[50%] Histograms for cloud comparison

[pguth]

Hi,

An intern and I have been playing with Cloud Compare, and have been very impressed. We are looking to assess a lossy compression method for clouds. We have two independent clouds from OpenTopography.org (leaf on and leaf off for a region in Pennsylvania, and snow covered and snow free for a region in New Mexico), and have compressed both. Then I want to compare each of the four point clouds to the others--if the compressed versions differ no more than the two clouds from different times, then for many purposes the lossy compression is acceptable.

I have run all the comparisons in Cloud Compare, and make the following figure (reduced in size for the web)



The challenge is that the axes are different in each comparison (primarily the horizontal axis), making comparisons extremely difficult.

Would there be a way to force the x-axes to have the same scaling in a series of histograms like this? I know the source code is out there, but I code in Delphi (GIS program MICRODEM) and C++ is a bit foreign to me.

Peter Guth

[daniel]

Hi,

with the latest version, you can indirectly set the X boundaries as you wish:
- in the properties dialog of the point cloud, check the 'release boundaries' checkbox for the active scalar field
- then you can set the min and max 'displayed' values as you wish (by directly changing the value in the corresponding spinboxes). The histogram window uses these same values for the X axis.
--> so if you manually set common boundaries to all the scalar fields, you should get what you want (at least for the X axis).

The histogram display in CloudCompare is clearly not very advanced. For now, the best way to get much nicer results is to export the cloud(s) to an ASCII file and then import this file either in excel (with space as separator) or matlab. All scalar fields of a cloud are exported at once in the same ASCII file (as one column per SF).

Best,

[pguth]

Hi Daniel,

Thanks for the quick reply.

The export of the data works great. My files are too big for Excel, but I can import it into my GIS software, which in my view is better anyway.

The next time you're working with the code, I would like to put it a vote for the first line in the export file to have column names. I prefer that for Excel, Matlab, or the GIS, as it can be a drag to try to insert in a huge file (120 MB for the one I did, 1.4 million points). Not a big problem, but I think it could be a nice addition.

Peter Guth

[vanb@umd.edu]

Hi Daniel,

I'm attempting to find the checkbox you described above to 'release boundaries' in the properties dialog; however, I don't see it anywhere :( is this functionality available in the latest version of cloudcompare (2.11.alpha)? I'd like to change the x axis as you described, such that all of my histogram graphs have the same axes across different point cloud comparisons.

Thank you for the help!

Sarah

[daniel]

Ah! That disappeared a long time ago. You can now create your own custom and 'absolute' color scale that you can apply to all your clouds (see https://www.cloudcompare.org/doc/wiki/i ... es_Manager)

[abajwa28]

Hi Daniel,

Is there a way of exporting the Histogram as a Tiff file with higher resolution and DPI?

The attachments are the two different sizes of the histograms :

Rscan1 is the standard default export and the RIght loop is the export I made by maximizing the dialog box and then saving it.

But still it does not have the correct resolution or DPI that I would like. Any suggestions on how to export it with better quality?

Thank you

[abajwa28]

It is showing me errors to upload the pictures.

[daniel]

To make an export image with higher DPI, I guess if enlarging the dialog and copying it (with ALT+print screen on Windows for instance) then the only solution is to re-create it with Excel or any other tool (this is the purpose of the export to csv).