Display normals for point cloud and colour each point based on a third value

[jkissi]

Hi!
I want to display the normals for a point cloud and modify the colours of the points based on a third value (the mean curvature), and add a colourbar. I have the points, normals and curvature already, just need some instruction on how to put it all together. If you could give an example, that would be great also!

Thanks
J

[marcomusy]

E.g.:

import vedo
import numpy as np

pts = vedo.Points(np.random.randn(1000000,3))
pts.pointdata['Curvature'] = pts.points() # whatever, i'm making this up!

# add an array which corresponds to the z component
pts.pointdata['ZCurvature'] = pts.pointdata['Curvature'][:,2] # numpy array

pts.cmap('jet', "ZCurvature").addScalarBar()

vedo.show(pts, axes=1)

[jkissi]

Thanks for your response. Your example helped me alot and made me realise that my curvature estimation function isn't working. Hahaha, so thanks for your help there!

Was wondering if there's a way to display a numpy array of point normal values? I tried using the NormalLines() function for meshes to generate normals and display them:

    vpts1 = ve.Points(points, r=3)
    vpts2 = ve.Mesh(points)
    norms = ve.NormalLines(vpts2, atCells=False)
    vpts1.pointdata['Curvature'] = avg_curv
    vpts1.cmap('jet', "Curvature").addScalarBar()
    ve.show(vpts1, norms, axes=1).close()

and I got this:

You can see that all the normals are sideways. Any idea why this could be happening? If you can suggest a way to add and display them manually, I can add pre-calculated normals that I know are correct. As always, cheers in advance for you help.

Cheers
j

[marcomusy]

HI, just use your numpy vectors of normals and add them to the point positions, then use Lines().
let me know if you need help,
marco

[jkissi]

Hi
Think I managed to get it down:

Here's the code:

    vp = ve.Plotter(bg='black', sharecam=1)

    # generate to the Points obj
    vpts = ve.Points(points,  r=3)

    # generate normal lines
    norms = ve.Lines(vpts, normals, alpha=0.25, scale=0.02)

    vp.show(vpts, norms, axes=1).close()

I think It's pretty efficient looking.... but as always, any suggestions welcome.

Cheers!!
J

[marcomusy]

looks soft and fluffy i love it :)

so what you call normals should be vpts + normals i guess..

one thing you can do is to use vmin and vmax in cmap() to adjust the range of the scalar visualization if needed..