摘要
从单幅图像获得物体的表面高度是计算机视觉中的一个重要研究领域,迭代算法的计算精确度高,但收敛速度较慢。该文对于几种常用的共轭梯度优化算法进行了详细分析,提出了在三维表面重建过程中实现共轭梯度算法的具体步骤和计算方法,并评价了算法的性能和优缺点。对合成图像进行仿真,并将表面恢复结果和算法收敛速度与传统的变分迭代方法比较,验证了算法的可行性和实时性。
Extracting surface depth from a shaded image(SFS)is one of the classic problems in computer vision.Many iterative algorithms correspond to a feasible surface,but converge slowly.In this paper,several conjugate gradient algorithms used in the3D shape recovery are analyzed carefully,furthermore,the concrete steps and calculation details are proposed and their advantages and disadvantages are discussed respectively.By recovering surface height of a synthetic image and comparing the result and converging speed with traditional iterative method,the feasibility of the approach is verified.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第35期24-26,56,共4页
Computer Engineering and Applications
基金
国家自然科学基金项目(编号:60141002)
南昌航院测控中心开放实验室基金项目(编号:KG200104001)
关键词
由阴影恢复形状
迭代计算
共轭梯度下降
表面重建
Shape from shading,Iterative calculate,Conjugate gradient descent ,Surface reconstruction
