Acceleration Structure Building on GPU

This thesis focuses on the construction of octree for three-dimensional triangular models on graphics hardware. This acceleration structure was chosen based on a comparison of the most commonly used acceleration structures. The approach to the construction of this structure is based on the method of construction of a sparse voxel octree which is derived from the former. The process uses voxelization accelerated by the graphics card's rendering pipeline. To facilitate this process and enable the integration of the proposed procedure into future applications, this thesis also includes design of a C++ language library