You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
Coastal erosion driven by wind-generated waves is a phenomenon that unfolds slowly but steadily, which is why it's often only noticed once the damage is already significant. This project builds on my diploma thesis, which studied erosion at three specific coastlines in Corfu, Greece, and the methods for protecting them. The goal is to build a system that combines satellite imagery — both current and historical, spanning 20-30 years — with artificial intelligence, to detect, measure, and predict coastline erosion at scale, starting in Greece, then expanding to Europe, and eventually worldwide.
The core goals are:
Develop an AI model that compares satellite images across different time periods and automatically detects shoreline shifts.
Build a database/map of "high-risk" erosion points, starting with the three Corfu coastlines studied in my thesis.
Validate the model against field measurements and historical records, so results are reliable rather than purely visual.
Gradually expand coverage to more Greek coastlines, then Europe, and eventually internationally, via an open platform/API.
We'll achieve this by combining freely available satellite archives (e.g. Landsat, Sentinel, historical aerial photography) with computer vision and deep learning techniques for change detection, building on the foundation already established by my academic research.
Computing power (GPU/cloud) for training and running the AI models on large volumes of satellite imagery
Access to premium/historical satellite archives where free data isn't sufficient
Field visits and measurements at the three Corfu coastlines for validation (ground truthing)
Development of a platform/app where results will be accessible (to authorities, municipalities, researchers)
Possibly, compensation for data engineering or GIS support if the scope grows
My name is Giorgos Tsimpoulis, and I'm graduate Civil Engineer from the Aristotle University of Thessaloniki, currently working at a technical engineering office for the past two years. At the moment, I'm working on this project on my own. However, if funded, i would be supported by my thesis advisor, who is an expert in this specific research field, holding a PhD and an extensive recod of conference presentations. I don't yet have direct experience combing AI with this domain, which is one of the areas I'm looking to build through this project. In terms of track record, my thesis was presented and evaluated by a three-member committee at my university, as well as reviewed by the Technical Chamber of Greece (TEE)
Likely causes of failure:
Insufficient quality or availability of historical satellite imagery for certain areas, limiting comparison accuracy
Difficulty generalizing the AI model beyond the initial three coastlines (different geomorphology elsewhere)
Limited computing/financial capacity to scale beyond Greece
Likely outcomes:
The project remains at a local/research level (e.g. limited to the three Corfu coastlines) rather than scaling across Europe or globally
The findings still retain academic value and usefulness as a case study, even without reaching a production-ready tool