TerraMesh Masks

TerraMesh-Masks-Eval TerraMesh-Masks-Eval is a human-verified benchmark dataset to evaluate open-vocabulary segmentation models on satellite imagery. This dataset provides binary segmentation masks with captions togehther with input samples from TerraMesh. We also provide a training dataset, called TerraMesh-Masks. Examples from the evaluation subset: Usage Download the data loading code from GitHub and install requirements with pip install -r… See the full description on the dataset page: https://huggingface.co/datasets/ibm-esa-geospatial/TerraMesh-Masks-Eval.

TerraMesh A planetary‑scale, multimodal analysis‑ready dataset for Earth‑Observation foundation models: TerraMesh merges data from Sentinel‑1 SAR, Sentinel‑2 optical, Copernicus DEM, NDVI, and land‑cover sources into more than 9 million co‑registered patches ready for large‑scale representation learning. You find more information about the data sampling and preprocessing in our paper: TerraMesh: A Planetary Mosaic of Multimodal Earth Observation Data. Samples from the TerraMesh… See the full description on the dataset page: https://huggingface.co/datasets/ibm-esa-geospatial/TerraMesh.

Dataset Card for Multi-Temporal Crop Classification Dataset Summary This dataset contains temporal Harmonized Landsat-Sentinel imagery of diverse land cover and crop type classes across the Contiguous United States for the year 2022. The target labels are derived from USDA's Crop Data Layer (CDL). It's primary purpose is for training segmentation geospatial machine learning models. Dataset Structure TIFF Files Each tiff file covers a… See the full description on the dataset page: https://huggingface.co/datasets/ibm-nasa-geospatial/multi-temporal-crop-classification.

Overview: OPERA Level-3 Dynamic Surface Water Extent (DSWx) product from Harmonized Landsat-8/9 and Sentinel-2A/B/C (HLS), processed and developed by OPERA (JPL/USGS). Values for B01 WTR are a masked interpreted water classification layer. This represents pixel-wise classification into one of three water classes (not water, open water, and partial surface water), HLS snow/ice, HLS cloud/cloud shadow and adjacent to cloud/cloud shadow, or no data classes.   This dataset contains Level-3 Dynamic OPERA surface water extent product version 1. The data are validated surface water extent observations beginning April 2023. Known issues and caveats on usage are described in the Product Documentation. The input dataset for generating each product is the Harmonized Landsat-8 and Sentinel-2A/B/C (HLS) product version 2.0. HLS products provide surface reflectance (SR) data from the Operational Land Imager (OLI) aboard the Landsat 8 satellite and the MultiSpectral Instrument (MSI) aboard the Sentinel-2A/B/C satellite. This EGIS service offers access to the water classification data layer.   More information about the dataset is available at https://www.earthdata.nasa.gov/data/catalog/pocloud-opera-l3-dswx-hls-v1-1.0 OPERA was informed by the Satellite Needs Working Group (SNWG), an interagency effort of the U.S. Government dedicated to identifying and addressing Earth observation needs across U.S. civilian federal agencies. Specifications: Temporal coverage: January 1, 2023 – Present Version: 1.0 Spatial Resolution: 3,660 rows and 3,660 columns at 30-m pixel spacing Processing Level: 3 Spatial coverage/reference: Near Global (N: 84 S: -84 E: 180 W: -180) coverage distributed over a projected map coordinate using the Universal Transverse Mercator (UTM) projection (109.8 km × 109.8 km) (EPSG:3857). Sensor: The input dataset for generating each product is the Harmonized Landsat-8 and Sentinel-2A/B/C (HLS) product version 2.0. HLS products provide surface reflectance (SR) data from the Operational Land Imager (OLI) aboard the Landsat 8 satellite and the MultiSpectral Instrument (MSI) aboard the Sentinel-2A/B/C satellite. Time interval: Daily Time standard: Coordinated Universal Time (UTC) Update frequency: Updated daily (based on ~3-day satellite acquisition latency)   Fields  objectid: The unique, internal database ID for the specific image. name: The granule/filename of the scene. minps, maxps, lowps, highps: "Pixel Spacing (ps)" metrics used by the ArcGIS server's backend to decide at what zoom levels the image should be visible on a map. category: A backend GIS tag indicating the type of raster (1 for primary data/source raster & 2 for overviews). centerx & centery: The precise center point (Longitude and Latitude) of the image footprint. shape: The geographic footprint of the image. processingdate: The date and time the image was uploaded and processed into the image service. tileid: The specific tile identifier of the product, which is also the tile ID of the input HLS product in the Military Grid Reference System (MGRS). (e.g., T11SPA or T12STF). datetime: The acquisition date and time derived from the input HLS granule name. productgenerationdatetime: The date and time at which the product was generated by OPERA. cloud_coverage: The percentage of pixels that is obscured by clouds. spacecraft_name: The name of the sensor platform (e.g., Landsat-8, Sentinel-2A, Sentinel-2B). downloadurl: The direct web link to download the individual data file (requires NASA Earthdata authentication). urldisplay: A secondary URL field sometimes used by web maps to generate clickable links in pop-up windows. Colormap: Values for the B01 WTR layer are a masked interpreted water classification layer. This represents pixel-wise classification into one of three water classes (not water, open water, and partial surface water), HLS snow/ice, HLS cloud/cloud shadow and adjacent to cloud/cloud shadow, or no data classes.  Intended Use: This service provides the ability to perform temporal trend analysis, zonal statistics, and to track dynamic water volume/extent changes over time.    Limitations of Use: If opening with the MapViewer in the browser, due to the size of the underlying data, it can take a few minutes for the data to render on the map. Tips: zoom into your area of interest on the map; and/or change time slider settings - see the PODAAC Cookbook tutorial.  Known issues and caveats on usage are described in the Product Documentation. Contact: If you have feedback on this service, please provide them via this form. For questions about this service, please contact podaac@podaac.jpl.nasa.gov For tutorials on using this service, please visit the PO.DAAC Cookbook.

Dataset Card for Amfitrite-Open-Waters-HAB-Sentinel2 This dataset contains multispectral Sentinel-2 satellite imagery tiles focused on open water and coastal marine environments, classified by the potential presence of Harmful Algal Blooms (HABs). It is designed to train Machine or Deep Learning models (like CNNs) for large-scale environmental monitoring and ocean anomaly detection. Dataset Details Dataset Description… See the full description on the dataset page: https://huggingface.co/datasets/kostaspic/amfitrite-open-waters-hab-sentinel2.

Sentinel-2 cloud and cloud-shadow masks outside Africa 1,640 Sentinel-2 Level-1C patches of 509 x 509 pixels (10 m, 13 bands, uint16 top-of-atmosphere reflectance x 10,000), one from each of 1,640 locations outside Africa, each with an expert mask: 0 clear, 1 thick cloud, 2 thin cloud, 3 cloud shadow. Packed as tar shards (train-NN.tar, members train/images/.tif and train/masks/.tif), with their sha256 in SHA256SUMS. Source and licence Every patch… See the full description on the dataset page: https://huggingface.co/datasets/Emulated-Inc/sci-cloud-mask-sentinel2-new-region.