Good Overlap
-
- I have made 10-20 posts
- Posts: 16
- Joined: Wed Jun 27, 2018 7:48 am
- 5
- Full Name: nhoy lee
- Company Details: nta surveying
- Company Position Title: surveyor
- Country: qatar
- Linkedin Profile: No
Good Overlap
Hi Guys,
May I know what is Good overlap means in laser Scanning and how to check it in the site if you had a Good overlap
May I know what is Good overlap means in laser Scanning and how to check it in the site if you had a Good overlap
- Matt Young
- Honorary Member
- Posts: 3929
- Joined: Fri Nov 30, 2007 3:03 pm
- 16
- Full Name: Matt Young
- Company Details: Baker Hicks
- Company Position Title: CAD-BIM Lead
- Country: UK
- Linkedin Profile: No
- Has thanked: 45 times
- Been thanked: 40 times
Re: Good Overlap
In general terms it's the amount of overlap between each scan. Ideally you would want at least 20 to 25% of overlap between each scan, in order to run cloud to cloud registration algorithms. This applies to all scanners and software generally.
If you don't see that there is nothing, then you are kidding yourself.
-
- I have made 10-20 posts
- Posts: 16
- Joined: Wed Jun 27, 2018 7:48 am
- 5
- Full Name: nhoy lee
- Company Details: nta surveying
- Company Position Title: surveyor
- Country: qatar
- Linkedin Profile: No
Re: Good Overlap
Thank You Matt, but please tell me the checking points and how the overlap justified before leaving the site
- Scott.Warren
- V.I.P Member
- Posts: 364
- Joined: Tue Jan 30, 2018 3:14 pm
- 6
- Full Name: Scott Warren
- Company Details: WSP Canada
- Company Position Title: Project Manager - Geomatics Engineer
- Country: Canada
- Linkedin Profile: Yes
- Location: Dartmouth, Nova Scotia, Canada
- Has thanked: 43 times
- Been thanked: 111 times
- Contact:
Re: Good Overlap
Heyo,
It depends on the equipment you use, as some equipment may have a screen or companion tablet that you can visually check to see that you have 20%-25% of common data between scans. If your scanner doesn't have a way to view the data you collect in the field, then it comes down to experience. I would recommend for each static scan taken, you place a mark (paint, tape, chalk, etc) and write the scan number where the scanner was setup on the ground. It helps visualize what information was collected, and helps avoid lack of overlap, or missing key features of your site.
The goal is to collect 20%-25% of the same data that was collected by 1 or more other scans, for each scan you take.
Cheers,
Scott
It depends on the equipment you use, as some equipment may have a screen or companion tablet that you can visually check to see that you have 20%-25% of common data between scans. If your scanner doesn't have a way to view the data you collect in the field, then it comes down to experience. I would recommend for each static scan taken, you place a mark (paint, tape, chalk, etc) and write the scan number where the scanner was setup on the ground. It helps visualize what information was collected, and helps avoid lack of overlap, or missing key features of your site.
The goal is to collect 20%-25% of the same data that was collected by 1 or more other scans, for each scan you take.
Cheers,
Scott
- Daniel Wujanz
- V.I.P Member
- Posts: 227
- Joined: Fri Aug 24, 2018 11:26 am
- 5
- Full Name: Daniel Wujanz
- Company Details: technet GmbH
- Company Position Title: 3D Laser Scanning Specialist
- Country: Germany
- Linkedin Profile: Yes
- Location: Berlin
- Has thanked: 206 times
- Been thanked: 188 times
- Contact:
Re: Good Overlap
Dear Lee,
I should add that just having 20 to 25% of overlap between point clouds is not enough. What you also need to compute transformation parameters by ICP-algorithms or plane-based approaches like Scantra is "geometric contrast". That means that the overlapping region has to contain geometric information that is distributed in at least three cardinal directions. Otherwise you will not be able to register these datasets.
Here's an example:
Let's say you've scanned a very long straight rectangular shaped tunnel. That means you can only use the walls, the ceiling and the floor for registration. Even though you may have an overlap of 50% or more you will still not be able to register these datasets since one, so called degree of freedom, cannot be solved - namely the translational component along the tunnel's axis. You can imagine this effect by folding two pieces of paper in a way that they form a right angle. If you now place these pieces of paper on top of each other, you can still shift them along the folded edges. In technical terms the ICP would converge in a local minimum - which yields in misaligned point clouds.
Cheers
Daniel
I should add that just having 20 to 25% of overlap between point clouds is not enough. What you also need to compute transformation parameters by ICP-algorithms or plane-based approaches like Scantra is "geometric contrast". That means that the overlapping region has to contain geometric information that is distributed in at least three cardinal directions. Otherwise you will not be able to register these datasets.
Here's an example:
Let's say you've scanned a very long straight rectangular shaped tunnel. That means you can only use the walls, the ceiling and the floor for registration. Even though you may have an overlap of 50% or more you will still not be able to register these datasets since one, so called degree of freedom, cannot be solved - namely the translational component along the tunnel's axis. You can imagine this effect by folding two pieces of paper in a way that they form a right angle. If you now place these pieces of paper on top of each other, you can still shift them along the folded edges. In technical terms the ICP would converge in a local minimum - which yields in misaligned point clouds.
Cheers
Daniel
-
- V.I.P Member
- Posts: 183
- Joined: Mon Apr 27, 2009 3:38 pm
- 14
- Full Name: GrahamM
- Company Details: AVEVA
- Company Position Title: Software Engineer
- Has thanked: 1 time
- Been thanked: 6 times
Re: Good Overlap
I guess it depends on what you are trying to deliver a point cloud that registers or one where the site has been adequately captured.
- Daniel Wujanz
- V.I.P Member
- Posts: 227
- Joined: Fri Aug 24, 2018 11:26 am
- 5
- Full Name: Daniel Wujanz
- Company Details: technet GmbH
- Company Position Title: 3D Laser Scanning Specialist
- Country: Germany
- Linkedin Profile: Yes
- Location: Berlin
- Has thanked: 206 times
- Been thanked: 188 times
- Contact:
Re: Good Overlap
Overlap usually refers to the common area covered by two or more scans while completeness refers to... well the level of completeness (or adequately captured as you called it) ; )
- smacl
- Global Moderator
- Posts: 1409
- Joined: Tue Jan 25, 2011 5:12 pm
- 13
- Full Name: Shane MacLaughlin
- Company Details: Atlas Computers Ltd
- Company Position Title: Managing Director
- Country: Ireland
- Linkedin Profile: Yes
- Location: Ireland
- Has thanked: 627 times
- Been thanked: 657 times
- Contact:
Re: Good Overlap
Good explanation on the need for distinct geometry in overlapping area, though only a requirement if you're registering using a feature matching algorithm. i.e. not a problem if you have targets which would hopefully usually be the case in tunnels.Daniel Wujanz wrote: ↑Fri Sep 14, 2018 10:17 am Overlap usually refers to the common area covered by two or more scans while completeness refers to... well the level of completeness (or adequately captured as you called it) ; )
- Daniel Wujanz
- V.I.P Member
- Posts: 227
- Joined: Fri Aug 24, 2018 11:26 am
- 5
- Full Name: Daniel Wujanz
- Company Details: technet GmbH
- Company Position Title: 3D Laser Scanning Specialist
- Country: Germany
- Linkedin Profile: Yes
- Location: Berlin
- Has thanked: 206 times
- Been thanked: 188 times
- Contact:
Re: Good Overlap
True that, Shane! However, it would be rather unusual to report the overlap on the targets' surface.
If you're really into using targets you probably would report which targets are visible from a certain viewpoint in order to ensure that you're able to register the lot later on.
Cheers
Daniel
If you're really into using targets you probably would report which targets are visible from a certain viewpoint in order to ensure that you're able to register the lot later on.
Cheers
Daniel