Leveraging satellite data with the help of AI
Recently the BigEarth Project, led by Prof. Dr. Begüm Demir ended. The EU Community Research and Development Information Service (CORDIS) has now featured the BigEarth Project on their website. The BigEarth project has been supported by the European Research Council (ERC) and is at the forefront of integrating AI into satellite data archives. The Remote Sensing Image Analysis Group developed innovative machine learning algorithms to enhance the way these massive data repositories are searched and utilized, making it easier for researchers and professionals to access the information they need.
One of the standout innovations of the BigEarth project is the image-based querying technique. It allows users to search vast satellite data archives using images, a feature that promises to revolutionize fields such as search and rescue operations, climate science, and mineral exploration.
Read the full story.
All code developed for constructing BigEarthNet v2.0 is released open source. This enables users to reproduce BigEarth. It provides the tools to extend it to any geographical area of interest, fostering further advancements and collaboration within the scientific community.
BigEarthNet v2.0 Update
Webpage: https://bigearth.net
Paper: https://arxiv.org/abs/2407.03653
Pretrained model weights: https://huggingface.co/BIFOLD-BigEarthNetv2-0