Video and photo analytics. Object detection and analysis. The implementation can be either a built-in “identification” code module into the customer’s system or a stand-alone web interface. Objects for identification -people, cars, license plates, animals, buildings, other objects. The model operates using a video camera and can be used autonomously, i.e. not tied to a UAV or robotic platform.

The task is to detect and identify objects of previously known types. Additionally – the ability to clarify the characteristics of objects (size, color, position, etc.)

Main characteristics:

• ability to work at altitudes from 30 to 3000 meters;

• travel speed from 0 to 130 km/h;

• work in poor visibility conditions (light fog, moderate snow, heavy rain);

• the ability to recognize with limited visibility of the object, including overlap of more than 25%;

• real-time operating mode (at least 5 frames per

second) with a resolution of at least Full HD on Jetson Nano computers.