How to show a single Volume Pixel by vol.legosurface?

[CCHLENG]

this example code can show this:

I add

data_matrix[1, 1, 40] = 3

and then

show([vol, lego], N=2, axes=1).close()

I get this:

point data_matrix[1, 1, 40] can be seen in left side but any Volume Pixel in right side.

if I use invert=True in vol.legosurface, this fig is shown. Point data_matrix[1, 1, 40] is missed if inverted. why this volume pixel can't be shown when invert=False?

how to show a single Volume Pixel by vol.legosurface? thanks!

[marcomusy]

Hi, this is a bit tricky but the value from the numpy array are not associated to the voxel center but to the vertices.
Note that in your last image you see 4 little cubes missing (not just one voxel or cell).
This normally is irrelevant as one deals with large dataset, but may become relevant for very small datasets.
See eg:

from vedo import *

data_matrix = np.zeros([70, 80, 90], dtype=float)
data_matrix[0:30,   0:30,  0:30] = 1.
data_matrix[30:50, 30:60, 30:70] = 2.
data_matrix[50:70, 60:80, 70:90] = 3.
data_matrix[1, 1, 40] = 3.

vol = Volume(data_matrix, c=['white','b','g','r'], mode=1)

print(vol.pointdata.keys())
print(vol.celldata.keys())

pts = Points([[0,0,0],[30,30,30], [50,60,70], [50,60,70]]).ps(10)

lego = vol.legosurface(1.5,3.5, invert=0, boundary=1)

show(vol.box(), lego, pts, axes=1).close()

[CCHLENG]

Thanks for your reply!

In your demo, boundary=1 in vol.legosurface make the little cube visible.
Enlighten by numpy array are associated to the vertices, I can use

data_matrix[0, 0, 0] = 3
 vol = Volume(data_matrix, mode=1, origin=[-0.5, -0.5, -0.5])

to show a single cube like this:

But I observed that the color of legosurface is error when I add lego.cmap('hot_r', vmin=1, vmax=3.5).add_scalarbar3d() after lego = vol.legosurface(1.0, 3.5, boundary=1) .The middle large dataset and the single cube at [0,0,0] should be yellow. What can I do to show the right color?

Actually, what I want to visualize is a 3D occupancy grid map which created from a points cloud.
For example, if there is any points in cube(x[-0.5,0.5],y[-0.5,0.5],z[-0.5,0.5]), the cube which center is [0,0,0] should be shown.
The final result may like this, many cube flock together, and some cube is single.
I realized that I also can draw it by box , but it might be hard to calculate every center and color of the box. Is there a easy way to show a 3D occupancy grid map?

[marcomusy]

Hi what about this:
examples/pyplot/histo_3d.py

[CCHLENG]

It's like what I want! Thanks!
The link in vedo gallery need to be fix, I have found that figure but the link is error. The 'D' of 'histo_3D.py' in the link should be lowercase.

In histo_3D.py ,

pcloud = Volume(counts).topoints()
marker = Cube()
glyphed_pcl = Glyph(pcloud, marker, color_by_scalar=True)

Glyph creates marker at every points of pcloud.
What's different between marker is cube and box? I found that when I set marker as box ,the surface is teared, but cube works as expect.
Ang how can I create pcloud directly? from my pointclould and value? In the demo, pcloud is created by .topoints() method, which return a vedo Points.

    ...
    def topoints(self):
        ...
        mpts = vedo.Points(v2p.GetOutput())
        return mpts
    ...

Which method should I use to set the value of every point?

[marcomusy]

Hi,
thanks I will fix the link!
You can create a pointcloud directly by passing 2D/3D coordinates, e.g.

from vedo import *

pts = np.random.rand(100,3)
pcloud = Points(pts)

# add any data to each point
pcloud.pointdata["myarray1"] = np.sin(pts[:,0])
# pcloud.pointdata["myarray2"] = np.sin(pts[:,1])

marker = Cube().scale(0.1)
# pcloud.pointdata.select("myarray1")
glyph = Glyph(pcloud, marker, scale_by_scalar=True)
glyph.cmap('rainbow').add_scalarbar('counts')

show(glyph, axes=1)

[CCHLENG]

It works! Thanks!
Do you have a donate link?

[marcomusy]

Do you have a donate link?

I will set up one! :)

Let me know if need more help.