I have a project where i have scanned indoors but got alot of points of the outside through the windows. After processing the Scans in SCENE 6.2 they are marked as scan point clouds. If i just open the 3d views of the Scans and use the polygontool to remove the Points of the outside that i dont want to be used in the cloud to cloud registration, is that enough or do i have to load the Scans and remove the Points? Because when running Cloud to Cloud registration they are loaded, more or less depending on the subsampling setting?
Does anyone who know which points SCENE are using in Cloud to Cloud registartion?
//Erik
Scene 6.2 Cloud to Cloud reg, which points are used?
- sealund
- Forum Supporter
- Posts: 114
- Joined: Fri Jan 08, 2010 12:47 pm
- 14
- Full Name: Erik Hagestrand
- Company Details: Bjerking AB
- Company Position Title: Surveyor - Laserscan technician
- Country: Sweden
- Linkedin Profile: Yes
- Location: Uppsala
- Contact:
- flohrd
- V.I.P Member
- Posts: 287
- Joined: Wed Jun 27, 2012 4:06 pm
- 11
- Full Name: Daniel Flohr
- Company Details: FARO
- Company Position Title: VP of Software Engineering
- Country: Germany
- Linkedin Profile: No
- Location: Stuttgart, Germany
- Has thanked: 2 times
- Been thanked: 7 times
- Contact:
Re: Scene 6.2 Cloud to Cloud reg, which points are used?
You can remove the points in the way you are describing as scans are updated to be in sync with the point clouds once you save the project.
The result is identical to doing it the other way around where you load the scans and delete the points. In this case the point clouds would be updated to be in sync with the scan.
So either way it will remove the points permanently and the registration will no longer use them.
The result is identical to doing it the other way around where you load the scans and delete the points. In this case the point clouds would be updated to be in sync with the scan.
So either way it will remove the points permanently and the registration will no longer use them.
- sealund
- Forum Supporter
- Posts: 114
- Joined: Fri Jan 08, 2010 12:47 pm
- 14
- Full Name: Erik Hagestrand
- Company Details: Bjerking AB
- Company Position Title: Surveyor - Laserscan technician
- Country: Sweden
- Linkedin Profile: Yes
- Location: Uppsala
- Contact:
Re: Scene 6.2 Cloud to Cloud reg, which points are used?
Ok, that´s good to know!
Thanks for fast reply!
//Erik
Thanks for fast reply!
//Erik
-
- V.I.P Member
- Posts: 401
- Joined: Mon Mar 23, 2009 4:04 pm
- 15
- Full Name: SAttaya
- Company Details: Sev1Tech
- Company Position Title: Sr Software Analyst-RemoteSensingEngr
- Country: USA
- Linkedin Profile: No
- Has thanked: 127 times
- Been thanked: 17 times
Re: Scene 6.2 Cloud to Cloud reg, which points are used?
Daniel,
Can you answer Erik's other ?:
and register those together.
But based on the number of points compared in the scanManager results,
I don't think that is correct.
Can you answer Erik's other ?:
I thought that C2C was basically find vertical and horizontal planesDoes anyone who know which points SCENE are using in Cloud to Cloud registartion?
and register those together.
But based on the number of points compared in the scanManager results,
I don't think that is correct.
- flohrd
- V.I.P Member
- Posts: 287
- Joined: Wed Jun 27, 2012 4:06 pm
- 11
- Full Name: Daniel Flohr
- Company Details: FARO
- Company Position Title: VP of Software Engineering
- Country: Germany
- Linkedin Profile: No
- Location: Stuttgart, Germany
- Has thanked: 2 times
- Been thanked: 7 times
- Contact:
Re: Scene 6.2 Cloud to Cloud reg, which points are used?
Cloud 2 Cloud actually uses all scan points and does not rely on special geometry!
It basically looks at all the scan points and minimizes the three-dimensional position error between all the points for each scan pair.
Afterwards, a bundle adjustment is applied to minimize error propagation.
It basically looks at all the scan points and minimizes the three-dimensional position error between all the points for each scan pair.
Afterwards, a bundle adjustment is applied to minimize error propagation.
-
- V.I.P Member
- Posts: 401
- Joined: Mon Mar 23, 2009 4:04 pm
- 15
- Full Name: SAttaya
- Company Details: Sev1Tech
- Company Position Title: Sr Software Analyst-RemoteSensingEngr
- Country: USA
- Linkedin Profile: No
- Has thanked: 127 times
- Been thanked: 17 times
Re: Scene 6.2 Cloud to Cloud reg, which points are used?
Daniel, so under ScanManager > scan point tensions > Used points
the number (typically 50K-70K-100K-120K) that shows is just
the number of points used in computing the statistics that are shown,
NOT the points used to do the C2C?
==================================================================================
<
Maybe I should have asked if the c2c was between multiple "adjoining" scans.
For example:
* 5 scans
-- 1 center scan (1/1)
-- 4 other scans (1/4) at 90°, 180°, 270° & 360°/0°
Assume:
* no overlap between the 1/4 scans that are 180° opposite each other
* 50% overlap between 1/1 to each of the other 4 scans
* 50% overlap between each 1/4 and it's two adjointing (+/- 90°) scans
so at a minimum
* 1/1 scan is adjacent to other 4 scans (4 C2Cs)
* scan at 90° is adjacent to 1/1, 360° and 180°, but not 270° (2 new C2Cs)
* scan at 180° is adjacent to 1/1, 90° and 270°, but not 360° (2 new C2Cs)
* scan at 270° is adjacent to 1/1, 180° and 360°, but not 90° (2 new C2Cs)
* scan at 360° is adjacent to 1/1, 270° and 90°, but not 180° (2 new C2Cs) (closes the loop)
does 6.2.2.7 do all of those C2Cs (4+2+2+2+2)?
is there some implied assumption of 4mm of noise/ranging error?
>
PS- thanks for the quick reply to my previous
the number (typically 50K-70K-100K-120K) that shows is just
the number of points used in computing the statistics that are shown,
NOT the points used to do the C2C?
==================================================================================
What is the bundle adjustment applied to if the c2c is between two (2) clouds (scans)?Afterwards, a bundle adjustment is applied to minimize error propagation.
<
Maybe I should have asked if the c2c was between multiple "adjoining" scans.
For example:
* 5 scans
-- 1 center scan (1/1)
-- 4 other scans (1/4) at 90°, 180°, 270° & 360°/0°
Assume:
* no overlap between the 1/4 scans that are 180° opposite each other
* 50% overlap between 1/1 to each of the other 4 scans
* 50% overlap between each 1/4 and it's two adjointing (+/- 90°) scans
so at a minimum
* 1/1 scan is adjacent to other 4 scans (4 C2Cs)
* scan at 90° is adjacent to 1/1, 360° and 180°, but not 270° (2 new C2Cs)
* scan at 180° is adjacent to 1/1, 90° and 270°, but not 360° (2 new C2Cs)
* scan at 270° is adjacent to 1/1, 180° and 360°, but not 90° (2 new C2Cs)
* scan at 360° is adjacent to 1/1, 270° and 90°, but not 180° (2 new C2Cs) (closes the loop)
does 6.2.2.7 do all of those C2Cs (4+2+2+2+2)?
is there some implied assumption of 4mm of noise/ranging error?
>
PS- thanks for the quick reply to my previous