Skip to content

0.8.1

Choose a tag to compare

View all tags
@github-actions github-actions released this 04 Aug 14:16
· 39 commits to main since this release
0.8.1
15b059a

SAE GEO Mapper

A SAE stage that maps object locations from camera / pixel space to geo-coordinate space. For that it needs to be configured with a number of optical and geometrical camera parameters.\

Geo Mapper can be run in two fundamental modes: static and dynamic. Static means, that for every videosource (camera) next to necessary mapping data position data is configured in settings.yaml. In dynamic mode Geo Mapper will take camera position from video frames which are inserted by video source component.

Quick Setup

Due to cameratransform being stuck on Python 3.11 (due to Pandas 1.x), we need to use pyenv for this repository.

  • Make sure pyenv is installed
  • Install the latest Python 3.11 (e.g. Python 3.11.9 at the time of writing): pyenv install 3.11.9
    • If you encounter installation error, check that you have all Python build dependencies installed (refer to pyenv README)
  • Set the installed Python version for this directory, e.g. pyenv local 3.11.9
  • Run poetry install, it should automatically pick up the correct Python version and use it for the virtualenv

How to Build

See dev readme for build & package instructions.

Input/Output

  • Input message must be a SaeMessage. The geo-mapping is done on each Detection message within. If there are no Detection messages, the processing is effectively a no-op.
  • Output is the input SaeMessage with geo-coordinates added to every Detection. All other fields are preserved.

Changelog

0.7.0

  • Upgrade vision-api to 3.1.0
  • Add SaeMessage.frame.camera_location if pos_lat and pos_lon is set (independent of passthrough mode)
Assets 4
Loading