摘要
由单目图像获得物体表面高度 (由阴影恢复形状 )是计算机视觉中的一个重要研究领域 .迭代计算算法的准确度在合成图像的表面恢复上得到了验证 ,在遥感图像的表面高度提取中也得到了应用 .文中对一种新的由单目图像恢复物体表面高度和梯度的算法进行了详细分析 ,并提出了实现算法的具体步骤和计算方法 .对合成图像进行迭代计算 ,获得了准确的表面高度 。
Surface height reconstruction from a single image (shape from shading) is an important research domain of computer vision. The accuracy of iterative calculus was verified by recovering surface height of an object from its synthetic image, and the application of the calculus has been extended to reconstruct surface height from remote sensing image. In this paper, a new calculus of height and gradient from a single image is analyzed carefully, furthermore, the concrete steps and calculation details are proposed. The surface height and gradient can be reconstructed accurately from a synthetic image by the new calculus, which provides a feasible approach for extracting the surface shape of weld joint from its image.
出处
《计算机学报》
EI
CSCD
北大核心
2000年第2期147-152,共6页
Chinese Journal of Computers
基金
国家自然科学基金项目!( 5 963 5 160 (重点 )
5 95 75 0 5 7)
