Comparative Study of Diverse Techniques for Flaw Segmentation in TOFD Images of Austenitic Stainless Steel Weld

• Published in: Indian journal of science and technology 2015-11, Vol.8 (30), p.1-1
• Main Authors: Theresa Cenate, C. F., Sheela Rani, B., Sangeetha, D. N., Venkatraman, B.
• Format: Article
• Language: English
• Subjects: Algorithms; Austenitic stainless steels; Crystal defects; Discontinuity; Segmentation; TOFD; Welded joints
• ISSN: 0974-6846; 0974-5645
• DOI: 10.17485/ijst/2015/v8i1/78322

Grain structure of Austenitic Stainless Steel (ASS) weld causes difficulty in defect detection while inspecting using Time of Flight Diffraction (TOFD) method. This study aims at comparing segmentation methods to overcome this difficulty based on defect characterization. This study makes use of TOFD and Radiographic images of ASS weld pads fabricated with defined linear and volumetric defects. The Region-Based Level Set algorithm and Discontinuity Based Segmentation algorithm were explored for achieving flaw segmentation and quantitative characterization and validation of the result with that of standard radiographic results. The efficiency of the algorithms was analyzed by comparing and validating the size of defect with that of standard radiographic results in the form of error percentage. The consistency of error percentage in defect sizing achieved by Region-Based Level Set algorithm for all the test images given in the database indicate that, this algorithm is the best as compared to Discontinuity Based Segmentation algorithm for defect segmentation and characterization in TOFD images.