query point normals

[ThomasD]

Hi Daniel,

Perhaps I've missed something, but is it possible to play on the normal rendering and querying?

This follows up on geological plane attitude computation.
Playing on the colour rendering of the point normal would enable visualizing plane families easily. A practical colour scheme for doing this is the HSV colour space. The geological strike info is attached to H colour coordinate and dip info attached to S coordinate. The V colour coordinate may be left alone or manually set to a constant value.

The logical next step would be to querying point clouds for a given orientation, and offer to put selected points in a subset. The query may not always result in subsets, it may just be to isolate plane sets visualy. Before extracting, it may also be useful to dynamically paint points with a "target" colour bar using the plane parameters. Again use HSV colour bar for representing subset candidates: H set to "red" (0) and "S" set to the tolerance parameter in reverse (saturated 1 is close to given values and desaturated is far from given values). Beyond the tolerance don't display the points. These methods are already dynamically implemented in the properties control pannel. It would "just" need porting to a dialog box.

Geologists will appreciate if the normal query parameters (desired dip and strike each with two tolerance values attached) were presented in a geologist-readable format, rather than unit vector components. The same numerical conversion as for displaying geological planes info applies.

The naming of the point subset could be called something like "dip(+/- x) - strike(+/-y)" e.g. "N088°E(+/-2°) - 052°SE(+/- 5°)", just like you named geological planes but adding both tolerance infos.

Cheers,

Thomas

[daniel]

Hi Thomas,

this shouldn't be too complicated and I kinda like this idea. This tool would not only be useful for geologists, but also for other people. Thanks to EDF support, CloudCompare should get several new features in the very near future (May/June 2012). If I have enough time I'll try to add such a tool in the TODO list.

By the way, I was also about to implement one of your old request (to get the distance components in 3 different scalar fields, one for each dimension). Maybe the vector field also.

[ThomasD]

Hi Daniel,

Thanks for this. And how about using the V value in HSV coordinates to emphasize the interactive selection visualization. Tune down unselected points by diminishing their V value to a level like 0.5, while selected points keep a value of 1.

One thing bothers me a little about the H colour continuum. The human eye has little change perception between some of the colours. Looking into a perceptual colour scheme might be useful. But this is something for when the tool works well and we (I) understand a little better about perceptual colour detection in the xy CIE Lab colour space. Any thoughts along those lines?

All the best,

Thomas

[daniel]

Hi Thomas,

Finally, I've been really short on time so I couldn't do everything you suggested. I only implemented the very first part (converting normals to HSV) in the last online version (06/24/2012). But the remaining part should follow ... soon?