基于亚像素级几何图元覆盖检测算法  

Coverage Detecting Algorithm Based on Sub-Pixel Geometric Primitives

作　　者：李华[1,2] 杨华民 韩成[1,2] 赵建平 

机构地区：[1]长春理工大学计算机科学技术学院,长春130022 [2]特种电影技术及装备国家地方联合工程研究中心,长春130000

出　　处：《系统仿真学报》2017年第11期2723-2730,2741,共9页Journal of System Simulation

基　　金：吉林省重点科技攻关项目(20140204050GX;20140204009GX);吉林省重点科技成果转化项目(20130303011GX);长春市重大科技攻关计划(14KG008);吉林省省级产业创新专项资金项目(2016C091)

摘　　要：为减少、消除由于光照空间和相机空间采样密度不匹配造成的走样现象,提出一种亚像素级准确的硬阴影反走样绘制方法。该方法从光源视角抽取场景前表面的几何图元,利用保守光栅化方法存储固定大小、局部近似的场景可见三角形;计算纹素的顶点坐标,检测跨纹素几何图元,精确重建阴影测试深度。测试实验表明,几何图元覆盖检测算法能准确检测出纹素中包含的90%以上的隐含三角形,算法有较好的适应性,能有效减少阴影图算法的走样问题。A hard shadow anti-aliasing rendering algorithm with sub-pixel accuracywas proposed. The method focused on addressing aliasing artifacts of the shadow map due to the sample density mismatching between light space and camera space. The method extracted geometric primitives from the scene＇s front view from a light source and storedscene visible triangles with fixed-size and local approximation using conservative rasterization; An algorithm was proposed to detect the covered plurality primitives in one texel based on the texel＇s vertices, and reconstruct testing depth of shadow ray precisely. The experiment results showed that covered primitives＇ detection algorithm was able toidentify more than 90% triangles implicated in one texel. This method reduced jagged shadows effectively compared with standard shadow mapping.

关 键 词：阴影图 硬阴影 亚像素 反走样 图元 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]