Improved estimation of radial and axial gap of fuel pins from radiography images using image processing tools

Abstract

We report a new methodology for the precise estimation of radial gap from radiography images using image sharpening tools. The image sharpening is done by differentiation of pixels over a region, using a gradient approach. The gradient approach assumes relatively large gradient values for prominent edges in an image, small values for smooth regions and zero for regions that have constant grey values. Applying a first-order derivative to an image demarcates the boundaries and enhances the visibility of small steps and other finer details. This new approach facilitates enhanced detectability of radial gaps. Using this approach, radial gaps of 100 microns and above have been measured, which was not possible without employing image enhancement tools. The procedure has been successfully applied to the estimation of pellet-to-clad gap and pellet-to-pellet axial gap in mixed uranium-plutonium oxide ((U, Pu) O₂) fuel pins. The minimum radial gap (pellet-to-clad gap) measured using this approach is 64 microns. It has been found that the image processing tools can substantially reduce the unsharpness of the image due to scattering components and detector characteristics. This methodology is considered to be useful for the estimation of internal dimensions with better accuracy in assembled components using X-ray radiography images.