Skip to Main Content
In non-destructive testing (NDT) field, noise suppression, or denoising is a permanent topic. In general, the NDT signal shows transient characteristics and the defect component varies in time. The conventional methods, such as Fourier analysis and filtering, linear averaging or simple thresholding can hardly reduce noise without losing the defect information. This paper introduces a technique based on thresholding by wavelet transform. The proposed technique compromises the continuous wavelet soft (shrinkage) and hard thresholding techniques. We explore the use of the continuous wavelets transform as powerful feature detection tools for data analysis. Our experimental results show its effectiveness on both noise removal and defect information preservation.