摘要
本文介绍了奇异值分解SVD滤波原理，并用不同的具有代表性的合成地震记录进行了试验、分析，得出了影响SVD滤波效果的主要因素，从而提出了SVD滤波的改进方法─—基于SVD的倾角扫描叠加和自组织特征映射神经网络确定SVD滤波参数。将此法用于实际地震资料处理，效果明显。
The principle of singular-value decomposition (SVD)filtering is sketched out here. Trial analysis of different typical synthetic seismograms is made to find out the major factors which influence SVD filtering effect,so that the measures for improving SVD filtering are advanced,which are dip scanning stack based on SVD,and SVD filtering parameter determination using self-organized characteristic mapping neural network. This method has been applied in real data processing to result in good effect.
出处
《石油地球物理勘探》
EI
CSCD
北大核心
1994年第6期783-792,共10页
Oil Geophysical Prospecting
关键词
地震数据处理
SVD滤波
数据处理
滤波
seismic data processing, SVD filtering, pattern recognition, scanning, stack, neural network
