Skip to content

A computer vision system for tracking the speed of passing cars using AWS Deep Lens

Notifications You must be signed in to change notification settings

kylefritz/speed-cam

Repository files navigation

Speed Camera

I mounted an AWS Deep Lense camera at my house. I'm using this repo to develop a computer vision system for tracking the speed of passing cars.

setup

pip install -r requirements.txt
honcho start
# setup STORAGE_ROOT, hardcoded to /mnt/usb-sd

data model

track ->> regions
frame ->> regions
frame ->> track

Next steps

  1. Use tracker to associate sucessive frames
    1. could start by just associating with any track already have and then killing tracks after 5 seconds
  2. Compute speed on tracks
  3. Post tracks to the rails app when reaped 1. need a worker process to upload reaped tracks
  4. Display tracks
  5. Integration test for tracker
    1. Need to collect a bunch of live frames for ground truth. ideally cruise control through the scene

env suggestions

  • use pyenv
  • pyenv install will install the right python from the .python-version file

About

A computer vision system for tracking the speed of passing cars using AWS Deep Lens

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published