Cloud-to-Cloud Distance

[vered waksman]

Hi, I wanted to ask some questions about this tool

1. If my "compared" cloud is denser then the "reference" one -
A. If I use "General parameters" - how are the points with no "matching" points in the reference cloud are taken into account?
B. Would it be always recommended using "local modeling"?
2. I am most interested in Z axis - therefor I want to use the "split x, y, ...". Is there a way of exporting this data of distances of each point to her reference point in a CSV file? when I try to press the CSV button I get the attached error message.

Thank you in advanced,
Vered.

[daniel]

1A. They are matched to the nearest neighbor (which can be quite far however). You can play on the 'max distance' parameter to avoid looking for points too far away
1B. Yes definitely. Another option is to use the more robuste M3C2 plugin (see wiki)

2. You can export the cloud (with its distance scalar fields) as an ASCII file. Setting the export parameters correctly, it can correspond to a nice CSV file, with the distances in separate columns.

[vered waksman]

Hi Daniel! Thank you for your response :)

Generally I am trying to compare my generated point cloud to a point clout exported from a Lidar sensor (my reference).
The parameter that interests me mostly is the "distance" to objects (Automotive field) therefor I thought looking at Z axis will be as close to what I am looking for.
My PC is much denser than the reference PC therefor I user "local modeling", "Quadric",
1. Should I use the default radius?
2. Looking at the results in "C2C absolute distance (Z) meaning - the distance between each point to her reference in Z axis? (because the numbers do not match what is seen with the human eye - see attachment)

3. Grayed points are points without matching with reference cloud?
4. Is the Scale in meters?

Thank you very much for your help!
Vered.

[daniel]

1. Maybe not, the default value is really a default one, base don the cloud bounding-box. Depending on your density you may have to increase or decrease it (hard to tell just by looking at this snapshot).

2. That's the Z component of the vector between the 2 points: the point from the compared cloud, and it's nearest point in the other PC. Which is not necessarily the nearest point along Z! The C2C distance algorithm is quite "simple". Maybe using M3C2 would help here (you can look for the nearest points along normal directions, hence you can control this direction by tweaking your normals, or choosing purely vertical or horizontal normals, etc. - rotating the clouds might help as well).

3. In this case it would be more the points for which it was not possible to compute a local model (i.e. the radius was too small)

4. It will be in the same units as your cloud coordinates (units are sadly never saved in most PC files)

[vered waksman]

Hi
Thank you for your detailed answers :) I am using the M3C2 plugin as you advised.

Another question - is there an option to display a continues movie on CC?
If not, is there another program you know and recommend that enables presenting two PC in continues frames?

Thank you and best regards!
Vered.

[daniel]

Well, not sure to fully understand your need, but there's the qAnimation plugin? (https://www.cloudcompare.org/doc/wiki/i ... n_(plugin))

(For now, you'll have to use the 2.11.3 version as it seems there's a small issue when rendering animations with the 2.12.alpha version)

[vered waksman]

Thanks, actually I believe I didn't explained myself well.
I am looking for an option to display few frames in a row - for instance of a video I filmed using Lidar sensor.
Is there an option to load frame after frame and display it?