Potential Anomaly Separation Using Genetically Trained Multi-Level Cellular Neural Networks

BİLGİLİ E., UÇAN O. M., ALBORA A. M., GÖKNAR Z. C.

Cellular Neural Networks and their Application CNNA2002, Germany, 1 - 04 July 2002, vol.7, pp.391-398

  • Publication Type: Conference Paper / Full Text
  • Volume: 7
  • Country: Germany
  • Page Numbers: pp.391-398
  • Istanbul University Affiliated: Yes

Abstract

 

In this paper, Multi-Level Genetic Cellular Neural Networks
(ML-GCNN) are applied to the geophysical problem
of potential anomaly separation and satisfactory results are
obtained, compared to classical deterministic approaches.
ML-GCNN is a stochastic image processing technique
which is based on template optimisation using
neighbourhood relationships of the pixels. The residual
anomaly separation used in location decisions is one of the
main problems in geophysics. The method proposed here is
used in evaluating the Dumluca iron ore region of Turkey.