摘要
在参考谷歌文件系统分布式处理技术基础上,利用三维空间下八叉树结构与编码的快速空间定位机制,实现对三维地震数据的分块存储;采用中间文件形式进行子块切分与传输,减少本地时间开销;同时实现了基于地震道的一级缓存和基于子块的二级缓存,提升了数据访问效率。进一步基于分布式八叉树结构,设计实现了地震属性的分布式映射归并计算方法。研究方法为大数据背景下三维数据体的高效存储与处理分析提供了技术支持。
Based on the distributed processing technology of Google file system, the fast spatial positioning mechanism of Octree structure and coding in three-dimensional space is used to realize the block storage of three-dimensional seismic data;the sub block is segmented and transmitted in the form of intermediate file to reduce the local time overhead;at the same time, the primary cache based on seismic channel and the secondary cache based on sub block are realized to improve the efficiency of data access. Further, based on the distributed Octree structure, the distributed mapping and merging calculation method of seismic attributes is designed and implemented. The research method provides technical support for efficient storage, processing and analysis of 3 D data volume under the background of big data.
作者
彭成
PENG Cheng(Petroleum Exploration and Production Research Institute,SINOPEC,Beijing 100083,China)
出处
《智能计算机与应用》
2022年第10期169-175,共7页
Intelligent Computer and Applications
基金
国家页岩油重大专项“陆相页岩油资源和选区评价技术与软件实现”(2017ZX05049-001-007)
中科院A类战略性先导科技专项“深层油气勘探定量评价软件平台”(XDAXX010405)。
