改进的实时反走样透视阴影图算法及实现

摘要

为解决透视阴影图算法中,物体被非线性比例导致愈加明显的自阴影或表面波纹现象,提出一种基于GPU的实时反走样透视阴影图改进算法.该算法结合透视阴影图算法以及D.Weiskopf的dual depth layers算法,通过采用GPU深度剥离技术采样场景的双层深度值信息,并引入自适应偏移量Zbias来计算场景的深度偏移量.该算法采用C语言实现,同时使用GPU顶点渲染器和像素渲染器编程有效避免程序复杂度.实验结果表明,该改进算法较好解决了透视阴影图的自阴影问题,提高了渲染阴影质量.%Because objects are scaled non-uniformly in perspective shadow maps algorithm, the self-occlusion or surface acne becomes more severe. To solve this problem, an improved real-time anti-aliased perspective shadow maps algorithm based on GPU is presented. The algorithm combines perspective shadow maps algorithm and D. Weiskopf s dual depth layers algorithm, samples the two closest surface depth values by using depth peeling technology, and an adaptive bias is introduced to calculate the depth bias of scene. This algorithm is implemented in C, the complexity of program is avoided effectively by programming in vertex shader and fragment shader of GPU. The results show that the improved algorithm efficiently addresses the self-occlusion problem, and the quality of real time shadow rendering is improved efficiently.