Point Cloud Curve-Help!
-
- I have made <0 posts
- Posts: 2
- Joined: Thu Jul 30, 2020 8:53 pm
- 3
- Full Name: Colin KAiruz
- Company Details: Linea
- Company Position Title: Director
- Country: South Africa
- Linkedin Profile: No
Point Cloud Curve-Help!
Hi Everyone,
I have a two BLK 36o's that we have been using for over two years now and recently the point clouds have started to come our "curved" as in the images attached. It seems to be more prominent when many scans are required in a straight line (like a building facade or long passage), but all of our point clouds are being registered this way.
Has anyone experienced anything similar to this before?
We use Recap Mobile to capture the data in the field on our ipads and then Recap Pro for the importing and registration.
Regards
Colin
I have a two BLK 36o's that we have been using for over two years now and recently the point clouds have started to come our "curved" as in the images attached. It seems to be more prominent when many scans are required in a straight line (like a building facade or long passage), but all of our point clouds are being registered this way.
Has anyone experienced anything similar to this before?
We use Recap Mobile to capture the data in the field on our ipads and then Recap Pro for the importing and registration.
Regards
Colin
You do not have the required permissions to view the files attached to this post.
- jcoco3
- 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: Point Cloud Curve-Help!
Hi Colin,
Yes, all of us have experienced this. This appears to be just a very common registration error due to "stacking." Old surveyors would probably call this "breaking chain." This is precisely why long linear chains of scan registration meander and curve. You can try to combat the problem by closing numerous loops creating a network that your registration software may be able to bundle adjust in an attempt to compensate, but I think many people here would agree that it would be easier and probably better to add some form of survey control when capturing larger projects like this. Even if you get something that is flat or appears to be correct from your registration software without some additional control it will be exceedingly difficult to know whether or not the entire project is globally accurate. I don't mean globally as in where it sits on the planet, but for example the entire project could be tilted by many inches or even feet on the ends.
This is a brief explanation, but I am sure others will chime and point you towards the revelavant educational material.
Hope this helps
Yes, all of us have experienced this. This appears to be just a very common registration error due to "stacking." Old surveyors would probably call this "breaking chain." This is precisely why long linear chains of scan registration meander and curve. You can try to combat the problem by closing numerous loops creating a network that your registration software may be able to bundle adjust in an attempt to compensate, but I think many people here would agree that it would be easier and probably better to add some form of survey control when capturing larger projects like this. Even if you get something that is flat or appears to be correct from your registration software without some additional control it will be exceedingly difficult to know whether or not the entire project is globally accurate. I don't mean globally as in where it sits on the planet, but for example the entire project could be tilted by many inches or even feet on the ends.
This is a brief explanation, but I am sure others will chime and point you towards the revelavant educational material.
Hope this helps
-
- I have made <0 posts
- Posts: 2
- Joined: Thu Jul 30, 2020 8:53 pm
- 3
- Full Name: Colin KAiruz
- Company Details: Linea
- Company Position Title: Director
- Country: South Africa
- Linkedin Profile: No
Re: Point Cloud Curve-Help!
Hi Jcoco,
Thanks for your feedback, it is greatly appreciated.
How would i go about doing this..using targets?
Thanks for your feedback, it is greatly appreciated.
How would i go about doing this..using targets?
- jcoco3
- 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: Point Cloud Curve-Help!
You might have to clarify what you mean in your question for me a bit, because it could mean two different things. Yes, you could add a bunch of targets and improve the registration without adding survey control, but my previous comments about global error would remain.
My preference would be to throw up 3 to 5 checkerboards every so often and shoot them in with a total station for control. The frequency in which you do so is dependant on how fast your scanner and scanning situation accumulates error balanced against the accuracy needs of your project. I would love to give you a general rule of thumb like every couple of hundred feet, but that is dangerous and leaves out too much needed information. With a modern Faro scanner (like we use), under moderate accuracy requirements for simple modeling practices, every couple hundred feet would likely work. With a more accurate scanner I might would be more confident in going farther. I don't have any experience with the blk360, but everything I do know about it would suggest that the distance in question might need to be shortened a good bit.
Nothing beats testing this out for yourself and verifying your own results...not to mention its fun.
My preference would be to throw up 3 to 5 checkerboards every so often and shoot them in with a total station for control. The frequency in which you do so is dependant on how fast your scanner and scanning situation accumulates error balanced against the accuracy needs of your project. I would love to give you a general rule of thumb like every couple of hundred feet, but that is dangerous and leaves out too much needed information. With a modern Faro scanner (like we use), under moderate accuracy requirements for simple modeling practices, every couple hundred feet would likely work. With a more accurate scanner I might would be more confident in going farther. I don't have any experience with the blk360, but everything I do know about it would suggest that the distance in question might need to be shortened a good bit.
Nothing beats testing this out for yourself and verifying your own results...not to mention its fun.
- Leandre Robitaille
- 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: Point Cloud Curve-Help!
I havent try this on a blk, but I think there is no way to "break"a cloud 2 cloud approach such as a blk360 survey with total station surveyed points. It will most likely do the U shaped from the c2c, then do a best fit translation to those coordinates. Only way I know now to control the error path with a any 3d scan is to do a survey with targets between each scans (or many clusters of small c2c registration). You might want to consider spheres as well.
Or, your registration might just be horribly bad and you might want to rework that.
Or, your registration might just be horribly bad and you might want to rework that.
-
- V.I.P Member
- Posts: 352
- Joined: Thu Jan 14, 2016 12:26 pm
- 8
- Full Name: Simeon Herrod
- Company Details: Terra Measurement Limited
- Company Position Title: 3D Data Manager
- Country: UK
- Linkedin Profile: No
- Has thanked: 73 times
- Been thanked: 110 times
Re: Point Cloud Curve-Help!
Survey control with targets and/or spheres.
For something linear I don't like to go much further than 25m between groups of 3-6 targets and for a square/rectangular building exterior, I like targets on at least two of the corners, with at least one scan on the corner catching two walls.
Probably overkill, but I know it works for us and we rarely have any registration issues.
I can think of only a handful of jobs in the last 5 years where our clients haven't wanted everything related to Ordnance Survey coordinates, so survey control is part of our normal workflow.
For something linear I don't like to go much further than 25m between groups of 3-6 targets and for a square/rectangular building exterior, I like targets on at least two of the corners, with at least one scan on the corner catching two walls.
Probably overkill, but I know it works for us and we rarely have any registration issues.
I can think of only a handful of jobs in the last 5 years where our clients haven't wanted everything related to Ordnance Survey coordinates, so survey control is part of our normal workflow.
- tadol
- V.I.P Member
- Posts: 126
- Joined: Wed Dec 14, 2011 12:32 am
- 12
- Full Name: Tad Laird
- Company Details: Koppa Target Spheres
- Company Position Title: Owner
- Country: USA
- Linkedin Profile: No
- Location: Oakland, California, USA
- Has thanked: 2 times
- Been thanked: 4 times
- Contact:
Re: Point Cloud Curve-Help!
I’ve been making a simple mini-prism mount that aligns with the center point on our targets, and mounts identically, either by magnet or ⅝” thread - it’s simple, fast, and quite inexpensive. If you need target spheres and a survey-capable solution, send me an email and I’d be glad to help you get what you need -
http://www.KoppaTargets.com
Inexpensive Laser Scanning Targets & Accessories
for Architecture, Engineering, & Forensics
"We're right on targets!"
Inexpensive Laser Scanning Targets & Accessories
for Architecture, Engineering, & Forensics
"We're right on targets!"
- stevenramsey
- Global Moderator
- Posts: 1937
- Joined: Sun Aug 12, 2007 9:22 pm
- 16
- Full Name: Steven Ramsey
- Company Details: 4DMax
- Company Position Title: Technical Specialist Scanning
- Country: UK
- Skype Name: steven.ramsey
- Linkedin Profile: Yes
- Location: London
- Has thanked: 30 times
- Been thanked: 72 times
- Contact:
Re: Point Cloud Curve-Help!
I would try using Cyclone reg360 rather than recap. Recap isn't known for good strong registrations
-
- V.I.P Member
- Posts: 544
- Joined: Fri Feb 24, 2017 10:47 am
- 7
- Full Name: Martin Graner
- Company Details: PointCab GmbH
- Company Position Title: Research and Development
- Country: Germany
- Linkedin Profile: No
- Has thanked: 160 times
- Been thanked: 175 times
- Contact:
Re: Point Cloud Curve-Help!
I can only recommend to read the taming errors series from Daniel, he tackles this in a part and explains why this can happen:ColinInCapeTown wrote: ↑Fri Jul 31, 2020 9:04 pm Hi Jcoco,
Thanks for your feedback, it is greatly appreciated.
How would i go about doing this..using targets?
viewtopic.php?f=153&t=14984
There are currently 7 parts.