Pre-stack Seismic Facies Analysis via Waveform Sparse Representation

Geophysics ◽  
2020 ◽  
pp. 1-77
Author(s):  
Kunhong Li ◽  
Zhining Liu ◽  
Bin She ◽  
Jiandong Liang ◽  
Guangmin Hu
Keyword(s):  
Spatial Structure ◽  
Dictionary Learning ◽  
Seismic Data ◽  
Facies Analysis ◽  
Intensity Variation ◽  
Seismic Facies ◽  
Distance Metric ◽  
Structure Variation ◽  
Waveform Data ◽  
Seismic Facies Analysis

Seismic facies analysis based on pre-stack data is becoming popular. Vertical elastic transitions produce the spatial structure variation of pre-stack waveforms, while lateral elastic transitions produce the amplitude intensity variation. In the stratigraphic seismic facies analysis, more attention should be paid to waveform spatial structure than amplitude intensity. Conventional classification methods based on distance metric are difficult to adapt to stratigraphic seismic facies analysis because a distance metric is a comprehensive measure of waveform structure and amplitude intensity. A dictionary learning method for pre-stack seismic facies analysis is proposed herein. The proposed method first learns several dictionaries from labeled pre-stack waveform data, and these dictionaries consist of several normalization vector bases. The pre-stack waveform spatial structure is therefore embedded in these learned dictionaries, and the amplitude intensity is eliminated by the normalization process. Afterward, these dictionaries are used to sparsely represent pre-stack seismic data. Seismic facies are classified and determined according to representation error. A source error separation method is used to improve the anti-noise performance of dictionary learning by iteratively segmenting the noise out in the training data. The results on synthetic and real seismic data show that the proposed method has a stronger tolerance to noise, and the obtained seismic facies boundary is more accurate and clearer. This demonstrates that the proposed method is an effective seismic facies analysis technique.

Artificial immune-based self-organizing maps for seismic-facies analysis

Geophysics ◽  
2012 ◽  
Vol 77 (4) ◽  
pp. O45-O53 ◽  
Author(s):  
Puneet Saraswat ◽  
Mrinal K. Sen
Keyword(s):  
Immune System ◽  
Data Reduction ◽  
Seismic Data ◽  
Artificial Immune System ◽  
Facies Analysis ◽  
Seismic Facies ◽  
Artificial Immune ◽  
Self Organizing Maps ◽  
Data Set ◽  
Seismic Facies Analysis

Seismic facies, combined with well-log data and other seismic attributes such as coherency, curvature, and AVO, play an important role in subsurface geological studies, especially for identification of depositional structures. The effectiveness of any seismic facies analysis algorithm depends on whether or not it is driven by local geologic factors, the absence of which may lead to unrealistic information about subsurface geology, depositional environment, and lithology. This includes proper identification of number of classes or facies existing in the data set. We developed a hybrid waveform classification algorithm based on an artificial immune system and self-organizing maps (AI-SOM), that forms the class of unsupervised classification or automatic facies identification followed by facies map generation. The advantage of AI-SOM is that, unlike, a stand-alone SOM, it is more robust in the presence of noise in seismic data. Artificial immune system (AIS) is an excellent data reduction technique providing a compact representation of the training data; this is followed by clustering and identification of number of clusters in the data set. The reduced data set from AIS processing serves as an excellent input to SOM processing. Thus, facies maps generated from AI-SOM are less affected by noise and redundancy in the data set. We tested the effectiveness of our algorithm with application to an offshore 3D seismic volume from F3 block in the Netherlands. The results confirmed that we can better interpret an appropriate number of facies in the seismic data using the AI-SOM approach than with a conventional SOM. We also examined the powerful data-reduction capabilities of AIS and advantages the of AI-SOM over SOM when data under consideration were noisy and redundant.

A visual data-mining methodology for seismic facies analysis: Part 1 — Testing and comparison with other unsupervised clustering methods

Geophysics ◽  
2009 ◽  
Vol 74 (1) ◽  
pp. P1-P11 ◽  
Author(s):  
Iván Dimitri Marroquín ◽  
Jean-Jules Brault ◽  
Bruce S. Hart
Keyword(s):  
Data Mining ◽  
Pattern Recognition ◽  
Seismic Data ◽  
Facies Analysis ◽  
Clustering Algorithms ◽  
Unsupervised Clustering ◽  
Seismic Facies ◽  
Visual Data ◽  
Visual Data Mining ◽  
Seismic Facies Analysis

Seismic facies analysis aims to identify clusters (groups) of similar seismic trace shapes, where each cluster can be considered to represent variability in lithology, rock properties, and/or fluid content of the strata being imaged. Unfortunately, it is not always clear whether the seismic data has a natural clustering structure. Cluster analysis consists of a family of approaches that have significant potential for classifying seismic trace shapes into meaningful clusters. The clustering can be performed using a supervised process (assigning a pattern to a predefined cluster) or an unsupervised process (partitioning a collection of patterns into groups without predefined clusters). We evaluate and compare different unsupervised clustering algorithms (e.g., partition, hierarchical, probabilistic, and soft competitive models) for pattern recognition based entirely on the characteristics of the seismic response. From validation results on simple data sets, we demonstrate that a self-organizing maps algorithm implemented in a visual data-mining approach outperforms all other clustering algorithms for interpreting the cluster structure. We apply this approach to 2D seismic models generated using a discrete, known number of different stratigraphic geometries. The visual strategy recovers the correct number of end-member seismic facies in the model tests, showing that it is suitable for pattern recognition in highly correlated and continuous seismic data.

Unsupervised seismic facies analysis via deep convolutional autoencoders

Geophysics ◽  
2018 ◽  
Vol 83 (3) ◽  
pp. A39-A43 ◽  
Author(s):  
Feng Qian ◽  
Miao Yin ◽  
Xiao-Yang Liu ◽  
Yao-Jun Wang ◽  
Cai Lu ◽  
...  
Keyword(s):  
Feature Extraction ◽  
Seismic Data ◽  
Facies Analysis ◽  
Recognition Task ◽  
Seismic Facies ◽  
Self Organizing Maps ◽  
Convolutional Autoencoder ◽  
Unsupervised Deep Learning ◽  
Rich Information ◽  
Seismic Facies Analysis

One of the most important goals of seismic stratigraphy studies is to interpret the elements of the seismic facies with respect to the geologic environment. Prestack seismic data carry rich information that can help us get higher resolution and more accurate facies maps. Therefore, it is promising to use prestack seismic data for the seismic facies recognition task. However, because each identified object changes from the poststack trace vectors to a prestack trace matrix, effective feature extraction becomes more challenging. We have developed a novel data-driven offset-temporal feature extraction approach using the deep convolutional autoencoder (DCAE). As an unsupervised deep learning method, DCAE learns nonlinear, discriminant, and invariant features from unlabeled data. Then, seismic facies analysis can be accomplished through the use of conventional classification or clustering techniques (e.g., K-means or self-organizing maps). Using a physical model and field prestack seismic surveys, we comprehensively determine the effectiveness of our scheme. Our results indicate that DCAE provides a much higher resolution than the conventional methods and offers the potential to significantly highlight stratigraphic and depositional information.

A visual data-mining methodology for seismic facies analysis: Part 2 — Application to 3D seismic data

Geophysics ◽  
2009 ◽  
Vol 74 (1) ◽  
pp. P13-P23 ◽  
Author(s):  
Iván Dimitri Marroquín ◽  
Jean-Jules Brault ◽  
Bruce S. Hart
Keyword(s):  
Data Mining ◽  
Seismic Data ◽  
Clustering Analysis ◽  
Facies Analysis ◽  
Seismic Facies ◽  
Visual Data ◽  
Visual Data Mining ◽  
Data Mining Approach ◽  
Seismic Facies Analysis ◽  
3D Seismic Data

A visual data-mining approach to unsupervised clustering analysis can be an effective tool for visualizing and understanding patterns inherent in seismic data (i.e., seismic facies). The unsupervised clustering analysis is completely data-driven, requiring no external information (e.g., well logs) to guide the seismic-trace classification. We demonstrate the application of the visual data-mining approach to seismic facies analysis on a real 3D seismic data volume. We select two stratigraphic intervals, the first including a Devonian pinnacle reef system and the second containing a Jurassic siliciclastic channel system. Both analyses show major stratigraphic features that can be defined in horizon slices or other types of visualization. However, the visual data-mining approach creates seismic facies maps with improved visual detail, distinguishing seismic trace-shape variability in the data. We also compare the facies maps with those obtained from a commercial package for seismic facies classification. Both approaches created similar facies maps, but the visual strategy better depicts subtle stratigraphic changes in the bodies being imaged, offering insight into the nature of these features.

Seismic facies analysis based on deep convolutional embedded clustering

Geophysics ◽  
2019 ◽  
Vol 84 (6) ◽  
pp. IM87-IM97
Author(s):  
Yanting Duan ◽  
Xiaodong Zheng ◽  
Lianlian Hu ◽  
Luping Sun
Keyword(s):  
Seismic Data ◽  
Facies Analysis ◽  
Clustering Algorithms ◽  
Feature Space ◽  
Seismic Facies ◽  
Clustering Methods ◽  
Model Data ◽  
Geophysical Model ◽  
Facies Classification ◽  
Seismic Facies Analysis

Seismic facies classification takes a two-step approach: attribute extraction and seismic facies analysis by using clustering algorithms, sequentially. In general, it is clear that the choice of feature extraction is critical for successful seismic facies analysis. However, the choice of features is customarily determined by the seismic interpreters, and so the clustering result is affected by the difference in the seismic interpreters’ experience levels. It becomes challenging to extract features and identify seismic facies simultaneously. We have introduced deep convolutional embedded clustering (DCEC), which aims to simultaneously learn feature representations and cluster assignments by using deep neural networks. Our method learns mapping from the data space to a lower dimensional feature space in which it iteratively optimizes a clustering objective by building a specific loss function. We apply the method to the Modified National Institute of Standards and Technology (MNIST) data, geophysical model data, and field seismic data. In the MNIST data, the DCEC method shows better latent space of clustering results than traditional clustering methods. In the geophysical model data, the accuracy of waveform classification based on DCEC method is higher than traditional clustering methods. The results from the seismic data demonstrate that selection of input data and method has an important effect on the clustering result. In addition, our method is helpful for improving the resolution of seismic facies edges and offers the richer depositional information than the traditional clustering methods.

Quantitative lithostratigraphic interpretation of seismic data for characterization of the Unayzah Formation in central Saudi Arabia

Geophysics ◽  
2002 ◽  
Vol 67 (5) ◽  
pp. 1372-1381 ◽  
Author(s):  
Frédérique Fournier ◽  
Pierre‐Yves Déquirez ◽  
Costas G. Macrides ◽  
Marty Rademakers
Keyword(s):  
Pattern Recognition ◽  
Saudi Arabia ◽  
Seismic Data ◽  
Facies Analysis ◽  
Quality Criteria ◽  
Seismic Facies ◽  
Data Set ◽  
Good Convergence ◽  
Average Porosity ◽  
Seismic Facies Analysis

A lithostratigraphic interpretation of seismic data sets covering the Unayzah fluviatile formation in central Saudi Arabia has allowed us to map the sandstone distribution and to characterize the average porosity of the formation. First, sandstone distribution was predicted through seismic facies identification and interpretation with well information. Seismic facies analysis was performed with statistical pattern recognition applied to the portions of traces at the reservoir level, these traces being characterized by a series of seismic attributes. A good convergence of results from unsupervised and supervised pattern recognition was observed. This increases the confidence in the interpretation of sandprone facies. Second, using a statistical relationship between the reservoir average porosity defined at the wells and selected amplitudes of adjacent traces, the porosity was predicted all over the area covered by the seismic information. The model was based on a multivariate linear regression, showing satisfying quality criteria (correlation coefficient, residuals, etc.). The porosity variation predicted from the seismic data complements the sandstone distribution, derived from the seismic facies analysis. In particular, some areas where sandstones are predicted do not appear as porous as could have been suspected from their lithological content, perhaps as the result of diagenetic effects. Last, seismic facies analysis with pattern recognition applied to 2‐D exploratory lines, partly intersecting the 3‐D data set, led to the identification of potential prospects (Unayzah interval with a high sand–shale ratio).

Sign in / Sign up

Export Citation Format

Share Document