ADVERTISEMENT

Artificial Intelligence Improves Seismic-Image Reconstruction

You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.

To ensure continued access to JPT's content, please Sign In, JOIN SPE, or Subscribe to JPT

Seismic imaging provides vital tools for the exploration of potential hydrocarbon reserves and subsequent production-planning activities. The acquisition of high-resolution, regularly sampled seismic data may be hindered by physical or financial constraints, which lead to undersampled, sparse seismic data. However, if seismic data are available at a higher resolution and sampled evenly throughout the region of interest, the generated 3D models of petrophysical properties could be improved. Such improvements would show potential benefits through the successive steps of reservoir modeling and production planning.

Traditional Approaches

Traditional methods used to overcome the previously mentioned data-quality issues can be divided broadly into three categories.

  • Wave-equation-based methods. These methods use physics-based wave-propagation equations, using velocity models to reconstruct missing seismic traces.
  • Domain-transform methods. These are data-driven methods that involve transformation of data between different domains, such as time and frequency.
  • Prediction-error filters. These methods use a filter that learns from the known seismic data and constructs missing seismic data.
  • Given recent advances in the field of artificial intelligence (AI), it is worth examining whether AI methods also can be useful in the task of seismic-data reconstruction.

Generative Adversarial Networks (GANs)

GANs are a recent addition to the field of solution techniques. A variant of GANs, conditional generative adversarial networks, was used to test the efficacy of GANs for seismic-data reconstruction. Fig. 1 shows a schematic of a GAN model for seismic-data reconstruction. To effectively reconstruct a seismic image, a large number of 2D seismic images are fed into the GAN during the training process.

Fig. 1—Schematic of a GAN for seismic interpolation.

 

A GAN model consists of two main components:

  • Generator. This is a deep convolutional neural network (CNN) that uses random noise as an input and generates an image by expanding the input through a series of deconvolutions. As shown in Fig. 1, the generator network is fed 2D seismic images, with a portion of these masked to replicate missing traces. The generator network tries to reconstruct these masked portions of the 2D seismic images.
  • Discriminator. This is also a deep CNN that is shown the images reconstructed by the generator network and the original seismic images from which some parts were masked. The discriminator’s job is to distinguish between the images reconstructed by the generator from the original seismic images.

As training progresses, the generator network tries to create images that the discriminator will not be able to differentiate from the original seismic images. At the same time, the discriminator’s ability to distinguish the generated images from the original images continuously improves. As a result, the generator begins producing reconstructed seismic images, which look similar to the original ones.

Sample Results

The authors used a variation of the GAN architecture and trained it using 2D seismic images extracted from 3D seismic surveys available in the public domain. An important aspect of GANs is the intensive computational requirements, which were addressed by using graphical-processing-unit-accelerated cloud computing machines. At the end of training, 2D seismic images were reconstructed for varying widths of the masked portions in the original images. Fig. 2 demonstrates some of the results obtained from the trained GAN. The seismic interpolation performed by the GAN is capable of reconstructing the missing traces. A closer observation of the images in the second row of Fig. 2 reveals that a fault was reconstructed satisfactorily by the GAN, making the reconstructed image appear almost identical to the original image. As the width of the mask was increased, a slight deterioration in the quality of reconstruction was observed. However, the third row of Fig. 2 demonstrates the GAN’s ability to reconstruct prominent features of a seismic image, even in scenarios in which a significant number of traces are missing.

Fig. 2—The left column shows 2D seismic images with a portion masked to simulate missing traces in the seismic image. The central column shows the 2D seismic images reconstructed by GAN-enabled seismic interpolation. The right column shows original 2D seismic images. Each image is 256×256 pixels in size.

Conclusions

In this case study, the authors have assessed and demonstrated the capability of a deep learning approach based on GANs in reconstructing seismic images with missing traces. GANs not only reconstruct the missing data in simple scenarios but also show efficacy in reproducing complex features in seismic images such as faults. The image-­reconstruction concept is based on the context-aware interpolation of seismic data, a task learned by the GANs through intensive training over tens of thousands of seismic images. This concept can also be extended to generating finer-resolution seismic images for surveys where distance between subsequent inlines and crosslines is large because of financial or operational constraints. The high-resolution, equally sampled data can be helpful in reducing uncertainty in reservoir models, leading to better decisions regarding the location of reserves and more-efficient well placement. Hence, the discussed methodology can improve exploration and production decision-making significantly through use of reconstructed seismic data.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of the open-submission paper “Artificial Intelligence for Seismic-Image Reconstruction,” by Yogendra Narayan Pandey, SPE, and Govind Chada, Prabuddha, and Tejas Karmarkar, Oracle Cloud Infrastructure. The paper was not presented at an SPE conference and has not been peer-reviewed.

Artificial Intelligence Improves Seismic-Image Reconstruction

01 October 2019

Volume: 71 | Issue: 10

No editorial available

ADVERTISEMENT


STAY CONNECTED

Don't miss out on the latest technology delivered to your email weekly.  Sign up for the JPT newsletter.  If you are not logged in, you will receive a confirmation email that you will need to click on to confirm you want to receive the newsletter.

 

ADVERTISEMENT

No editorial available

ADVERTISEMENT

ADVERTISEMENT

ADVERTISEMENT