中文摘要：

传统的从明暗恢复形状（Shape from Shading,SFS）技术通常假定物体表面为Lambert表面,然而对于实际的物体表面来说,上述简化造成了重构结果误差较大。近年来,许多学者开始关注于非Lambert表面下的SFS技术的研究。结合作者自身的研究,从成像过程建模、图像辐照度方程的建立及求解数值算法出发,介绍了非Lam-bert表面从明暗恢复形状技术的研究进展。首先,简要介绍了非Lambert表面下典型的SFS技术;其次,以Oren-Nayer模型为例详细说明了非Lambert表面SFS方法的关键技术;最后,对非Lambert表面SFS技术的发展趋势进行了探讨。

英文摘要：

Most traditional shape from shading（SFS） techniques assume that the surfaces of an object are Lambertian surfaces.For real surfaces,however,the above simplification result in a higher reconstruction error.Recently,some authors have attempted to study the SFS techniques for non-Lambertian surfaces.Starting with the modeling of the image formation process,the formulation of the image irradiance equation and the solving numerical algorithm,this paper introduces the research progress on SFS techniques for non-Lambertian surfaces combining with the authors＇ research.First,the classical SFS techniques for non-Lambertian surfaces are introduced simply.Second,taking the Oren-Nayer model as an example,the key technologies of SFS methods for non-Lambertian surfaces are explained in detail.Finally,the further development trend of SFS techniques for non-Lambertian surfaces is also discussed.