Haibin Di

Multi-attribute K-means Cluster Analysis for Salt Boundary Detection

Robust detection of salt bodies has been the recent focus of hydrocarbon exploration and production from 3D seismic surveying in the last decade. This study presents a new salt-boundary detection method based on multi-attribute k-means cluster analysis, which consists of two major components. First, a suite of seismic attributes is selected and computed from the seismic volume, from which the salt boundaries can be readily differentiated from the surrounding non-boundary features in various ways. Second, the k-means cluster analysis is performed in the attribute domain and generates a probability volume, which depicts the boundaries of salt domes observed in the original seismic amplitude. The proposed method is verified through applications to the F3 seismic dataset of multiple salt bodies over the Netherlands North Sea. The results demonstrate not only good match between the detected salt boundaries and the seismic images, but also great potential for accurate salt surface/body extraction to assist structural framework modeling in the zones of geologic complexities due to salt domes.

Seismic Multi-attribute Classification for Salt Boundary Detection - A Comparison

Accurate delineation of salt bodies is one of the major tasks of hydrocarbon exploration and production from 3D seismic surveying. With the increasing demand of high-resolution seismic interpretation, the size of 3D seismic volumes as well as the number of available seismic attributes has been rapidly rising, which adds the difficulties for interpreters to examine and interpret every vertical line and time slice in a seismic volume. Various machine learning techniques have been introduced from the field of image/video processing to help address this limitation; however, little effort has been devoted to fair comparisons between these techniques. This study implements six commonly-used classification techniques and compares their capabilities for salt-boundary detection, including logistic regression, decision tree, random forest, support vector machine, artificial neural network, and k-means clustering, through applications to the F3 seismic dataset of multiple salt bodies over the Netherlands North Sea. The good match between the detected salt boundaries and the original seismic images indicates that based on well-selected attributes, all six classification techniques are capable of providing reliable salt detection from 3D seismic data to assist structural framework modelling in the presence of salt domes.

Multiscale fusion for seismic geometric attribute enhancement

In this abstract, we propose a multiscale fusion technique to enhance seismic geometric attributes, such as dip and curvature, which are very sensitive to noise present in seismic data. For a give seismic section, first, we construct a Gaussian pyramid that allows us to generate the seismic attribute at different resolutions (scales). Then, all attributes at the different scales are fused together to form the proposed multiscale enhanced attribute. Applications to the 3D seismic dataset over the Great South Basin in New Zealand demonstrate that the proposed method is capable of improving both the resolution and noise robustness of the first-order dip and the second-order curvature attributes, compared to existing methods and algorithm. Such improvement indicates the great potential of our multiscale fusion technique for enhancing the quality of more multitrace seismic attributes, such as coherence, flexure, and GLCM.

Weakly Supervised Seismic Structure Labeling via Orthogonal Non-Negative Matrix Factorization

Summary With the growing demand of high-resolution subsurface characterization from 3D seismic surveying, the size of 3D seismic datasets has been dramatically increasing, and correspondingly, the process of interpreting a seismic dataset is becoming more time consuming and labor intensive. In addition, supervised machine learning has proved to be very successful for many applications in computational seismic interpretation. However obtaining training labels for large volumes of seismic data is a very demanding task. Furthermore, while the amount of data is continuously growing, the ability of human experts to label data remains limited. In this work, we propose a weakly-supervised framework for labeling seismic structures using Non-Negative Matrix Factorization (NMF) with additional sparsity and orthogonality constraints. We show that weakly-supervised learning requires a much smaller number of labels. Furthermore, we show that “rough” image-level labels of specific seismic structures can be mapped into finer more localized locations within the seismic volume. Results obtained by labeling fault regions and salt dome boundaries from the Netherlands F3 block prove to be very promising.

Multiscale Fusion for Improved Instantaneous Attribute Analysis

Summary In this abstract, we propose a multiscale approach for enhancing the resolution of instantaneous attributes, such as cosine of phase and phase dip, both of which tend to be highly sensitive to noise in seismic data. In particular, we use a multiscale representation, namely the Gaussian pyramid, to exploit seismic features at different resolutions followed by multiscale fusion to enhance both the quality and reduce noise sensitivity of these attributes. The value of the proposed technique is demonstrated through application on the Netherlands offshore F3 block, indicating its potential for improving more seismic attributes, such as coherence and curvature.