Sensor residual self-coding network seismic data denoising method

Abstract

The invention discloses a sensor residual self-coding network seismic data denoising method, and aims to overcome the defect that a traditional denoising method cannot meet the high precision exploration requirement on the aspects of generalization ability, denoising fidelity and the like. Characteristic data coded and reconstructed by a convolution self-coding network are output by a decoding part after passing through a multilayer sensor residual convolution self-coding block constructed by a multi-layer sensor convolution layer, a multi-scale layer, a BN layer and an auto-encoder, and combined with an optimized convolution core, so that the sensor residual self-coding network seismic data denoising method based on deep learning is put forward. The seismic data denoising method disclosedby the invention can still remove multiple waves and random noise at the same time while completely retaining the local details of seismic data and generating no false impression.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to a perceptron residual self-encoding network seismic data denoising method in seismic data denoising. Background technique [0002] Seismic data noise can be divided into coherent noise and random noise. Traditional denoising needs to select different methods for denoising according to the characteristics of signal and noise. The traditional methods for removing multiples can be divided into two categories, one is the filtering method based on the difference between effective waves and multiples, among which the Radon transform method and predictive deconvolution method are more commonly used, and the other is based on The method of wave theory, in which the wave equation extrapolation method and feedback method are more commonly used, Radon transform denoising will produce more artifacts at the same time; the prediction deconvolution method is more dependent on parameter se...

AI Technical Summary

Problems solved by technology

Among them, based on Auto-Encoder (AE) network, Convolutional Neural Network (CNN), Generative Adversarial Networks (GAN) are widely used in the field of denoising, but the earthquake denoising method based on deep learning Less, the random noise removal algorithm based on the residual convolutional neural network has stronger denoising performance; based on the convolutional neural network denoising model, it can remove seismic random noise with unknown variance; the residual convolutional neural network image The denoising method applied to seismic data denoising can effectively remove random noise, but the above deep learning method can only remove random noise and cannot suppress multiple waves

Images