摘要
目前采用的从明暗恢复形状(shape from shading,SFS)法存在对物体表面光滑度要求高,对噪声敏感等问题,为此,提出了一种基于数学形态学的SFS法,通过数学形态学提取图像灰度函数的峰、谷、脊、沟,鞍等形状特征,并采用球状点假设法确定物体的表面方向,恢复物体三维表面。实验表明,该方法具有更好的精度和抗噪性能。
For the shape from shading algorithm which is commonly used at present is demanding the smooth and it is very sensitive with the noise, this paper proposed an improved SFS algorithm based on the mathematical morphology, it extracted the characteristic points such as the apex, vale, ridge, channel, saddle and so on from the gray function of the image by using the mathematical morphology, and then determinated the surface direction by using the globular point hypothesis method, then recovered the 3D surface. The experiment shows that this method has better accuracy and noiseproof feature.
出处
《计算机应用研究》
CSCD
北大核心
2009年第10期3941-3944,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(40874094)
关键词
数学形态学
球状点假设
从明暗恢复形状
mathematical morphology
globular point hypothesis
shape from shading（SFS）
