BayWT Image Fusion Method for Enhancement of Eddy Current Sub-surface Defect Images

• Published in: SN computer science 2023-11, Vol.4 (6), p.837, Article 837
• Main Authors: Soni, Anil Kumar, Soni, Aradhana, Tamrakar, Chandan
• Format: Article
• Language: English
• Subjects: Advances in Machine Vision and Augmented Intelligence; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Computer vision; Data Structures and Information Theory; Decomposition; Defects; Eddy current testing; Eddy currents; Entropy; Image enhancement; Information Systems and Communication Service; Methods; Nondestructive testing; Original Research; Pattern Recognition and Graphics; Performance evaluation; Probability; Signal to noise ratio; Software; Software Engineering/Programming and Operating Systems; Standard deviation; Statistical inference; Steel plates; Surface defects; Vision; Wavelet transforms
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-023-02310-1

Eddy current (EC) testing is one of the mostly used non-destructive evaluation (NDE) techniques for detection and imaging of defects in conducting material. The process of combination of the multiple images in to a single image to get the clear information is called image fusion. In this paper, wavelet transform and Bayesian principle-based image fusion method (BayWT) is proposed for enhancement of detectability and signal-to-noise ratio (SNR) of EC defect images. BayWT is used for fusion of EC sub-surface defect images generated at different frequencies from AISI type 304L stainless steel plate. The proposed fusion method is evaluated using different image metrics (i.e. SNR, entropy, standard deviation and fusion mutual information) and achieved three-time additional SNR. The performance of the BayWT fusion method is also compared with the commonly used NDE image fusion methods.