首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于LDPC矩阵采样方法的三维地震数据重构
引用本文:常傲,韩立国,于江龙,张良.基于LDPC矩阵采样方法的三维地震数据重构[J].世界地质,2018,37(2):636-645.
作者姓名:常傲  韩立国  于江龙  张良
作者单位:1. 吉林大学 地球探测科学与技术学院, 长春 130026;2. 中国石油新疆油田分公司 勘探开发研究院, 新疆 克拉玛依 834000
基金项目:国家重点研发计划课题(2017YFC0307405).
摘    要:按照以往的测量矩阵进行实际勘察无法较好地实现地震数据的重构,大多数矩阵设计复杂且元素多样,不符合实际标准。笔者将LDPC矩阵应用到地震勘探采样中,用较少的地震数据来重构完整的地震记录。采用K-SVD字典对数据进行稀疏变换,FISTA算法进行重构。对比以往的随机采样、离散均匀采样和jitter采样,LDPC矩阵更适合三维环境且恢复重构效果更好。对比重构数据的信噪比,三维模拟数据和实际数据都证实了LDPC矩阵信噪比值最高,重构出的数据效果最好。

关 键 词:压缩感知  数据重构  测量矩阵  LDPC矩阵
收稿时间:2017-10-31
修稿时间:2018-03-06

3-D seismic data reconstruction based on LDPC matrix sampling scheme
CHANG Ao,HAN Li-guo,YU Jiang-long,ZHANG Liang.3-D seismic data reconstruction based on LDPC matrix sampling scheme[J].World Geology,2018,37(2):636-645.
Authors:CHANG Ao  HAN Li-guo  YU Jiang-long  ZHANG Liang
Institution:1. College of Geo-exploration Science and Technology, Jilin University, Changchun 130026, China;2. PetroChina Xinjiang Oilfield Branch Exploration and Development Research Institute, Kelamayi 834000, Xinjiang, China
Abstract:The existing sampling matrices cannot realize good seismic data reconstruction results, and most of matrices are complicated and have different elements, which are not suitable for seismic prospecting. The authors applied LDPC matrix to seismic data acquisition, using less data to realize complete data reconstruction. The K-SVD dictionary was used for sparse transform, and FISTA reconstruction algorithm was used to recover original data. Compared with previous random sampling, discrete regular sampling and jitter sampling, the LDPC (Low Density Parity Check) matrix was more suitable in 3-D environment and achieved better performance. Comparing SNRs of reconstructed data of different methods, both 3-D simulated data and field data proved that LDPC matrix has the best SNR and performance.
Keywords:compressed sensing  data reconstruction  sampling matrix  LDPC matrix
本文献已被 CNKI 等数据库收录!
点击此处可从《世界地质》浏览原始摘要信息
点击此处可从《世界地质》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号