基于小波多级分解的数字散斑相关搜索方法  

Wavelet-transform-based hierarchical search algorithm for the digital speckle correlation method

作　　者：简龙晖[1] 马少鹏[1] 张军[1] 金观昌[1] 姚学锋[1] 

机构地区：[1]清华大学工程力学系,北京100084

出　　处：《清华大学学报（自然科学版）》2003年第5期680-682,共3页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(19902008;19972033);航天科技创新基金资助项目(041810013)

摘　　要：为提高数字散斑相关方法(DSCM)的测量精度和计算速度,提出一种基于小波多级分解的数字散斑相关搜索方法。该方法利用二维离散正交小波技术对图像滤波去噪同时进行多级分解,从最低的一级开始进行相关搜索计算,然后逐级回溯。与传统的相关搜索方法相比,该方法可提高运算速度至少1个数量级并提高测量精度,有更强的适用性,可以对大位移、有噪声的图像进行数字相关位移场的处理。The measurement accuracy and calculational speed of the digital speckle correlation method (DSCM) was enhanced with a wavelettransformbased hierarchical search algorithm. The 2D discrete orthogonal wavelet technicque can generate hierarchical images with less noise by using wavelet transforms. The correlation search began from the lowest level and then proceeds to the higher level images. Compared with conventional correlation search methods, this hierarchical search algorithm increases the computational speed by the order of more than one, has high accuracy and is more widely applicable. The method can be applied to large displacement and high noise images.

关 键 词：小波多级分解 数字散斑相关搜索方法 变形测量 图像 滤波 去噪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]