Method and system for seismic anomaly detection

Abstract

A method and system for seismic anomaly detection is disclosed. Hydrocarbon prospecting relies on accurate modeling of subsurface geologic structures and detecting fluid presence in the geologic structures. For example, a seismic survey is gathered and processed to create a mapping of the subsurface region. The processed data is then examined, such as by comparing pre- or partially-stacked seismic images, in order to identify subsurface structures that may contain hydrocarbons. Instead of relying on engineered image attributes, which may be unreliable and biased, to identify anomalous features, an unsupervised machine learning framework is used to learn the relationships among partially-stack images or among pre-stack images to detect the anomalous features, and in turn hydrocarbon presence.

Dependencies are inferred from the claim text.