Infrared Thermography for Defect Detectability in Structural Health Monitoring of Concrete Structures: Heat Transfer Physics, Resolution Limits, and AI-Assisted Diagnostics

Infrared thermography (IRT) has emerged as a vital tool for structural health monitoring (SHM), yet the quantitative links between heat transfer physics and defect detectability in reinforced concrete remain underdeveloped. This review establishes a physics-oriented framework to clarify the mechanisms, operational parameters, and technological constraints governing IRT-based diagnostics. We analyze the interplay of thermal diffusivity, emissivity, and transient conduction in generating defect-induced contrasts. Quantitative synthesis identifies a minimum thermal contrast of 0.5 °C as a requirement for reliable detection, with active thermography proving effective for defects within a 5–10 cm depth range. Furthermore, we quantify how moisture-induced emissivity variations exceeding 10% introduce critical uncertainties in field measurements. The study underscores the necessity of integrating heat transfer theory with advanced signal processing and artificial intelligence (AI) to overcome current resolution limits. By defining these physical boundaries, this review provides a robust foundation for automated, intelligent infrastructure diagnostics, emphasizing that detectability is fundamentally governed by diffusion physics rather than raw temperature measurements.