Review on optimization strategies of probabilistic diagnostic imaging methods

Abstract

With the continuous development of intelligent infrastructure, structural health monitoring (SHM) and non-destructive testing (NDT) have become major research focuses. Ultrasonic-guided wave imaging technology not only integrates the global impact of damage on structures but also provides intuitive localization and severity characterization of the damage. Probabilistic diagnostic imaging (PDI) methods, which do not require direct interpretation of guided wave signals and can achieve high-quality imaging with sparse arrays, have garnered increasing attention. This paper introduces the principles, general processes, and technical advantages of PDI methods. Based on the process of the PDI, existing optimization strategies are categorized into two types: internal process optimizations, which include sensor layout, damage indices optimization, construction of the distribution weight function, and data fusion; and external process optimizations, which include spurious image suppression, on-site environment detection, and integration of methodologies, each analyzed in detail. With the affirmation of the value of these strategies, this paper also highlights the current issues within these methods and explores potential future developments by integrating emerging technologies such as intelligent sensing, big data, and artificial intelligence. These insights provide valuable reference suggestions for the continued optimization of these methods.