ADF based anisotropic diffusion model applied to radiography image enhancement

Weld joints quality control with radiography technique is widely used. Unfortunately, the poor quality of these images and the presence of various kind of noise make processing of radiography film an intricate task. This paper presents a new anisotropic diffusion filtering method for radiography image noise reduction and sharpening. The proposed diffusion scheme is guided with the adaptive diffusion flow vector ADF. The use of the ADF vector flow permits to preserve perfectly weak image boundaries while removing noise. Experimental results on different synthetic and real welding radiography images confirm the efficiency and robustness of our model in comparison with other diffusion methods.