Powder Bed Fusion of Metal using Laser Beams (PBF-LB/M) as an additive manufacturing process is raising a great interest in many industrial sectors. However, the limited process understanding of the highly complex physical phenomena has hindered the further development of PBF-LB/M. Various build defects during the process have been found to be associated with the instability of the melt pool, whose dynamics and stability are driven by its temperature field. Accordingly, melt pool temperature monitoring is particularly crucial for product quality control. Nonetheless, the major difficulties in the melt pool temperature monitoring are the localized and extremely rapid heating and cooling, as well as the unknown radiation emissivity during the process. As a result, the understanding of these process fundamentals is limited.

We therefore aim to develop a multispectral imaging method to determine the absolute surface temperature of the melt pool. A multispectral camera records the light reflected from the surfaces in eight spectral bands. With a suitable spectral emissivity model, the absolute surface temperature and emissivity can be obtained simultaneously by solving an optimization problem that minimizes the spectral distance between the measured signal and the reconstructed signal according to Plank`s law. This integrated approach requires knowledge from the disciplines of laser technology, material, image processing, computer vision as well as function optimization. With a clear understanding of the surface temperature field, the melt pool stability and dynamics can be controlled and thus improving the productivity and quality of the PBF-LB/M.