Application of frequency domain processing to X-ray radiographic images of welding defects

Abstract

X-ray radiographic technique is widely used as Nondestructive Testing (NDT) for inspection of welding joints. The quality of X-ray images is frequently sensitive for noise, lighting and working place surroundings. Therefore, such images need to be carefully examined to identify welding discontinuities. However, the human interpretative are usually contradictory and may cause confusion. Present work has proposed an image processing technique based on frequency domain processing to clarify the X-ray radiographic images taken for different welding discontinuities. The images are captured using low-resolution video camera and high-resolution digital camera. Preprocessing imaging tasks are performed to reduce the noise in the low-resolution video sequence images. Subsequently, low-pass followed by emphasis high-pass frequency filters are applied to enhance the images with appropriate cut-off frequencies. It is found that high-frequency domain image filtering is enhancing the morphological features of the X-ray images. The enhanced images become visually cleared and the welding discontinuities are possibility accessible.