Topological Similarity Measurement: A Methodology For Fast Full-Field Deformation Estimation

The correlation-based image subset matching methods such as digital image correlation (DIC) are widely used for measuring the deformation fields. The specimen surface must be covered with speckle patterns, or more exactly, random intensity distributions with sufficient contrast, which act as a carrier of deformation information, to acquire reliable and accurate matching in the correlation calculation. The requirement of the artificial speckles on the specimen surface inevitably contaminates the specimen. To maintain the cleanliness of the specimen surface, an in situ deformation estimation method without speckle fabrication is developed and its result is demonstrated. Topological similarity measurement (TSM) is a feature vector-based displacement estimating algorithm for full-field deformations that uses feature vectors that encode the relative positions of neighboring mark spots. Compared with the DIC method, this approach makes up for its deficiency in displacement estimation of speckle-free flat surface by incorporating a topological arrangement-based feature vector, subpixel center location method, and biharmonic spline interpolation method. Experimental results indicate that the proposed TSM method is feasible to become the preferred displacement estimation method for uniform deformation. As a complement to the DIC method, the TSM method is more suitable for uniform deformation distribution, which requires high measurement accuracy but cannot use DIC. (c) 2020 Society of Photo-Optical Instrumentation Engineers (SPIE)