摘要
从明暗恢复形状(SFS)是计算机视觉中三维形状恢复问题的关键技术之一,其任务是利用单幅灰度图像中明暗变化来恢复表面三维形状。快速步进法(FMM)作为求解Eikonal偏微分方程的有效手段,可以用于处理从明暗恢复形状问题。本文提出了一种基于改进的Godunov格式的多模板快速步进法(MSFMM)算法,解决了多源FMM的波动交汇问题,具备了处理多源SFS问题的能力。从实验结果可以看出,本文方法可以得到比MSFMM更高精度的结果,可以有效处理多源SFS。
Shape from shading (SFS) is one of the critical techniques on shape recovery in computer vision, which can obtain 3- D shape of the visible surface of an object from only one image of it using the shading knowledge in the given picture. Fast Marching Method (FMM) is an efficient method for Eikonal Equation and can be adopted in the SFS problem. In this paper, we propose a multi- stencil fast marching method based on the improved Godunov format. In this algorithm the improved Godunov format is used to solve the problem which happens when wave fronts converge. Experiments demonstrate that our algorithm can obtain more accuracy than MSFMM does and yield good performance for multi-source SFS problem.
出处
《信号处理》
CSCD
北大核心
2009年第6期905-910,共6页
Journal of Signal Processing