Mirrroring surface
-
- I have made 10-20 posts
- Posts: 18
- Joined: Fri Mar 20, 2015 11:06 am
- 9
- Full Name: Pier Carlo Ricci
- Company Details: Artificio Digitale Snc
- Company Position Title: Artificio Digitale Snc
- Country: Italia
- Skype Name: Pier Carlo Ricci
- Linkedin Profile: Yes
- Been thanked: 3 times
Mirrroring surface
Just a doubt: how about cleaning mirroring false-positive scans point in a mobile solution? In standard terrestrial I clean any single scans cut away bad portions (boring but very easy) ... in a mobile solution I don’t think is possible.
I really would see a data scan in a usual inside house with quite a mirror in every room
Piero
I really would see a data scan in a usual inside house with quite a mirror in every room
Piero
-
- I have made 10-20 posts
- Posts: 18
- Joined: Fri Mar 20, 2015 11:06 am
- 9
- Full Name: Pier Carlo Ricci
- Company Details: Artificio Digitale Snc
- Company Position Title: Artificio Digitale Snc
- Country: Italia
- Skype Name: Pier Carlo Ricci
- Linkedin Profile: Yes
- Been thanked: 3 times
Re: Mirrroring surface
really? Nobody can give me an idea about what happen when mobile slam pass nerby a mirror??
- 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: Mirrroring surface
When a mobile scanner passes a mirror you'll get the reflected points placed as if they were on the other side of the mirror rather than reflected by the mirror. The distance will be the sum of d1 = the distance from the scanner to the mirror plus d2 = the distance of the reflected object to the mirror. The point will be place on the line from the scanner to the mirror at a distance d1+d2 from the scanner. In the data, the mirrored data looks a bit like a window. It is more tricky to clean than mirrored data from static scanning, particularly if you don't have an option to get the scanner position for any given point in many systems. Most pragmatic solution would be to spray chalk or cover the mirrors prior to scanning. There could possibly be a noise signature in the distance and intensity that might help isolate these rogue points but it would be manufacturer dependent.
-
- I have made 10-20 posts
- Posts: 18
- Joined: Fri Mar 20, 2015 11:06 am
- 9
- Full Name: Pier Carlo Ricci
- Company Details: Artificio Digitale Snc
- Company Position Title: Artificio Digitale Snc
- Country: Italia
- Skype Name: Pier Carlo Ricci
- Linkedin Profile: Yes
- Been thanked: 3 times
Re: Mirrroring surface
exactly, point mirrored "data looks a bit like a window" but in a static scan data are easy to select and delete ... but using a mobile handled slam I think there a mass of point "like viewed trough" and, if I walk around the wall, they are added together good ones on the other side ... ins't it?
I think if there isn't a software trick to automatic delete that "mirro effects" in a usual civil house every walked scan is very chaotic data (there are quite a mirror every 2-3 rooms).
I think if there isn't a software trick to automatic delete that "mirro effects" in a usual civil house every walked scan is very chaotic data (there are quite a mirror every 2-3 rooms).
- 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: Mirrroring surface
Difficult to do from software even if you had the mirror outline as you need to link every scan point back to the scanner position to determine if has passed through a mirror. Most scan formats don't include this information for SLAM data, e.g. LAS, E57, Recap etc... Maybe ask the guys at Leica if they have a solution here. Using spray chalk or similar to cover the mirrors prior to scanning could be a more pragmatic solution.piero wrote: ↑Tue Mar 23, 2021 6:58 am exactly, point mirrored "data looks a bit like a window" but in a static scan data are easy to select and delete ... but using a mobile handled slam I think there a mass of point "like viewed trough" and, if I walk around the wall, they are added together good ones on the other side ... ins't it?
I think if there isn't a software trick to automatic delete that "mirro effects" in a usual civil house every walked scan is very chaotic data (there are quite a mirror every 2-3 rooms).
- Carbix
- V.I.P Member
- Posts: 236
- Joined: Sat Mar 16, 2019 1:36 am
- 5
- Full Name: Daniel Loney
- Company Details: Excelsior Measuring
- Company Position Title: Owner
- Country: Canada
- Linkedin Profile: Yes
- Location: Canada
- Has thanked: 33 times
- Been thanked: 50 times
- Contact:
Re: Mirrroring surface
Why not have the software define normals of a plane and clean past them.
Mirror solution, if you have 2 plane normals parallel to each other. Passing the opposing plane would result in deletion. As it’s technically impossible. You would need to set the plane variable as mirror or solid wall.
Window solution, same as mirror solution but with an offset variable for window pane location. This could also be assisted with an ultimates inspection of drastic intensity drop. Isolated intensity filter past normal plane.
Honestly all of this shouldn’t be that hard to code if your already working in this space... as a coder.
Mirror solution, if you have 2 plane normals parallel to each other. Passing the opposing plane would result in deletion. As it’s technically impossible. You would need to set the plane variable as mirror or solid wall.
Window solution, same as mirror solution but with an offset variable for window pane location. This could also be assisted with an ultimates inspection of drastic intensity drop. Isolated intensity filter past normal plane.
Honestly all of this shouldn’t be that hard to code if your already working in this space... as a coder.
-
- I have made 10-20 posts
- Posts: 18
- Joined: Fri Mar 20, 2015 11:06 am
- 9
- Full Name: Pier Carlo Ricci
- Company Details: Artificio Digitale Snc
- Company Position Title: Artificio Digitale Snc
- Country: Italia
- Skype Name: Pier Carlo Ricci
- Linkedin Profile: Yes
- Been thanked: 3 times
Re: Mirrroring surface
Thanks for replyies,
So the trouble is real... I’m very very glad if somebody can upload a screenshot of a mobile scan (blk2go or Geoslam, or so on...) passing in front of a mirror ... I think there is a lot of false positive point over the plane over the other good point of the rear room, pratical impossible to cut away by a user
So the trouble is real... I’m very very glad if somebody can upload a screenshot of a mobile scan (blk2go or Geoslam, or so on...) passing in front of a mirror ... I think there is a lot of false positive point over the plane over the other good point of the rear room, pratical impossible to cut away by a user
-
- I have made <0 posts
- Posts: 2
- Joined: Mon Mar 22, 2021 7:24 pm
- 3
- Full Name: Ivona Hubova
- Company Details: GeoSLAM
- Company Position Title: Solution Architect
- Country: UK
- Skype Name: hubova.ivona_1
- Linkedin Profile: Yes
- Been thanked: 1 time
Re: Mirrroring surface
Hi there. Mirrors will cause reflections. The advantage of using GeoSLAM handheld is that the system can be carried in a way that it does not face the mirrors directly. You could walk into a room with the back in or simply position the system lower/ higher so it would not scan the mirrors. It is possible to capture cleaner data this way. However, when the mirrors or other reflective surfaces are in the pointcloud, intensity can also help. Reflected points have lower intensity values which can be used for further data cleaning. Hope this helps. I can also provide some pointcloud examples by email.
-
- I have made 10-20 posts
- Posts: 18
- Joined: Fri Mar 20, 2015 11:06 am
- 9
- Full Name: Pier Carlo Ricci
- Company Details: Artificio Digitale Snc
- Company Position Title: Artificio Digitale Snc
- Country: Italia
- Skype Name: Pier Carlo Ricci
- Linkedin Profile: Yes
- Been thanked: 3 times
Re: Mirrroring surface
OK, I think the only way to solve mirrors is to “crop” a dataset based on acquisition time, “from ... to”, isolating moment passing in front of that... and this dataset would be easy to cut away false positive point mirrored ...
Anybody know if software can “crop” on acquisition time?
P.S. Some sample dataset would be great to test... most on alignment on complex walks...
Anybody know if software can “crop” on acquisition time?
P.S. Some sample dataset would be great to test... most on alignment on complex walks...
- 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: Mirrroring surface
I don't think cropping on time will work. For any given time range that includes the mirror you will also have many more scan points that aren't on the mirror, i.e. all the points in the spherical range of the scanner at that position. I have a couple of ideas for algorithms that could potentially work really well but would also need some test data, ideally including raw photography and instrument position matched to scan points.piero wrote: ↑Wed Apr 21, 2021 8:59 pm OK, I think the only way to solve mirrors is to “crop” a dataset based on acquisition time, “from ... to”, isolating moment passing in front of that... and this dataset would be easy to cut away false positive point mirrored ...
Anybody know if software can “crop” on acquisition time?
P.S. Some sample dataset would be great to test... most on alignment on complex walks...