| 摘要 |
Data defining one or more graphic objects, such as buildings and vegetation, are automatically generated in a repeatable manner within each of a plurality of visual tiles of a simulated space. A seed is determined for a tile as a function of a location of the tile in the space. The seed is used to pseudo-randomly determine visually apparent characteristics of the graphic objects, such as position, height, and texture. A tile is preferably characterized by one or a combination of texture classes, such as vegetated and urban. Any tile can be associated with an annotation that specifies other characteristics of an object, such as its footprint, or of a sub-area within which objects will be depicted in the tile and rendered with the randomly determined characteristics. For multi-class tiles, the annotations are used to mask automatically generated graphic objects from undesired portions of the tile.
|
