Advanced Image-Processing Techniques for Automatic Nondestructive Evaluation of Adhesively-Bonded Structures Using Speckle Interferometry

摘要: In conventional optical nondestructive evaluation (NDE) of structures using shearography or electronic speckle pattern interferometry (ESPI), results are typically provided in the form fringe patterns deformation contour plots. However, order to fully automate process defect detection, it is desirable obtain simpler which easier interpret. We present here one such system based on additive–subtractive shearography/ESPI. This processes and provides sizes locations defects as disbonds adhesively-bonded composite structures. achieved by exciting structure under inspection an acoustic stressing mechanism sweeps a range vibration frequencies structure. Since defective areas have different mechanical properties from their neighboring regions, varying complex obtained at frequencies. propose algorithm enables automatic identification selection relevant that pertain only localized deformations associated with defects, excludes images any overall modes entire The also includes pixel-by-pixel adjustable thresholding scheme compensates for intensity variations due nonuniform reflectivity unpainted dirty test objects. Morphological processing then performed extract shapes processed clusters. Various structures, simple aluminum specimens simulated honeycomb-based aviation repair patch specimen, been successfully evaluated this system.