Sphere based registration is not accurate, I need some help

Discuss FARO hardware here.
Post Reply
Augusto 3D
V.I.P Member
V.I.P Member
Posts: 436
Joined: Sat Dec 06, 2014 5:26 am
9
Full Name: C-Augusto
Company Details: 3d scanning
Company Position Title: USA
Country: US
Has thanked: 115 times
Been thanked: 30 times

Sphere based registration is not accurate, I need some help

Post by Augusto 3D »

Good morning,
I purchased a set of spheres from a reputable vendor and have been TRYING to use them to the best of our knowledge based on our research and FARO training (when we bought our scanner) and I personally cannot seem to get desirable results.

When using target-based registration, the spheres do not line up to produce round spherical objects, instead, they are off ever so slightly enough for walls and other important features not to align. However, if we forget about target-based registration and run C2C everything lines up (and the spheres are more offset from each other)

I understand that targets can move, however this particular time we used magnet mounts and mounted them to columns and places that would not be moving. The scan resolution was also high to capture plenty of points on the spheres.

The image below is what I am getting - notice how out of alignment the scans are on the round column shown - the same happens with other walls etc.
I am going to use C2C to get this project done since I verified our site measurement matches the results in the point cloud if we go this route.

I m just really annoyed to be honest, that we can't get the spheres and target to work to make things simple, especially considering that this is a proven workflow over the years.

I need some help and If someone feels confident enough and knows the subject in and out to provide some training on spheres, target, and control am happy to pay a reasonable fee for your assistance, send me a DM.

I am also wondering if it could be a scanner issue, maybe we need to do a site compensation - we were advised that we didn't need to go that by Faro since the unit is new and it is site-specific (the scanner is 2 months old). Could there be something wrong with our spheres?

help please, thank you
Screenshot 2022-06-18 152727.jpg
You do not have the required permissions to view the files attached to this post.
User avatar
Leandre Robitaille
V.I.P Member
V.I.P Member
Posts: 542
Joined: Sat Aug 03, 2019 1:53 am
4
Full Name: Leandre Robitaille
Company Details: Cima+
Company Position Title: Civil Technician - Surveyor
Country: Canada
Has thanked: 61 times
Been thanked: 246 times

Re: Sphere based registration is not accurate, I need some help

Post by Leandre Robitaille »

To me it could be a series of many factors;
-not enough spheres are used, you need at least 3 between each scans...I run 18 spheres PER scanner, this allows 6+ spheres detected between scans
-bad sphere postionning, you want large triangles, spread them appart within acceptable distances
-wrong sphere diameter is detected, change this in scene.
- mirrored spheres threw windows /false positives not deleted
- i would download taglabs extended toolbox app and delete badly detected spheres, use default values. This will get rid of any yellow spheres...or manually delete the bad spheres using ATS quality manager and open each scans sort by pts number and delete under 80 points.
- bad density used on scanner and spheres arent detected correctly. Outside you will want some 1/2 scans, often 1/4 is enough but outside you want larger sphere spread and 1/2 does help.
1/4 is 14m max for 145mm spheres
1/2 is about 25m max for 145mm spheres
I would not keep anything above 25m for spheres, use large checkerboards above 25m.
- Bad loops. You want to close loops, a long elipse you want spheres /targets in the middle to help balance everything. You can make c2c clusters to help close loops
- sphere moved on site (wind, fences, poles that move, etc)
- bad sphere positionning that hurts algorythm (in grass, threw fences, threw windows)
-scanner is out of calibration

Using targets requires some understanding of the limits of the device and it is threw trial and error you will perfect your on site techniques. Most important thing you want is closing loops, triangles with the targets and a good distance from the scans to spheres.
Augusto 3D
V.I.P Member
V.I.P Member
Posts: 436
Joined: Sat Dec 06, 2014 5:26 am
9
Full Name: C-Augusto
Company Details: 3d scanning
Company Position Title: USA
Country: US
Has thanked: 115 times
Been thanked: 30 times

Re: Sphere based registration is not accurate, I need some help

Post by Augusto 3D »

Thank you for the response. You have just validated that my process is no wrong, albeit I didnt use as many spheres but you can still see 3 or 4 and the scans were done at 1/2 -x2 or x4 (I forgot) since I also found that thats the best ROI with our plus series.

The scanner is new so I am not sure that it is out of calibration. I have never done an onsite comp. Is that something that I should have done?

Could the spheres be changing shape with the heat?

PS. What type of large checker boards do you use?
User avatar
landmeterbeuckx
V.I.P Member
V.I.P Member
Posts: 1616
Joined: Tue May 01, 2012 5:19 pm
11
Full Name: Lieven Beuckx
Company Details: Studiebureau Beuckx
Company Position Title: Owner
Country: Belgium
Linkedin Profile: Yes
Has thanked: 183 times
Been thanked: 548 times

Re: Sphere based registration is not accurate, I need some help

Post by landmeterbeuckx »

Check the 3d distances from 2 scans where you have the same spheres in it. This already gives some ideas.
LSBbvba
Surveying services - 3D Laserscanning
Tel : +32477753126
www.lsbbvba.be
[email protected]
User avatar
Leandre Robitaille
V.I.P Member
V.I.P Member
Posts: 542
Joined: Sat Aug 03, 2019 1:53 am
4
Full Name: Leandre Robitaille
Company Details: Cima+
Company Position Title: Civil Technician - Surveyor
Country: Canada
Has thanked: 61 times
Been thanked: 246 times

Re: Sphere based registration is not accurate, I need some help

Post by Leandre Robitaille »

Augusto 3D wrote: Mon Jun 20, 2022 1:10 am Thank you for the response. You have just validated that my process is no wrong, albeit I didnt use as many spheres but you can still see 3 or 4 and the scans were done at 1/2 -x2 or x4 (I forgot) since I also found that thats the best ROI with our plus series.

The scanner is new so I am not sure that it is out of calibration. I have never done an onsite comp. Is that something that I should have done?

Could the spheres be changing shape with the heat?

PS. What type of large checker boards do you use?
- I would run a accuracy check on the scanner before. This is how you do it ; https://knowledge.faro.com/Hardware/3D_ ... 0the%20bar.
Do the one with ''compare scans taken with different orientation''

-I don't know what spheres you are using, if its the Flexi spheres yes they can change size but it is not noticeble at temperatures human can survive at.

- As for ''large'' targets we had some custom 12'' stickers made here that we put on aluminium plates that we can hang/nail (I also have some that I put on the boxes we use to cary the spheres with), we just asked them to custom print the size https://tdr360.com/en/product-category/ ... r-scanner/
It's very rare we use these but I do find myself using them in those super odd projects where I want a link to close a loop but the shot is far, I do a 1/1 4x focus scan on them and can detect them up to 45ish meters away. Or I just set up the boxes for the spheres at the start around the project and I can detect them every now and then if I need more targets on a scan.
User avatar
msivil
I have made 80-90 posts
I have made 80-90 posts
Posts: 86
Joined: Tue Oct 19, 2010 10:12 am
13
Full Name: Marko Sivil
Company Details: Sivil 3d Oy
Company Position Title: CEO
Country: Finland
Linkedin Profile: Yes
Been thanked: 1 time
Contact:

Re: Sphere based registration is not accurate, I need some help

Post by msivil »

Hi,

If you created the spheres automatically, I would check the sphere fittings visually one by one.
Best Regards,
Marko Sivil

www.sivil3d.fi
Augusto 3D
V.I.P Member
V.I.P Member
Posts: 436
Joined: Sat Dec 06, 2014 5:26 am
9
Full Name: C-Augusto
Company Details: 3d scanning
Company Position Title: USA
Country: US
Has thanked: 115 times
Been thanked: 30 times

Re: Sphere based registration is not accurate, I need some help

Post by Augusto 3D »

Thanks for all the help guys, I am not too sure what I happnieng. I compared the position on some spheres from different scan locations (quick views) and there are essentially in the same place with less than 1/8" difference. See below:
What is condisered "too much" of a difference?

ST1 ST2
SCAN A -3632.201715 -521.231127 -2053.531601 -2936.501783 -793.515933 -2052.991495
SCAN B -3632.127294 -521.210718 -2053.555923 -2936.541948 -793.826836 -2052.636204

I have a question, if I set anti-correspondence for some spheres and I start to get scans failing, how do I associate them again?
User avatar
msivil
I have made 80-90 posts
I have made 80-90 posts
Posts: 86
Joined: Tue Oct 19, 2010 10:12 am
13
Full Name: Marko Sivil
Company Details: Sivil 3d Oy
Company Position Title: CEO
Country: Finland
Linkedin Profile: Yes
Been thanked: 1 time
Contact:

Re: Sphere based registration is not accurate, I need some help

Post by msivil »

I would recommend checking the sphere fittings visually one by one on 3d view.
Best Regards,
Marko Sivil

www.sivil3d.fi
User avatar
Leandre Robitaille
V.I.P Member
V.I.P Member
Posts: 542
Joined: Sat Aug 03, 2019 1:53 am
4
Full Name: Leandre Robitaille
Company Details: Cima+
Company Position Title: Civil Technician - Surveyor
Country: Canada
Has thanked: 61 times
Been thanked: 246 times

Re: Sphere based registration is not accurate, I need some help

Post by Leandre Robitaille »

Leandre Robitaille wrote: Mon Jun 20, 2022 11:02 am It's very rare we use these but I do find myself using them in those super odd projects where I want a link to close a loop but the shot is far, I do a 1/1 4x focus scan on them and can detect them up to 45ish meters away. Or I just set up the boxes for the spheres at the start around the project and I can detect them every now and then if I need more targets on a scan.
20220620_190935.jpg
Just came back to share my large target solution lol
You do not have the required permissions to view the files attached to this post.
INV_Scanning
I have made 40-50 posts
I have made 40-50 posts
Posts: 46
Joined: Mon Nov 11, 2019 10:18 am
4
Full Name: Sebastian Winkler
Company Details: INVERS Industrievermessung und Systeme
Company Position Title: Project Engineer
Country: Germany
Linkedin Profile: Yes
Location: Essen, Germany
Has thanked: 5 times
Been thanked: 6 times

Re: Sphere based registration is not accurate, I need some help

Post by INV_Scanning »

After looking at your target tensions I would give it another run without inclinometer in use. Especially for your scan ...001
Post Reply

Return to “FARO”