摘要
曲面拟合法求解亚像素位移是数字图像相关技术中的主要方法之一。它具有抗噪声能力强和精度高等特点。本文着重讨论了移动最小二乘三次曲面拟合方法,并与传统的三次曲面拟合法进行了比较。通过模拟和实验分析表明,移动最小二乘三次曲面拟合方法具有更高的亚像素定位精度,且稳定性良好。
Curved Surface fitting algorithm is one of the main methods in sub-pixel displacement location by digital image correlation method. The method has many advantages, such as good noise-tolerance, high precision. The method of cubic surface fitting is mainly discussed in this paper, and the idea of moving least-square(MLS) is introduced into the cubic surface fitting to increase the location precision for Sub-pixel Displacement. This novel method is compared with traditional cubic surface fitting algorithms in the location precision. Both simulation and experimental results indicate that MLS-Cubic surface fitting algorithm has higher precision of sub-pixel displacement location and better stability.
出处
《固体力学学报》
CAS
CSCD
北大核心
2008年第S1期224-227,共4页
Chinese Journal of Solid Mechanics
基金
国家自然科学基金项目(10672065)资助
关键词
曲面拟合
数字图像
移动最小二乘
相关技术
亚像素
surface fitting, digital image, moving least-square, correlation method, sub-pixel displacement