Machine Vision & Surveillance - @Target
Automatic Target Detection and Recognition (ATD/R) algorithms for Unmanned Aerial Vehicles (UAV). These algorithms facilitate the recognition and tracking of moving targets by using information from an on-board video subsystem. The proposed system places special emphasis on the use of low complexity video analysis technique to achieve real-time operation. Dependable object detection, even in cases of image distortion and poor visibility.
  • Automatic target detection;
  • Automatic target tracking;
  • Resistance to various visual distortions;
  • Rotational and scale invariant;
  • Tracking of partially hidden objects;
  • Simultaneous monitoring of several objects;

These algorithms can be used for fully unmanned UAV navigation and targeted surveillance and monitoring in automatic and semi-automatic modes.

The solution consist of three basic modules - objects tracking module, video streaming module and interface module to on-board navigation system. Objects tracking part is responsible for targets detection and recognition. Streaming module can be used to compress the video and pass it to the control center. Interface part is used to integrate the solution with other on-board equipment.

Apart of UAV applications the solution proposed can be used in varios environments for detection and tracking of moving aerial and on-ground targets from stationary or moving on-ground platforms.
©Avarex Ltd., 2008-2012