Aldera
Soft

Detect and track pedestrians in video footage

The system input is: - video sequences from cameras; - information about the location of the cameras, other meta information; - geographic location of smart tags. The system, using a neural network, highlights people on the video sequence; using meta information determines the coordinates of these people and compares with the geographical location of the marks. People without tags are marked on the video sequence as violators.

  • Web
[a]Detect[/a] and [a]track[/a] pedestrians in video footage

Goal And Tasks

Develop a neural network that will track people in the winery

[a]Detect[/a] and [a]track[/a] pedestrians in video footage
  • 01Developing a neural network
  • 02Designing the interaction of a neural network with cameras
  • 03Processing visitors' faces
  • 04Matching people's faces using a neural network
  • 05Neural network training

User

The security service specialist can control the videos processed using the neural network to determine access rights to a certain territory, based on the fare of the purchased ticket. When buying a ticket, a photo of the visitor is taken, which is then compared with the incoming video to recognize the person. In case of violation of visiting a certain territory, the system sends the coordinates and personal data of the visitor to the security specialist.

User

Features

  • 01Processing video recordings from cameras in real-time
  • 02Matching faces with photos using a neural network
  • 03Visitor face recognition
  • 04Determination of the accessibility zone depending on the purchased ticket fare
  • 05Tracking site visitors
Features

Technologies used

  • Python
  • Celery
  • Postgresql
  • Pytorch
  • Opencv
  • Pycuda
  • TensorRT
  • Sqlalchemy
  • Grpc
  • Fastapi
  • Zeromq
  • Redis
  • Docker
  • Ffmpeg
  • Webrtc (Janus)
  • Websockets
  • Firebase

More projects

Do you have a project or an idea for implementation?

Tell us, fill out the form below and we will discuss the details

By clicking the "Submit" button, I agree to the privacy policy and the processing of personal data

Submit