DynamicEarthNet (video-compressed): Daily PlanetFusion Land Cover Change Cubes, 2018-2019

Provider
Toker et al. (TUM, DLR, Planet); video repackaging by ISP, Universitat de Valencia
License
CC-BY-SA-4.0 (source)
Size
4 files
Stored at
Hugging Face (https://huggingface.co/datasets/isp-uv-es/DynamicEarthNet-video)
Access
Free and public. No account with the data provider is needed.
Category
Earth Observation & Geospatial
Source
https://arxiv.org/abs/2203.12560

About this dataset

DynamicEarthNet is a benchmark for multi-class land-use and land-cover change segmentation from daily satellite time series (CVPR 2022). It covers 75 areas of interest worldwide, each 1024x1024 pixels at about 3 m, imaged every day from 1 January 2018 to 31 December 2019 as four-band PlanetFusion surface reflectance (red, green, blue, near infrared), with monthly pixel-wise labels in seven classes: impervious surface, agriculture, forest and other vegetation, wetlands, soil, water, and snow and ice. This repository re-encodes the original 525 GB of GeoTIFF stacks as 12-bit H.265 video with the xarrayvideo library, in the TACO format, at two quality levels: a 9.0 GB and a 2.2 GB tacozip. The repackagers report that change-segmentation scores from U-TAE, U-ConvLSTM and 3D-UNet are statistically unchanged on the compressed cubes. The original data are published by TUM on mediaTUM under CC BY-SA 4.0, which carries over to this derivative; cite Toker et al. 2022.

How to load this dataset in Python

from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="isp-uv-es/DynamicEarthNet-video", filename="dynamicearthnet-video-71psnr.tacozip", repo_type="dataset")
print(path)

Data structure

Files

Example file: dynamicearthnet-video-71psnr.tacozip

Analyze this dataset without downloading it

On Scigantic, DynamicEarthNet (video-compressed): Daily PlanetFusion Land Cover Change Cubes, 2018-2019 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