从明暗恢复形状方法的原理分析及常用解法综述  

Review： on Principle Analysis and Typical Algorithms of Shape from Shading

作　　者：朱良勇[1] 苏红旗[1] 杨彬[2] 晏磊[2] 

机构地区：[1]中国矿业大学(北京)机电与信息工程学院,北京100083 [2]北京大学空间信息集成与3s工程应用北京市重点实验室,北京100871

出　　处：《中国科技博览》2013年第27期510-512,共3页China Science and Technology Review

摘　　要：从明暗恢复形状（Shapefromshading，简称SFS）是计算机视觉中三维形貌恢复问题的关键技术。设计了基于单幅图像恢复三维形貌的算法，其核心思想是根据单幅图像的灰度信息进行亮度分析和转换，利用倾角和偏角计算表面点法矢，进而得到表面点高度值。文章通过对SFs基本原理的分析和几种常见的解法的简述与性能评价，说明了SFS算法还需要在现有的研究基础上，对过去的SFS算法的实现方法进行有效的改进，进一步提高SFS算法的实用性。Shape from shading（SFS） is the critical technique to 3D shape recovery in computer v -ision. A algorithms of 3D shape recovery based on single image is developed. The key technique is to make use of shading value to process lure inance analysis and conversion, thus to calculate the surface normal by slant and tilt, finally to get the height value. In this paper, the basic principle of SFS is analyzed, and the solution of fotlr typical algorithms is described in simple terms, and the performance is evaluated. It is illustrated that, it could not only make it possibleto overcome the basis of past research, improve the method of algorithms of SFS, but also benefit the performance of the algorithms of SFS.

关 键 词：图像处理 从明暗恢复形状 辐照量 辐射量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]