Modified correlation criterion for digital image correlation considering the effect of lighting variations in deformation measurements

The Newton-Raphson (N-R) algorithm based on the sum of squared differences (SSD) function used in the digital image correlation (DIC) may obtain unreliable results in deformation measurements because light source variation is not taken into account. Here, a modified N-R algorithm that includes the effect of light source variations on the measurement results is presented. Instead of involving the mean value and square sum of the gray value of the object, variables were employed in the proposed algorithm. Thus, we refer to it as variables-based sum of squared differences (VSSD) function. VSSD has been validated by comparing the computer generated speckle images, which indicates that it could give rise to more accurate results in calculating deformation than SSD when the surrounding light source changes. In addition, VSSD is compared with the zero-normalized sum of squared differences (ZNSSD) function that is often used to remove the light source variation. The results show that VSSD can achieve the same accuracy within in less time compared to ZNSSD. (C) 2012 Society of Photo-Optical Instrumentation Engineers (SPIE). [DOI: 10.1117/1.OE.51.1.017004]