Iterative image deconvolution for resolution enhancement in X-ray non-destructive testing

Abstract

X-ray is a widely used technique in industrial and medical imaging. Especially for non-destructive testing of safetycritical components, high spatial resolution is crucial for the reliable detection of subtle details. To this end, iterative image deconvolution is a common approach in photography and microscopy to retrospectively improve spatial resolution. In this work, the iterative Richardson-Lucy (RL) deconvolution algorithm is applied to X-ray images for various models of the point spread function (PSF), based on the slanted edge method. The image quality is evaluated using the interpolated basic spatial resolution (iSRb) based on the ASTM E2002 standard, allowing for RL parameter optimization. Using optimized damping and iteration parameters, a resolution improvement of over nine percent with limited artifact formation is achieved. The results show that RL deconvolution is a promising approach for improving spatial resolution, offering high potential for industrial X-ray.