Tomographic reconstruction and 3D thermal modeling of laser powder bed fusion using infrared thermography

Laser powder bed fusion (LPBF) is widely regarded as one of the most promising additive manufacturing technologies. However, the influence of varying process conditions on component quality and performance has been a major barrier to its broader adoption. This study proposes and validates a tomographic reconstruction method for LPBF using infrared thermal imaging. By externally mounting an infrared thermal imager and acquiring thermal radiation signals through an infrared-transparent germanium glass observation window, the research mitigates image distortion through the development of a custom solid circular array calibration plate for camera parameter calibration and image rectification. To extract valid thermal signals, an adaptive image segmentation process based on an improved Otsu algorithm was developed, enabling effective isolation of laser-affected regions from complex backgrounds. Subsequently, image synthesis techniques were employed to reconstruct infrared signals during manufacturing, generating corresponding three-dimensional thermal models. Validation was conducted using two representative structural samples: inverted inclined plane (P-θ) and inverted conical surface (R-θ). Results demonstrate that the reconstruction method effectively characterizes surface conditions of non-visible internal walls, correlates thermal model chromatic distribution with surface roughness, and detects periodic thermal fluctuations during fabrication. The methodology maintains efficacy under simulated high-noise conditions representing powder interference, revealing top-layer defects including pores and localized overheating zones. Residual stress analysis further indicates that regions of elevated thermal accumulation within the thermal model correspond to reduced residual stresses, confirming the technique’s potential for indirect assessment of material property evolution. This approach establishes an effective visualization framework for LPBF process analysis, particularly enabling in-situ monitoring and quality assessment of complex internal architectures.