Home / Technology / Drone AI Spots Objects in Real-Time
Drone AI Spots Objects in Real-Time
24 Dec
Summary
- New drone AI processes data and identifies objects instantly.
- Technology offers 20-30% time savings in area monitoring.
- Future updates will enable night searches and offline operation.
A Russian student startup, DronOko, has pioneered a new computer vision technology for unmanned aircraft systems. This innovative solution facilitates real-time data processing and object identification directly on the drone. By automating search and control processes, the technology significantly reduces the time required for area monitoring, with initial estimates suggesting a 20-30% efficiency gain.
The DronOko system has already demonstrated its practical utility through rigorous testing, including simulations in challenging environments like forests. Its high accuracy is attributed to extensive training on a vast dataset of aerial photographs. Currently, the system can reliably detect people in visible light, proving its readiness for deployment in industrial settings and agricultural applications.
Looking ahead, the startup aims to further enhance the technology by training its neural network to interpret infrared thermal imaging, enabling searches in low-visibility conditions such as at night or through smoke. Plans also include optimizing the system for small onboard computing modules for offline drone operation and developing features for tracking moving targets and mapping their precise locations.
