Abstract
Abstract. The transient electromagnetic method (TEM) is extremely important in geophysics. However, the secondary field signal (SFS) in the TEM received by coil is easily disturbed by random noise, sensor noise and man-made noise, which results in the difficulty in detecting deep geological information. To reduce the noise interference and detect deep geological information, we apply autoencoders, which make up an unsupervised learning model in deep learning, on the basis of the analysis of the characteristics of the SFS to denoise the SFS. We introduce the SFSDSA (secondary field signal denoising stacked autoencoders) model based on deep neural networks of feature extraction and denoising. SFSDSA maps the signal points of the noise interference to the high-probability points with a clean signal as reference according to the deep characteristics of the signal, so as to realize the signal denoising and reduce noise interference. The method is validated by the measured data comparison, and the comparison results show that the noise reduction method can (i) effectively reduce the noise of the SFS in contrast with the Kalman, principal component analysis (PCA) and wavelet transform methods and (ii) strongly support the speculation of deeper underground features.
Highlights
Through the analysis of the secondary field signal (SFS) in the transient electromagnetic method (TEM), the information of underground geological composition can be obtained and has been widely used in mineral exploration, oil and gas exploration, and other fields (Danielsen et al, 2003; Haroon et al, 2014)
The whole process realizes the noise reduction of the secondary field actual signal based on the secondary field theoretical signal, and the model maps the singular points to locations where there is a high probability of occurrence, which is similar to the most estimative method based on observations and model predictions by Kalman filtering
We can find from the figure model of noise reduction based on the SFSDSA model of secondary field data that the SFSDSA model is better than Kalman filter, wavelet transform and principal component analysis (PCA) methods
Summary
Through the analysis of the secondary field signal (SFS) in the transient electromagnetic method (TEM), the information of underground geological composition can be obtained and has been widely used in mineral exploration, oil and gas exploration, and other fields (Danielsen et al, 2003; Haroon et al, 2014). It is necessary to make full use of the characteristics of the secondary field signal to reduce the noise in the data and increase the effective range of the data. Many methods have been developed for noise reduction of the transient electromagnetic method These methods can be broadly categorized into three groups: (1) Kalman filter algorithm (Ji et al, 2018), (2) wavelet transform algorithm (Ji et al, 2016; Li et al, 2017) and (3) principal component analysis (PCA) (Wu et al, 2014). The use of the deep learning model to reduce the noise of geophysical signals has not been applied. In this paper, the SFSDSA (secondary field signal denoising stacked autoencoders) model is proposed to reduce noise, based on a deep neural network with SFS feature extraction. AEs with noise reduction capability (denoising autoencoders, DAEs) (Vincent et al, 2008) have been widely used in image denoising (Zhao et al, 2014), audio noise reduction (Dai et al, 2014), the re-
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
