Where ordinary cameras are of little use—in complete absence of visibility—it is now possible to find a person. Russian scientists have created a surveillance system that works even in difficult conditions, can be deployed in a few hours, and costs many times less than thermal imagers.
The development was presented by scientists from the Murom Institute of Vladimir State University. The system uses artificial intelligence and can "see" through smoke and dust.
Moreover, no complex equipment is required for operation. Standard Wi-Fi equipment is sufficient. The entire system can be deployed in 2–4 hours and operates autonomously.
Thermal imagers detect infrared radiation and allow you to see in the dark, but they are expensive. The new development costs 15–20 times less, while solving the problem of detecting a person in conditions of poor visibility.
The technology is based on the analysis of the radio signal between two points. Only two routers are needed to operate—a transmitting and a receiving one—and a computing module, such as a server or Raspberry Pi.
When a person moves, the module changes the signal parameters—its amplitude and phase. The neural network records these changes and uses them to reconstruct the silhouette. In fact, the system "draws" a person without seeing them directly.
The laboratory prototype showed a recognition accuracy of one person at 91%. The developers are now refining the system so that it can detect several people simultaneously.
Читайте ещё материалы по теме:
- БПЛА Supercam S350 фиксирует сотни целей на передовой: оператор назвал камеру одной из лучших среди аналогов
- Система «Безопасный город» поумнела: NtechLab установила 1600 камер с распознаванием лиц в Пензенской области
- Разработчики модернизировали беспилотник «Скат-350М» для борьбы с зенитными дронами