TRAINING OF CELLULAR NEURAL NETWORKS AND APPLICATION TO GEOPHYSICS

Özet

In this study, to determine horizontal location of subtle boundaries in the gravity anomaly maps, an image processingmethod known as Cellular Neural Networks (CNN) is used. The method is a stochastic image processing methodbased on close neighborhood relationship of the cells and optimization of A, B and I matrices known as cloning templates.Template coefficients of continuous-time cellular neural networks (CTCNN) and discrete-time cellular neuralnetworks (DTCNN) in determining bodies and edges are calculated by particle swarm optimization (PSO) algorithm.In the first step, the CNN template coefficients are calculated. In the second step, DTCNN and CTCNN outputs arevisually evaluated and the results are compared with each other. The method is tested on Bouguer anomaly map ofSalt Lake and its surroundings in Turkey. Results obtained from the Blakely and Simpson algorithm are comparedwith the outputs of the proposed method and the consistence between them is examined. The cases demonstrate thatCNN models can be used in visual evaluation of gravity anomalies.