Remote sensing data from various space missions as well as geodata on the state of the environment are increasingly available free of charge in immense quantities.At the same time, social media platforms such as Flickr, Twitter or Open Street Map provide access to further textual or visual information about the environment. Nevertheless, the acquisition of high-quality and quantitative training data for training the algorithms is still the biggest challenge, which is why many AI earth observation applications fail or are only applicable in a very localized way.
Therefore, the overall goal of TreeSatAI is the prototypical development of AI methods for the monitoring of forests and tree inventories on local, regional and global scales. Based on freely accessible geodata from different sources (remote sensing, administration, social media, mobile apps, monitoring libraries, open image databases) prototypes for deep learning based extraction and classification of tree and stand features for four different use cases in the field of forest, nature conservation and infrastructure monitoring will be developed.
Partners
- TU Berlin: TU Geo
- TU Berlin: RSiM
- LiveEO GmbH
- LUP GmbH
- Vision Impulse GmbH
