Fusion of multi-frequency eddy current signals-by using wavelet analysis method

Abstract

This paper presents a novel scheme to fuse one-dimensional multi-frequency eddy current signals by using multiresolution discrete wavelet analysis method. This technique consists of three steps. First, raw signals are preprocessed and decomposed into approximations and details at different resolution levels. The discrete wavelet transform is adopted at this stage. Then, several fusion processes are implemented in the coefficients domain. Finally, the inverse discrete wavelet transform is achieved to get the fusion result. In this technique, we develop a new mask-signal-modulated fusion algorithm to fuse in coefficients domain. It is performed and compared with other fusion methods based on the criteria of signal-to-noise ratio of the fused result. From our experiments, it demonstrates a better performance and shows a promising application on the two-dimensional multi-frequency eddy current signals.