Compressive Geospatial Analytics

Compressive sensing is a randomized data acquisition method that linearly samples sparse or compressible signals at a rate much below the Nyquist-Shannon sampling theorem, and outperforms traditional signal processing techniques through performing both sensing and size reduction tasks simultaneously. Edge-computing is a decentralization approach that provides several properties (specifically reducing the need for moving a large volume of data) via pushing the computation towards the edge of the network. we adapt and integrate underlying concepts in compressive sensing and edge-computing to propose a novel data acquisition framework for earth science applications. The proposed model leverages compressive sensing to address not only big-data aspects of geospatial applications but also noise, uncertainty, and consistency of the measurements. In addition, our approach provides flexible and low-cost data visualization, integration, and analysis properties. Our model also subsides the need for utilizing advanced technologies (e.g., high-performance computing for machine learning applications) which is crucial for democratizing/improving the access to data and broadening the user communities—specifically in underdeveloped and developing countries. Furthermore, the proposed framework tries to address not only FAIR data principles but also open-Data concerns.