We are creating this topic to ask you about your opinion regarding calibration approaches and to compare notable approaches!
You can find the survey here: https://forms.gle/zdQj2Hk282vWF4Dm6
To contextualize, we are the developers of ATOM, which is a set of calibration tools for multi-sensor, multi-modal, robotic systems, based on the optimization of atomic transformations as provided by a ROS based robot description. Moreover, ATOM provides several scripts to facilitate all the steps of a calibration procedure. We also gave a ROSCon 2022 talk.
We are conducting this survey to try to figure out if ATOM is helpful for users to calibrate their systems.
**If you did not use ATOM before [you definitely should :-)], you can still give us helpful information by answering the survey. **
It should take less than 5 minutes to complete this survey.
Thanks a lot for your important effort! I definitely have ATOM on my todo list. I really liked the list of calibration tools you’ve compiled for the survey, so I re-post it here so that it does not get forgotten once the survey is closed =)
maybe slightly off-topic but is there any calibration solution for finding the extrinsic calibration of a camera (or two if that helps) on a mobile robot (not a robotic arm), preferably on ROS2? If not, even just a paper/algorithm to achieve this would be appreciated. Thanks!
We are currently working on the expansion of ATOM to that case. Theoretically, ATOM should be able to solve the problem, so we started running experiments since last year. The tests we carried out were successful, which means ATOM is able to solve this problem with several sensors, not just a single RGB camera.
Specifically, ATOM can estimate a transformation between a sensor and the base_link of a mobile robot.
Right now we are writing a paper on this topic to submit in the near future, so I can’t go into details.
However, if you would like you can create an issue in ATOM and we will try to guide you on how to calibrate for that case.
Apart from ATOM I think this package is also aimed at this problem:
@miguelriemoliveira thank you very much for your answer and offer to help! I will take a look at the package and likely take you up on your offer of writing an issue in ATOM for more guidance.