account_name='radiantearth', anon=True)
fs.ls('mlhub/marida/')[:10]

Lists the files with anonymous access. Runs on any machine, with no cloud account.

Data structure

File types in sampled files: JSON

Files

Example file: .source/metadata.json

Analyze this dataset without downloading it

On Scigantic, MARIDA: Marine Debris Archive for Sentinel-2 Semantic Segmentation opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Earth Observation & Geospatial datasets