A near-infrared camera (NIR) is employed to monitor layer-by-layer in-situ electron beam powder bed fusion (E-PBF). We aim at demonstrating that such a device allows various kinds of flaws to be reliably captured thanks to qualitative thermal information. Defects such as distortions or uneven top surfaces can be detected in the NIR-images. Information about temperature heterogeneities can also be linked with some distortions occurring near overhangs. Some lack-of-fusion pores are revealed in the NIR-images. Unexpected localized defects appearing as isolated pores elongated along the building direction, and referred to as chimney pores, are also detected in the NIR-images. Pore-like defects can be described in 3D based on the NIR-images. Post-processing non-destructive characterization performed using X-ray computed tomography (XCT) is carried out to validate the information provided by the NIR-images. Comparisons between the 3D images obtained from the NIR-images acquired in-situ and processed using image analysis, and the 3D images obtained through XCT demonstrate that the morphology and spatial distribution of the defects can be reliably captured by NIR-imaging.

使用近红外摄像机（NIR）来监视逐层原位电子束粉末床熔化（E-PBF）。我们旨在证明这种设备可以通过定性的热信息可靠地捕获各种缺陷。可以在NIR图像中检测到诸如变形或顶面不平坦的缺陷。有关温度异质性的信息也可能与突出端附近发生的某些变形有关。在NIR图像中发现了一些融合不足的毛孔。在NIR图像中也检测到了意外的局部缺陷，这些缺陷出现为沿构造方向拉长的孤立孔，称为烟囱孔。可以在基于NIR图像的3D中描述类似毛孔的缺陷。进行使用X射线计算机断层扫描（XCT）执行的后处理非破坏性表征，以验证NIR图像提供的信息。从就地获取并使用图像分析处理的NIR图像获得的3D图像与通过XCT获得的3D图像之间的比较表明，可以通过NIR成像可靠地捕获缺陷的形态和空间分布。