Projects
Contact
AI-based spatio-temporal forecasting of satellite driven environmental big data
This project focuses on developing an advanced artificial intelligence framework for spatio-temporal forecasting using satellite-derived environmental data such as ocean parameters, meteorological parameters, or Arctic sea ice concentration. By leveraging state-of-the-art computer vision techniques and deep learning architectures, we aim to analyze time-series satellite imagery to predict short-term environmental changes. The methodology incorporates convolutional neural networks combined with recurrent neural networks or transformer-based models to effectively capture both spatial patterns and temporal dependencies in the data. Our approach addresses fundamental challenges in environmental monitoring, including irregular sampling, varying spatial resolutions, and complex nonlinear relationships between variables. The predictive models will be evaluated using robust metrics that assess both spatial accuracy and temporal consistency. This project may contribute to the growing field of AI-driven environmental monitoring and provide valuable tools for understanding and predicting dynamic earth system processes, which can inform climate studies, disaster management, and environmental policy decisions.
Presentation(slide), demo video, and source code at Github
3 students
6 months