The Otsu method for assessing porosity and distribution of structural defects in the image of computed tomography of SLM products

The method of observation and statistical analysis of the number and distribution of structural defects (discontinuities of various morphologies and non-metallic inclusions) in the volume of loaded samples of powder stainless and high-strength alloys of Fe-Cr-Ni, Fe-Cr-Ni-Mo systems manufactured according to additive technology SLM is described.
The task of assessing porosity from a computed tomography (CT) image is to determine the presence of elements (objects) in the image with certain individual characteristics. This characteristic is the brightness of the image elements. The simplest and most natural way to detect an object/objects is to select a brightness threshold or threshold classification (thresholding). In the work, the Nobuyuki Otsu method, developed in 1979, was used for these purposes.
A comparison of the informative value of porosity results for the analysis of images obtained using X-ray computed tomography and metallographic analysis (optical, SEM) techniques is presented. The highest porosity content in the CT scanned sample was $\sim$0.61-0.82%. The pores in the samples of SLM alloys in both cases (optical microscopy and computed tomography) are unevenly distributed. The problems affecting the amount of information and quantitative characteristics of stress concentrators recorded in objects are analyzed.
The prospects of using the Ots algorithm for processing CT images and estimating the porosity distribution in samples of additive manufacturing are shown, which made it possible to visualize in detail the internal pores in samples in 3D without their physical and chemical destruction compared with metallographic preparation.