Structure-oriented direction generalized total variation regularization method for denoising seismic data

Abstract

The invention discloses a structure adaptive direction generalized total variation regularization method for seismic data denoising, which includes the following steps: (1) Based on the gradient structure tensor, a new seismic event space-varying dip is proposed Calculation method; (2) put l 2 ‑The constant angle θ in the DTGV regularization model is generalized to the space variable angle Θ i,j , and then constructed a new structure-adaptive direction generalized total variation regularization model (l 2 ‑SADTGV regularization model); (3) using the Chambolle‑Pock primal‑dual algorithm for l 2 ‑SADTGV regularization model for solving. The method proposed by the invention is suitable for processing seismic data with complex geological structure. Compared with existing technologies, SADTGV regularization technology has the following advantages: (1) It can effectively improve the signal-to-noise ratio of noisy seismic data; (2) It can enhance the lateral continuity of seismic events and improve the vertical resolution of seismic sections And maintain geological structure characteristics such as fault information of seismic data; (3) It has better amplitude preservation ability.

Description

Technical field: [0001] The invention belongs to the technical field of seismic exploration, and relates to a Structure-Adaptive Directional Total Generalized Variation (SADTGV) regularization method for denoising seismic data. Background technique: [0002] Noise suppression plays a very important role in seismic data processing, and is the primary task to improve the accuracy of seismic data interpretation. Random noise is randomly distributed in the temporal and spatial domain of seismic data, covering the effective frequency band of seismic data. Therefore, random noise removal is an important issue in seismic data processing. Seismic data denoising methods commonly used in the industry include wavelet transform method, median filter, KL transform, singular value decomposition, etc. [0003] In recent years, the variational regularization method for solving inverse problems has been applied to suppress random noise in seismic data, and achieved certain results. Among t...

AI Technical Summary

Problems solved by technology

[0020] The purpose of the present invention is to provide a structure-adaptive directional generalized total variation (Structure-Adaptive Directional Total Generalized Variation, SADTGV) regularization method for seismic data random noise removal, to overcome the "not suitable for processing with Seismic data with spatially variable direction information"

Images