Skip to Main Content


SEISIS is a knowledge-based system for the automatic segmentation of seismic sections into large regions of common textural properties. Such regions are believed to contain geologic information which can be related to large scale tectonic events, such as salt diapirs and shale ridges, or to different depositional environments of the constituent sediments. Human expert knowledge is introduced in SEISIS in order to resolve uncertainties in the numerical data and to help in making segmentation decisions. This domain-dependent knowledge, however, may be stated using terms having imprecise or fuzzy meanings, such as “unlikely”, “usually”, “seldom”, etc. Furthermore, the conditions in the IF part of expert rules normally refer to and should be matched with the information/facts collected during the segmentation process. These facts generally are associated with certain types of “uncertainties” to reflect their relative truthfulness. How to integrate all this information and knowledge, which is typically of diverse sources and with different scales, to reach a final classification decision is a crucial problem in the development and actual implementation of SEISIS’s knowledge-based segmentation process. We shall discuss in detail the knowledge-based reasoning process used in SEISIS. We use a small piece of a stacked seismic section from the Gulf of Mexico and a sample expert rule relevant to that particular section to demonstrate (1) the actual numerical and symbolic computations involved in SEISIS, and (2) the integration of information and knowledge using a Certainty Factor vector updating procedure with a probability based plausible reasoning model.

You do not currently have access to this chapter.

Figures & Tables





Citing Books via

Close Modal
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close Modal
Close Modal