Seismic data interpolation method combining Gabor feature extraction and support vector regression

Abstract

The invention discloses a seismic data interpolation method combining Gabor feature extraction and support vector regression. The method comprises the steps of pre-interpolation, Gabor feature extraction, feature vector extraction, regression function training and seismic image reconstruction which are executed in sequence. The method comprises the following steps: performing pre-interpolation complementation on a missing seismic image lacking a seismic trace to obtain a low-resolution seismic image, performing Gabor filtering on the low-resolution seismic image to obtain a feature image, obtaining a prediction feature vector by utilizing the low-resolution seismic image and feature image transformation, and training a regression function by training the feature vector and a label; and finally, inputting the prediction feature vector into a trained regression function, and obtaining a reconstructed seismic image through regression reconstruction and transformation. According to the invention, the seismic data interpolation algorithm combining Gabor filtering and SVR is designed, the feature image obtained by Gabor feature extraction and the regression reconstruction capability of SVR are fully utilized, and a clear and complete seismic image is obtained.

Description

technical field [0001] The invention belongs to the field of seismic data processing, in particular to a seismic data interpolation method combined with Gabor feature extraction and support vector regression. Background technique [0002] Seismic exploration is the most effective exploration method in petroleum exploration, but field seismic data acquisition is affected by complex geological conditions and acquisition environment, which tends to make the acquired seismic data incomplete or irregularly distributed. Seismic data interpolation technology is used to solve the negative impact caused by insufficient spatial sampling in the process of data acquisition. A dense seismic record is necessary for many subsequent seismic processing steps. [0003] At present, traditional seismic data interpolation methods are mainly divided into the following categories: prediction-based methods, such as Spitzs f-x prediction filtering method; low-rank-based methods, such as singular sp...

AI Technical Summary

Problems solved by technology

Sparse transformation-based methods can only achieve good performance in certain sparse domains

Images