为了解决海洋数据的多源异构问题,使用户对大量复杂的海洋数据进行有效管理,设计并实现一套海洋信息管理系统。首先介绍系统的体系结构,然后对各模块进行详细的阐述,并对海洋数据进行了详细的分析。针对海洋数据存在数据不一致、数据空缺和数据冗余等问题,使用基于累积变化量的时间序列异常检测方法,有效检测出海洋数据中的异常点,再对异常点进行修正,保证了海洋数据的质量,从而建立一套完善的海洋信息管理系统,使得海洋相关部门的业务流程更加科学化和规范化。
In order to solve the multi-source heterogeneous problem of ocean data and make users to manage a large number of complex ocean data efficiently, we design and implement a set of ocean information management systems. First, we introduce the architecture of the systems, then give a detailed exposition on various functional modules, and carry on detailed analysis for the ocean data. Aiming at the problems in ocean data such as data inconsistencies, data gaps and data redundancy and so on, we use time series anomaly detection method which is based on the cumulative variation, it can detect outliers in ocean data effectively and then amend them, thus ensures the quality of ocean data, thereby a set of comprehensive ocean information management systems have been established, they guide the business process of relevant marine denartments to be more scientific and standardised.