PresentationPDF Available

Interactive Rendering and Stylization of Transportation Networks using Distance Fields



Presentation of Research Paper "Interactive Rendering and Stylization of Transportation Networks using Distance Fields"
Interactive Rendering and Stylization of Transportation
Networks using Distance Fields
Matthias Trapp, Amir Semmo, Jürgen Döllner
Hasso Plattner Institute, University of Potsdam, Germany
Towards Cartography-oriented Stylization
Paris Street Map 1780
P1: contour lines surround fine-textured fills or solid colors
to add visual contrast and improve figure-ground perception
P2: primary streets overlap secondary or tertiary streets in
hierarchical representations of street networks; wavy or fuzzy
to express uncertainty
P3: names follow principal line directions and are placed within
streets or outside line segments
London Street Map 1913
London Underground Map 1933
Modern Tourist Map of Paris
P7: yellow established as a conventional color tone for main
streets, with a discrete gradation towards grey and white
shading for tertiary roads
P6: streets are tinted using qualitative color schemes to represent
street classes and distinguish them from the underlying terrain
P5: a hierarchy of emphasis is drawn among reference elements,
such as different line weights and colors to portray different
grades of roads
P4: dynamic filtering and scaling of geometric features improves
perception of roads at high view distances and avoids
Conceptual and Technical Requirements
Pre-processing of discrete level-of-details consume additional memory and yield
incoherent rendering when switching between these levels during zooming or within
perspective projection.
Levels-of-Detail should be computed during rendering based on viewing settings
Increasingly detailed networks require high amounts of main/video memory.
Network representation should exhibit a small memory footprint and fast updates
View-dependent cartographic stylization of transportation networks are key features for a
number of applications.
Rendering technique should provide a sufficient parameterization, i.e., covering level-of-detail
rendering, interactive filtering, and highlighting
Related Work :: Overview
Street Rendering Approaches
Object-space Approaches
Geometry-based Approach
Texture-based Approach
Screen-Space Approaches
Stencil-based Approach
Discard-based Approach
Geometry-based Approach
Basic Principle:
Pre-compute geometry at different LoD
Forward rendering geometry
Most flexible approach w.r.t. stylization
High memory consumptions for complex networks
No transitions between static LoD
High rendering costs
Texture-based Approach
„Interactive 3D Visualization of Vector Data in GIS”
O. Kersting, J. Döllner, ACM GIS 2002
Basic Principle:
Generate texture trees from geometry
Off-screen rendering at different resolutions
Use for texturing during scene rendering
Relies on data pre-processing
Requires intermediate representation
Suffers from texturing artifacts
Stencil-based Approach
„High-Quality Cartographic Roads on High-Resolution DEMs”
M. Vaaraniemi, M. Treib, R. Westermann, WSCG Journal 2011
Basic Principle:
Generate shadow volume per street segment
Avoid cracks between segments using cap cones
Enables distance-based scaling of street segments
Rendering using shadow volume approach
Requires additional data structure (volumes)
Limited styling capabilities (color + outline)
Discard-based Approach
„A screen-space approach to rendering polylines on terrain”
D. Ohlarik, P. Cozzi; SIGGRAPH Poster 2011
Basic Principle:
Extrude polylines to walls
Compute intersection with terrain
Discard fragments accordingly
Avoid dashing and smearing
Stylization with single color only
No distance-dependent segment width
Distance-fields for Stylization Parametrization
„ Real-Time Rendering of Water Surfaces with Cartography-Oriented Design”
A. Semmo, J. E. Kyprianidis, M. Trapp, J. Döllner; CAe 2014
Overview of Approach
Forward Rendering Pass Deferred Rendering Pass
Overview of Distance-Field Generation
Geometry Generation
Attributed Vertex Cloud
Geometry Shader
Vertex Pulling
Encoding of Distance using Texturing
dresult = min(dsource , ddestination)
Blending Modes
No Blending Min-Blending
Distance Field Colored Distance Field Colored
Result: 2D Texture-Array
Memory consumptions: #layer
precision (3 * 32 Bit = 12 Byte)
Evaluation of Distance Fields :: Procedural Textures
Procedural textures evaluated on a per-fragment basis:
Deferred texturing based on distance fields
Application of procedural and image textures possible
Bottom-up compositing based on street rank
“Improved Alpha-Tested Magnification
for Vector Textures and Special Effects”
Chris Green; SIGGRAPH 2007
Final Compositing Step
Bottom-up evaluation of each layer (street category) and blending
Application Examples :: Distance-based Evaluation
Application Examples :: Stylization Variants
Application Examples :: Regions-of-Interest
Application Examples :: Regions-of-Interest
Performance Evaluation
Test data sets of different complexity
from Open Street Map (OSM) data base
Approach is fill-limited w.r.t. number of
street categories to render
Data Set
# Nodes
Berlin 1
Berlin 2
390 x 260 670 x 450 1280 x 800
1 Category 3 2,9 3 3 2,9 3 25,5 25,4 25,3
2 Categories 3,2 3,3 3,4 3,2 3,3 3,4 29 29 29,2
4 Categories 4,1 4,1 4,2 4,1 4,2 4,2 36,1 26,2 36,2
8 Categories 5,5 5,4 5,5 5,7 5,6 5,8 50,1 50,1 50,2
Distance-field generation:
Memory consumptions for large numbers of categories
Future Work :: Geometry Draping
Digital Elevation Model Result
Planar Network Geometry
Future Work :: Geometries
Generate alternative geometric representations
View-dependent adaptation of geometric representations
A concept for high-quality cartographic rendering exemplified
for complex street networks.
Interactive hardware-accelerated rendering technique having
minimal memory footprint for network representation.
Interactive stylization and colorization using deferred texturing
based on distance fields generated on per-frame basis
Potentials for future research
Questions & Comments ?
Matthias Trapp /
Amir Semmo /
Jürgen Döllner /
This work was funded by the Federal Ministry of Education and Research (BMBF),
Germany within the InnoProfile Transfer research group "4DnD-Vis".
Conference Paper
Full-text available
Transportation networks are important components of digital maps. The display of streets and railroad lines are a key for orientation and navigation using traditional and digital maps. While 2D digital maps enable effective communication, their visualization within 3D digital maps, based on virtual 3D city models, is often subject to occlusion that hinders the effective perception. There are a number of approaches to partially resolve occlusion effects within 3D virtual environments such as ghosted and cut-away views as well as interactive multi-perspective views. This paper proposes occlusion hints to emphasize occluded parts of transportation networks. In contrast to existing occlusion management techniques, these present only hints towards occluded parts to the viewer.
With the extensive application of virtual geographic environments and the rapid development of 3D visualization analysis, the rendering of complex vector lines has attracted significant attention. Although there are many rendering algorithms in 3D geographic information system (GIS), they are not sufficiently flexible to meet the requirements for rendering linear symbols composed of diverse colors and shapes. However, the interactive rendering of a scene and the accuracy of the symbols are important components for large-scale, complex vector lines. In this paper, we propose a graphics processing unit (GPU)-accelerated algorithm for rendering linear symbols on 3D terrain. Symbol rendering is embedded within the terrain-rendering process, and vector lines are encoded in a 3D texture and then transferred to the GPU. A set of visual properties are used to enrich the expression of symbols with the help of geometric operations in the fragment shader. A series of experiments demonstrate that the proposed method can be utilized for drawing various pixel-exact linear symbols and can achieve real-time rendering efficiency.
ResearchGate has not been able to resolve any references for this publication.