摘要
针对使用朗伯反射模型描述漫反射表面的形状恢复算法存在较大误差的问题,提出了一种偏角的计算和特定方向的选择相关的点光源参数估计的从明暗恢复形状的算法。首先假定摄像机采用正交投影,方向与光源方向一致,建立适合漫反射表面的图像辐照度方程,然后将方程转化为包含物体深度信息的偏微分方程,使用点光源估计方法应用于SFS技术的线性化方法,进而得到物体表面的三维形状。合成人脸的PGM格式的图像实验结果表明,采用改进的算法成像准确,且深度信息值域比较宽,效果满足要求。
Focusing on the high error of shape from shading（SFS） for diffuse surfaces, a SFS method that correlation of angle calculation and selection of a specific direction based on Oren - Nayar reflection model （ FSFS - ON） was proposed. The image irradiance equation for diffuse surfaces was established under assumptions that the camera performs orthographic projection and its direction is the same as the direction of the light source. The equation was then transformed into the partial differential.equation that includes the shape information of the surface. Point light source estimation method was applied to the SFS technology linearization method so that the heights of three - dimensional surface was obtained. Experimental results and comparisons on face the PGM format image show that using the improved algorithm can obtain the aecurate contour and relatively wide range of depth information.
出处
《计算机仿真》
CSCD
北大核心
2012年第12期317-320,共4页
Computer Simulation
基金
陕西省自然科学基础研究计划项目(2007D07)
陕西省自然科学基金资助项目(2009JM7007)
