Seismic data processing using DUnet
A DUnet engine produces a processed image of seismic data acquired over an underground formation. The DUnet engine includes: a contractive path that performs multilayer convolutions and contraction to extract a code from the seismic data input to the DUnet, an expansive path configured to perform multilayer convolutions and expansion of the code, using features provided by the contractive path through skip connections, and a model level that performs multilayer convolutions on outputs of the contractive path and expansive paths to produce the processed image and/or an image that is a difference between the processed image and the seismic data. A fraction of the seismic data may be selected for training the DUnet engine using an anchor method that automatically extends an initial seismic data subset, based on similarity measurements. A reweighting layer may further combine inputs received from layers of the DUnet model to preserve signal amplitude trend.
1 . A seismic exploration method for denoising and/or deghosting seismic data, the method comprising:
obtaining seismic data acquired over an underground formation; and
generating a denoised image of the underground formation by reconstructing seismic signals from a noisy input domain of the seismic data using a DUnet engine,
wherein the DUnet engine reconstructs the seismic signals via:
a contractive path that performs multilayer convolutions and contraction to extract a code and contractive features from the seismic data input to the DUnet engine,
an expansive path configured to perform multilayer convolutions and expansion of the code to produce expansive features, using the contractive features provided by the contractive path through skip connections,
a model level that performs multilayer convolutions on a highest level of the contractive features and a highest level of the expansive features, the highest level of the contractive features and the highest level of the expansive features having a size as the seismic data input, and outputs of the multilayer convolutions are averaged to generate a model output; and
a reweighting layer comprising a locally connected layer that receives the model output and applies at least one of spatial and temporarily smooth coefficients and additional multilayer convolutions to preserve a signal amplitude trend of the seismic data to generate the denoised image.
2 . The method of claim 1 , further comprising training the DUnet engine using training data, which includes a fraction of the seismic data or other seismic data, and processed data corresponding to the fraction of the seismic data or the other seismic data.
3 . The method of claim 2 , wherein the fraction of the seismic data is selected using an anchor method that automatically extends an initial seismic data subset to the fraction of the seismic data, based on similarity measurements.
4 . The method of claim 2 , wherein parameters of the DUnet engine are optimized during training using an Adam algorithm.
5 . The method of claim 2 , further comprising:
validating the training by comparing the processed image produced by DUnet engine and processed data generated without using the DUnet engine.
6 . The method of claim 1 , wherein the seismic data is acquired during a marine survey with a variable depth streamer, and DUnet engine performs deghosting to produce the processed image.
7 . The method of claim 6 , wherein the DUnet engine employs generating mirror data from said seismic data, transforming said seismic data and said mirror data into Tau-P domain seismic data, dividing said Tau-P domain seismic data into a plurality of data windows, generating a ghost delay time for each of said plurality of data windows, and/or generating deghosted seismic data based on said ghost delay time to perform the deghosting.
8 . A seismic data processing apparatus for exploring an underground formation, the apparatus comprising:
an interface configured to obtain seismic data acquired over an underground formation;
a data processing unit configured to generate a denoised image of the underground formation by reconstructing seismic signals from a noisy input domain of the seismic data using a DUnet engine,
wherein the DUnet engine reconstructs the seismic signals via:
a contractive path that performs multilayer convolutions and contraction to extract a code and contractive features from the seismic data input to the DUnet engine,
an expansive path configured to perform multilayer convolutions and expansion of the code to produce expansive features, using the contractive features provided by the contractive path through skip connections,
a model level that performs multilayer convolutions on a highest level of the contractive features and a highest level of the expansive features, the highest level of the contractive features and the highest level of the expansive features having a size as the seismic data input, and outputs of the multilayer convolutions are averaged to generate a model output; and
a reweighting layer comprising a locally connected layer that receives the model output and applies at least one of spatial and temporarily smooth coefficients and additional multilayer convolutions to preserve a signal amplitude trend of the seismic data to generate the denoised image.
9 . The seismic data processing apparatus of claim 8 , wherein the data processing unit is further configured to perform training of the DUnet engine using training data, which includes a fraction of the seismic data or other seismic data, and processed data corresponding to the fraction of the seismic data or the other seismic data.
10 . The seismic data processing apparatus of claim 9 , wherein the data processing unit automatically extends an initial seismic data subset to the fraction of the seismic data using an anchor method, based on similarity measurements.
11 . The seismic data processing apparatus of claim 9 , wherein the data processing unit optimizes parameters of the DUnet engine during the training using an Adam algorithm.
12 . The seismic data processing unit of claim 9 , wherein the data processing unit is further configured to validate the training by comparing the difference between the processed image produced by DUnet engine and processed data that generated without using the DUnet engine.
13 . The seismic data processing apparatus of claim 8 , wherein the seismic data is acquired during a marine survey with a variable depth streamer, and DUnet engine performs deghosting to produce the processed image.
14 . The seismic data processing apparatus of claim 13 , wherein the DUnet engine employs generating mirror data from said seismic data, transforming said seismic data and said mirror data into Tau-P domain seismic data, dividing said Tau-P domain seismic data into a plurality of data windows, generating a ghost delay time for each of said plurality of data windows, and/or generating deghosted seismic data based on said ghost delay time to perform the deghosting.
15 . A non-transitory computer readable recording medium storing executable codes that when executed by a computer make the computer perform a seismic exploration method for denoising and/or deghosting seismic data, the method comprising:
obtain seismic data acquired over an underground formation; and
generate a denoised image of the underground formation by reconstructing seismic signals from a noisy input domain of the seismic data using a DUnet engine,
wherein the DUnet engine reconstructs the seismic signals via:
a contractive path that performs multilayer convolutions and contraction to extract a code and contractive features from the seismic data input to the DUnet engine,
an expansive path configured to perform multilayer convolutions and expansion of the code to produce expansive features, using the contractive features provided by the contractive path through skip connections,
a model level that performs multilayer convolutions on a highest level of the contractive features and a highest level of the expansive features, the highest level of the contractive features and the highest level of the expansive features having a size as the seismic data input, and outputs of the multilayer convolutions are averaged to generate a model output; and
a reweighting layer comprising a locally connected layer that receives the model output and applies at least one of spatial and temporarily smooth coefficients and additional multilayer convolutions to preserve a signal amplitude trend of the seismic data to generate the denoised image.
16 . The computer readable recording medium of claim 15 , wherein the method further comprises training the DUnet engine using training data, which includes a fraction of the seismic data or other seismic data and processed data corresponding to the fraction of the seismic data or the other seismic data.
17 . The computer readable recording medium of claim 16 , wherein the fraction of the seismic data is selected using an anchor method that automatically extends an initial seismic data subset to the fraction of the seismic data, based on similarity measurements.