The Study Of Hierarchical Uncertain Fusion Model And TheApplication In Multi-sensor Sea Surface Temperature

The era of Big Data is coming with the information overload on the Internet. By theend of2011, the amount of information had reached to1.8ZB, which would bedoubled in every two years. In2020,33%of this information is valuable and only0.4%can be analyzed The Big Data means not only the scale but also the descriptivecapacity. Under these characteristics, the method for analyzing plays an important rolein information processing. So it is in the marine information. This paper does someresearches on new information fusion model, which is used to deal with themulti-satellite Sea Surface Temperature (SST).The main contents and innovations of the thesis:1) Big Data is used to describe complexity accurately in its large amounts,all-dimensional and high-density. And in this paper, it is defined as Holographic Data.Holographic Fusion is a method to cognize complexity comprehensively based onHolographic Data. During this processing, the uncertainty of information has twomeanings, the first one is the description process, and the deeper one is the cognitionprocess. The latter plays more and more important role in the era of Big Data. In thispaper, the conflict, consistency and coordination among multi-sensor measurementsare quantified and used in the fusion model.2) In fusion model, information losing exists in the process of the pretreatment,quantification and fusion. The differences and mistakes would emerge from thisinformation losing, and then the accuracy and reliability would decrease. To solve thisproblem, this losing is quantified by moving relative entropy in this paper.3) In this paper, cognition principle is simulated to cognize big data by thehierarchical and multi-granularity holographic fusion model. In the fusion model, theaccuracy and the reliability of the fusion result will be enhanced as the amount ofinformation increases and granularity is refined. During this process, coupling matrixfusion model, coordination adaptive fusion model and spatial cognition retrieval model are used to resolve these problems.In the context of multi-sensor Sea Surface Temperature (SST) fusion, this paperquantifies uncertainty, conflict, consistency, coordination and information losingduring the fusion process, constructs a hierarchical and multi-granularity holographicfusion model to simulate the cognitive processing of humanity. In this fusion model,the accuracy and reliability of the result will be enhanced as the information increasesand its granularity is refined. In this paper, group decision model, hierarchical andmulti-granularity model are also discussed. At the end, the proposed modelaccomplishes the information fusion of extended information and unstructured data.However, this study is only an abecedarian research. More researches are needed,for example: the formal description of holographic data and holographic fusion model,the collection and quantification of cognition principle, transformation amongdifferent information granularities. Many studies have shown that in complexinformation system，the group decision would have a more accurate result thanindividual decision. It is feasible that further researches on the holographic datacognitive can be done with swarm intelligence and group decision.