Heuristic Enhancement of Magneto-Optical Images for NDE. (28th September 2010)

• Record Type: Journal Article
• Authors: Cacciola, Matteo; Megali, Giuseppe; Pellicanò, Diego; Calcagno, Salvatore; Versaci, Mario; Morabito, Francesco Carlo
• Other Names: Rebello João Marcos A. Academic Editor.
• Is Part Of: EURASIP journal on advances in signal processing. Volume 2010(2010)
• Publication Date: 2010-09-28
• Subjects: Signal processing -- Periodicals; Traitement du signal; Signal processing; Periodicals; 621.3822
• DOI: 10.1155/2010/485695
• Languages: English
• ISSNs: 1687-6172

Abstract

Abstract : The quality of measurements in nondestructive testing and evaluation plays a key role in assessing the reliability of different inspection techniques. Each different technique, like the magneto-optic imaging here treated, is affected by some special types of noise which are related to the specific device used for their acquisition. Therefore, the design of even more accurate image processing is often required by relevant applications, for instance, in implementing integrated solutions for flaw detection and characterization. The aim of this paper is to propose a preprocessing procedure based on independent component analysis (ICA) to ease the detection of rivets and/or flaws in the specimens under test. A comparison of the proposed approach with some other advanced image processing methodologies used for denoising magneto-optic images (MOIs) is carried out, in order to show advantages and weakness of ICA in improving the accuracy and performance of the rivets/flaw detection.