Method and system for extracting low-frequency part of seismic data by using deep learning method
A technology of deep learning and seismic data, applied in seismology, seismic signal processing, scientific instruments, etc., can solve problems such as low reliability and difficulty in obtaining low-frequency information, and achieve convenient promotion and application, improved accuracy and stable inversion sexual effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0034] This embodiment discloses a method for extracting low-frequency parts of seismic data using a deep learning method, including the following steps:
[0035] S1 simulates the pre-trained seismic wave signals to generate forward seismic data, and uses the forward seismic data to form a training data set.
[0036] The pre-trained seismic wave signals are forward simulated by selecting velocity models and wavelet waveforms, usually using the finite difference method. Velocity models include actual subsurface models and classical velocity models in the field of geophysics, such as the Marmousi model. Velocity models can be derived from sea models or land models. If the velocity model is an ocean model, the boundary condition of the ocean model adopts the absorbing boundary condition to prevent the ghost wave from suppressing the low frequency components. The wavelet adopts broadband Yu's wavelet, and the parameters P and Q of the wavelet are 1 and 20 respectively. The spec...
Embodiment 2
[0044] Based on the same inventive concept, this embodiment discloses another method for extracting the low-frequency part of seismic data using the deep learning method. No more details will be given, only the differences from Embodiment 1 will be described.
[0045] Get the picture of the high-frequency part of the Marmousi model data, such as Figure 5 shown. Carry out standardization processing to the high-frequency part of Marmousi model data, wherein the standardization processing process is the same as embodiment one, the picture of the high-frequency part of Marmousi model data through standardization processing is as follows Figure 6 shown. Bring the picture of the high-frequency part of the standardized Marmousi model data into the final deep learning model, and Figure 6 Extract the low frequency part in the middle. The extracted low-frequency part of the picture is as follows Figure 7 shown.
[0046] The image of the high-frequency part of the standardized ...
Embodiment 3
[0049] Based on the same inventive concept, this embodiment discloses a system for extracting low-frequency parts of seismic data using deep learning methods, including:
[0050] The data set forming module is used to generate forward modeling seismic data by simulating the pre-trained seismic wave signal, and utilize the forward modeling seismic data to form a training data set;
[0051] The model training module is used to establish an initial deep learning model, and train the initial deep learning model through the training data set to obtain the final deep learning model. The deep learning model is used to extract the feature map of the low frequency part of the seismic wave signal;
[0052] The low-frequency extraction module is used to bring the seismic wave signal to be extracted into the final deep learning model, and extract the low-frequency part of the seismic wave signal to be extracted.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More