数字散斑相关方法亚像素求解的一种混合方法

亚像素求解是数字散斑相关方法(DSCM)中最重要的技术。亚像素求解算法的精度和效率直接影响DSCM的精度和计算速度。目前广泛应用的两种亚像素求解方法在处理高质量散斑图时都存在一定的缺陷:基于灰度插值的散斑场亚像素恢复方法,其计算精度较高,但计算消耗太大;基于相关系数分布的拟合方法,其计算消耗小,但计算精度较低。在分析两种亚像素求解方法各自优缺点的基础上,提出了一种结合两种算法优势的混合算法。混合算法在保证亚像素求解精度的基础上,可以大大降低整个算法的计算消耗。

A hybrid method for subpixel registration of digital speckle correlation method

Subpixel registration is the most important technique in digital speckle correlation method(DSCM).The accuracy and efficiency of the subpixel registration algorithms directly affect the accuracy and efficiency of DSCM.The currently most often used algorithms for subpixel registration could be divided into two categories: the Gray-value Interpolation based on subpixel reconstruction(GISR) method,and the correlation coefficient distribution based on fitting(CCDF) method.However,when processing the high-quality speckle images using DSCM,one could find that the GISR method was with high accuracy but with very large computational consumption,while the CCDF method with rather low computational consumption but with low accuracy.A hybrid method based on the combination of GISR and CCDF is proposed.Using the hybrid method,the computational consumption decreases greatly while the accuracy of the algorithms holds.The flow chart to realize the hybrid method is introduced and then an experiment is completed to verify the efficiency and accuracy of the method.