Magmatic rocks are frequently encountered during hydrocarbon exploration in rift-related sedimentary basins. As magmatic rocks may contribute positively and negatively to the hydrocarbon systems, their spatiotemporal distribution and structural elements are crucial for exploration in frontier basins. With the proliferation and increased density of seismic reflection data, various subsurface magmatic features can be discriminated and illuminated via conventional interpretation approaches, such as attribute extraction, opacity rendering, or geo-body extraction. However, these manual interpretation techniques are labor-intensive, subject to interpreter bias, and often bottleneck with respect to time data delivery. A supervised machine learning approach could efficiently resolve these issues by amalgamating suitable seismic attributes, such as energy, reflection strength, texture, and similarity, and automatically delineating these magmatic features in 3D seismic reflection data. Our machine learning neural network (NN) classified igneous features from nonigneous features in two different seismic surveys within the natural laboratory of the offshore Otway Basin, southeast Australia. This multilayer perception NN designed in this study resulted in an optimized igneous probability meta-attribute cube that could effectively reveal the extension and distribution of igneous features and several structural elements in the study area. We have developed the detailed workflow of this artificial neural network and observed the efficiency of this approach in different seismic surveys. These results illustrate the potential of the NN in imaging other complex igneous features from 3D seismic data in the Otway Basin and worldwide.

You do not have access to this content, please speak to your institutional administrator if you feel you should have access.