The Local Z-Buffering Rendering

the large complex scenes is presented in this paper. The local Z-buffering originates from the depth complexity based scene decomposition and following multistage rendering. In a naïve local Z-buffering, a large complex scene is decomposed into a low depth complexity scene and high depth complexity scene by computing a partition plane in the view space; the low and high depth complexity scenes are rendered by the Z-buffering and classic ray casting respectively. To achieve an output sensitive real-time rendering algorithm, a series of algorithms are developed to optimize the naïve version from aspects of general hardware support, empty pixel removal and fast potential visible primitive identification. Finally, the parallel version local Z-buffering rendering is introduced. Compared with similar types of rendering algorithms, the LZB algorithm is easy to implement, requires the least preprocessing time and achieves the satisfactory speed-up in the large complex scenes rendering.