Geometric Element Preserving Reconstruction of a Geological Surface

Abstract

A reconstruction of geological surfaces clearly displays the shape and three-dimensional (3D) spatial distribution of the geological surfaces, such as horizons and faults. It is the basis for understanding the geological structure and establishing reservoir models. Because of deficiencies in the acquisition, conversion, and interpretation, the obtained data inevitably contain a certain degree of noise and some outliers, which usually lead to the blurring or even the extinction of the original surface geometry elements (such as ridges, valleys, and fold hubs) that determine the shape of the geological surface; thus, the accuracy of geometric elements of the reconstructed geological surface is not high. Therefore, the extraction and preservation of geometric elements is one of the key factors in geological surface reconstruction, which is of great significance to reducing the risks associated with hydrocarbon exploration and development caused by structural uncertainty. In this paper, we propose a geometric element preserving reconstruction framework for the geological surfaces of two important elements: ridges and valleys. First, we use the classification algorithm to identify the candidate ridge and valley points within the seismic interpretation data, and then, we generate ridge and valley curves according to the support vector regression. Finally, we construct a geological surface reconstruction model using the geometric element curves as constraints. Our framework can be easily extended to various 3D surface reconstructions with other geometric elements. The experimental results obtained using a real data set show that the proposed framework effectively preserves the geomorphological features of the reconstructed surface.