使用多层深度采样的屏幕空间环境光遮挡

Screen space ambient occlusion using multi depth sampling

为解决传统的基于水平线的屏幕空间环境光遮挡算法不能准确计算被遮挡物体的环境光遮挡值的问题,提出了一种使用多层深度采样的屏幕空间环境光遮挡算法。该算法使用桶型深度剥离算法获取场景中各层次的深度信息,并使用这些信息为场景中所有物体计算环境光遮挡值,有效解决了算法不能为被遮挡物体计算环境光遮挡值的问题。该算法的计算过程完全在图形处理器中完成。实验结果表明,该算法相对于原算法具有更好的真实感。

To deal with the inaccurate ambient occlusion issue for occluded objects in the traditional horizon-based screen-space ambient occlusion algorithm, screen-space ambient occlusion using multi depth sampling algorithm is proposed. In our method, bucket depth peeling algorithm is exploited to capture and sort multiple fragments＇ depth per pixel. By peeling these depth layers and computing the ambient occlusion value for every object in the scene, the artifacts of inaccurate ambient occlusion for occluded objects is solved. The computation of the algorithm is completely performed with graphic processing unit （GPU）. Experimental results show that this method achieves promising reality.