A denoising method for magnetotelluric signal based on noise discrimination

Abstract

The invention discloses a method for denoising magnetotelluric signals based on noise discrimination, which includes: calculating the approximate entropy and LZ complexity of each electromagnetic signal sample; using the approximate entropy, LZ complexity and category value training of each electromagnetic signal sample The preset classification model is used to obtain the noise discrimination classification model; the magnetotelluric signal to be processed is obtained, and then the noise discrimination is performed on the magnetotelluric signal to be processed according to the noise discrimination classification model to obtain the electromagnetic signal segment of non-strong interference and the electromagnetic signal segment of strong interference; The electromagnetic signal segment with strong interference is combined with complementary set empirical mode decomposition and wavelet threshold method for noise suppression processing; the electromagnetic signal segment after denoising suppression processing is merged with the electromagnetic signal segment without strong interference to obtain the reconstructed magnetotelluric signal. The present invention can more accurately discriminate strong interference and non-strong noise interference data segments through the above method, retain the real magnetotelluric signal, and improve the denoising effect of the magnetotelluric signal.

Description

Technical field [0001] The invention belongs to the technical field of magnetotelluric signal processing, and specifically relates to a method for denoising magnetotelluric signals based on noise discrimination. Background technique [0002] Magnetotelluric (MT) method is an electrical exploration method that uses natural field sources. It studies the electrical properties and electrical properties of underground rock formations by observing natural alternating electromagnetic fields with regional and global distribution characteristics on the ground. Distribution characteristics. This method has large detection depth, simple construction and low cost, and has been widely used in many fields of geophysics. However, due to the weak signal of the natural magnetotelluric field and the wide frequency band, the data collected in the field will inevitably be disturbed to varying degrees, especially the increasingly serious human noise interference, which leads to the deterioration of ...

AI Technical Summary

Problems solved by technology

Such as far reference method, Robust method, least square method, wavelet transform, Hilbert-Huang transform, morphological filter, etc., can suppress noise to a certain extent and improve the quality of magnetotelluric data, but these methods have certain limitations, and Focus on overall processing, lack of noise screening, the result often results in over-processing of magnetotelluric data and loss of a large amount of low-frequency slow-change information, resulting in poor denoising effect

Images