Distributed Dynamic Z-Buffer Visibility

摘要

This paper presents an approach to determine visible objects in dynamic environments. The idea behind the algorithm is to make use of Z buffer, temporal coherence and network computing to estimate which objects are visible for a given viewpoint, and achieving running rendering and visibility on the same time. For each frame, our scheme uses a separate machine (server) to predisplay the visible objects of the last frame and read the Z buffer. Then the server builds a Z buffer pyramid and uses it to cull the fully invisible objects. Tests show that the algorithm is suited to dynamic environments with high depth complexity.