一种针对海洋数据的空间抽样方法  

A SPATIAL SAMPLING APPROACH FOR MARITIME DATA

作　　者：王睿晗[1] 黄冬梅[1] 王振华[1] 周雪楠 

机构地区：[1]上海海洋大学信息学院,上海201306

出　　处：《计算机应用与软件》2015年第5期228-230,245,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(61272098);上海高校优秀青年教师培养项目(B-5409-11-0012);国家重点基础研究发展计划项目(2012CB316200);海洋赤潮灾害立体监测技术与应用国家海洋局重点实验室开放研究基金课题(MATHAB201307)

摘　　要：抽样是提高海量海洋数据环境调研、数据挖掘、可视化展示等处理运算效率的有效手段。经典的抽样方法(简单、系统、分层、簇)是针对传统工业产品建立起来的,直接应用于海洋数据则存在效率低,信息冗余、空间分布不均等问题。针对这些问题,提出一种针对海洋数据的空间抽样方法,该方法首先利用海洋数据的相关性对研究区域分层;接着基于海洋数据的空间变异性,设计优化的系统空间抽样方法。实验结果表明,新的空间抽样方法在降低样本信息冗余的同时,兼顾了样本点的空间分布。该方法能有效地获取分布均匀的样本点位信息,且兼顾了样本点之间的变异性,较好地保留了区域内的特征变化信息。Sampling method is an effective means in improving the efficiency of processing and operation of environmental investigation, data mining and visualisation display in regard to massive maritime data. However, the classical sampling methods （ such as simple, system- atic, stratified and clustering and so on） are set up targeted at the traditional industrial products, when directly applied in maritime data, there will be the problems such as low efficiency, redundant information and uneven spatial distribution, etc. To solve these problems, we propose a spatial sampling method suitable for the maritime data. In this method, we first stratify the studying regions using the correlation of maritime data; Then, based on the spatial variation of maritime data, we design the optimised systematic spatial sampling method. Experi- mental results show that the new spatial sampling method balances the decrease of sample information redundancy and the spatial distribution of sample points. The method can effectively obtain the information of uniformly distributed positions of sample points, and takes the variability between the sample points into account as well, it also well retains the characteristics change information in the region.

关 键 词：海洋数据 分层抽样 空间系统抽样 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]