BigEarth

06/06/2024

Our European Research Council (ERC) funded project have been successfully completed with more than 70 journal and conference papers (with their publicly available code), 3 benchmark datasets (e.g., ) and 1 prototype 'query by image' based search engine ( ). For details, please see: https://bigearth.eu

With the ERC grant, my journey in Berlin started: I built the Remote Sensing Image Analysis (RSiM) group at Technische Universität Berlin in 2018, and the Big Data Analytics for Earth Observation group at BIFOLD - Berlin Institute for the Foundations of Learning and Data in 2022, where I’m extremely proud to work together with outstanding researchers and students.

When I received the grant, I expressed my feeling with a verse from Nazim Hikmet:
“I do not want to listen the songs anymore, I want to sing”
Together with my team at and , we will continue singing by conducting frontier research in the field of processing and analysis of remote sensing images for Earth observation with interdisciplinary approaches associated to , and .

The photo below is from ’s Foundation Models Tutorial day.

24/12/2020

Big news for !
of Technische Universität Berlin has been enriched by including images. The new version contains 590,326 pairs of and image patches to support research on multi-modal/cross-modal image classification, retrieval and search.
This work is supported by the European Research Council (ERC) and by (Berlin Institute for the Foundations of Learning and Data).

http://bigearth.net

Special thanks go to Arne De Wall (a researcher at the Remote Sensing Image Analysis Group and the Database Systems and Information Management Group at the Technische Universität Berlin) for the hard-work to construct the BigEarthNet with Sentinel-1 patches. His codes are publicly available at http://bigearth.net.

BigEarthNet is a benchmark archive, consisting of 590,326 pairs of Sentinel-1 and Sentinel-2 image patches. The first version (v1.0-beta) of BigEarthNet includes only Sentinel 2 images. Recently, it has been enriched by Sentinel-1 images to create a multi-modal BigEarthNet benchmark archive (called....

18/12/2020

If you are wondering how to reduce the negative impact of land-use and land-cover annotations, check our publicly available software and methods on our site.

Rightwards arrow
https://noisy-labels-in-rs.org

Annotating RS images with multi-labels at large-scale to drive DL studies is time consuming, complex, and costly in operational scenarios. To address this issue, existing thematic products (e.g., Corine Land-Cover map) can be used, however the land-use and land-cover labels through these products ca...

06/09/2020

thanks Jose Luis Holgado Alvarez for his great contribution to our team and wishes great success in his future endeavors in France🦾🛰

The paper and codes of his method: gitlab.tubit.tu-berlin.de/rsim/S2-cGAN

21/08/2020

If you are wondering how to use image captions for information discovery from big EO data archives, please see our new paper:
Toward Remote Sensing Image Retrieval Under a Deep Image Captioning Perspective
https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9154525

26/07/2020

There are several researchers and colleagues all around the world, contributing to advance through their related software tools and datasets. We appreciate their excellent work and effort-->
bigearth.eu/externals

09/07/2020

S2-cGAN: Self-Supervised Adversarial Representation Learning for Binary Change Detection in Multispectral Images

Code&Short Paper: https://gitlab.tubit.tu-berlin.de/rsim/S2-cGAN

18/06/2020

We support the public availability of the software! Here you can find all our codes publicly available for private use and research only:

22/05/2020

MAML-RSIC: A Deep Multi-Attention Driven Approach for Multi-Label Remote Sensing Image Classification
Code&Paper: https://gitlab.tubit.tu-berlin.de/rsim/MAML-RSIC

17/04/2020

If you want to download large-scale Sentinel products, take a look to our publicly available 'parallelCollGS' toolchain (https://gitlab.tubit.tu-berlin.de/rsim/parallelCollGS) for parallel queries to download Sentinel data from a variable number of collaborative ground segments.

ZECM - GitLab Dienst

03/04/2020

The Remote Sensing Image Analysis (RSiM) Group funded by is 2 years old! We are so proud to be members of RSiM and , and celebrate it online!

08/02/2020

Do you still use DL models trained on ImageNet for Sentinel-2 images classification/search and retrieval!? We highly suggest DL models available @ bigearth.net!

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 ...