| As the focus and difficulty of Situational Awareness,object grouping is the important basis for determining the relationship between the objects and it is the foundation of the realization of the data fusion system.In this thesis,we mainly study the algorithm of object grouping,and algorithm improvement and concrete realization for its defects are given.The main work of this thesis is as follows:Firstly,in-depth analysis of data fusion,Situational Assessment and Situational Awareness of the three hierarchical structures based on the relationship between the specific,points out that the key and difficult problems of the object grouping in Situational Awareness,the existing object grouping methods are studied.A hierarchical clustering algorithm of the proposed object clustering algorithms is proposed.Secondly,the characteristics and functions of the classical hierarchical clustering algorithm are analyzed.The advantages and disadvantages of Rock algorithm,Cure algorithm and Chameleon algorithm in similarity computation are analyzed.The Chameleon algorithm which has obvious advantages in the similarity calculation is analyzed,and the basic concepts,mathematical models and methods are studied in detail.The limitations of Chameleon algorithm are found by theoretical analysis and simulation experiments.Finally,according to the limitation of Chameleon algorithm,the DPC algorithm based on density peak is introduced into the first stage of Chameleon algorithm,and the community structure is introduced into the second stage of Chameleon algorithm.Two-stage improvement algorithm which can be used to solve the problem of object grouping is proposed.Then this thesis introduces the mathematical model and process of the improved Chameleon algorithm in detail.The experimental results show that the Chameleon improved algorithm proposed in this thesis has the advantages of robustness to the input parameter and dealing with the multi-shape well in solving the target grouping problem. |
PDF Full Text Request