Abstract:The common reflection surface stack with theoutput imaging scheme(CRS-OIS) is an importantrevision for conventional CRS stack method.Inthis paper,we present a robust graphic processingunit(UPU)based 3D-CRS-OIS algorithm usingthe large-scale thread-level parallel computing ar-chitecture of UPU.With the advantages of UPUplatform in the aspects of high memory bandwidth,multi-registers and multi-processors,the majorcomputing burden of 3D-CRS-OIS,attribute pa-rameter searching,has been executed upon UPU,which improves greatly the efficiency of 3D-CRS-OIS algorithm.The synthetic and field data exam-ples show that the proposed method is much moreefficient than CPU-based algorithm,which furtherextends the practical value of 3D-CRS-OIS.
倪瑶, 杨锴. 基于GPU计算平台实现三维输出道方式的共反射面元(3D-CRS-OIS)叠加[J]. 石油地球物理勘探, 2013, 48(1): 49-57.
Ni Yao, Yang Kai. 3D common reflection surface stack algorithm based GPU with the output imaging scheme(3D-CRS-OIS). OGP, 2013, 48(1): 49-57.