摘要
目前对海洋涡旋的分析主要采用水团、近海海流分析和数值模拟方法,这些方法结合不同的区域资料已经有过很多成功的研究。但面对日益增长的多源、大面积海洋遥感数据,在快速自动化提取海洋涡旋的空间信息方面还存在很多难题。因此该文针对海洋涡旋空间信息的快速定量分析问题,提出了基于案例推理(Case-BasedReasoning,CBR)的海洋涡旋空间相似性研究方法。该方法首先采用多源数据(遥感数据、遥感反演数据、观测数据等)进行历史涡旋案例定量分析,并构建基于CBR的海洋涡旋案例表达模型和历史案例库;继而采用“基于重心矢径的序列分析方法”进行历史涡旋空间相似性分析,并结合典型的海洋涡旋区域给出具体实例。研究结果表明该方法是可行的。
The current analysis on ocean eddies, which mainly focuses on water mass & current analysis and numerical simulation approaches, has been proved to be successful by various applications in different regions. However, with the blowup of multi-source and large region covering ocean remote sensing data, there are still many difficulties in fast and automatic extraction of spatial information of ocean eddies. In this paper, a new approach of automatic extraction for the spatial information of ocean eddies based on Case-Based Reasoning (CBR) was proposed. It includes three basic steps. First, the information of the eddy’s spatial structure and attributes is recognised from the original remote sensing data. Then, a library of historical eddy cases is built using the case expression model. Finally, a Radius Vector Serial Analysis Model Based on Barycentre (RVSAMB) is provided to analyse the spatial similarity between historical cases for further forecasting and dynamic analysis.
出处
《热带海洋学报》
CAS
CSCD
北大核心
2005年第3期1-9,共9页
Journal of Tropical Oceanography
基金
国家高科技863计划海洋监测主题项目(2003AA634040
2002AA639460)
