Hi, I'm getting started with the point cloud. I am currently working on the Lecia RC360 scanner and mainly scan entire buildings. I start working with the CC program and try to use the Clean> SOR filter or Clean> Noise filter tool. Using these tools, the SOR filter removes most of the noise, but sometimes it also removes the corners of the rooms and even more often a large part of the facade. On the other hand, the Noise filter removes a lot of points but I don't see any difference, the noise is still there.
What advice do you have for this situation? What tool parameters to use or do you know better ways to clean up noise?
clearing the point cloud
-
- I have made <0 posts
- Posts: 1
- Joined: Mon Aug 01, 2022 7:47 am
- Full Name: Krystian Cholewa
- Company Details: Firma
- Company Position Title: pracownik
- Country: Polska
- Linkedin Profile: No
- Has thanked: 1 time
- Florentin.Drogeanu
- I have made <0 posts
- Posts: 4
- Joined: Wed Nov 08, 2017 1:43 pm
- 5
- Full Name: Florentin Drogeanu
- Company Details: Geomatics
- Company Position Title: Principal Surveyor
- Country: United Kingdom
- Linkedin Profile: Yes
- Has thanked: 4 times
- Been thanked: 3 times
- Contact:
Re: clearing the point cloud
If the scanned subjects are fairly regular in shape I would manually remove noise by using an Orthometric view and fence/polygon cutting it out. Once the bulk of it is done you should try the filter.
With the SOR filter, try to increase the number of points used in computing the mean distance (I would start with a minimum of 50) and run it a few times by working your way down from a higher standard deviation (maybe start from 1.5).
Whilst these filters are great for certain scenarios, most of the time they cannot be used as a one off and have to work in multiple passes with various parameters.
With the SOR filter, try to increase the number of points used in computing the mean distance (I would start with a minimum of 50) and run it a few times by working your way down from a higher standard deviation (maybe start from 1.5).
Whilst these filters are great for certain scenarios, most of the time they cannot be used as a one off and have to work in multiple passes with various parameters.