Defect Detection Using Hidden Markov Random Fields

Date
2005-01-01
Authors
Dogandžić, Aleksandar
Eua-Anant, Nawanat
Dogandžić, Aleksandar
Zhang, Benhong
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Journal Issue
Series
Department
Electrical and Computer Engineering
Abstract

We derive an approximate maximum a posteriori (MAP) method for detecting NDE defect signals using hidden Markov random fields (HMRFs). In the proposed HMRF framework, a set of spatially distributed NDE measurements is assumed to form a noisy realization of an underlying random field that has a simple structure with Markovian dependence. Here, the random field describes the defect signals to be estimated or detected. The HMRF models incorporate measurement locations into the statistical analysis, which is important in scenarios where the same defect affects measurements at multiple locations. We also discuss initialization of the proposed HMRF detector and apply to simulated eddy‐current data and experimental ultrasonic C‐scan data from an inspection of a cylindrical Ti 6‐4 billet.

Comments

The following article appeared in AIP Conference Proceedings 760 (2005): 704 and may be found at doi:10.1063/1.1916744.

Description
Keywords
Citation
DOI
Collections