| Summary: | Virtual battlefields devoid of vegetation deprive soldiers of valuable training in the critical aspects of terrain tactics and terrain-based situational awareness, but creating realistic landscapes by hand is notoriously expensive; requiring both proprietary tools and trained artists, hampering rapid scenario generation and limiting reuse. GENETICS is a new object placement scheme where the arduous task of placing vegetation objects is reduced to finding readily-available source data and setting a few parameters. Our approach constructs large-scale natural environments at run-time using a procedural image-based algorithm without the need for skilled artists or proprietary tools. The resulting vegetation-laden terrain looks realistic, and the algorithm can be extended to incorporate a wide variety of environmental factors. GENETICS offers researchers the ability to quickly and easily build consistent large-scale synthetic natural environments to examine vegetation clutter requirements necessary to accomplish distributed mission training tasks. This dissertation presents and implements the GENETICS algorithm, compares it against other vegetation placement schemes, and outlines how simulationists can use GENETICS to quickly and cheaply build large-scale natural environments. It also touches upon level of detail algorithms, ecotype modeling and how GENETICS can be used to generate land cover data where none exists.
|
|---|
