In the last few years, we have analyzed the best alternatives for acquiring and processing visual knowledge with the goal of supporting problem solving. We call
the set of mental models that support the process of reasoning over information that comes from the spatial arrangement and visual aspects of entities. Also, visual knowledge is implicit, meaning that it is difficult to be explicitly represented solely with propositional constructs. In this paper, we describe a representational approach that helps geologists in capturing and applying this kind of knowledge, in order to support software development applied to interpretation tasks in Petroleum Geology applications. Our approach combines propositional constructs with visual pictorial constructs in order to model visual knowledge of geologists. These constructs are proposed in a strong formal model, founded by Formal Ontology concepts. Based on these constructs, we develop a full ontology for stratigraphic description of sedimentary facies. The Formal Ontology background and the approach are detailed and evaluated through the paper.