由单幅二维灰度图像重构物体表面形状  

3D reconstruction from single 2D image of object

作　　者：高月芳[1] 罗飞[1] 曹建忠[1] 

机构地区：[1]华南理工大学自动化科学与工程学院,广州510640

出　　处：《计算机应用研究》2007年第8期195-197,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(69684001)

摘　　要：鉴于目前很少有论文讨论完整的由单幅二维灰度图像重构物体表面形状的算法,包括它的控制参数的估计及算法的实现,介绍了一种完整的SFS算法。它在考虑自遮掩影响的情况下,有效地估计了SFS算法中涉及的各种控制参数,并引入亮度约束、灰度梯度约束和可积性约束,计算出表面高度和表面向量,实现三维重构。与传统的算法相比,本算法无论是在速度还是在精度方面都达到了比较高的水平,具有一定的应用前景。最后还指出了在MATLAB中实现需要注意的问题。Since there have been few papers on a whole SFS algorithm, this paper introduced a SFS algorithm which includes the estimation of the control parameters and the implementation of the algorithm. Taking into consideration of the effect of selfshading, it can effectively estimate the control parameters used in the SFS algorithms. It implemented the brightness con- straint, the intensity gradient constraint and the integrability constraint to compute the surface height and orientation and to reconstruct the surface shape. This algorithm performed better than a great deal of conventional algorithms in both speed and accuracy. This paper finally mentioned some problems that should be paid attention to in MATLAB.

关 键 词：三维重构 由明暗恢复形状 参数估计 变分法 分层算法 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]