امین روشندل کاهو - موسسه ژئوفیزیک دانشگاه تهران
عبدالرحیم جواهریان - موسسه ژئوفیزیک دانشگاه تهران
بابک نجاراعرابی - دانشکده برق و کامپیوتر، پردیس فنی دانشگاه تهران

Deconvolution is one of the most important seismic data processing steps which improves the quality of seismic data. Classic methods of deconvolution constructed based on assumptions which violated in real world. Noise free data is one of the mentioned assumptions. Using of Hopfield neural network in seismic deconvolution to estimate earth reflectivity is one of deconvolution methods. In this paper, sensitivity of this method to noise and frequency band of source wavelet is analyzed. This method estimates the reflectivity series in two stages: ١. reflectivity location detection, ٢. reflectivity magnitude estimation.

واهمامیخت، سری بازتاب زمین، شبکة عصبی هاپفیلد