Automated Defect Detection Algorithm Applied to Shearography in Composites

The aeronautic industry is progressively introducing composite material in its products. Nowadays the fabrication begins to be mature, but huge efforts remains to be done on the experimental validation of these materials. More specifically, these composite materials introduce new type of defects (such as delamination, glue defects…) which need to be detected during the fabrication and maintenance to prevent failures. Moreover, the widespread use of these materials impose to the future detection instruments to be cheap, robust and easy to manipulate by non-specialists. Currently shearography is widely used for the detection of defects in the composite material. It is cheap, and relatively robust and flexible, but requires special qualification to interpret the shearograms and identify the defects. Therefore, efforts need to be carried to ease the interpretation of stereographic results for non-specialist of the technique. For this, we propose to automate the identification of defects in shearograms by the development of specific algorithms.
In this paper, we present an innovative algorithm which realizes this operation. The process inspects the wrapped phase map obtained to identify high local fringe density. These potential detects are then analyzed and selected based on a set of parameter criteria such as dimensions, shape, and detection threshold. Successful results obtained on thermally solicited composite sample observed in industrial conditions are shown. We also discuss the limitations and perspective of this approach.