.png){kind=icon}
.png){kind=icon}
Return
Adaptive Regularization-Based Reconstruction Algorithm for Neutron Exterior Computed Tomograph
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ndteint.2026.103715.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In neutron computed tomography (NCT) for non-destructive testing, the exterior problem often arises when the neutron beam cannot fully penetrate the sample, the detector size is limited, or only the outer region of the object is required for inspection. The resulting truncation of projection data leads to distortions and artifacts in reconstructed images. To address this issue, a dynamic-weighted anisotropic relative total variation (DW-ARTV) model is proposed, built upon the polar-coordinate ARTV (P-ARTV) framework. The model introduces a dynamic weighting mechanism that adaptively adjusts the regularization strength based on each pixel’s radial position and local gradient magnitude, effectively balancing central-region smoothing with edge preservation. Validation using two-dimensional numerical phantoms and real NCT data demonstrates that DW-ARTV surpasses conventional methods in suppressing radial artifacts, enhancing edge sharpness, and improving quantitative metrics, confirming its effectiveness and superiority in exterior NCT reconstruction.
Keywords:
Neutron computed tomography
Exterior problem
Dynamic-weighted anisotropic total variation
Image reconstruction
Artifact suppression
Journal
N
IF:
4.5
Papers:
181
Citations:
0