Database Fundamentals

1
Database Fundamentals
2
What Well Cover

• Processes for building visual databases
• Terrain, culture, features models
• Level of detail mechanisms
• Geo-typical Imagery-based databases
• Evolution of database format standards

3
Real-time Graphics Database

• The Database stores the environmental
  representation for terrain, features, and models
  during simulation exercises

Host Computer
GFX Computer/ Image Generator
I/O Devices
Display 1
Environment Database
Display N
4
Database Complexity

• Terrain
• visual
• sensor
• 2D roads, rivers
• 3D buildings
• 3D powerlines
• point lights
• light strings
• moving models
• healthy
• damaged

• clutter models - trees, bushes, rocks
• bridges (navigable)
• geospecific or geotypical imagery for ground
  features

5
Database Development Overview

• Sources
• Elevation grids, vector features, imagery
• Final Product
• Polygon mesh for terrain, polygonal objects,
  texture maps
• Issues
• Large databases 100 km x 100km at 30 meters
  requires 20 million polygons attributes
  texture
• Extensive use of image texture results in large
  database storage requirements
• Placement and features on objects can be very
  time consuming even if a standard process is
  followed

6
Database Development Process

• Select consistent coordinate system(s) for
  database
• Acquire and prepare imagery for textures
• Mesh the Terrain
• Regular Grid or TIN (Triangulated Irregular
  Network)
• Plant 2D Cultural Features
• Roads, rivers, lakes, forests
• Add 3D models for buildings, bridges, trees, etc
• Geo-specific model texture and geometry
• Examine and iterate to fix errors

7
Terrain Meshing Options
TINs (Triangular Irregular Networks) put density
where it is needed
Colored according to elevation
Wireframe of terrain network
8
Level of Detail Treatment Terrain

• Terrain Skin LODs
• Polygon budget is critical to obtain maximum
  performance
• Adjust polygon density for application

Low resolution LOD
High resolution LOD
9
Plant 2D Cultural Features

• Use vector sources such as NIMA DFAD
• Popular tools have mechanisms for cut-in of 2D
  culture (Terrex, Terrasim)
• Planting generic features such as trees can be
  automated to an extent (stoicastic placement)
• Look for tools that allow iterative alignment of
  culture overlaid on imagery

10
Level of Detail Treatment Models

• Real-time modeling tools have decimation routines
  to begin model LOD creation
• Each model or feature has a range-based LOD
  switch out assigned

LOD switches as eye to object range increases
(increasing right)
11
What is Scene Complexity?

• Also known as Polygon Density
• Total number of polygons in the renderable
  viewing volume
• Includes transparent and occluded polygons
• Includes polygons with at least one vertex in the
  viewing frustum
• Includes any polygon sent to graphics H/W

12
What is Depth Complexity?

• Number of times a pixel is written to
• Averaged across entire framebuffer
• Independent of drawing order
• Includes occluded pixels and background

13
What is Depth Complexity?

• Depth complexity is the number of times a pixel
  is overwritten
• Typically discussed as an average across entire
  frame buffer
• It is independent of drawing order
• Includes occluded pixels and background

wireframe illustrating depth of features
pixel overwrites blue low, white high
14
Typical Depth Complexities
15
Quality of Imagery-based Visuals
Sharpness and detail imagery can contain high
resolution detail Complexity of scene Scene can
contain much more detail than LODs allow with 3D
models
16
Varying Resolution by Interest Area

• Concentrate imagery acquisition on areas of
  interest
• Seamlessly blend insets into global cliptexture

17
Cliptexture

• Hardware accelerated support for large-scale
  textures
• Automatic paging of textures by Performer
• Cliptexures built using freeware SGI tool
  (clipgen)

18
DBGS Tools TerraSim

• Accepts GIS, imagery, DTM, and feature data
• Automates the dataflow of the DBGS process
• Script language for large databases
• Parallel processing of project graph
• Constrained TIN algorithms

www.terrasim.com
19
DBGS Tools MPI Creator

• Standard toolset for military training and
  simulation industry
• Feature Modules
• High-fidelity roads
• TerrainPro
• GIS import
• Generates Clipmap textures for SGI hardware
• Large Installed Base

www.multigenparadigm.com
20
DBGS Tools Terrex

• Source Folder and intuitive wizards accelerate
  database development
• Interactive editor allows modification of source
  data
• Legacy database conversion tools available
• Clipmap generation is supported
• Windows only

www.terrex.com
21
Database Process Flow
Step 8

Step 4

Step 3

Step 5

Step 1

Step 2

Compile
Enhance
Add 3D
Create

Acquire
Register
for IG

Source

Terrain
Features

Source

Source

Polygon
Skin

Feature locations

Step 6

Step 7

Construct
Moving
Build
Mod
els

Clipmap
Texture


22
Summary

• Source and tools exist for processing data for
  visualization and analysis
• Commercial Database tools have come a long way
  recently
• Database density must be controlled because it
  affects run-time performance
• Stay within polygon and depth complexity
  capacities
• Use imagery instead of polygons (texture is
  cheapest)
• Imagery-based databases are proliferating