We recently completed work on a long-term computer vision project in the counter-UAV (anti-drone) sector. As part of a very small ‘agile’ team, we have developed a visual drone-detector that runs on very small hardware (a Jetson NX) with an off-the-shelf camera, and is able to process frames at 30fps, which is sufficient to feed target information into an on-board guidance system to guide the ‘take-down’ or ‘interceptor’ drone towards the target drone.
The project was complex and challenging, including building the whole visual system into a simulator to allow test ‘flights’ to take place there while the visual and guidance systems were developed.
The system has now been moved into the real world, and has been tested successfully numerous times in the field (literally). Video is not ours to share at the moment, but hopefully soon.
Skip to content