Image-Based Ultrasound Speed Estimation in Isotropic Materials

Abstract

In Nondestructive Evaluation (NDE), ultrasound imaging is a useful tool for detecting flaws or measuring corrosion in metal pipes. Phased-array systems are becoming standard in NDE, as well as post-processing imaging techniques such as the Total Focusing Method. These techniques require knowledge of parameters such as distance between transducer elements, times of arrival and sound speed to produce sharp images. A deviation in sound speed, for instance, may compromise image accuracy. In this paper, we propose a sound speed estimator based on image quality metrics. The approach consists in generating several reconstructed images from the same acquisition signals, but using a set of candidate sound speeds. The speed value that yields the sharpest image is then chosen as the estimated sound speed. Estimates made on data obtained by simulation of a typical scenario presented relative combined uncertainties as low as 0.0166%. The results, obtained in both simulated and real scenarios, show good agreement with expected speeds. The proposed method can be considered a reference-free sound speed calibration method aimed at field applications.