Camera_info for 3d camera not working for rectification

[sam-tj]

Hello Team,

I need a rectified depth image so as to use it for distance measurement, but to do so I pass it through image_proc package. That package doesn't work with the camera_info topic published from the ROS driver from ifm. Now if I run the camera calibration using python or ros, then I get a set of camera params which when published as a new topic work for image rectification.

Additionally, there is a script in documentation repo of ifm which has rectify function for python toolbox. When that is used the the depth image is rectified same as with new params and image_proc package.

I hope the explanation above was not too confusing.

I can share the new camera params if required.

[desengph]

Hello @sam-tj

The statement about non-published camera_info topic is correct. This is due to the calibration models used by the O3R heads:

• O3R225 heads use a fish eye model which may be supported out of the box by standard ROS image rectification pipelines
• O3R222 heads use a non-standard Bouguet model which is not supported out of the box by standard ROS image rectification pieplines

If I understand you correctly, you have prepared a separate ROS package which is based of the image rectification and 2D RGB to 3D TOF point cloud image mapping on ifm3d.com examples?

We would be very happy to review such a tool that is subpixel accurate based on the models and tools mentioned above. Feel free to tag me of @lola-masson in the respective GH repository / pull request.

Thank you.