This dataset contains ground reference crop type labels and multispectral and synthetic aperture radar (SAR) imagery from multiple satellites in an area located in Western Cape, South Africa. There are five crop types from the year 2017: Wheat, Barely, Canola, Lucerne/Medics, Small grain grazing. The AOI is split to three tiles. Two tiles are provided as training labels, and one tile will be used for scoring in the competition. Input imagery consist of time series of Sentinel-2, Sentinel-1 and Planet Fusion (daily and 5-day composite) data. You can access each source from a different collection. The Planet fusion data are made available under a CC-BY-SA license. As an exception to the AI4EO Terms and Conditions published on the competition website, you confirm, by participating in it, that you agree that your results will be made public under the same, open-source license. The Western Cape Department of Agriculture (WCDoA) vector data are supplied via Radiant Earth Foundation with limited distribution rights. Data supplied by the WCDoA may not be distributed further or used for commercial purposes. The vector data supplied are intended strictly for use within the scope of this remote sensing competition - for the purpose of academic research to our mutual benefit. The data is intended for research purposes only and the WCDoA cannot be held responsible for any errors or omissions which may occur in the data.

• undefined
• undefined
• Data link
• undefined

MLHUB Short Name: A Fusion Dataset for Crop Type Classification in Western Cape, South Africa Version ID: 1 Unique ID: C2781412697-MLHUB

Georeferenced crop yield prediction is a valuable tool for agronomists and policymakers. One challenge with many existing datasets is that of location accuracy. GPS locations for fields can end up offset from the true location due to sensor inaccuracies or from locations being collected at the edges of fields rather than the field centers. This makes it harder to connect remote-sensed data to the yield values. The goal of this project was to produce a method that can help correct these location offsets by finding the most probable field center given an input location.

• undefined
• undefined
• Data link
• undefined

MLHUB Short Name: East Africa Agricultural Field Centers Version ID: 1 Unique ID: C2781412176-MLHUB

Floating marine debris is a global pollution problem which leads to the loss of marine and terrestrial biodiversity. Large swaths of marine debris are also navigational hazards to ocean vessels. The use of Earth observation data and artificial intelligence techniques can revolutionize the detection of floating marine debris on satellite imagery and pave the way to a global monitoring system for controlling and preventing the accumulation of marine debris in oceans. This dataset consists of images of marine debris which are 256 by 256 pixels in size and labels which are bounding boxes with geographical coordinates. The images were obtained from PlanetScope optical imagery which has a spatial resolution of approximately 3 meters. In this dataset, marine debris consists of floating objects on the ocean surface which can belong to one or more classes namely plastics, algae, sargassum, wood, and other artificial items. Several studies were used for data collection and validation. While a small percentage of the dataset represents the coastlines of Ghana and Greece, most of the observations surround the Bay Islands in Honduras. The marine debris detection models created and the relevant code for using this dataset can be found [here](https://github.com/NASA-IMPACT/marine_debris_ML).

• undefined
• undefined
• Data link
• undefined

MLHUB Short Name: Marine Debris Dataset for Object Detection in Planetscope Imagery Version ID: 1 Unique ID: C2781412735-MLHUB

The PlanetScope Level 1B Basic Scene and Level 3B Ortho Scene full archive products are available as part of Planet imagery offer. The Unrectified Asset: PlanetScope Basic Analytic Radiance (TOAR) product is a Scaled Top of Atmosphere Radiance (at sensor) and sensor corrected product, without correction for any geometric distortions inherent in the imaging processes and is not mapped to a cartographic projection. The imagery data is accompanied by Rational Polynomial Coefficients (RPCs) to enable orthorectification by the user. This kind of product is designed for users with advanced image processing and geometric correction capabilities. Basic Scene Product Components and Format Product Components: - Image File (GeoTIFF format) - Metadata File (XML format) - Rational Polynomial Coefficients (XML format) - Thumbnail File (GeoTIFF format) - Unusable Data Mask UDM File (GeoTIFF format) - Usable Data Mask UDM2 File (GeoTIFF format) Bands: 4-band multispectral image (blue, green, red, near-infrared) or 8-band (coastal-blue, blue, green I, green, yellow, red, Rededge, near-infrared) Ground Sampling Distance: - Approximate, satellite altitude dependent - Dove-C: 3.0 m-4.1 m - Dove-R: 3.0 m-4.1 m - SuperDove: 3.7 m-4.2 m Accuracy: <10 m RMSE The Rectified assets: The PlanetScope Ortho Scene product is radiometrically-, sensor- and geometrically- corrected and is projected to a UTM/WGS84 cartographic map projection. The geometric correction uses fine Digital Elevation Models (DEMs) with a post spacing of between 30 and 90 metres. Ortho Scene Product Components and Format Product Components: - Image File (GeoTIFF format) - Metadata File (XML format) - Thumbnail File (GeoTIFF format) - Unusable Data Mask UDM File (GeoTIFF format) - Usable Data Mask UDM2 File (GeoTIFF format) Bands: 3-band natural color (red, green, blue) or 4-band multispectral image (blue, green, red, near-infrared) or 8-band (costal-blue, blue, green I, green, yellow, red, RedEdge, near-infrared) Ground Sampling Distance: - Approximate, satellite altitude dependent - Dove-C: 3.0 m-4.1 m - Dove-R: 3.0 m-4.1 m - SuperDove: 3.7 m-4.2 m Projection: UTM WGS84 Accuracy: <10 m RMSE PlanetScope Ortho Scene product is available in the following: PlanetScope Visual Ortho Scene product is orthorectified and color-corrected (using a colour curve) 3-band RGB Imagery. This correction attempts to optimise colours as seen by the human eye providing images as they would look if viewed from the perspective of the satellite PlanetScope Surface Reflectance product is orthorectified, 4-band BGRN or 8-band Coastal Blue, Blue, Green I, Green, Yellow, Red, RedEdge, NIR Imagery with geometric, radiometric and corrected for surface reflection. This data is optimal for value-added image processing such as land cover classifications. PlanetScope Analytic Ortho Scene Surface Reflectance product is orthorectified, 4-band BGRN or 8-band Coastal Blue, Blue, Green I, Green, Yellow, Red, RedEdge, NIR Imagery with geometric, radiometric and calibrated to top of atmosphere radiance. _$$Planet Explorer Catalogue$$ https://www.planet.com/explorer/ can be accessed (Planet registration requested) to discover and check the data readiness. It is worth to mention that the data distribution is based on Data Voucher, corresponding to maximum amount of square kilometers can be ordered and downloaded by the project in a maximum period of 15 moths (this duration cannot be extended) starting from the project proposal acceptance date. Each Date Voucher includes PlanetScope tile view streaming access for a total of 20,000 tiles per calendar month during the project period. All details about the data provision, data access conditions and quota assignment procedure are described in the _$$Terms of Applicability$$ https://earth.esa.int/eogateway/documents/20142/37627/Access-to-ESAs-Planet-Missions-Terms-of-Applicability.pdf available in Resources section.

• Data link
• Data link
• PlanetScope.Full.Archive

ESA Short Name: PlanetScope.Full.Archive Version ID: NA Unique ID: C1965336933-ESA

This dataset contains field boundaries for smallholder farms in eastern Rwanda. The Nasa Harvest program funded a team of annotators from TaQadam to label Planet imagery for the 2021 growing season for the purpose of conducting the Rwanda Field boundary detection Challenge. The dataset includes rasterized labeled field boundaries and time series satellite imagery from Planet's NICFI program. Planet's basemap imagery is provided for six months (March, April, August, October, November and December). The paired dataset is provided in 256x256 chips for a total of 70 tiles covering 1532 individual fields.<br><br>Input imagery consists of a time series of planet Basemaps from the NICFI program (monthly composite) data.<br><br>Imagery Copyright 2021 Planet Labs Inc. All use subject to the Participant License Agreement.

• undefined
• undefined
• Data link
• undefined

MLHUB Short Name: Rwanda Field Boundary Competition Dataset Version ID: 1 Unique ID: C2781412768-MLHUB