Image processing methodology for detecting delaminations using infrared thermography in CFRP-jacketed concrete members by infrared thermography

Earthquake-prone countries like Japan pay close attention to structures’ seismic performance to improve the integrity of buildings and infrastructure during earthquakes. For decades, carbon fiber reinforced polymers (CFRPs) have been widely used to improve the seismic performance of the existing structures by enhancing their bending resistance and shear capacity. However, as CFRP layers deteriorate with age, the occurrence of delaminations between the protected mortar, concrete surface and CFRP sheets reduces their reinforcement effects, making them more susceptible to seismic loads. Because delaminations are often hard to detect manually via visual observations, it is imperative to develop effective strategies for inspecting the structures and detecting any defects. As such, the affected structural members can be repaired within the appropriate timeframe.

Among the non-destructive methods for visualizing structural defects, infrared thermography has drawn considerable research attention as an appropriate inspection method for CFRP-reinforced structures. This can be attributed to their convenience, time efficiency and capability to gather required data from a broad surface area. However, the application of infrared thermography is limited by the misevaluation of delaminated areas. Solving this challenge requires improving the accuracy of infrared thermography results using algorithms, which has increasingly become an area of interest amongst many researchers.

Herein, Dr. Jiancheng Gu and Dr. Shigeki Unjoh from Tohoku University proposed an image processing-based methodology for delamination detection in CFRP-jacketed concrete structures via infrared thermography. This work was an extension of their previous work published in another paper. Specimen used in this work were the same as those used in that work. But in this study, all the specimens were tested under natural conditions, which means in the outdoor and no heaters. Their main aim was to enhance the detection capability of this method and extend its application to long-term monitoring of CFRP-jacketed concrete members subjected to harsh environmental conditions. Generally, the present work was based on the previous study and was more close to the final target: applied to in-service structures. Their work is currently published in the journal, Composite Structures.

In the study, four specimens with 16 artificial delaminations that varied according to parameter values similar to existing CFRP-jacketed structures were evaluated under different weather conditions: summer, winter, rainy and sunny conditions. The artificial delaminations considered various experimental parameters, including size, surface cover mortar, depth and the water content of the delamination void. The testing surface temperature was measured throughout the experiment because delamination detection was based on the differences in the variation of surface temperature during the test period. In addition, a comparative test to use AI to automatically detect delaminated areas in original thermal images and process images, respectively, was conducted.

Results revealed that processing data using the proposed method could detect delaminations that were undetectable by original thermal images, such as delamination voids containing water and small-sized delimitations. Nevertheless, those with extremely small voids remained undetectable due to the absence of temperature differences between the delaminations and adjacent regions. The proposed method was also applicable in harsh natural conditions, allowing the determination of shapes and locations of most delaminated areas. Furthermore, results also revealed that the ambient temperature exhibited no significant effects on the results, but the time periods for testing and data collection intervals significantly affected the accuracy of the results, prompting the authors to recommend appropriate time values from the results of parametric analysis. Moreover, the findings also showed that the proposed method was more accurate and efficient when coupled with deep learning for automatic detection of delaminated areas.

In summary, the researchers presented a reliable methodology for improving the accuracy and efficiency of automatically detecting internal delamination in CFRP-jacketed structures using infrared thermography. In a statement to Advances in Engineering, Dr. Jiancheng Gu observed that the study findings could contribute to developing a more advanced structural inspection system consisting of data collection, analysis and delamination detection technologies for regular structural inspection.