Struktury pro členění prostorových dat

Abstract

Raytracing is very high computing method. If it used high resolution and scene complexity, it is need to use high computing time, to the scene get drawed. If we had to compute every ray intersection with all triangles, of which is scene created, calculation it would take a lot of computing time. For this reasons for optimalization is used space partitioning data structures, which triangles stored to his own structure, and this structure is used for traverse. Each of these structures have his own advantages and disadvantages, accordig to scene settings. For another acceleration is used parallel computing in scene drawing or in generation of structure. For options of parallel computing is used processor (OpenMP) or graphic card performance. Graphic cards used interface Cuda (brand Nvidia) or OpenCL (brand AMD).