Optimizing algorithm for orientation determination in spatially resolved acoustic spectroscopy

Spatially Resolved Acoustic Spectroscopy (SRAS) is a nondestructive ultrasonic technique for measuring surface acoustic wave (SAW) velocity spectra across individual grains. It has been demonstrated that SRAS, when coupled with computational analysis of acoustic inverse problems, can estimate crystallographic orientation maps and elastic constants in both isotropic and anisotropic materials. While SRAS enables rapid scanning and SAW velocity mapping over large areas (e.g., 30 × 50 mm), the computational inversion step for orientation determination remains a bottleneck. This step typically relies on a brute-force 2-D cross-correlation between experimental SAW velocity spectra and a precomputed library of theoretical velocity spectra within the material’s fundamental zone (i.e., its symmetrically district orientations). The resulting optimization landscape is non-convex, making conventional search strategies computationally intensive. To address this, we investigate the performance of several optimization algorithms to accelerate orientation estimation while maintaining accuracy. Because orientation extraction is a prerequisite for elastic constants inversion, improvements at this stage potentially offer downstream benefits for the broader SRAS workflow. This approach aims to enhance the practicality of SRAS for high-throughput materials characterization by reducing latency in both orientation and elastic constant inversion.