Only released in EOL distros:
The v4r_ellipses package contains a computer vision library which is able to detect ellipses within images. The package is able to estimate the pose of the circle related to the ellipse the circle diameter as well as the camera parameter are known. A dynamic reconfigure interface allows the user to tune the parameter of the system to ones needs. But be aware that the pose of a projected circle within a image (ellipse) has two solutions and only one is published as TF. A costom message (v4r_msgs) publishes all the data as well as the pose ambiguity.
- Maintainer: Markus Bader <markus.bader AT tuwien.ac DOT at>
- License: BSD
- Source: git https://github.com/v4r-tuwien/v4r_ros.git (branch: hydro-devel)
This package detects ellipses in camera images and computes the 3D pose of the related circle if the radius of the circle is given. The current 3D pose estimation is based on the [Chen2004]. Varius Parameter and algorithm used to find ellipses can be tuned via ros shared parameters or by using the dynamic reconfigure intervade.
Chen2004) Chen, Q.; Wu, H. & Wada, T. Pajdla, T. & Matas, J. (Eds.) Camera Calibration with Two Arbitrary Coplanar Circles Computer Vision - ECCV 2004, Springer Berlin Heidelberg, 2004, 3023, 521-532, http://link.springer.com/chapter/10.1007/978-3-540-24672-5_41
Detected Ellipses with projected normals of the two possible pose solutions.
Dynamic reconfigure interface
First of all you need a camera publisher with calibration information (see MonocularCalibration to calibrate your camera) and lets assume you ar publishing your camera under /camera/image_raw and /camera/camera_info
Secound you can start the ellipse detection
rosrun v4r_ellipses v4r_ellipses_node image:=/camera/image_raw camera_info:=/camera/camera_info
- running the reconfigure module allows you to tune parameters
rosrun rqt_reconfigure rqt_reconfigure