基于GPU的移动立方体算法及其应用-模式识别与智能系统专业论文.docxVIP

AbstractAbstract Abstract Abstract Surface rendering and its applications is one of the hot research topics in Visualization in Scientific Computing which relates tO digital image processing， computer graphics，computer visualization and virtual reality．Surface rendering produces clean—cut iso·surface，which Can exhibit the three dimensional information effectively with high rate when rendering．Therefore，how to further improve the extracting speed and rendering quality of surface rendering，how to apply it to practice are meaningful both in theory and in application． Traditional Marching Cubes has low speed in iso．surface extraction and poor real-time performance when rendering．To solve the problems，this thesis uses powerfully parallel computing capacity of GPU tO extract iso．surface and save the equivalent points in video memory．The equivalent points Can be accessed directly from the video memory,which prevents overloading I／O operation between CPU and GPU． Experimental results show that compared with the traditional algorithm，the Marching Cubes algorithm based on GPU does improve the extraction speed of iso．surface and real-time performance significantly． In practical applications，in order to fully mine the information contained in data fields，it also needs interactive processing towards data．In this thesi3，a method of real-time painting and clipping based on GPU Marching Cubes algorithm，with a surface marking rule，is presented．Experimental results show that the proposed method can enhante the perception of surface model．In order to overcome the non．unif01 111 phenomenon and the dependency on orientation when rendering，a block rule by marking the region of interest in objective space is presented．Experimental results demonstrate that the proposed method can resolve those problems resulting from the surface marking rule effectively． Ke)avord：GPU Marching Cubes Real-time painting Real—．time clipping 西安电子科技大学 西安电子科技大学 独创性(或创新性)声明 秉承学校严谨的学风和优良的科学道德，本人声明所呈交的论文是我个人在 导师指