Is there a plug-in for LiDARView that will allow IMU/INS data to integrate with the LiDAR data to increase accuracy?
Also has loop closure been implemented yet in LiDARView?
Pretty new to this, I was looking through the plug-ins today but didn’t see anything that hinted at IMU/INS.
Although in the SLAM webinar from last year this was mentioned as a project in progress.
Thanks for your interest in those topics.
IMU/INS integration in our SLAM library is still a work in progress.
In our latest SLAM release, we have integrated the use of external poses as tight optimization contraint.
Raw IMU data imntegration is not yet present in that release, but is a work in progress.
This SLAM release will be integrated quite soon in the next LidarView release ( 4.3-RC4 ).
Loop closure ( manual for now, meaning we have to select which frames are corresponding to neighbouring locations ) is also a work in progress, but not yet part of that release.
Auto-detection of loop closure will then be a next step.
Thanks for the response. It is great to know that IMU/INS integration is in the works! I would like to experiment with using a small INS such as the RUG-3-IMX-5-DUAL or even something more affordable such as the VectorNav VN-100. To integrate with VLP-16 time clock via PPS NMEA for location correction/collaborations.
Do you have a rough estimate on when the new release of Lidarview might have those features? I wish I was a coder and could help but I am just an engineer very interested in this tech and building my own SLAM scanner. Thanks again!
You can find the last LidarView release (v4.3rc4) with SLAM enabled here.
Awesome thank you Timothée! I appreciate that!
If you process the IMU to get the external poses, it is already doable.
It’s unclear from the RUG-3-IMX-5 documentation, but it seems if you connect a GNSS, you get embedded sensor fusion with X,Y,Z,roll,pitch,yaw output. But maybe only the highend RUG-3-IMX-5-RTK does it.
Bastien Jacquet, PhD
Former LidarView creator & Lead,
Now helping companies deliver their R&D projects @ Perception4D