A denoising method for artificial source electromagnetic observation signal

Abstract

The invention discloses a method for denoising observation signals by artificial source electromagnetic method. The method includes: first performing fast Fourier transform on the noise-containing signal, removing power frequency interference in the frequency domain signal, and performing inverse Fourier transform Then return to the time domain; then perform EEMD decomposition on the time domain signal after removing the power frequency interference, and attenuate the signal baseline drift; finally, based on the shift invariant sparse coding, further correct the signal baseline drift and remove other human noises. The present invention can effectively remove humanistic noises such as power frequency interference noise, baseline drift noise, and pulses whose amplitude is much larger than the useful signal without losing the useful signal, and allows strong noise to exist in the whole process from the beginning to the end of the observation , so that the artificial source electromagnetic method can be observed in a continuous and strong interference environment.

Description

technical field [0001] The invention belongs to the field of exploration geophysical signal processing, and relates to a denoising method for artificial source electromagnetic observation signals, in particular to a denoising method for periodic signals, including pseudo-random signals. Background technique [0002] In artificial source electromagnetic prospecting, the transmitter sends a controllable electromagnetic signal through a grounded electrode or an ungrounded return line, and then observes the signal through a receiver at a certain distance to obtain underground electrical structure information. Compared with the natural field source electromagnetic method, the signal-to-noise ratio of the artificial source electromagnetic method is significantly improved. However, in actual exploration, the signals collected by the receiver are still affected by various noises to varying degrees. Therefore, the collected signals are removed. Noise is still an essential task. [0...

AI Technical Summary

Problems solved by technology

[0003] Although the existing data processing methods such as data segment selection and robust estimation can achieve certain results, if the noise lasts for a long time, or even exists in the entire observation period from the beginning to the end, these methods will not be effective

Wavelet transform, morphological filtering and other methods are not ideal due to the loss of useful signals.

[0004] In addition, for the ubiquitous power frequency interference noise, existing instruments usually use hardware notch filters or use DSP and other devices to build digital notch filters. The circuit is too bloated, and because the performance of the notch filter is not ideal, there are problems of damaging useful signals to varying degrees

Although the digital notch filter simplifies the circuit scale, it has higher requirements on the performance of DSP and other devices, and also has the problem of damaging useful signals.

Images