Computer Vision for Military Anti-Drone Defence

As aerial threats like drones to NATO and Latvian borders evolve, air defense systems must provide rapid, reliable threat detection and response. Manual identification and targeting processes can't match the speed required for effective counter-drone operations. A new approach is needed - one that can identify, track, and neutralize aerial threats in real-time without human intervention delays. Furthermore, the solution must be cost-effective and manufacturable with  available industrial hardware components.

WeMPS, one of the leading industrial hardware engineering companies, engineered the hardware platform for this military-grade defense system. We partnered with WeMPS to establish the software architecture that would bring autonomous threat detection to life. Working side-by-side with WeMPS engineers to develop the software foundation that powers the system, Cube helped expand WeMPS's in-house software development capabilities. Together, we designed and implemented specialized convolutional neural networks trained for visual drone identification, integrated with real-time processing on a hardened Linux OS for mission-critical operations. The system performs instant ballistic calculations to ensure precision engagement while maintaining the speed necessary for effective defense.
‍
To accelerate development and reduce costs, a comprehensive virtual simulation environment was created - a digital twin of the physical hardware solution. This allows for extensive testing and training scenarios without the expense and logistical complexity of live-fire exercises with physical drones and ammunition.

The result is a robust defense capability that helps protect critical infrastructure and borders with speed and reliability. By eliminating human reaction time delays and providing 24/7 autonomous monitoring, the system significantly enhances defensive posture while reducing the burden on military personnel.