大数据分析技术的高光谱图像数据库查询策略研究  

Research on query strategy of hyper-spectral image database based on big data analysis technology

作　　者：王颖 陈曦 WANG Ying;CHEN Xi(Computer and Information Engineering College,Guizhou University of Commerce,Guiyang 550014,China)

机构地区：[1]贵州商学院计算机与信息工程学院,贵阳550014

出　　处：《激光杂志》2020年第8期101-104,共4页Laser Journal

基　　金：贵州省普通高等学校工程研究中心;贵州省大数据应用工程研究中心(No.黔教合KY字【2017】022)。

摘　　要：以从高光谱图像数据库中高效获取所需高光谱图像为最终目标,研究大数据分析技术的高光谱图像数据库查询策略,利用大数据分析技术,将高光谱图像数据库分层数据层、服务层、应用层以及展示层4个层次,采用基于时空二维联合的高光谱图像数据库挖掘方法,有机结合数据库特征集与高光谱图像背景,获取最优分组数量,利用Apriori算法逐步过滤高光谱图像层间关系,避免数据库低层形成的待选数据库数量较多问题,消除数据存储空间影响,引入空间维度,完成各种监测点的层次划分,实现高光谱图像数据库高效查询。实验结果表明:所提策略可实现所需高光谱图像查询,且具备较高查询效率和查询精度。With the ultimate goal of efficiently obtaining the required hyper-spectral images from hyper-spectral image database,the hyper-spectral image database query strategy of big data analysis technology is studied.The big data analysis technology is used to divide the hyper-spectral image database into the data layer,the service layer,the application layer and the presentation layer.Adopting the hyper-spectral image database mining method based on twodimensional joint of space-time,organically combining the database feature set and the hyper-spectral image background to obtain the optimal number of groups.Using the Apriori algorithm to gradually filter the relationships among hyper-spectral image layers which can avoid the problem of a large number of candidate databases formed by lower database layers.Eliminating the impact of data storage space,introducing the spatial dimensions,completing the hierarchical division of various monitoring points to achieve efficient query of hyper-spectral image databases.The experimental results show that the proposed strategy can achieve required hyper-spectral image query,and has higher query efficiency and accuracy.

关 键 词：大数据分析 高光谱图像 数据库 查询策略 数据挖掘 效率 

分 类 号：TN297[电子电信—物理电子学]