EYE-Sense: empowering remote sensing with machine learning for socio-economic analysis

Abstract

EYE-Sense is a Web-GIS platform which allows for easy access to valuable socio-economic insights from Earth Observation (EO) data by offering a code-less approach. The platform enables users to access various EO parameters, such as atmospheric and water quality indexes, and night-light activity. Moreover, the platform's pre-trained computer vision models (Faster-RCNN, Mask-RCNN, and YOLO) empower users to detect objects such as, e.g., airplanes, ships, containers, and beach umbrellas, to address specific user-based tasks. To provide cost-efficiency, scalability, flexibility, and easy maintenance, EYE-sense adopts a serverless architecture, leading to up to 50.4% processing cost reduction when compared to traditional server-based solutions. By bridging the gap between data gathering and processing, EYE-Sense extends the reach of Earth observation data to a broader audience.