Introspecting Local Binary Feature Vectors for Classification of Radiographic Weld Images

Abstract

AbstractWith the advent of technology and industrialization, welding has emerged as a potential tool. The process of welding gives a permanent joint to the material which is joined together, but it also affects the properties of the constituents. Any kind of deficiency in the welding process gives rise to weld defects. There should be a mechanism for accurate inspection of welded materials in order to maintain the superiority of design of the welded material and smooth operation. A proper welded material assures proper protection and trustworthiness. Non – Destructive Inspection is one of the significant features for accurate identification of the weld defects. Nowadays, this technique is extensively used for weld flaw detection as it doesn’t alter with the property of the welded objects. As computer technology has paved its way in all domain of engineering, the current research work focusses on discovery of a scientific solution for exact identification and classification of imperfections in welding. In the present work, the image database has been created from Welding research laboratory, Department of Mechanical and Industrial Engineering, Indian Institute of Technology Roorkee, Roorkee and consists of 79 radiographic weld images with 8 types of flaws and one without flaws. The present work explores the various local binary pattern feature vectors for the classification of flaws in radiographic weld images. It gives a concise description of the various existing feature extraction techniques for the proposed database.KeywordsNon – destructive testingWeld defectsLocal binary pattern feature vectors