提出了一种基于图像空间的焦散实时逼真绘制算法,算法将焦散的计算过程进行了分解并采用动态组织浮点纹理的方式将计算所需的几何数据从几何空间转换到了图像空间,可完全通过GPU编程实现实时逼真绘制。算法的主要优点是适用于可变形物体,可模拟由物体双面折射形成的焦散现象,算法效率受场景规模影响小并可方便地与阴影等绘制效果相结合。实验表明,GPU的数据处理能力和可编程特性得到了充分挖掘,与经典焦散绘制算法相比具有明显的效率优势。
An image space based real-time and realistic caustics rendering algorithm was proposed. The computing process of caustics was decomposed, and in order to convert data from geometry space to image space, the render- to-texture method was employed to dynamically organize the geometric data into float textures. So the real-time and realistic caustics rendering can be completely implemented on GPU. Especially, the algorithm is suitable for deformable objects, can approximately simulate two-side refraction and can also be simply combined with shadow mapping algorithm, meanwhile the scale of virtual scene has little effect on the algorithm efficiency. The experimental results show that the algorithm can fully exploit the powerful computing capability of GPU, which has much higher efficiency than traditional caustics rendering algorithms.