Toward practical guidelines for infrared thermography of concrete bridge decks: A preliminary investigation across U.S. climate zones

Delamination detection in concrete bridge decks employing IRT relies heavily on selecting favorable environmental conditions to produce the desired detection accuracy. Many studies have focused on improving IRT detection by optimizing the data collection or analysis techniques. However, there exists limited and conflicting literature on recommended data collection times. Therefore, this study evaluated favorable detection windows across the five climate zones in the U.S. using field and numerical simulation data. First, the field data was collected on two separate days to verify the numerical simulation results by developing a simulation with identical conditions. The verified numerical simulation results indicated that the simulations were representative of real-life detection patterns. So, numerical simulations for each zone and season were developed. The data collected from the numerical simulations were analyzed using the Sørensen–Dice Coefficient (SDC), comparing the known delamination location (ground-truth) to the predicted delamination through a threshold-based segmentation technique to predict favorable detection windows based on obtained accuracy. The contributions of this preliminary study are as follows: (1) a report of favorable detection windows per zone and season, which, in turn, potentially improves detection accuracy; The favorable conservative reported window for each zone was determined to be from noon to early evening, (2) provide a replicable numerical simulation process, enabling interested agencies to utilize the approach to conduct internal studies to evaluate favorable detection windows for their specific location, and (3) statistically evaluate the combined effects of ambient temperature and solar irradiance values and patterns on detection accuracy. • Uses Infrared Thermography to predict best delamination detection times across U.S. • Reported times aim to improve detection accuracy. • Statistically evaluates effect of temperature and solar irradiance on detection. • Replicable, validated method to predict detection windows in any setting or region. • Leverages IRT for cost-effective, efficient, and accurate delamination detection.