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