Resampling and ploting multiple slices from a vtkPolydata

Support
1

I have a vtkPolydata with 3D coordinates of the vertex of the grid as points and Temperature at each point stored as point field data. I want to plot the Temperature distribution at different elevations. I am using Python for the operation. I have created a pipeline, and I am able to get the required slice from the vtkPolydata. However, when I rendered it, the plot looked like a scatter plot as shown below:

Slice Plot
Slice Plot679×672 5.96 KB

I want to plot it without any white space in between even when zoomed in. Please help me with this. I think I have to interpolate or create contours from the data, but I have no idea how to do it even after searching for a couple of days.

The code I used to generate the plot is given below.

The vtkPolydata I used to generate the above plot can be downloaded from here: vtkPolydata

import vtk
from vtk.util import numpy_support
import numpy as np

# Load vtkPolydata function
def loadvtp(fname):
    reader = vtk.vtkXMLPolyDataReader()
    reader.SetFileName(fname)
    reader.Update()
    data=reader.GetOutput()
    return data

# Load vtkPolydata
polydata=loadvtp("TestPolydata.vtp")

# Set the elevation at which data need to be plotted
ele=12

# %% Clip polydata based on the elevation
# Create plane in +Z to cut the polydata below elevation
plane = vtk.vtkPlane()
plane.SetOrigin(0, 0, ele)
plane.SetNormal(0, 0, 1)

# create polydata clipper
clipper = vtk.vtkClipPolyData()
clipper.SetInputData(polydata)
clipper.SetClipFunction(plane)
clipper.Update()

out_data=clipper.GetOutput()

# Create plane in -Z to cut the polydata above elevation
plane = vtk.vtkPlane()
plane.SetOrigin(0, 0, ele+0.5)
plane.SetNormal(0, 0, -1)
clipper = vtk.vtkClipPolyData()
clipper.SetInputData(out_data)
clipper.SetClipFunction(plane)
clipper.Update()

# Cliped polydata
sampled_data=clipper.GetOutput()

# Get the "Temperature" Array from the polydata
point_scalars = sampled_data.GetPointData().GetArray("Temperature")

# Create a lookup table to map the point field data to colors
lut = vtk.vtkLookupTable()
lut.SetNumberOfTableValues(2)
lut.SetTableValue(0, 1, 0, 0, 1) # red
lut.SetTableValue(1, 0, 1, 0, 1) # green
lut.Build()

# Create a mapper and actor to display the cliped plane
mapper = vtk.vtkPolyDataMapper()
mapper.SetInputConnection(clipper.GetOutputPort())
mapper.SetScalarModeToUsePointFieldData()
mapper.ScalarVisibilityOn()
mapper.SetColorModeToMapScalars()
mapper.SelectColorArray(0)
mapper.SetLookupTable(lut)
mapper.SetScalarRange(point_scalars.GetRange())

actor = vtk.vtkActor()
actor.SetMapper(mapper)

# Create the RenderWindow, Renderer
colors = vtk.vtkNamedColors()

ren1 = vtk.vtkRenderer()
renWin = vtk.vtkRenderWindow()
renWin.AddRenderer(ren1)
iren = vtk.vtkRenderWindowInteractor()
iren.SetRenderWindow(renWin)

ren1.AddActor(actor)

ren1.SetBackground(colors.GetColor3d('Gainsboro'))
renWin.SetSize(650, 650)
renWin.SetWindowName('Test')

style = vtk.vtkInteractorStyleTrackballCamera()
iren.SetInteractorStyle(style)

ren1.TwoSidedLightingOff()

ren1.ResetCamera()

renWin.Render()
iren.Start()
2

It seems to me that your data file contains a vtkPolyData with only points for sample support. From your data file:

    <Piece NumberOfPoints="6363101"  NumberOfVerts="6363101" NumberOfLines="0" NumberOfStrips="0" NumberOfPolys="0">

Hence the data rendered as points.

3

You can increase point size to give it a grid-like look:

actor.GetProperty().SetPointSize( 10 )

The example above will set your points to squares or circles (depends on your OpenGL backend) 10-pixel wide/tall.

4

Thank you @Paulo_Carvalho for the suggestion. I tried it, but the plot looks weird as we zoom in. Is there any other way to interpolate the point field data or create contours from the data? I have tried to use the probe filter to sample data to a plane, but no luck.

5

Hello,

You can also try vtkPointGaussianMapper as your mapper. It renders the points as something called an impostor in computer graphics jargon. An impostor is a visual representation more complex than the underlying data (e.g. a billboard is an impostor that gives the illusion of immersion by using camera-facing 2D images). Examples here: https://python.hotexamples.com/examples/vtk/-/vtkPointGaussianMapper/python-vtkpointgaussianmapper-function-examples.html The idea here is to render the points like grid cells without actually building a mesh.

take care,

Paulo

1 Like
6

Thank you very much @Paulo_Carvalho. I will try this.

1 Like