RBF-BASED LASER SPECKLE PATTERN DIGITAL IMAGE CORRELATION METHOD FOR SURFACE STRAIN MEASUREMENTS

摘要

This paper introduces an innovative technique that integrates a genetic algorithm based digital image correlation with laser speckle photography (LSP) for the measurement of surface displacements in structures. The images (before and after deformation) are digitized using a digital camera, and the grayscale intensity matrices are read and processed by Matlab image processing toolbox. The two matrices of the images are then inputted into an iterative program based on the genetic algorithm that utilizes an advanced cross correlation technique to determine the surface displacements. Additionally, the strains are computed by radial basis function (RBF) differentiation. The computed displacements are compared with simulated results obtained by finite element analysis. Close agreement between the two results proved the validity of the developed non-contact technique for accurately measuring surface displacements. The experimentally measured displacements can be directly used in an inverse technique to detect and characterize subsurface cavities in structures.