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

•An efficient nondestructive technique using infrared thermography for delamination detection in concrete structures strengthened by the CFRP jacketing.•Considered the effects of various delamination features that exist in real structures on detection accuracy.•Proposed a methodology that includes a standard flow to process images for detecting boundaries and determining delamination regions.•Verified the effectiveness of the proposed methodology based on experiments under several different natural conditions.•Showed the proposed methodology to have good potential for automatic delamination detection based on deep learning. This study presents a methodology for detecting delaminations in carbon fiber reinforced polymer (CFRP)-jacketed concrete structures by infrared thermography. Four specimens with artificial delaminations were evaluated through passive experiments under different weather conditions including winter, summer, sunny, and rainy conditions. The test parameters considered for the artificial delaminations in the specimens included size, depth, surface cover mortar, and the water content in the delamination void. The methodology detected delamination regions by boundary recognition based on the differences in surface temperature variations during a period. It could detect delaminations more efficiently and accurately than visual assessments based on thermal images. Furthermore, a few delaminations that were undetectable by thermal images were detected after image processing with the proposed methodology. In addition, the accuracy of the results was significantly affected by the time period for testing and the data-collection intervals. We discuss the recommended values obtained by parametric analysis and implement an application example using the proposed method and deep learning based on the experimental data.