Pictured here are screen shots from different places throughout the GVP
(Global Visualization Process) database. GVP is a real-time
visualization engine with complete object management capabilities.
Developed by the Navy Crewstation Technology Lab over the last two
years, it is ideal for out-the-window scene generation (i.e. flight
simulation), command and control systems, mission rehearsal, and just
about anything else you might want to simulate in the real world.
The top left image, shows the entire earth as rendered in GVP. Unlike
most visualization systems, GVP uses round world geometry (WGS84). This
eliminates many of the problems encountered in distributed simulation
with engines using flat-world databases. Also, most graphics engines
use fog to obscure the far clipping plane. In GVP, we are using the
curvature of the earth to limit the visual field. For those interested,
the far clipping plane is set at 30,000,000 meters, and the near
clipping plane is set at 1 meter. Additionally, we have a model of the
entire earth (as demonstrated in the picture) at varying resolutions.
For the sake of simplicity, we used a map over parts of the US instead
of terrain -- it makes things a little more interesting. The imagery
here has a feature resolution of 2km.
The next image, counter-clockwise, shows part of our Washington DC
database, Reagan-National Airport. You can see two planes on approach
to the runway. Actually, there is just one plane captured at two
different instances in time -- they would not likely be landing aircraft
that closely. The next image is from the same database. It is the
Ellipse (the White House is in the bottom of the image). The resolution
of this piece of the database is 1.25 meters.
The mountains (center-right) simply show that we are not just pasting
imagery on a flat ball. We are building a terrain database using DTED
(digital terrain elevation data). The imagery resolution here is 5
In the lower and upper right, you can see sample images from our 3 inch
database. In the top image, you can see two people standing on a runway
with their shadows stretching out from them. In the lower image, you
can see that cars are clear to the point that make and model can easily
be discerned. Anyone want to guess what the red car is?
All told, the imagery in this system comprises about 180 GB. We are
looking at ways of using image compression techniques to make that
database about 36 GB.
Not shown in this image is the object management capability of GVP. We
can simulate thousands of entities all at once, and have multiple
viewpoints attached to the database at once. This allows for unique
interactions like dogfighting and head-to-head mission planning. Think
Since this system is currently under patent investigation, I cannot make
too many comments on specific implementation details, though I will
happily share what I can. We do not have a web page, but I did produce
a DVD which showcases this system rather well. If anyone is interested
in a DVD, please let me know.