A Noise Reduction Method for Seismic Data

Abstract

The present invention discloses a seismic data noise reduction method. For the seismic data noise reduction problem, the method comprises the steps of: S1, employing a robust Huber norm to replace anerror item of a target function to obtain the target function based on the Huber norm; S2, introducing a schatter p<-> norm into the target function obtained in the steps to obtain the target functionbased on the schatter p<-> norm; and S3, performing solution of the target function obtained in the step S2 to obtain robust estimation of a Hankel matrix. The seismic data noise reduction method caneffectively remove regular noise while suppressing random noise.

Description

technical field [0001] The invention belongs to the field of seismic data processing, in particular to a seismic data noise reduction technology. Background technique [0002] Seismic exploration is an important means to develop resources such as mineral deposits, oil and natural gas. In the process of exploration, the data will inevitably be affected by various factors, resulting in the presence of various types of noise in the collected data: for example, the amplitude and frequency of seismic signals gradually decay during the propagation process, and during the data collection process Noise always exists, which will lead to the decrease of the signal-to-noise ratio of the real signal, so that the useful signal is lost in the noise, which will affect our analysis of geological information. Therefore, we need an effective means of noise reduction to improve the quality of seismic data and improve the reliability of seismic exploration. [0003] The existing research work...

AI Technical Summary

Problems solved by technology

[0003] The existing research work is mainly integrated in the noise reduction of random noise. There are a lot of research work in this area, which can be divided into two categories: (1) Time-domain filtering methods: such as classical structure-oriented filtering, median filtering, edge-preserving filtering etc.; (1) Transform domain method: including commonly used tau-p transform, Radon transform, f-x domain filter, eigenimage filter, EMD, decomposition based on wavelet transform, radial trace domain and curvelet transform for removal. The operation on the transform domain coefficients will produce obvious side effects after the inverse transformation, such as the obvious vermicomposting phenomenon on the filtered gather records, etc.

More importantly, the above two frameworks cannot remove regular noise very well, so in actual seismic data processing, it is necessary to establish a processing flow to suppress Gaussian and non-Gaussian noise: first use the above method to remove Gaussian white noise, and then use a special On the one hand, it is very time-consuming and labor-intensive; on the other hand, the smoothing effect of random noise will affect the removal of regular noise

Images