BigEarthNet

A New Large-Scale Sentinel-2 Benchmark Archive

About BigEarthNet

The BigEarthNet archive was constructed by the Remote Sensing Image Analysis (RSiM) Group and the Database Systems and Information Management (DIMA) Group at the Technische Universität Berlin (TU Berlin). This work is supported by the European Research Council under the ERC Starting Grant BigEarth and by the German Ministry for Education and Research as Berlin Big Data Center (BBDC).

The BigEarthNet is a new large-scale Sentinel-2 benchmark archive, consisting of 590,326 Sentinel-2 image patches. To construct the BigEarthNet, 125 Sentinel-2 tiles acquired between June 2017 and May 2018 over the 10 countries (Austria, Belgium, Finland, Ireland, Kosovo, Lithuania, Luxembourg, Portugal, Serbia, Switzerland) of Europe were initially selected. All the tiles were atmospherically corrected by the Sentinel-2 Level 2A product generation and formatting tool (sen2cor). Then, they were divided into 590,326 non-overlapping image patches. Each image patch was annotated by the multiple land-cover classes (i.e., multi-labels) that were provided from the CORINE Land Cover database of the year 2018 (CLC 2018).

The BigEarthNet is significantly larger than the existing archives in remote sensing and opens up promising directions to advance research for the analysis of large-scale remote sensing image archives. It is also very convenient to be used as a training source in the context of deep learning. For the details about the BigEarthNet, please see our paper:

G. Sumbul, M. Charfuelan, B. Demir, V. Markl, BigEarthNet: A Large-Scale Benchmark Archive for Remote Sensing Image Understanding, IEEE International Conference on Geoscience and Remote Sensing Symposium, Yokohama, Japan, 2019.

Downloads

If you use BigEarthNet archive, please cite our paper given below:

G. Sumbul, M. Charfuelan, B. Demir, V. Markl, BigEarthNet: A Large-Scale Benchmark Archive for Remote Sensing Image Understanding, IEEE International Conference on Geoscience and Remote Sensing Symposium, Yokohama, Japan, 2019.

News

We are looking for highly motivated PhD candidates to join the Remote Sensing Image Analysis (RSiM) group at the Faculty of Electrical Engineering and Computer Science, Technische Universität Berlin, Germany. The research of the PhD candidates will aim at developing innovative machine learning techniques (with a special focus on deep learning) for the analysis of big data from space. The main topics include: 1) developing deep neural network models that can overcome the data imbalance problems for satellite image classification; and 2) developing active learning methods that are applicable to the designed deep neural networks. MSc degree is required in computer engineering or computer science with experience in computer vision, deep learning for image understanding. Very good command of German and English is required. For details, please visit this link.
We prepared a script that returns the original Level-1C tile information of each Sentinel-2 image patch given in the BigEarthNet. To download the script, see: https://gitlab.tubit.tu-berlin.de/rsim/bigearthnet-tools. Note that all the considered Level-1C tiles were atmospherically corrected by the Sentinel-2 Level 2A product generation and formatting tool (sen2cor) before creating the patches. Thus, the script is useful to have access to the original Level-1C tile of each patch.
In the BigEarthNet, we have kept the image patches with seasonal snow cover. This is to allow tag based search and retrieval studies independently from seasons. However, soon we will provide the list of patches with seasonal snow that can be important for the scene classification and content based image search/retrieval studies.
BigEarthNet has been released at the Big Data from Space (BIDS'19) Conference in Munich, Germany.

FAQ

Yes, some Sentinel-2 patches include seasonal-snow. We kept them in the archive to allow tag-based search and retrieval studies independently from seasons. However, if you work on scene classification, content-based image retrieval and search only by using Sentinel-2 image patches, we suggest not to include patches fully covered by seasonal-snow for training and test stages of the machine/deep learning algorithms. The list of these patches is provided in the ‘Download’ page.
Sentinel-2 tiles in the BigEarthNet are associated to cloud cover less than 1%, whereas in some cases the cloud cover are populated within some patches. We kept these patches in the BigEarthNet to allow multi-modal learning studies (soon Sentinel-1 images will be also included in BigEarthNet). However, if you work on scene classification, content-based image retrieval and search only by using Sentinel-2 image patches, we suggest not to include these patches for training and test stages of the machine/deep learning algorithms. The list of patches fully covered by cloud and cloud shadow is provided in the ‘Download’ page.
The BigEarthNet Archive is licensed under the Community Data License Agreement – Permissive, Version 1.0.