基于Curvelet域的位场多源数据融合方法
|
摘要
如何将大面积低精度数据与局部高精度数据进行融合,统一编制图件,是地球物理综合研究面临的重要问题。位场数据融合主要对低、高精度数据重叠区进行融合,为了提高融合效果,保持重叠区高精度信息,本文提出了基于Curvelet域的结合区域梯度的边界衰减方向加权融合算法。该法利用尺度和梯度调节衰减函数对不同方向的Curvelet系数进行加权处理。理论分析及中国南方中扬子地区的区域航磁数据融合实例表明,本文算法可以在保持局部信息的前提下,能够以较小的重叠区域,实现不同精度数据间的融合,重叠区数据精度高,同时避免了高频震荡、异常扭曲等边界效应。
To fuse the low-precision data of large area with the local high-precision data for unified compilation is an important issue for integrated geophysical research.The key of this job is the fusion for overlap area of low-precision,high-precision data.In order to enhance the effectiveness of fusion data and preserve information of high-precision data,we propose here a new fusion algorithm,called boundary attenuation and direction weighted with regional gradient in curvelet domain.We use the scale and gradient to impact the weighted operator and then apply it on the curvelet coefficients as weighted calculation.Both theoretical analysis and applications on aeromagnetic data fusion in South China indicate that the algorithm can maintain the local information and get a good fusion result with a small overlap area.Data Accuracy at overlap area is high and especially it avoids the boundary effect,such as high-frequency vibration,abnormal distortions,etc.
To fuse the low-precision data of large area with the local high-precision data for unified compilation is an important issue for integrated geophysical research.The key of this job is the fusion for overlap area of low-precision,high-precision data.In order to enhance the effectiveness of fusion data and preserve information of high-precision data,we propose here a new fusion algorithm,called boundary attenuation and direction weighted with regional gradient in curvelet domain.We use the scale and gradient to impact the weighted operator and then apply it on the curvelet coefficients as weighted calculation.Both theoretical analysis and applications on aeromagnetic data fusion in South China indicate that the algorithm can maintain the local information and get a good fusion result with a small overlap area.Data Accuracy at overlap area is high and especially it avoids the boundary effect,such as high-frequency vibration,abnormal distortions,etc.
引文
[1]E D C,Shen Q,Xu Y et al.High-accuracy topo-graphical information extraction based on fusion ofASTER stereo-data and ICESat/GLAS data in Ant-arctica.Science in China(Series D:Earth Sciences),2009,52(5):711~722
[2]张强,郭宝龙.基于Curvelet变换的多光谱图像与全色波段图像融合.系统工程与电子技术,2006,28(12):1786~1789Zhang Qiang,Guo Bao-long.Fusion of multispectraland panchromatic images based on the curvelet trans-form.Systems Engineering and Electronics,2006,28(12):1786~1789
[3]张丽莉,吴健生,王家林.图像处理在地球物理学中的应用.石油地球物理勘探,2003,38(3):317~323Zhang Li-li,Wu Jian-sheng and Wang Jia-lin.Appli-cation of image processing techniques to geophysics.OGP,2003,38(3):317~323
[4]吴健生,刘苗.基于小波的位场数据融合.同济大学学报(自然科学版),2008,36(8):1133~1137Wu Jian-sheng,Liu Miao.Potential field data fusionbased on wavelet multi-resolution decomposition.Journal of Tongji University(Natural Science),2008,36(8):1133~1137
[5]Deighan A J,Watts D R.Ground-roll suppressingusing the wavelet transform.Geophysics,1997,62(6):1896~1903
[6]刘天佑,杨宇山,李媛媛等.大型积分方程降阶解法与重力资料曲面延拓.地球物理学报,2007,50(1):290~296Liu Tian-you,Yang Yu-shan,Li Yuan-yuan et al.The order-depression solution for large-scale integralequation and its application in the reduction of gravitydata to a horizontal plane.Chinese J Geophys(in Chi-nese),2007,50(1):275~281
[7]刘天佑.位场勘探数据处理新方法.北京:科学出版社,2007
[8]Zhang H L,Liu T Y and Zhang Y C.Denoising ofseismic data via multi-scale ridgelet transform.Earthquake Science,2009,22(5):493~498
[9]张恒磊,刘天佑,宋双等.基于Curvelet域的叠前地震资料去噪方法.石油地球物理勘探,2008,43(5):508~513Zhang Heng-lei,Liu Tian-you,Song Shuang et al.Curvelet domain-based prestack seismic data denoisemethod.OGP,2008,43(5):508~513
[10]Mallat S.A theory for multiresolution signal decom-position:The wavelet representation.IEEE Trans-actions on Pattern Analysis and Machine Intelli-gence,1989,11(7):674~693
[11]Candès E J,Donoho D L.Curvelet manuscript.http:∥www-stat.standford.edu/~donoho/Reports/1999/curvelets.pdf,1999
[12]Candès E J,Donoho D L.Continuous Curvelet trans-form.Appl Comput Harmon Anal,2005,19(2):198~222
[13]Neelamani R,Baumstein A I,Gillard D G et al.Co-herent and random noise attenuation using the Curve-let transform.The Leading Edge,2008,27(2):240~248
[14]Herrmann F J,Hennenfent G.Non-parametric seis-mic data recovery with curvelet frames.Geophysics JInt,2008,173(1):233~248
[15]Herrmann F J,Wang D L,Hennenfent G et al.Cur-velet-based seismic data processing:A multiscale andnonlinear approach.Geophysics,2008,73(1):A1~A5
[16]Douma H,Hoop M de.Leading-order seismic ima-ging using curvelets.Geophysics,2007,72(6):S231~S248
[17]Nencinia F,Garzellia A,Baronti S et al.Remotesensing image fusion using the curvelet transform.Information Fusion,2007,8(2):143~156
[18]刘天佑.米仓山、大巴山前缘综合地球物理方法研究.中国石化勘探南方分公司,2008
[2]张强,郭宝龙.基于Curvelet变换的多光谱图像与全色波段图像融合.系统工程与电子技术,2006,28(12):1786~1789Zhang Qiang,Guo Bao-long.Fusion of multispectraland panchromatic images based on the curvelet trans-form.Systems Engineering and Electronics,2006,28(12):1786~1789
[3]张丽莉,吴健生,王家林.图像处理在地球物理学中的应用.石油地球物理勘探,2003,38(3):317~323Zhang Li-li,Wu Jian-sheng and Wang Jia-lin.Appli-cation of image processing techniques to geophysics.OGP,2003,38(3):317~323
[4]吴健生,刘苗.基于小波的位场数据融合.同济大学学报(自然科学版),2008,36(8):1133~1137Wu Jian-sheng,Liu Miao.Potential field data fusionbased on wavelet multi-resolution decomposition.Journal of Tongji University(Natural Science),2008,36(8):1133~1137
[5]Deighan A J,Watts D R.Ground-roll suppressingusing the wavelet transform.Geophysics,1997,62(6):1896~1903
[6]刘天佑,杨宇山,李媛媛等.大型积分方程降阶解法与重力资料曲面延拓.地球物理学报,2007,50(1):290~296Liu Tian-you,Yang Yu-shan,Li Yuan-yuan et al.The order-depression solution for large-scale integralequation and its application in the reduction of gravitydata to a horizontal plane.Chinese J Geophys(in Chi-nese),2007,50(1):275~281
[7]刘天佑.位场勘探数据处理新方法.北京:科学出版社,2007
[8]Zhang H L,Liu T Y and Zhang Y C.Denoising ofseismic data via multi-scale ridgelet transform.Earthquake Science,2009,22(5):493~498
[9]张恒磊,刘天佑,宋双等.基于Curvelet域的叠前地震资料去噪方法.石油地球物理勘探,2008,43(5):508~513Zhang Heng-lei,Liu Tian-you,Song Shuang et al.Curvelet domain-based prestack seismic data denoisemethod.OGP,2008,43(5):508~513
[10]Mallat S.A theory for multiresolution signal decom-position:The wavelet representation.IEEE Trans-actions on Pattern Analysis and Machine Intelli-gence,1989,11(7):674~693
[11]Candès E J,Donoho D L.Curvelet manuscript.http:∥www-stat.standford.edu/~donoho/Reports/1999/curvelets.pdf,1999
[12]Candès E J,Donoho D L.Continuous Curvelet trans-form.Appl Comput Harmon Anal,2005,19(2):198~222
[13]Neelamani R,Baumstein A I,Gillard D G et al.Co-herent and random noise attenuation using the Curve-let transform.The Leading Edge,2008,27(2):240~248
[14]Herrmann F J,Hennenfent G.Non-parametric seis-mic data recovery with curvelet frames.Geophysics JInt,2008,173(1):233~248
[15]Herrmann F J,Wang D L,Hennenfent G et al.Cur-velet-based seismic data processing:A multiscale andnonlinear approach.Geophysics,2008,73(1):A1~A5
[16]Douma H,Hoop M de.Leading-order seismic ima-ging using curvelets.Geophysics,2007,72(6):S231~S248
[17]Nencinia F,Garzellia A,Baronti S et al.Remotesensing image fusion using the curvelet transform.Information Fusion,2007,8(2):143~156
[18]刘天佑.米仓山、大巴山前缘综合地球物理方法研究.中国石化勘探南方分公司,2008
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心