Skip to content

A proof of concept of an Extrinsic transformation of the detections in an object detection application

Notifications You must be signed in to change notification settings

alexbarnadas/SceneProcessing

Repository files navigation

Scene Procerssing

Extrinsic transformation of the detections in an object detection application. With multiple cameras parallel processing. This project is made for the pilot use case in SERMAS for the ODIN project from the Horizon2020.

Assumptions: The camera is still and in a low abrupt angle.

Installation

python -m venv venv

pip install -r requirements.txt

And you are ready to run main.py.

Camera calibration

  1. The first frame of the input video will apperar.
  2. Click accurately in a squared area of the scene framed
  3. Press esc

Run Scene understanding

  1. Specify the IP adresses of the cameras to be processed in real time in the main.py script.
  2. Specify the conditions of the alerts in alerts.py.
  3. Run main.py

The video will reproduce and also a window with the bird view coordinates of the detections will be drawn.

Every detection will be tracked and the orientation of the track will be shown with an arrow. Also the velocity in pixels per second will be plot frame by frame next to every detection

About

A proof of concept of an Extrinsic transformation of the detections in an object detection application

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages