A bidirectional predictive interpolation method for seismic data based on improved bp neural network

Abstract

The invention provides a seismic data bidirectional prediction interpolation method based on an improved BP (Back-Propagation) neural network. The method comprises the following steps: step I, obtaining seismic data information, determining a transverse interpolation or a longitudinal interpolation according to the actual condition of loss of the seismic data information, and establishing a corresponding bidirectional prediction mode; step II, improving the BP neural network by introducing an additional momentum method and a self-adaptive learning rate method, and establishing an improved BP neural network; step III, based on the bidirectional prediction mode established in the step I and the improved BP neural network established in the step II, respectively carrying out seismic data interpolation prediction from two opposite directions, and outputting prediction results; and step IV, fusing the seismic data interpolation prediction results in the two opposite directions obtained in the step III to obtain a final seismic data interpolation result. According to the method in the invention, multi-channel seismic information is fully utilized, and the interpolation precision and effect are improved.

Description

technical field [0001] The invention belongs to the technical field of seismic data processing, and relates to a bidirectional predictive interpolation method for seismic data based on an improved BP neural network. Background technique [0002] In the process of seismic data acquisition, missing seismic traces and insufficient spatial sampling are common. Therefore, seismic data interpolation technology is a commonly used method in the process of seismic data processing. [0003] At present, the commonly used seismic data interpolation technology is based on BP (Back-Propagation) neural network algorithm. The BP neural network algorithm is a back-propagation (Back-Propagation) learning algorithm proposed by Rumelhart in 1985 on the basis of the error back-propagation theory. The construction of BP neural network is based on multi-layer feed-forward network, which consists of input, output and hidden layers. The input signal is transmitted between the neurons of each layer...

AI Technical Summary

Problems solved by technology

[0004] However, since the BP neural network adopts the steepest descent method in nonlinear programming and modifies the weight according to the negative gradient direction of the error function, there are usually the following problems: low learning efficiency, slow convergence speed, and easy to fall into a local minimum state

Therefore, if the BP neural algorithm is used to directly predict earthquake interpolation, the interpolation result must have a certain deviation from the actual

[0005] In addition, the currently commonly used interpolation methods are mainly polynomial interpolation and spline interpolation. Polynomial interpolation is simple and approachable. However, it only considers the integrity and ignores local independence; spline interpolation is a piecewise polynomial interpolation algorithm. The polynomials on each adjacent segment have a certain connection, which not only maintains the simplicity and approximation feasibility of the polynomial, but also maintains the local independence between each segment, but it cannot make full use of multi-channel seismic information, which limits Its Application in Seismic Data Processing

Images