Amoeba Based Novel Post-Processing Algorithm for Industrial Computed Tomography Scans Enhancement

Industrial Computed Tomography (ICT) for flaw profiling of critical systems is a highly efficient and widely used non-destructive testing (NDT) technique. Prompt image reconstruction is required while scanning industrial components; offering many benefits, such as increased throughput, reduced system downtime, etc. This can be accomplished by minimizing the projection views. However, it degrades the image quality which may introduce flaw misinterpretation. Thus, sophisticated post-processing methods can be implemented for image quality enhancement. A novel amoeba-based algorithm is proposed for ICT images enhancement from sparsely acquired projections. In the proposed scheme, the conventional amoeba filter is modified for amoeba shaping. The proposed scheme is validated by applying it on six different flaw profiles of a ball bearing phantom. The proposed scheme significantly improves the severely degraded and noise-contaminated reconstructed ball bearing phantom images. The effectiveness of the scheme is evaluated against traditional scheme for efficient crack profiling in ICT images.