A Comparative Study of Region Growing & Morphological Edge Detection Segmentation Techniques to Detect Tine Flaws in Weldments

Abstract

A comparative study of morphological edge detection and region growing segmentation technique on NDT radiographic images of weldment is presented in this paper. With the help of these techniques the extraction of flaws, produced during the welding process is done, which is further used to measure the dimensions of flaw. This paper will describe the segmentation methods which are based on the discontinuity and homogeneity criteria of gray levels and intensity of pixels of an image. In edge detection method different morphological methods like dilation and erosion are' use to obtain best results. Similarly in region growing, multiple seed technique is used to find best results. These methodologies are compared and concluded to be effective for different types of weld flaws extraction after being successfully tested with 30 images (Standard defect radiographic images of welds in steel obtained from EURECTEST, International scientific Association Brussels, Belgium).