Full text loading...
-
Detecção De Refletores Sísmicos Por Rede Neural Discreta
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 6th International Congress of the Brazilian Geophysical Society, Aug 1999, cp-215-00314
Abstract
The artificial neural networks have proven to be a powerful tool to solving a wide variety of optimization<br>problems. In this work we develop a recurrent network with no self-feedback loops and no hidden neurons for<br>seismic signal processing where this neural network gives us reflectivity location and reflectivity magnitude<br>estimation. The most important advantage of this neural network is the use a type of activation function which<br>permits three possible states of neuron to estimate the position of the seismic reflectors in such way to<br>reproduce its true polarities. The operational evaluation of this neural network architecture is accomplished in<br>synthetic data obtained through the ray theory.