图像相关方法中采用多种子区进行位移测量

在图像相关位移测量方法中一般采用正方型子区搜索中心点的位移.当测量靠近裂纹或缺陷部位的位移时中心型子区失去效果,为了解决这一问题提出了非中心型子区的方法.证明了方法的有效性.认为在相关搜索时测量点既可选定在子区的中心位置,也可选定在非中心位置.进而提出了将测量点确定在子区角点的多种子区的方法.除通常的中心型子区外,至少还有4种子区可以采用,分别是右下角型、左下角型、右上角型和左上角型,这4种子区不但适用于表面裂纹或缺陷部位的位移测量,而且增加了图像有效信息的提取面积.

Using Multi-Subregions in Image Correlation Method for Displacement Measurement

Generally, image correlation method for displacement measurement employs a square subregion to determine the displacements of its centroids.However, this kind of subregions is ineffective while measuring displacements near edges of holes or cracks. To solve the problem,non-central subregions are presented,and the feasibility is confirmed.The non-central subregions in images are used to calculate the displacements by correlation between deformed and undeformed states.In fact, the measurement points can be either at the centroid of a subregion or at the other points.Therefore, corner subregions are brought forward. At least four corner subregions can be used, i.e. left-bottom type, right-bottom type,left-up type and right-up type.These new type subregions can both increase the valid area for extracting deformed information and be more suitable for measuring displacements on the surfaces that have cracks or defects.The validity of the new method has been confirmed by experiments.