New AI model could enable real-time maritime surveillance onboard satellites
A new AI model could transform maritime surveillance by enabling satellites to detect and classify vessels much faster than is currently possible, passing information to Earth within minutes rather than hours, and helping to monitor illicit activity including illegal fishing, smuggling, trafficking, and piracy.
The new technology, unveiled in a presentation today at the Symposium on Edge Computing in Washington D.C. by researchers from the Alan Turing Institute, demonstrates for the first time that high-accuracy Synthetic Aperture Radar (SAR) vessel detection can run autonomously on low-power hardware suitable for satellites.
The use of SAR satellite images to rapidly detect “dark” ships (those that disable their Automatic Identification Systems) is critical for maritime surveillance, but requires analysis of images in real-time, ideally within minutes of data acquisition, to allow human operators to respond or other satellites to capture additional data.