Detecting Object Behaviour of Interest Using Deep Learning
MUSLS: Persistent Multi-Sensor Land Surveillance and Change Monitoring
MUSLS is an AI-powered prototype that integrates a suite of novel analytics engines developed in-house and proposes a foundation for developing the next generation of Geointelligence systems. It uses advanced Machine Learning and classical techniques to monitor, track and detect the behaviour of objects of interest, from large volumes of satellite imagery (Electro-Optical and SAR) acquired over an area of interest over time. This functionality is powered by a robust framework that seamlessly integrates cross-modality data, cutting-edge Deep Learning, classical SAR change detection, and statistical time series methods.
MUSLS advanced techniques can help to identify potential anomalies and changes, which in turn will facilitate informed decision making by relevant stakeholders. By leveraging the power of AI, advanced analytics, and vast volumes of multi-sensor satellite imagery, MUSLS forms the basis for a new analytics engine that can manage large volumes of data across multiple sensor platforms.