University of Tehran Press Journal of the Earth and Space Physics 2538-371X 32 2 2006 07 23 Estimation of earth reflection coefficient series using Hopfield neural network Estimation of earth reflection coefficient series using Hopfield neural network 125 138 80104 FA Amin Roshandel Institute of Geophysics, University of Tehran, P.O. Box 14155-6466, Tehran, Iran Abdolrahim Javaherian Institute of Geophysics, University of Tehran, P.O. Box 14155-6466, Tehran, Iran Babak Najar-Arabi Faculty of Engineering, University of Tehran, P.O. Box 14395-515 Journal Article 2021 03 02 The parallel processing of artificial neural networks makes them suitable for hardware implementations; therefore, using artificial neural networks for seismic signal processing problems has the potential of greatly speeding up seismic data processing. In this paper, a commonly used neural networks, Hopfield neural network, is used to implement deconvolution. The deconvolution procedure decomposes into two subprocesses: reflectivity locatin detection and reflectivity magnitude estimation. A Hopfield neural network is developed for each of subprocesses. The first neural network is developed to detect the reflectivity sequence. The second neural network is developed to determine the magnitudes of the detected reflections. These two neural networks are simulated for narrow-band wavelet, and the result is compared with that of using spiking deconvolution. With comparing the results, deconvolution of seismic data using Hopfield neural network: (1) unlike spiking deconvolution, is not sensitive to noise and (2) gives much better result than spiking deconvolution for a trace with noise. The parallel processing of artificial neural networks makes them suitable for hardware implementations; therefore, using artificial neural networks for seismic signal processing problems has the potential of greatly speeding up seismic data processing. In this paper, a commonly used neural networks, Hopfield neural network, is used to implement deconvolution. The deconvolution procedure decomposes into two subprocesses: reflectivity locatin detection and reflectivity magnitude estimation. A Hopfield neural network is developed for each of subprocesses. The first neural network is developed to detect the reflectivity sequence. The second neural network is developed to determine the magnitudes of the detected reflections. These two neural networks are simulated for narrow-band wavelet, and the result is compared with that of using spiking deconvolution. With comparing the results, deconvolution of seismic data using Hopfield neural network: (1) unlike spiking deconvolution, is not sensitive to noise and (2) gives much better result than spiking deconvolution for a trace with noise. https://jesphys.ut.ac.ir/article_80104_8b57232d8c5fafefadf3e2af190029e5.pdf