基于GPU计算的三维地震断层解释
|
摘要
应用图像处理方法对地震图像进行预处理后,可直接进行地震断层解释,但基于偏微分方程的图像扩散滤波方法计算效率低,不利于三维地震图像处理。基于GPU(Graphic Process Unit)计算的快速发展,应用GPU加速计算提取三维地震图像的结构张量,由结构张量导向进行地震图像扩散滤波增强预处理;基于三维结构张量识别的三维地震图像不连续区域,采用边缘检测和方差分析方法检测地层和断层区域,再通过设定阈值,提取三维空间中的断层。分别使用CPU和GPU计算处理了样点数为128×128×128和256×256×256的两个三维地震数据体,应用扩散增强滤波技术迭代10次,GPU计算速度分别是CPU计算速度的16.1倍和16.8倍。截取任意三维地震数据体的时间切片,对比了人工拾取断层和计算机自动拾取断层的效果,可以看出,对于较大尺度的断层,人工拾取和计算机自动拾取的结果基本一致,但是对于尺度较小的断层,计算机自动拾取结果更加准确,断层走向更加清晰。
Seismic images can be directly used for seismic fault interpretation after imaging diffusion filtering preprocessing.However,the computation efficiency is low,which is not beneficial for 3-D seismic image.The fast development of Graphic Process Unit(GPU) makes it possible to extract the structure tensor of 3-D seismic image by GPU acceleration computation.Then,by using the structure tensor,diffusion filtering enhancement preprocessing can be carried out on seismic images.Based on the discontinuous area on 3-D seismic image recognized by 3-D structure tensor,edge detection and variance analysis were used to detect the formations and faults. Furthermore,by setting threshold,the faults in 3-D domain were extracted.CPU and GPU computation was applied to process two 3-D seismic data volume with 128×128×128 and 256×256×256 samples.After ten times of iteration with diffusion enhancement filtering, the computation speed of GPU is 16.1 and 16.8 times of CPU computation.The effect of manually-picking and automatically -picking was compared by extracting arbitrary time slice of 3-D seismic image volume.The results show that for large-scale fault,the two results are almost the same,but for smaller-scale fault,automatically -picking is more accurate,and which can displays clearer fault strike.
Seismic images can be directly used for seismic fault interpretation after imaging diffusion filtering preprocessing.However,the computation efficiency is low,which is not beneficial for 3-D seismic image.The fast development of Graphic Process Unit(GPU) makes it possible to extract the structure tensor of 3-D seismic image by GPU acceleration computation.Then,by using the structure tensor,diffusion filtering enhancement preprocessing can be carried out on seismic images.Based on the discontinuous area on 3-D seismic image recognized by 3-D structure tensor,edge detection and variance analysis were used to detect the formations and faults. Furthermore,by setting threshold,the faults in 3-D domain were extracted.CPU and GPU computation was applied to process two 3-D seismic data volume with 128×128×128 and 256×256×256 samples.After ten times of iteration with diffusion enhancement filtering, the computation speed of GPU is 16.1 and 16.8 times of CPU computation.The effect of manually-picking and automatically -picking was compared by extracting arbitrary time slice of 3-D seismic image volume.The results show that for large-scale fault,the two results are almost the same,but for smaller-scale fault,automatically -picking is more accurate,and which can displays clearer fault strike.
引文
[1]Bahorich M,Farmer S.3-D seismic discontinuity for faults and stratigraphic features:the coherence cube [J].The Leading Edge,1995,14(10):1053-1058
[2]佘德平,曹辉,郭全仕.应用三维相干技术进行精细地震解释[J].石油物探,2000,39(2):83-88
[3]Marfurt K J,Kirlin R,Farmer SL,et al.3-D seismic attributes using a semblance-based coherency algorithm [J].Geophysics,1998,63(4):1150-1165
[4]Gersztenkorn A.Marfurt K J.Eigenstructure-based coherence computations as an aid to 3-D structural and stratigraphic mapping[J].Geophysics,1999,64 (5):1468-1479
[5]Trappe H.Hallmich G,Foell M Potential power in the application of seismic volume attributes[J].First Break,2000,18(9):397-402
[6]Bakker P.Image structure analysis for seismic interpretation [D].Delft Holland:technisch Universiteit, 2002
[7]H(o|¨)cker C,Fehmers G.Fast structural interpretation with structure-oriented filtering[J].Geophysics, 2003,68(4):1286-1293
[8]Gibson D,Spann M,Turner J.Automatic fault detection for 3d seismic data[C]//Proceedings of the VII digital image computing:techniques and applications. Sydney:Csiro Publishing,2003:821-830
[9]Rumpf M.Strzodka R.Nonlinear diffusion in graphics hardware[C]//Proceedings of EG/IEEE TCVG Symposium on Visualization VisSym.Ascona: Springer,2001:75-84
[10]Weickert J,Hagen H.Visualization and image processing of Tensor Fields[M].Berlin:Springer. 2006:17-47
[11]王绪松,杨长春.对地震图像进行保边滤波的非线性各向异性扩散算法[J].地球物理学进展,2006,21 (2):452-457
[12]Kass M,Witkin A.Analyzing oriented patterns [C]//Proceedings of the 9th international joint conference on artificial intelligence.San Francisco:Morgan Kaufmann Publishers,1985:944-952
[13]Perona P,Malik J.Scale-space and edge detection using anisotropic diffusion[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,1990,12 (7):629-639
[14]Weickert J.Anisotropic diffusion in image processing [M].Stuttgart:TeubnerVerlag,1998:1-3
[15]Canny J.A computational approach to edge detection [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(6):678-698
[16]Green S.GPU Physics[EB/OL].[2010-04-16].http: //gpgpa org/static/s2007/slides/15-GPGPU-physics. pdf
[17]刘红伟,李博,刘洪,等.地震叠前逆时偏移高阶有限差分算法及GPU实现[J].地球物理学报,2010,53 (7):1725-1733
[18]吴连贵,易瑜,李肯立.基于CUDA的地震数据相干体并行算法[J].计算机应用,2009,29(3):912-914
[19]NVIDIA.Cg Toolkit[EB/OL].[2010-04-16].http:// developer.nvidia.com/object/cg_toolkit.html
[2]佘德平,曹辉,郭全仕.应用三维相干技术进行精细地震解释[J].石油物探,2000,39(2):83-88
[3]Marfurt K J,Kirlin R,Farmer SL,et al.3-D seismic attributes using a semblance-based coherency algorithm [J].Geophysics,1998,63(4):1150-1165
[4]Gersztenkorn A.Marfurt K J.Eigenstructure-based coherence computations as an aid to 3-D structural and stratigraphic mapping[J].Geophysics,1999,64 (5):1468-1479
[5]Trappe H.Hallmich G,Foell M Potential power in the application of seismic volume attributes[J].First Break,2000,18(9):397-402
[6]Bakker P.Image structure analysis for seismic interpretation [D].Delft Holland:technisch Universiteit, 2002
[7]H(o|¨)cker C,Fehmers G.Fast structural interpretation with structure-oriented filtering[J].Geophysics, 2003,68(4):1286-1293
[8]Gibson D,Spann M,Turner J.Automatic fault detection for 3d seismic data[C]//Proceedings of the VII digital image computing:techniques and applications. Sydney:Csiro Publishing,2003:821-830
[9]Rumpf M.Strzodka R.Nonlinear diffusion in graphics hardware[C]//Proceedings of EG/IEEE TCVG Symposium on Visualization VisSym.Ascona: Springer,2001:75-84
[10]Weickert J,Hagen H.Visualization and image processing of Tensor Fields[M].Berlin:Springer. 2006:17-47
[11]王绪松,杨长春.对地震图像进行保边滤波的非线性各向异性扩散算法[J].地球物理学进展,2006,21 (2):452-457
[12]Kass M,Witkin A.Analyzing oriented patterns [C]//Proceedings of the 9th international joint conference on artificial intelligence.San Francisco:Morgan Kaufmann Publishers,1985:944-952
[13]Perona P,Malik J.Scale-space and edge detection using anisotropic diffusion[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,1990,12 (7):629-639
[14]Weickert J.Anisotropic diffusion in image processing [M].Stuttgart:TeubnerVerlag,1998:1-3
[15]Canny J.A computational approach to edge detection [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(6):678-698
[16]Green S.GPU Physics[EB/OL].[2010-04-16].http: //gpgpa org/static/s2007/slides/15-GPGPU-physics. pdf
[17]刘红伟,李博,刘洪,等.地震叠前逆时偏移高阶有限差分算法及GPU实现[J].地球物理学报,2010,53 (7):1725-1733
[18]吴连贵,易瑜,李肯立.基于CUDA的地震数据相干体并行算法[J].计算机应用,2009,29(3):912-914
[19]NVIDIA.Cg Toolkit[EB/OL].[2010-04-16].http:// developer.nvidia.com/object/cg_toolkit.html
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