Question about the quality of calibrations.

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Question about the quality of calibrations.

Mikor
Hi Mathieu!

I have a question, again. This time I want to learn more about the extrinsic calibration of between the sensors and how much accurate should they be. To be more specific I have an IMU & and a Lidar to calibrate, until now I've found this to be the most accurate calibration package so far but it does not provide any information about the translation as it says that the translations can be set as 0. And this made me wonder... would that be accurate enough for rtabmap ? Should I look for a 6DoF package? And finally is there any calibration package you recommend for 3d lidar - imu calibration and/or 3d lidar - 6DoF pose sensor (like T265) calibration?

If anyone has any recommendation for targetless calibration packages please share!
I've tried so far this by rectifying the T265 but I've had some problems and I stopped trying because the images are not the strong point of the IMU and I looked for techniques to calibrate the pose that the camera provides like this package, but the output was not consistent  and at the most cases not realistic. Then I used the package I mentioned at the previous paragraph. (That's my experience so far, it might be useful for someone.)  
Reply | Threaded
Open this post in threaded view
|

Re: Question about the quality of calibrations.

matlabbe
Administrator
Hi Mikor,

I think you have already more experience than me in lidar-imu calibration. Maybe someone else having tried similar approaches could come up and help you.

For rtabmap side, rtabmap uses imu only for orientation (the estimated quaternion). So if imu and lidar are fixed together, we don't care about translation between the sensors.

cheers,
Mathieu