Dynamic Mapping of RasterData For 3D Geovirtual Environments

1
Dynamic Mapping of Raster-DataFor 3D Geovirtual
Environments

• Matthias Trapp, Jürgen DöllnerHasso-Plattner-Inst
  itute, University of Potsdam, Germany

13th International Conferenceon Information
Visualisation14-17 July
2
motivation

• Using 3D geovirtual environments as scenery for
  visualization of dynamic geo-referenced thematic
  data

3
motivation

• Using 3D geovirtual environments as scenery for
  visualization of dynamic geo-referenced thematic
  data

Preprocessing Time days-hours
Preprocessing Time hours-minutes
Preprocessing Time minutes
4
wrap-up challenges goals

• Strong coupling of geometry and appearance
• Real-time rendering constraints
• Large-scale data sets require specific
  out-of-core renderer
• Results in (time-consuming) data pre-processing
• ? Cannot be used to visualize time-critical data
• Goals
• partially decoupling geometry and appearance to
• enable dynamic multi-texture mapping
• enable color transfer function
• for a large number of mapping configurations
• in real-time

5
overview dynamic texture mapping

• Given input data
• 3D point (fragment) of original model
• Data color layers
• Mapping parameterizations
• Requested final color
• Data color layers
• 1D/2D/3D image/textures
• data values, masks
• color gradients

6
texture coordinate generation

• Projective texture mapping
  Everitt 2001
• Orthogonal projection matrix

U
V
7
color transfer functions

• Describes color mapping of data values
• Define blending equations between mappings

8
rendering at a glance

• Extending OpenGL/DirectX multi-texturing pipeline
• Single-pass rendering
• Per-fragment dynamic texturing using shader
  programs
• GPU Data representation texture arrays
• Projective mappings encoded in texture-buffers
• Performance evaluation

9
application examples

• Hierarchical combination of raster data layers

Building category data
Traffic frequency data
10
(No Transcript)
11
application examples

• Object or route highlighting

Route
Landmark objects
Landmark objects
12
(No Transcript)
13
application examples

• Projective symbols to facilitate the design
  of interactive 3D digital maps

14
application examples

• Focus context visualization 2D magic lenses

15
conclusions

• Summary
• Extendable concept for dynamically mapping 2D
  raster-data
• Approach based on projective texture mapping
• Fully-hardware accelerated implementation
• Limitations
• Works only for 3D GeoVE that can be approximated
  by a plane
• Limited texture resolution
• Future Work
• Extend color transfer to style transfer functions

16
thank you...
Contact Matthias Trapp matthias.trapp_at_hpi.un
i-potsdam.de Jürgen Döllner
juergen.doellner_at_hpi.uni-potsdam.de
Computer Graphics System Group
www.hpi.uni-potsdam.de/3d Researchgroup
3D-Geoinformation www.3dgi.de