Study On Information Integration And Intelligent Management For Marine Comprehensive Observation System

The marine information observed is to be effectively dealt with not only to help the development of marine economy,protect the marine environment and forecast the marine disaster,but also to relate to the fundamental interests of national security and territorial integrity.By now,the observation technologies on marine information are remote sensing and mensurating,automatic observation,acoustic detection,satellites, aircraft,ship,diving equipments,buoys and observatories.Those observation ways achieve the three-dimensional observation of marine and the quick exploration of marine resources.With the diversification of observation ways,the marine information we got is increasingly large,multi-source heterogeneous and different format.The traditional application mode of data management is far not enough to meet the demand of the growth of marine data.Therefore,the development of system integration and intelligent management on the marine information becomes an inevitable trend.In this thesis,we analyses the characteristics of diversity,heterogeneous and time-space about marine observation information.The marine services would not be fully developed because the information could not be dealt with and made full use of. As a result,based on the pervasive environment,we design a network platform of marine integrated observation for the future of marine development.Then,we mainly study on the data integration and management based on the characteristics of multi-source marine information.First,we propose a fuzzy assimilation method for data based on that the marine observation information is multi-source,heterogeneous and rough.After the information to be fuzzy,and then to be assimilated,it can effectively extract the useful data and enhance the utilization percent.Then,based on the widespread application of neural network in the optimization, we promoted an algorithm of intelligent data fusion.By reducing the error,the algorithm firstly fuses the various indicators of different data sources in partial,and then it fuses the partial fusion results by weight.So,the fused data fully reflects out that the different data sources should have different weights in the fusion,and the effect of fusion result is lowed by reducing the error in the original data.Again,we present a model of information integratation based on XML for the multi-source and heterogeneous information.This model could effectively reduce the complexity of the information,unify the data format to be stored and provide the efficiency of marine service.Finally,we design an early-warning system for marine disaster based on fuzzy clustering on the network platform of marine integrated observation.The system can effectively excavate the useful information and provide the rapid service of early-warning.The research of the integrated and intelligent management of observation information can be maximized to extract the useful data from the original observation information,enhance and improve the performance of analysis and prediction system. On the basic of making an accurate prediction,forecasting and early-warning,it will help the researchers to deeply understand and master the law of formation and development of the marine disaster.And it will be very important to prevent and reduce the disaster in the coastal areas,protect the public health and social stability.The research can help to understand and grasp the environmental quality and the trend of its change,then be targeted to develop the work of protection for marine environment.It can provide the necessary information involved in legislation,planning,standards, assessment,management,governance,rehabilitation and construction in order to effectively protect the marine environments.