Key points are not available for this paper at this time.
The structural integrity of concrete infrastructure requires continuous monitoring to prevent catastrophic failures, yet traditional inspection approaches struggle with accuracy limitations and subjective interpretation. This comprehensive analysis examines how infrared thermography has evolved through computational integration to address these fundamental challenges in structural assessment. By systematically evaluating research developments from 2016 to 2025, this investigation bridges a crucial knowledge gap through an innovative framework that links advanced computational methodologies with real-world infrastructure applications. Our analysis diverges from existing literature by explicitly focusing on the transformative role of artificial intelligence in elevating thermal imaging from visual observation to sophisticated automated diagnosis. Through rigorous examination of carefully selected studies following established systematic review protocols, we reveal how emerging computational technologies are reshaping infrastructure inspection practices. This work introduces groundbreaking organizational structures for understanding image enhancement methodologies, provides a thorough evaluation of neural network architectures tailored to thermal analysis, and establishes the first systematic comparison of data enhancement techniques designed for thermographic applications. A particularly significant contribution involves our detailed examination of sensor fusion strategies, uncovering effective integration methods for combining thermal signatures with complementary imaging data across multiple processing stages. Our findings demonstrate how sophisticated computational frameworks effectively address historical limitations, including weather-dependent variability, subsurface detection constraints, and inconsistencies in human interpretation, thereby facilitating more reliable infrastructure evaluation. The research highlights critical disparities between controlled experimental validation and operational field performance, underscoring the need for adaptive systems that can maintain reliability across diverse environmental conditions. By identifying emerging technological convergences and delineating future research priorities encompassing physics-guided learning, autonomous knowledge discovery, and transparent decision systems, this analysis provides essential guidance for advancing infrastructure monitoring capabilities. The synthesis provides practitioners with evidence-based selection criteria for implementing thermal inspection solutions tailored to varying operational requirements, particularly valuable as global infrastructure ages while maintenance resources remain constrained.
Abdulridha et al. (Tue,) studied this question.