what we do?
Our pioneering Maritime Surveillance Sensor Suites enable a complete, persistent picture of maritime activity in and beyond territorial waters, delivering maritime situational awareness for government and other agencies interested in security of the seas.
Border security at sea is challenging. The vastness of the oceans and complexity of coastlines means beaches go unwatched and are vulnerable; criminal activity is unchecked; and cries for help may go unheard.
Traditional solutions of patrolling ships and aircraft are expensive and inefficient.
Big data and machine learning are enabling lower cost technologies to transform the way we monitor, analyze and respond to maritime activity.
Artificial intelligence is delivering a revolution in marine surveillance – automating the collection of data and the interpretation of results and so significantly reducing the cost of manning these systems while monitoring increasing amounts of data.
In our system, vessel data is collected from a variety of sensors.
Receiving data from both satellite and terrestrial sensors.
We have Optical and Thermal cameras for both day and night vision.
Our radar sensors can be mounted on the coast or at sea.
We have equipment to detect electromagnetic emissions from radar, radios and cellular phones.
This sensor information is then fused in real time to build an integrated picture.
Larger vessels are obliged to transmit an AIS signal and many others do so voluntarily which can be detected by satellite and by coastal receivers.
If the vessel is also detected on Radar, the two sources can be fused and a level of certainty can be applied to that contact.
If a contact is not transmitting on AIS but is detected on radar then it may warrant further investigation.
If it is a larger vessel it can be labelled as non-compliant.
Camera detections deliver an additional layer of information.
Once detected on radar our camera then search for the vessel and our computer vision system provides an automated visual classification and enables an inference to be made as to what the vessel is doing.
Finally, the detection of electromagnetic emissions provides further information on the vessel, providing a ‘fingerprint’ which can be used for identification.
Having collected the individual detections, we fuse the separate sensor feeds to provide rich and enhanced information on every contact.
This smart data can then be collated in the cloud , analysed and used to drive alerts of suspicious activity such as large vessels not transmitting on AIS.
Traditional surveillance methods utilising patrol aircraft and vessels are very expensive and non-persistent. So the ability to build a picture and provide specific cueing is key.
The system provides an overview of vessels in a sea area highlighting suspect vessels and helping co-ordinate assets for interdiction.
We take raw sensor data from radar and AIS...
...and integrate it with camera imagery to provide a visual context with which to overlay the track data.
Computer Vision enables insight from cameras at day or at night – differentiating between different types of vessel.
It can provide alerts on vessels that are in restricted areas, that are not reporting on AIS or even those approaching beaches of interest.
Our data is collected locally and then securely stored in a cloud environment meaning that it can be easily shared across many users.
Track data can be visualized within proprietary tools, or integrated into existing software solutions.
Systems can be monitored in real time during periods of heightened activity or where intelligence-led.
But at other times, targeted alerts of anomalous activity can be generated automatically and directed by the system to relevant operators who may otherwise be tasked on different activities.
Applications span the interests of multiple agencies.
Detecting small vessels with low radar cross section is challenging, but pattern of life algorithms can alert anomalous activity and allow sufficient time for interdicting assets to be directed to the scene.
These techniques enable the detection of small vessels even at night at long range.
Our algorithms are able to differentiate between activities without a visual reference point and without transponder reports, and so highlight anomalous vessel activity at scale.
These artificial intelligence algorithms are enabled by SiriusInsight’s unique combination of academic talent, artificial intelligence engineering and maritime domain expertise.
The ability to collate, process and analyse maritime data at scale, without human intervention, enables a wider, more persistent and more efficient coverage of the sea than was ever possible before.