Artificial intelligence algorithms for magnetic anomaly clusteringстатья
Информация о цитировании статьи получена из
Web of Science,
Scopus
Статья опубликована в журнале из списка Web of Science и/или Scopus
Дата последнего поиска статьи во внешних источниках: 22 декабря 2014 г.
Аннотация:The Euler deconvolution method provides estimates for the position of equivalent sources (Euler solutions) of potential field anomalies specified in a moving window. Simple theoretical examples show that these points are located on the contour of isolated anomaly-forming bodies and yield estimates of their depth. This method is very sensitive to many factors: the noise level, accuracy in calculating derivatives of the anomalous field, interference of signals from nearby sources, etc. In such cases, Euler points form fuzzy clouds, making it difficult to identify the contours of the bodies. On the other hand, calculations of theoretical models indicate that, even if Euler points do not form dense clusters near the contour of anomaly-forming bodies, the density of solutions is greater in the vicinity of these bodies. Therefore, the application of formal clustering methods to such problems appears natural and promising. This work uses the new RODIN algorithm designed to classify large data sets. This algorithm is based on the geometric approach to studying the concentration of objects in a finite metric space and on a new rigorous definition of a cluster. The free parameters inherent in the algorithm make it fairly flexible, enabling the identification of clusters with given properties within a wide range. In this work, the algorithm is used to select dense clusters on a set of Euler solutions. The test of this clustering method applied to model and real data showed that it is very efficient in determining the contours and depth of anomaly-forming bodies. In particular, the RODIN algorithm is more efficient that the conventional methods used for selecting Euler points. To demonstrate its performance, the algorithm is applied to the complex interference magnetic field near the Gulf of Saint-Maio in Brittany, France.