An Occlusion Culling Approach Based on Exploiting Multiple Hardware-Accelerated Occlusion Queries for Dynamical Scenes

Abstract

Occlusion culling has been studied extensively in computer graphics for years. In this paper, an occlusion culling approach based on exploiting multiple hardware-accelerated occlusion queries using the concept of eye-siding number for dynamic scene is proposed. Organizing the regular grid with overlapping voxels for the scene as an octree-like hierarchy, the actual position of dynamical objects can be updated efficiently. Based on the eye-siding number, the nodes occlusion front-to-back order enumeration can be done efficiently and the number of parallelizable occlusion queries for nodes in the hierarchy while traversing can be maximized efficiently and effectively. As experimental results shown, for all frames of the test walk-through in a dynamical environment, our approach does improve the overall performance.