VTK Tutorial

Data structures,
filtering and rendering
Stefano Perticoni – [email protected]
 Live material
http://notepad.stefanoperticoni.org

2
Prerequisites

3
 Prerequisites
The following Python 2.7 and vtk 5.10 execution environment for
Windows is available on your pc through the Spyder GUI:

Run Spyder (pythonxy gui) from startup icon or command line:

C:\Python27\Scripts\spyder.exe

4
Tools -> Preferences

5
 Exercise: learn vtkArray
#1 Make an array
import vtk
myArray = vtk.vtkDoubleArray()
list_dir(myArray)
help(myArray.SetValue)
print(myArray)
myArray.SetName('my first array')
myArray.SetNumberOfComponents(1)
myArray.SetNumberOfTuples(500*500) #going to make a 500x500 picture

#2 Fill it with data

from math import sin, cos
for x in range(0,500):
for y in range(0,500):
myArray.SetValue(x*500+y, 127.5+(1.0+sin(x/25.0)*cos(y/25.0)))

6
 Exercise: learn vtkArray
#1. Create the Data structure
id = vtk.vtkImageData()

#2. Define its Geometry

id.SetOrigin(0,0,0)
id.SetSpacing(1,1,1)

#3. Define its Topology

id.SetDimensions(500,500,1)

#4. Assign Data to the Structure, Geometry and/or Topology

id.SetScalarType(vtk.VTK_DOUBLE)
id.GetPointData().SetScalars(myArray)

#5. Inspect it
print(id)
print(id.GetPointData())
array = id.GetPointData().GetArray('my first array')
array.GetRange()

7
The VTK Graphics Subsystem

8
vtkRenderWindow

• SetSize() — set the size of the window

• AddRenderer() — add another renderer which draws into this

• SetInteractor() — set class to handles mouse/key events

- vtkRenderWindowInteractor->SetInteractorStyle()

• Render() — updates pipeline and draws scene

9
vtkRenderer

• SetViewport() - specify where to draw in the render window

• SetLayer() - set pane/depth in render window to draw on

• AddViewProp() - add objects to be rendered

• AddLight() - add a light to illuminate the scene
• SetAmbient() - set the intensity of the ambient lighting
• SetBackground() - set background color

• SetActiveCamera() - specify the camera to use to render the scene

• ResetCamera() - reset the camera so that all actors are visible

10
 vtkCamera
• Position - where the camera is located
• FocalPoint - where the camera is pointing
• ViewUp - which direction is "up"
• ClippingRange - data outside of this range is clipped
• ViewAngle - the camera view angle controls perspective effects
• ParallelProjection - turn parallel projection
on/off (no perspective effects)

• Roll, Pitch, Yaw, Elevation, Azimuth

move the camera in a variety of ways

• Zoom, Dolly - changes view angle (Zoom);

move camera closer (Dolly)

11
vtkCamera

12
 vtkActor (subclass of vtkProp)

• Visibility - is the actor visible?

• Pickable - is the actor pickable?

• Texture - a texture map associated with the actor

• SetOrigin/Scale/UserTransform - control where it is drawn
• GetBounds

• vtkProperty - surface lighting properties

13
Exercise: make a window
#1. Make a window
renwin = vtk.vtkRenderWindow()
renwin.SetSize(500,500)

#2. Make a renderer for that window

renderer = vtk.vtkRenderer()
renwin.AddRenderer(renderer)

#3. Control how it all looks

renderer.SetBackground2(1,1,1)
renderer.SetGradientBackground(1)

#4. Show it
renwin.Render()

14
 Exercise: show some data
#1. Access the data processing pipeline that has your data
mapper = vtk.vtkDataSetMapper()
mapper.SetInput(id)
mapper.ScalarVisibilityOff() # we'll talk about this soon

#2. Link that to the display system

actor = vtk.vtkActor()
actor.SetMapper(mapper)
renderer.AddViewProp(actor)
renwin.Render()

#3. Adjust the camera for a better view

renderer.ResetCamera()
renwin.Render()

15
Color control by vtkActor and vtkMapper

16
vtkProperty (Actor has)

• AmbientColor, DiffuseColor, SpecularColor — a different

color for ambient, diffuse, and specular lighting
• Color — sets the three colors above to the same
• Interpolation - shading interpolation method (Flat,
Gouraud)

• Representation — how to represent itself

(Points, Wireframe, Surface)

• Opacity — control transparency

17
vtkMapper (Actor also has)
• ScalarVisibilityOn()/Off()
- Color cells/points by data values or entire object by actor color

• Choose which array to color by

- SetScalarModeToDefault()
- SetScalarModeToUsePointData()
- SetScalarModeToUseCellData()
- SelectColorArray(array name)

• SetLookupTable(lut)
• SetScalarRange(min, max)
- range of data values for lut

• InterpolateScalarBeforeMappingOn()/Off()
- whether to interpolate colors across cells in color or data space

18
 vtkLookupTable (Mapper has)

• NumberOfColors - number of colors in the table

• TableRange - the min/max scalar value range to map
• If building a table from linear HSVA ramp:
- HueRange - mm/max hue range
- SaturationRange - min/max saturation range
- ValueRange - min/max value range
- AlphaRange - min/max transparency range

• If manually building a table

- Build (after setting NumberOfColors)
- SetTableValue( idx, rgba) for each NumberOfColors entries

19
Exercise : Visualize the topology
#1. Specify whole Prop color
actorProperty = actor.GetProperty()
actorProperty.SetDiffuseColor(0,1,1)
renwin.Render()

#2. Change from surface to edges rendering

actorProperty.SetRepresentationToWireframe()
renwin.Render()
renderer.GetActiveCamera().Zoom(10)
renwin.Render()

#3. Reset
actorProperty.SetRepresentationToSurface()
renderer.ResetCamera()

20
Exercise : Visualize the topology

#1. Turn on color from values

mapper.ScalarVisibilityOn()
renwin.Render()

#2. Match up lookuptable range

myArray.GetRange()
mapper.SetScalarRange(127,129)
renwin.Render()

21
Algorithms

22
 Read a data file, inspect and visualize
#1. Create a reader, tell it what file and run it
reader = vtk.vtkDataSetReader()
reader.SetFileName("c:/VTKSchool/Perticoni/MaterialeEsercitazioneVTK/data/Saint
HelenSP.vtk")

#2. Examine the result

id = reader.GetOutput()
print id.GetPointData().GetArray(0)
reader.Update()
print id.GetPointData().GetArray(0).GetRange()
mapper.SetInputConnection(reader.GetOutputPort())
mapper.SetScalarRange(682.0, 2543.0)
renwin.Render()
renderer.ResetCamera()
renwin.Render()

23
Pipeline execution model

24
 Demand Driven Pipeline
• Lazy evaluation
- Pipeline only produces results when you ask it to Update or Render()
- Changing a parameter or rearranging the pipeline doesn't do that.
- Each filter caches its most recent output

• Modified time
- Each filter keeps track of when it last produced data, and when its
parameters were last changed
- Pipeline only updates as far back as it has to
- Examples:
• Camera motion - data isn't reread, only mapper has to execute
• Change isovalue parameter
• Change filename

25
Exercise : manipulate the read in data

#1. Make filter to convert to a less constrained data structure

triangles = vtk.vtkDataSetTriangleFilter()

#2. Connect it
triangles.SetInputConnection(reader.GetOutputPort())

#3. Run it
triangles.Update()
print(reader.GetOutput().GetClassName())
print(triangles.GetOutput().GetClassName())

26
Exercise: manipulate the read in data

#1. Make and use a filter to change the geometry

warp = vtk.vtkWarpScalar()
warp.SetInputConnection(triangles.GetOutputPort())
warp.Update()
print(triangles.GetOutput().GetBounds())
print(warp.GetOutput().GetBounds())

#2. Show it
mapper.SetInputConnection(warp.GetOutputPort())
renwin.Render()

27
Exercise: manipulate the read in data

#3 Get a hold of window events

iren = vtk.vtkRenderWindowInteractor()
renwin.SetInteractor(iren)
iren.Initialize()
iren.Start()

# Press “e” to exit from the interaction

# Press “t” to selec camera
Trackball interactor

28
 Exercise: manipulate the data
#1. Make a clip filter and put it in pipeline
clip = vtk.vtkClipDataSet()
clip.SetInputConnection(warp.GetOutputPort())
mapper.SetInputConnection(clip.GetOutputPort())

#2. Make a source to orient clip filter with

plane = vtk.vtkPlane()
clip.SetClipFunction(plane)
plane.SetOrigin(560000,5120000,2000)

#3. Inspect the result

clip.Update()
print clip.GetOutput().GetBounds()
iren.Start()

29
 Interaction
• Events
- Instances of vtk classes can fire
events and watch events fired by
others
- watcher executes some code
whenever the event occurs

• Interactors
- Watch mouse, keyboard,
window system events to move
camera call render etc

• Widgets
- Special purpose classes that are
drawn in scene and watch events

30
Exercise: use a widget to interact with
the data

#1. Get a hold of window events

iren = vtk.vtkRenderWindowInteractor()
renwin.SetInteractor(iren)

#2. Make and initially place the widget

widget = vtk.vtkImplicitPlaneWidget()
widget.PlaceWidget(warp.GetOutput().GetBounds())
widget.SetOrigin([plane.GetOrigin()[x] for x in 0,1,2])
widget.SetNormal([plane.GetNormal()[x] for x in 0,1,2])

#3. Connect it to the renderwindow's events

widget.SetInteractor(iren)

31
Exercise: use a widget to interact with the data
#1. Connect the widget's events to our pipeline
def eventhandler(obj , event):
global plane
obj.GetPlane(plane)

widget.AddObserver("InteractionEvent", eventhandler)

#2. Configure the widget

widget.SetEnabled(1)
widget.DrawPlaneOn()
widget.TubingOn()

#3. Turn on interaction

iren.Initialize()
renwin.Render()
iren.Start()

32
 Exercises
From your browser open the Summary page
file:///C:/VTKSchool/Perticoni/MaterialeEsercitazioneVTK/index.html

33