ADAPTIVE CYCLIC RECOGNITION ALGORITHM IN GEOPHYSICAL STRUCTUROSCOPY

Abstract

This article explores the development and application of an adaptive cyclic algorithm for object recognition in geophysical rock structuroscopy. The author substantiates the advantage of recognizing groups of objects (layers, lithological variations) over identifying individual objects, enabling the integration of prior geological information and stratigraphic patterns into the analysis process. Using the example of the Kulmen landslide slope (Mailuu-Suu Mountain), the authors demonstrate the effectiveness of two procedures: a formal one (based on minimizing recognition errors) and a substantive one (taking into account step-by-step depth analysis). The use of cyclic recalculation of adaptive resistivity estimates enabled the transformation of the initial geoelectric section into a reliable geological-lithological model consistent with drilling data. The multi-stage adaptive procedure has been shown to provide high reliability (up to 85–97%) in the identification of soil structural classes.