Phase estimation via riesz transform in laser speckle interferometry for large-area damage imaging

Abstract

A direct phase estimation (DPE) via Riesz transform from a speckle pattern (images) for visualizing barely visible impact damage (BVID) occurring in layered composites is proposed in this paper. Rather than complex optical hardware setup for phase extraction in commercial digital speckle pattern interferometry (DSPI) or shearography (SG) systems, the proposed numerical (software) DPE implementation via Riesz transform with a log-Gabor filter enables a compact, portable, and cost-effective system for large-area real-time nondestructive inspection (NDI) due to its simplicity, stability, and efficiency.Using speckles in laser speckle interferometry (LSI) speckle pattern is intrinsically noisy; the correlation-based phase signatures derived from Riesz transform followed by the log-Gabor filter is implemented to supplant the conventional phase-difference algorithm to localize damage regions, especially in the case where the high-power diode laser (low coherence) can cause serious phase drift. By enveloping the monogenic signal through Riesz transform for direct phase retrieval is more intuitive than TPS or SPS. This yields visualization of the damage feature with higher fidelity. Consequently, a phase-correlation algorithm becomes a more robust operation to cope with speckle images from the interferometry.This full-field optical SG system was demonstrated and verified in an impacted damaged honeycomb composite panel using the proposed phase-correlation algorithm under thermal stress loading for large-area inspection in near real-time. The visualized damage images enhanced by principal component analysis (PCA) agreed with images from pulse laser/LDV and X-ray CT scan. The DPE algorithm has a promising potential to apply on any optical metrology for instantaneous phase estimation and can combine with selected imaging conditions for multi-damage detection.