Improved Utilization of Radiographs of Ancient Metal Objects under Heterogeneous Image Blurring

Abstract

ABSTRACT Availability of pertinent and detailed information for damaged ancient objects is vital for successful restorative, damage limitation intervention processes in archeology. Visual inspection and radiography testing (RT) constitute widely used non-destructive methods for conservators of such objects. Unfortunately, the quality of RT images is susceptible to blurring caused by the superimposed scattered X-rays. In this study, the no-black-pixel constraint (NBPC) algorithm on its own as well as in the form of an algorithm with the planar-scene assumption (NBPC + PA) was applied to RT images with a view to suppressing image blurring and enhancement of imaging contrast. The results of this study have shown, both qualitatively and quantitatively that the NBPC + PA and NBPC algorithms provide higher imaging contrast for identification and evaluation of damage in corroded ancient metallic relics than using unprocessed RT images alone. Contrast enhancement of up to about 120% and 78% was achieved in comparison with the original image by applying the NBPC+PA and NBPC algorithms, respectively. Practical application and effectiveness of using the processed images were evaluated by expert operators in the area and it was found that using the reconstructed images scored higher (on average by 88%) than the original RT images (on average by 73%) with the NBPC+PA scoring higher (on average by 91%) than the NBPC algorithm (on average by 85%). The operators have found utilizing the processed images leads to better inspection and evaluation of the object enabling better targeted informed for restorative intervention, damage limitation, and maintenance of such objects.