Remove the HFM noise by using Wavelet Neural Network

Abstract

During the HTEM signal processing we found that some kind of short-term noise till remained after the conventional denoising procedure. It is called high frequency motion-induced noise according to the research on its induced mechanism and time-frequency features. In this paper, we propose an approach to suppress the high frequency motion-induced noise based on wavelet neural network. The method uses high frequency motion-induced noise in dominance as a sample and then trains the wavelet neural network by it. After processing the field data, it is considered that the method proposed in this paper can effectively suppress the high frequency motion noise and the processing results build a good data foundation for subsequent processing.