长春理工大学学报(自然科学版)
長春理工大學學報(自然科學版)
장춘리공대학학보(자연과학판)
JOURNAL OF CHANGCHUN UNIVERSITY OF SCIENCE AND TECHNOLOGY(NATURAL SCIENCE EDITION)
2015年
1期
136-142
,共7页
反射阴影图%虚拟点光源%间接光照%自适应区域滤波
反射陰影圖%虛擬點光源%間接光照%自適應區域濾波
반사음영도%허의점광원%간접광조%자괄응구역려파
reflective shadow maps%virtual point light%indirect illumination%adaptive region filtering
提出了基于反射阴影图的间接光照绘制改进算法。利用间接光照变化平滑的特点对低空间分辨率间接光照的计算结果进行插值得到高分辨率间接光照的结果。由于边界区域间接光照的变化是高频的,必须对边界和非边界区域分别计算。利用可视场景点的几何属性图,检测场景中几何对象的边界,将可视场景区域分为边界区域和非边界区域。边界区域的间接光照在高分辨率下直接计算,其它区域则由低分辨率的间接光照结果插值得来,最后使用自适应区域滤波方法对高分辨率间接光照的结果进行滤波得到最终间接光照结果。实验结果表明,相比于原始的基于反射阴影图的间接光照绘制算法,本算法显著提高了绘制效率,并通过设置插值分辨率阈值来增强算法的适应性。
提齣瞭基于反射陰影圖的間接光照繪製改進算法。利用間接光照變化平滑的特點對低空間分辨率間接光照的計算結果進行插值得到高分辨率間接光照的結果。由于邊界區域間接光照的變化是高頻的,必鬚對邊界和非邊界區域分彆計算。利用可視場景點的幾何屬性圖,檢測場景中幾何對象的邊界,將可視場景區域分為邊界區域和非邊界區域。邊界區域的間接光照在高分辨率下直接計算,其它區域則由低分辨率的間接光照結果插值得來,最後使用自適應區域濾波方法對高分辨率間接光照的結果進行濾波得到最終間接光照結果。實驗結果錶明,相比于原始的基于反射陰影圖的間接光照繪製算法,本算法顯著提高瞭繪製效率,併通過設置插值分辨率閾值來增彊算法的適應性。
제출료기우반사음영도적간접광조회제개진산법。이용간접광조변화평활적특점대저공간분변솔간접광조적계산결과진행삽치득도고분변솔간접광조적결과。유우변계구역간접광조적변화시고빈적,필수대변계화비변계구역분별계산。이용가시장경점적궤하속성도,검측장경중궤하대상적변계,장가시장경구역분위변계구역화비변계구역。변계구역적간접광조재고분변솔하직접계산,기타구역칙유저분변솔적간접광조결과삽치득래,최후사용자괄응구역려파방법대고분변솔간접광조적결과진행려파득도최종간접광조결과。실험결과표명,상비우원시적기우반사음영도적간접광조회제산법,본산법현저제고료회제효솔,병통과설치삽치분변솔역치래증강산법적괄응성。
In this paper,an advanced algorithm based on reflective shadow maps (RSM) for rendering indirect illumina-tion is proposed. Due to the fact that indirect illumination changes smoothly,high-resolution indirect illumination can be calculated by interpolating the low-resolution results. However, the indirect illumination in regions of boundary changes sharply;thus the visible regions of scenes should be divided into areas of boundary and non-boundary via detection for the boundary of 3D objects with help of geometric attribute maps. The indirect illumination in the areas of boundary is computed directly, while others can be interpolated by low-resolution results. Finally, adaptive region filtering is used to smoothen the high-resolution indirect illumination results obtained by interpolation operations mentioned above. Exper-imental result shows that our algorithm is faster than the original RSM-based one. Its adaptability can be controlled by the interpolation resolution threshold.