X-ray computed tomography-based porosity segmentation in additive manufacturing: Comparison of traditional and AI approaches

¹ Research group 3D Innovation, Department of Physics, Stellenbosch University, South Africa
² Department of Mechanical and Mechatronic Engineering, Central University of Technology, South Africa, Free State
³ Comet Technologies Canada Inc, Montreal, Canada

^* Corresponding author: 25805886@sun.ac.za

Abstract

Additive manufacturing is increasing in popularity and several manufacturing industries are adapting to the technology. This is due to the benefits of the process such as allowing for complex designs using a variety of materials. However, the occurrence of defects such as porosity in the manufacturing process remains a major concern and an active area of research. In this study, we show how the detection and analysis of porosity using X-ray computed tomography images is performed using different state of the art methods. The methods are demonstrated and compared for Ti6Al4V cantilever samples with lack of fusion and gas porosity at varying levels and include global thresholding methods, as well as artificial intelligence approaches. The advantages and disadvantages of each approach are discussed.