New procedure for weld defect detection based-Gabor filter

Non-Destructive Testing (NDT) plays an important role in ensuring the reliable performance of welded components. Although, radiography is one of the most used techniques in weld defect inspection. It could become a challenging task since radiographic images are characterized by a low contrast, poor quality and uneven illumination. In this paper, we propose a new procedure of segmentation, based on Gabor filter and canny detector, to detect porosity. It consists of two stages: preprocessing step and edge detection step. Experiments are carried out on different weld radiographic images, collected from water pipes industry; show the good performance of the proposed approach in term of visual segmentation quality.