摘要
将Huber范数推广到最小二乘逆时偏移中,提升了算法的稳健性;同时考虑到最小二乘逆时偏移的计算量较大,将随机最优化相位编码技术引入到基于Huber范数的多震源最小二乘逆时偏移中,可有效提高计算效率,压制串扰噪声;最后通过常规有限差分正演模拟记录试算,验证了算法的有效性和对复杂模型的适用性。
We extend the Huber norm to the leastsquares reverse time migration(LSRTM)algorithm to promote its robustness.Then,considering a large amount of calculation of LSRTM,we introduce the stochastic optimization phase encoding ideas to multi-source LSRTM based on Huber norm.This can effectively improve the calculation efficiency and eliminate the noise.Finally,theoretical data modeled by conventional finite-difference method verifies the effectiveness and suitability of the proposed algorithm for complex models.
出处
《石油地球物理勘探》
EI
CSCD
北大核心
2017年第5期941-947,共7页
Oil Geophysical Prospecting
基金
国家油气重大专项项目(2016ZX05006-004)资助
关键词
最小二乘逆时偏移
相位编码
Huber范数
多震源
随机最优化
least-squares reverse time migration(LSRTM)
phase encoding
Huber norm
multisources
stochastic optimization
