Scan into a space without GPS

Discuss FARO SCENE software here.
User avatar
steves01x
V.I.P Member
V.I.P Member
Posts: 544
Joined: Tue Jul 21, 2015 3:58 pm
8
Full Name: Steve Smith
Company Details: Scanning
Company Position Title: Surveyor
Country: Scotland
Linkedin Profile: Yes
Has thanked: 31 times
Been thanked: 57 times

Re: Scan into a space without GPS

Post by steves01x »

DPT wrote: Wed Dec 18, 2019 2:08 pm - If I need to Scan two spaces separated by a steel watertight bulkhead, how can assambly these scans with the others without references between them?
Scan all the way through the ship so both spaces are in the one scanning job?

Any other way to join 2 spaces with no common connection in overlap of scans or fixed targets visible from each would just be a guess and giving you the wrong thickness...
User avatar
jcoco3
Global Moderator
Global Moderator
Posts: 1724
Joined: Sun Mar 04, 2012 5:43 pm
12
Full Name: Jonathan Coco
Company Details: Consultant
Company Position Title: Owner
Country: USA
Linkedin Profile: No
Has thanked: 70 times
Been thanked: 157 times

Re: Scan into a space without GPS

Post by jcoco3 »

Jonathan,
The imu doesn't do the registration, this is done with voxels (triangles) in riscan. The exact functioning is not al clear to me but the path how you went from one position to another is pretty good, even walking in a bend results in good initial positioning of the next position.
I'm a user, not the develper of course and speak of my own experience.

i would think that this is also a company secret how they combine everything. I know it works flawlessly 95% of the time. I'm sure that there is drift in the imu but this is compensated by the brilliant registration methods

EDIT : check this small video out, you see the scanpositions move the way they are positioned in comarison with the earlier one. After that it uses voxels for final adjustements.
[youtube]https://www.youtube.com/watch?v=WHmDIs_-V4M[/youtube]
Lieven, Thank you for the info and the video, that is a good enough explanation for me :D It seems like there is nothing too unusual about that IMU implementation. I suspect they are using the IMU to roughly derive the position of the next scan from the relative know position of the prior scan. I would imagine they are only using the IMU data after a scan is done and then you pick up the scanner and start moving, so any drift that occurred while stationary in the first scan is not all that relevant when trying to approximate the relative position of the next scan. Here is an example of what I am thinking. Let's say the first scan position is always at 0,0,0. During a scan the imu is on the drift starts to occur, by the time the scan is done the scanner's imu position is drifted to the moon. As soon as the scanner is picked up the IMU position is reset to 0,0,0. Then as you walk it to the next position the imu approximates you position along the way to roughly 10,10,0, and the you set it down. Then the process repeats with the next reset going back to 10,10,0. If that is how it works, then it is not magic but still neat :geek:
Like you I am not certain that my explanation is accurate, but regardless their IMU system does seems to work. I would love to see how this IMU data alone would compares to GPS position after a many scans. My anticipation is that the stacked error would be high, but I just don't know.

Darío,
I apologize for side tracking your thread a bit, so I will add that I also agree with many of the recommendations here. You really have your choice of scanning methodologies. Some will recommend targeted scanning while some would go with cloud to cloud approaches, but both would benefit from some addition external survey control reference points once the area you are scanning gets big enough. Big enough depends on the scanner you use, the choice of the above mentioned scanning methodologies, your skill level, and the accuracy mark you are trying to hit.

As far as your deliema of the two spaces separate by a bulkhead. I don't think there are currently any viable alternatives to GPS that can provide your position inside of the ship from external sources such as radio waves/Wi-Fi or any other means. Slam based mapping, scan registration, survey control networks, and the like will all provide methods for establishing those internal positions, but will all likely results in some accumulation of error since a loop or registration cannot be closed through the bulkhead. You can minimize that error by incorporating multiple paths and loops around those two areas (and that will likely suffice), but it will never be as good as obtaining a true measurement path through the bulkhead. If the bulkhead was one solid piece of plate steel the I would say you could use a d-meter (ultrasonic thickness gauge) to at least obtain the thickness of the plate to confirm your registration, but I suspect their is an empty space between the bulkhead walls prohibiting the use of a d-meter.

Edit:
Basically what Steve just said, just in my own more complicated and elaborate way :lol:
User avatar
Peyman Bashiri
V.I.P Member
V.I.P Member
Posts: 427
Joined: Sun Dec 13, 2009 6:32 pm
14
Full Name: Peyman Bashiri
Company Details: Aero Geometrics Canadian Z+F reseller
Company Position Title: Laser scanning and BIM Specialist
Country: Canada
Skype Name: pbashiri2003
Linkedin Profile: Yes
Location: Vancouver,BC
Has thanked: 56 times
Been thanked: 40 times
Contact:

Re: Scan into a space without GPS

Post by Peyman Bashiri »

The C2C method combine with limited targeting is the to go for boats and tight spaces like that. Of course a ZF 5016 or RTC360 is more desirable tool for the job.

Peyman
You do not have the required permissions to view the files attached to this post.
Post Reply

Return to “FARO SCENE”