Visual interpretation of events in petroleum exploration: An approach supported by well-founded ontologies

• We propose an ontology-based approach for visual interpretation tasks.
• Our approach applies ontological meta-properties of a foundational ontology.
• We propose a cognition-inspired representation structure for inferential knowledge.
• We apply our approach for interpreting depositional processes, in Petroleum Geology.

In visual domains, such as Medicine, Meteorology and Geology, the tasks are accomplished through intensive use of visual knowledge. In this work, we focus in an essential kind of task that is performed in many visual domains: the visual interpretation task. We call visual interpretation the expert reasoning process that starts with the perception of visual features of domain objects and results in the understanding of the scene. We propose a top-down approach for solving visual interpretation tasks, which is based on symbolic pattern matching supported by well-founded ontologies. Our approach combines well-founded domain ontologies; a meta-model for representing inferential knowledge, which is based on the notion of perceptual chunks; and a reasoning model for visual interpretation. The proposed model was applied for developing an expert system for automating the task of visual interpretation of depositional processes, which a crucial task for petroleum exploration.