AI based Fusion of Satellite/Airborne data for Biodiversity Change Characterization

Project objectives: With the support of the European Space Agency, HUN-REN SZTAKI and the Lechner Knowledge Centre (LTK) are collaborating to develop a new methodology for mapping and characterizing geographical areas based on the presence and temporal changes of water cover, in order to achieve different habitat management objectives. The development activity is based on the collection and analysis of key requirements and use cases of the participating stakeholders - national park experts and agricultural professionals. Based on the requirements, artificial intelligence methods are developed to perform time-series analysis and feature tracking of heterogeneous remote sensing measurements. The analysis is based on the effective fusion of satellite imagery and aerial high-resolution multispectral imagery, which allows for higher resolution in both time and space, ensuring the detection of changes in many features of areas affected by water impacts on different time scales. From a methodological perspective, both machine learning and geometric feature matching will be applied, relying on stochastic multilayer image segmentation methods and the latest deep neural network architectures. The method will be validated in collaboration with participating stakeholders, and implementing partners will make recommendations for operational application.