Learning and Recovery Algorithms for Multi-Sensor Data Fusion and Spectral Unmixing in Earth Observation

This project will explore machine learning and recovery algorithms for multi-sensor data fusion and spectral unmixing signal processing in earth observation. The satellite signals are modeled as being encoded by distributed and non-linear compressed sensing, where the non-linearities arise due to quantization and/or measurement matrices that are data-dependent. The data will be obtained from German Aerospace Center (DLR) instruments.