Technologies for remote identification using computer vision and artificial intelligence
Computer vision for unmanned transport

Infocom IT – is the first Ukrainian company developing an unmanned vehicle – breakthrough innovative solution in the field of computer vision and artificial intelligence for unmanned and remotely controlled land vehicles, robotic platforms and smart mobile applications.

Infocom IT Unmanned Vehicle concept:
  • using lidars, depth cameras and radars the system recognizes and tracks other vehicles, obstacles, markings and pedestrians
  • system analyses a statistical data + DNN
  • mathematical engine developed by Infocom LTD makes decision about next step.

Project based on tools, libraries, and computer vision frameworks: OPENCV, TensorFlow, ROS, video compression (H264/H265), image processing; objects detection; re-identification; Neural Networks; YOLO V3, semantic segmentation, Feature extraction, training regimes, multi-view geometry, bundle adjustment, SLAM.

Computer vision

OpenCV, Tensorflow, Depth Cameras, Radars, Lidars, SLAM algorithms, detection and recognition, tracking movements, determining coordinates in 3D space, recognition of size, color, voice.
Stack:C++, Python, CUDA, Yolo V3, ROS, DNN, Qt, OpenGl, DarkNet, SegNet
Driver Assistant

“Driver Assistant” – the assistant and navigator for every driver and corporate vehicle fleet. It`s a hardware-software system based on a smartphone, a mini PC, a mini controller, and an IR camera, which provides control of the driver’s condition and the environment on the road.

“Driver Assistant” concept:
  • The camera directed to the driver determines and analyzes how many times the driver has yawned, rubbed or closed his eyes for a long time. The camera also analyzes talking on the phone, smoking, atypical head position. In case of a driver’s fatigue or distraction, the system automatically gives a signal to increase driver’s attention or to stop and to rest.
  • The camera pointed at the road analyzes the position of the vehicles in front, calculates the distance to them, recognizes pedestrians and obstacles, road markings, road signs, and speed limit signs. In the case of a fast approach to another car, crossing a solid line, speeding, detecting a pedestrian in the middle of the road, the algorithm will give a warning signal to the driver about the danger and report the reason.