Implement of shade from shading method for improvement of weld defect detection in radiographic images

Abstract

Industrial Radiography is one of the most important methods to detect weld defects such as porosity, slag and crack. The interpretation of radiographic films characterizes the weld defects. The defect detection depends on radiographic film quality and the expert of interpreter. If the provided images of the industrial radiography are not clear, the detection of the defect can be difficult. But the radiographic image is usually noisy and has low quality. Thus, it is requred to use methods that can detect the defect accurately. The use of image processing techniques is available to achieve this aim. It is noticed that human eye is used to see objects in three dimensions and can also detect the depth. Thus, in this research, the shape from the shading (SFS) method is applied for two-dimensional radiographic images to extract three-dimensional ones. For the evaluation, the experts’ opinions in radiography have also been considered. The results of comments indicate that using the SFS method is valuable to improve the defects detection in weld radiography.

Keywords

• shape from shading
• image processing
• industrial radiography
• defect of weld