<p>In this paper, an automatic enhancement detection framework based on transient lock-in (TLI) thermography is proposed to achieve automatic enhancement inspection of thermal barrier coating (TBC) debonding defects. The framework encompasses thermal characteristic image extraction of TLI thermography, alpha blending strategy, and K-means clustering image segmentation. The TLI thermography is first used to address the artifact problem caused by the optical semitransparency of the ceramic topcoat in the TBC structure. A multi-frequency fusion method based on alpha blending strategy in TLI thermography is then proposed to enhance the visibility of TBC debonding defects. Finally, the K-means image clustering segmentation algorithm is applied to achieve the automatic non-destructive testing of TBC debonding defects. To demonstrate the superiority of the proposed method, four TBC specimens containing artificial defects with various ceramic topcoat thicknesses are prepared, and the corresponding characterization experiments of scanning electron microscope are also conducted. In addition, the effect of ceramic topcoat thickness, image fusion strategies, and post-processing methods on the detectability of TBC debonding defects are investigated in-depth, and the corresponding signal-to-noise ratio, detection rate, and runtime are also given in detail. Results reveal that the proposed method provides the highest signal-to-noise ratio, detection rate, and lowest runtime overall and can achieve automatic and rapid enhancement recognition of TBC debonding defects.</p>