Evaluation and optimization of digital image correlation processing variables using genetic algorithm. (August 2016)

• Record Type: Journal Article
• Authors: Gu, Xuesen; Liang, Yuan; Fu, Shan
• Is Part Of: Journal of strain analysis for engineering design. Volume 51:Number 6(2016)
• Page Start: 408
• Page End: 415
• Publication Date: 2016-08
• Subjects: Digital image correlation -- uncertainty -- simulated experiments -- genetic algorithm -- strain measurement; Strains and stresses -- Periodicals; Strain gages -- Periodicals; 624.176
• DOI: 10.1177/0309324716645704
• Languages: English
• ISSNs: 0309-3247

Abstract

In the surface strain measurement, different settings of the variables applied in digital image correlation procedure give rise to different measurement error levels. This article discusses how the variable settings affect the error, and meanwhile put forward a method to optimize the variables. The deformation of the synthetic reference image for digital image correlation procedure is determined by a finite element simulation of a tensile test under certain loading condition. The digital image correlation variables concerned in this article contain the subset size, step size, shape function and strain window size. Moreover, a genetic algorithm is developed to optimize the setting of these variables, and their sensitivities are further analysed to clarify the different impacts of each parameter on the total measurement error. As demonstrated by the results, the measurement can be improved to a more satisfactory level after the optimization to the digital image correlation processing variables.