Seismic wave impedance inversion method and device

Abstract

The invention discloses a seismic wave impedance inversion method and device, and the method comprises the steps: obtaining training seismic data which comprises labeled seismic data and unlabeled seismic data; establishing a closed-loop convolutional neural network CNN model, wherein the closed-loop CNN model comprises a forward modeling sub-model and an inversion sub-model, wherein the forward modeling sub-model is established according to a seismic forward modeling process, and the inversion sub-model is established according to a seismic inversion process; training a closed-loop CNN modelby using the labeled seismic data to obtain a converged closed-loop CNN model as an initial model; utilizing the label-free seismic data to train an inversion sub-model in the initial model to obtaina trained inversion sub-model; training a forward sub-model and a trained inversion sub-model in the initial model by using the training seismic data to obtain a seismic wave impedance forward-modeling and inversion model; and performing seismic wave impedance inversion by using the seismic wave impedance forward-modeling and inversion model. The seismic wave impedance inversion precision can be improved.

Description

technical field [0001] The present application relates to the field of geophysical exploration, in particular to a seismic wave impedance inversion method and device. Background technique [0002] Seismic wave impedance inversion is an important research content in seismic data interpretation. Wave impedance inversion predicts underground wave impedance parameters through observed seismic data, improves the resolution of seismic data, and helps engineers analyze the spatial structure and physical properties of underground strata. Traditional wave impedance inversion methods can be divided into inversion methods based on human experience and inversion methods based on data. The inversion methods based on human experience include recursive inversion, Dow integral inversion, generalized linear inversion, and logging constraint inversion, etc., which predict wave impedance data by adding sparsity constraints, building convolution models, and adding logging constraints . Data-...

AI Technical Summary

Problems solved by technology

[0004] The embodiment of the present application provides a seismic wave impedance inversion method to solve the problem of low accuracy of CNN-based seismic wave impedance inversion when the number of labeled samples is low. The method includes:

[0006] The embodiment of the present application also provides a seismic wave impedance inversion device to solve the problem of low accuracy of CNN-based seismic wave impedance inversion when the number of labeled samples is low. The device includes:

Images