PODIUMS is a nine-months project funded by ESA within the call ARCHITECTURE FOR OFFLINE PROCESSING AND MACHINE LEARNING IN MASS-MEMORIES. The project aims to conduct a comprehensive set of activities such as requirements analysis, design review, and feasibility assessment to lay the groundwork for the development of an artificial intelligence (AI)-based payload module for the analysis of multispectral data on-board satellites. Such a module aims to improve the efficiency of data processing and significantly enable rapid decision making by providing real-time analysis. To this end, an analysis among possible applications will be conducted to define examples of use case(s) that can best exploit AI-based processing on multispectral data, considering the limited resources available aboard satellites. Considering these potentials and constraints, the requirements and interface of the AI-High Performing Computing module will be determined, and a prototype will be designed and implemented, which will finally be analyzed in terms of performance using a specially selected set of images.

Eurac's role in the project is to leverage the expertise in remote sensing techniques of the Institute for Earth Observation to select use case(s), define the strategy for acquiring a suitable set of high-resolution images to be applied in the test campaign to evaluate the capabilities of the AI module, and to analyze the results obtained, thus supporting the project partners in evaluating the accuracy and efficiency of AI processing.

The project supports the further development of Eurac's expertise in AI and state-of-the-art high-resolution image processing techniques, as well as the growth of its partnership network.

Contact persons: Giovanni Cuozzo giovanni.cuozzo@eurac.edu