FOSS4G NA 2024

Tracking detected vessels at sea from unmanned surface vehicles
09-11, 14:00–14:30 (America/Chicago), Grand H

Saildrone manufactures and operates unmanned surface vehicles (USVs) to collect ocean data. For our Maritime Domain Awareness business, we built a cloud-based service to store and filter the ground tracks of other vessels detected by our fleet of USVs.


Saildrones are unmanned surface vehicles deployed for long-duration data collection missions on the ocean. Maritime Domain Awareness ensures the safety and security of waterways, with customers such as the US Navy. Saildrones operating in this area are equipped with sensors, including radar and machine learning-equipped cameras, that track the positions of vessels. Data from these sensors are processed by a mixture of proprietary software and open-source geo libraries and then sent to the cloud in real time. A service in AWS ingests that data into a PostGIS database, where it stores the location of these detected vessels over time. Users can view this data on a map, filter against a geographic region of interest, playback vessel tracks, as well as be notified of newly detected vessels that meet user-selected criteria. Heatmaps are also made using matplotlib for analytics of detections, and we are exploring open-source ways to present this dataset on web maps. Internally, images of the detected vessels are also ingested, not only to provide awareness to our customers, but also to improve machine learning detection models.