| 基于神经网络射线非直线传播的层析成像 |
| |
| 引用本文： | 唐雪原,黄华萍.基于神经网络射线非直线传播的层析成像[J].武汉大学学报(理学版),1996(1). |
| |
| 作者姓名： | 唐雪原 黄华萍 |
| |
| 作者单位： | 武汉大学电子信息学学院，武汉大学电子信息学学院 武汉，430072，武汉，430072 |
| |
| 摘 要： | 讨论了如何利用走时数据重建声波地震波或某些电磁波这类波长较长的射线慢度(波速的倒数)分布问题。首先根据费马原理,利用加权有向图求射线在给定介质中的最短传播路径,然后,基于Hopifield神经网络的联想功能,提出一种射线非直线传播的重建慢度分布的方法,计算机模拟表明,迭代是收敛的,并且能得到较好的慢度重建结果。
|
| 关 键 词： | 层析成像 慢度 神经网络 |
NONLINEAR TRAVELTIME TOMOGRAPHY BASED ON NEURAL NETWORK |
| |
| Tang Xueyuan,Huang Huaping.NONLINEAR TRAVELTIME TOMOGRAPHY BASED ON NEURAL NETWORK[J].JOurnal of Wuhan University:Natural Science Edition,1996(1). |
| |
| Authors: | Tang Xueyuan Huang Huaping |
| |
| Abstract: | Discusses the reconstruction of slowness (reciprocal of wave-speed) distributions of waves with long wave-length such as acoustic, seismic or some electromagnetic waves from first arrival traveltime data. First, according to Fermat's principle, we use a weighted directed graph to trace the least travetime paths of rays. Then, based on the associative function of Hopfield neural networks, a reconstructive appoach taking accounting of ray bending is proposed. The experiment results shows that the iteration converges and the quality of the reconstructive results are good. |
| |
| Keywords: | tomography slowness neural network |
| 本文献已被 CNKI 等数据库收录！ |
|
