An automated image processing algorithm to determine wrinkle characteristics from B-scans

This paper presents an automated algorithm which extracts wrinkle characteristics from B-Scan images of large composite parts. Out-of-plane ply misalignment is estimated by a bespoke fully automated algorithm based on the multiple field image analysis method (MFIA). The original algorithm is adapted to (1) specifically improve the method's robustness at handling low resolution images typical of B-scan data and (2) compute wrinkle characteristics such as wavelength, amplitude and location. The algorithm is first validated against a synthetic wrinkle created using a localised wave function, and shows wavelength characteristics computed to within ±1%. The algorithm is then demonstrated for real scans of out-of-plane wrinkles in the corners of a large-scale C-section component. Finally, we show how the algorithm is used to generate a finite 3D representation of the wrinkles. It is envisaged this model can be used to assess component knockdown factors without the need for expensive destructive testing.