Decomposing Strain Maps Using Fourier-Zernike Shape Descriptors

Shape analysis techniques such as geometric moment descriptors, Fourier descriptors, wavelet descriptors etc. have been used commercially in the fields of biometrics for finger print matching, iris matching and facial recognition for almost over a decade. These techniques are capable of decomposing high resolution images with 105 to 106 pixels into only a few hundred unique shape descriptors which are a true representation of the features in the corresponding images thus tremendously reducing the computational data and time required for image analysis and comparison. This paper explores the possibility of employing shape analysis techniques to facilitate the comparison of full-field data obtained from techniques such as digital image correlation, thermoelasticity and photoelasticity for the purpose of validation of computational models and damage assessment. A new shape descriptor is introduced which combines Fourier decomposition with Zernike moments. The Fourier-Zernike descriptor is shown to be capable of decomposing full-field strain distributions containing engineering features and discontinuities arising from damage by combining the desirable properties of its parent shape descriptors while eliminating their individual limitations.