RSiM group led by Prof. Begüm Demir is part of the Faculty of EECS at the TU Berlin. Our group performs research in the field of processing and analysis of remote sensing images for Earth observation with interdisciplinary approaches associated to remote sensing, machine learning, signal & image processing and big data management.
Gencer Sumbul successfully defended his PhD thesis titled “Deep Image Representation Learning for Knowledge Discovery from Earth Observation Data Archives” on May 9th and received the distinction 'summa cum laude'. Congratulations Gencer!
We have recently released ConfigILM that is a pytorch based library, enabling fast development of visual question answering systems. This open-source library provides a convenient implementation for seamlessly combining models from two of the most popular pytorch libraries, timm and huggingface.
Leonard Hackel was awarded the Rolf-Niedermeier-Award for the best master thesis in Computer Engineering of TU Berlin. In his thesis, he developed lightweight transformer-based visual question answering (VQA) models for Earth observation. For details, click here.
We will host "MagicBathy - Multimodal multitAsk learninG for MultIsCale BATHYmetric mapping in shallow waters" research project funded through the HORIZON Europe MSCA Postdoctoral Fellowships for the period 2023-2025. The project aims to establish an advanced framework for low-cost shallow water mapping.
Together with our partners within the TreeSatAI project funded by the Federal Ministry of Education and Research, we released the TreeSatAI benchmark archive that is a multi-sensor, multi-label dataset for tree species classification in remote sensing.
© ESA
RSiM is involved in a new project 'Demonstrator Precursor Digital Assistant Interface For Digital Twin Earth' funded by the European Space Agency for the period 2022-2024. In this project, we will collaborate with the e-GEOS, Italy and the University of Athens in Greece.
