Real-time seismic-image interpretation via deconvolutional neural network

Abstract

Seismic interpretation is now serving as a fundamental tool for depicting subsurface geology and assisting activities in various domains, such as environmental engineering and petroleum exploration. In the past decades, a number of computer-aided tools have been developed for speeding the interpretation process and improving the interpretation accuracy. However, most of the existing interpretation techniques are designed for interpreting a certain seismic feature (e.g., faults and salt domes) in a seismic section or volume at a time; correspondingly, the rest features would be ignored. Full-feature interpretation becomes feasible with the aid of multiple classification techniques. When implemented into the seismic domain, however, the major drawback is the low efficiency particularly for a large dataset, since the classification need to be repeated at every seismic sample. To resolve such limitation, this study proposes implementing the deconvolutional neural network (DCNN) for the purpose of real-time seismic interpretation, so that all the important features in a seismic image can be identified and interpreted both accurately and simultaneously. The performance of the new DCNN tool is verified through application of segmenting the F3 seismic dataset into nine major features, including salt domes, strong reflections, steep dips, etc. Good match is observed between the results and the original seismic signals, indicating not only the capability of the proposed DCNN network in seismic image analysis but also its great potentials for realtime seismic feature interpretation of an entire volume.

A machine-learning benchmark for facies classification

“…A few recent papers have illustrated the successful application of deconvolution networks for seismic interpretation applications (Alaudah et al, 2018;Di et al, 2018). Figure 7 illustrates the architecture of the deconvolution network used for both of our baseline models.…”

“…In recent years, there has been great interest in using fully-supervised deep learning models for seismic interpretation tasks such as facies classification (Shi et al, 2018;Dramsch and Lüthje, 2018;Waldeland and Solberg, 2017;Araya-Polo et al, 2017;Huang et al, 2017;Zhao, 2018;Di et al, 2018;Rutherford Ildstad and Bormann, 2017). Typically, deep learning models -such as convolutional neural networks (CNNs) -have millions of free parameters, and therefore require a large amount of annotated training data.…”

“…To address this problem, some researchers resort to annotating their own training and testing datasets. For example, in the Netherlands F3 block, Zhao (2018) annotated 40 inlines, Di et al (2018) annotated 12 inlines, while Rutherford Ildstad and Bormann (2017) only annotated a single inline for their model. The limited number of annotated sections is understandable given that the annotation process is time-consuming, requires subject matter expertise, and can be quite subjective.…”

Identification of Salt Deposits on Seismic Images Using Deep Learning Method for Semantic Segmentation

“…Motivated by good results in many areas, CNNs are becoming researchers' default choice for the segmentation of seismic images and identification of salt deposits. A huge number of papers [34][35][36][37][38][39][40][41] in 2018 and 2019 supports the claim. Dramsch and Lüthje [34] evaluated several classification deep CNNs with transfer learning to identify nine different seismic textures from 65 × 65 pixel patches.…”

“…Dramsch and Lüthje [34] evaluated several classification deep CNNs with transfer learning to identify nine different seismic textures from 65 × 65 pixel patches. Di et al [35] addressed the same problem using a deconvolutional neural network with three convolutional and three deconvolutional layers to speedup seismic interpretation. The same authors [36] addressed a problem of salt body delineation using a classification approach.…”