Compressive sensing of full field images in Lamb waves inspections

Lamb wave testing for structural health monitoring often relies on analysis of wavefields recorded through scanning laser Doppler vibrometers (SLDVs) or ultrasonic scanners. The informative content which can be extracted from these inspection is relevant. However, the acquisition process may be quite time-consuming. This paper presents a novel compressive sensing procedure to reduce the acquisition time while preserving the informative content. The proposed approach relies on the sparsifying properties of the Warped Curvelet transform. Such mathematical operator is particularly suited to represent multimode and dispersive Lamb waves. Experimental results show that the proposed approach allows to reduce the measurement locations required for exact signal recovery to less than 34% of the original sampling grid. The explanation of this success in data reduction lies in the curvelets ability to locally approximate the 3D geometry of the data.