Damage precursor detection in additively manufactured components with a multi-sensing and hybrid image fusion method

Additive manufacturing (AM) technology is pivotal in advancing aerospace manufacturing capabilities. Currently, multi-sensor damage precursor detection is a focal point of research aimed at identifying issues during the AM process. In this study, we developed a multi-source image acquisition system capable of simultaneously capturing both luminance and infrared intensity images, addressing critical challenges related to image quality and detection capability for damage precursors. By leveraging both visible and infrared imaging, we aim to detect a damage precursor that may elude single-sensor approaches. We propose a multi-sensor image fusion method, named NPP, which integrates the non-subsampled contourlet transform, principal component analysis, and pulse-coupled neural network frameworks to fuse visible and infrared optical image data for detecting damage precursors in complex environments. The performance of the fused images is compared and evaluated using various metrics. Experimental results indicate that our method excels in entropy information and achieves commendable results in edge strength, standard deviation, peak signal-to-noise ratio, and structural similarity. Comparative experiments demonstrate that our approach effectively enhances information richness, image contrast, and edge definition while preserving essential information from the source images. This study offers a novel solution, to the best of our knowledge, for multi-sensor information fusion within AM processes, presenting a breakthrough that addresses limitations hindering the broader application of AM technology in high-end aerospace applications.