多视纹理映射中无需设定偏差的快速可见性检测

Fast visibility detection without specifying the user-defined biases in multi-view texture mapping

在多视纹理映射中,Z缓冲区算法和光线追踪算法是两种主流的可见性检测算法。但其精度受限于人工设定的偏差。本文提出一种无需设定偏差的可见性检测方案。首先,根据倾斜摄影的投影参数设计了一种基于着色器编程的影像模拟算法,生成各影像视角下的初始可见性图(initial visibility map,IVM)。其中,完全不可见面元在模拟的过程中会被深度测试自动剔除。然后,基于矢量栅格化准则和像素深度对IVM中的可见面元执行投影区域优化(projection coverage refinement,PCR)。最后,提出惰性投影(lazy projection,LP)和迭代点边采样(iterative vertex-edge sampling,IVES)快速区分出IVM中的完全可见面元和非完全可见面元。利用两个数据集对本文算法进行了论证。结果表明本文算法的效率和精度均优于主流算法。

The Z-buffer and ray-tracing algorithms are the two popular approaches used in visibility handling for multi-view based texture mapping.However,the accuracies of the two algorithms are limited by the user-defined biases.We propose a solution in which no biases are specified.First,a shader-based rendering is designed according to the projection parameter of oblique photogrammetry to generate each view’s initial visibility map(IVM).The fully occluded primitives will be excluded by the depth test of graphics pipeline.Second,projection coverage refinement(PCR)is given to the visible primitives in IVM based on vector rasterization criterion and pixel depth.Last,lazy projection(LP)and iterative vertex-edge sampling(IVES)are proposed to distinguish the partially visible and fully visible primitives.We use two datasets to prove our method’s validity.The experimental results show that our method has a better performance than the mainstream algorithm.