MEMS中一种改进的移动最小二乘曲面拟合亚像素位移测量算法

曲面拟合法求解亚像素位移是数字图像相关测量MEMS亚像素位移的主要方法之一,现在研究热点之一的移动最小二乘曲面拟合法有时可能会产生病态甚至奇异,影响位移的测量精度和稳定性。文章提出了一种改进的移动最小二乘曲面拟合算法。实验表明:与传统的拟合法相比,该拟合曲面更光滑、亚像素精度更高、稳定性更好、抗噪能力更强,能够有效应用到MEMS的位移测量中。

An Improved Moving Least-squares Surface Fitting Algorithm for Sub-pixel Displacement Measurement in MEMS Displacement Measuremen

Surface fitting solution sub-pixel displacement is one of the main methods for the MEMS sub-pixel displacement of the digital image correlation method (DICM) measurement. As a hot research topic, the moving least square (MLS) surface fitting algorithm may generate the morbid equations and even singular phenomenon, thus affecting the accuracy and stability of displacement measurement. An improved moving least-square surface fitting algorithms is proposed in this paper. The experimental results show that the fitting surface is more smooth and the sub-pixel is more stable, and can be effectively applied in the MEMS displacement measurements.