split point cloud based on scalar field

Feel free to ask any question here
Post Reply
matknaak
Posts: 38
Joined: Wed May 04, 2016 12:43 pm

split point cloud based on scalar field

Post by matknaak »

Hi Daniel,

hope this has not been asked before - I have a point cloud with scalar fields for dip, dipdirection and an numeric index indicating that a cluster of points belong to a single plane/surface (in this case a rock fracture). So there is now many clusters in this point cloud with an individual cluster index and identical dip and dipdirection values. Ultimately, I like to generate for each cluster only one point/value with the given scalar fields. Is it possible to seperate/split the point cloud into many smaler point clouds based on the scalar field (cluster index)?

regards

mat
daniel
Site Admin
Posts: 7382
Joined: Wed Oct 13, 2010 7:34 am
Location: Grenoble, France
Contact:

Re: split point cloud based on scalar field

Post by daniel »

Yes you can with the 'Edit > Scalar fields > Filter by value' tool (one cluster at a time). Or you can do it with the command line version (http://www.cloudcompare.org/doc/wiki/in ... _line_mode) if you have too many clusters.
Daniel, CloudCompare admin
matknaak
Posts: 38
Joined: Wed May 04, 2016 12:43 pm

Re: split point cloud based on scalar field

Post by matknaak »

Hi Daniel,

thanks for your quick reply and sorry for the delay - I had to test your suggestions and had some administration problems when CC safes data from command line version. However, yes, I do have many clusters. So if I follow your suggested method, I wil generate a lot of seperate files. But what I realy want is a single point for each cluster or the statistics for each cluster. What I tried in the mean time is to use 'tools' + 'segmentation'+ 'lable connected comp.' with a high value for the octree level to seperate the clusters into seperate point clouds, and then select all these seperate point clouds and use 'edit' + 'subsample' with the largest possible min. space between points to generate a single point for each cluster. It seems to work quite well except that I am not sure if the 'lable connected comp.' segmentation tool will strictly seperate my clusters. Maybe for future it would be cool to have a segmentation tool based on scalar fields.
However, great software - there is allways a way to solve the problem.
cheers

mat
Post Reply