Edge Bundling Based Graph Visualization Research In 3D Space

Edge bundling is an important way to solve visual clutter in graph visualization,it is also commonly used in graph layout optimization.In the face of today's large-scale data,research of graph visulaization will inevitable face the problem of visual clutter,the problem has also become a hotspot of today's research.This thesis systematically analyzes the research of edge bundling at the present stage,for the existing algorithm did not distinguish important content in graph,proposes an content importance based edge bundling algorithm,and for the less of edge bundling algorithm in 3D and the weak support of geographic information data in 3D environment,proposes expectation based 3D force-directed edge bundling algorithm,which extends the advatages of above algorithm.Experiments show that,the algoritm can effectively ensure that different edge cluters do not effect each other while reducing visual clutter,and have a good bundling effect for 3D geographic information data in 3D environment.The main work and contribution of this thesis are as follows:(1)Existing visual clutter solve methods are summarized and analyzed: existing visual clutter solve methods are analysed,the general process of existing edge bundling algorithm are summarized,and the realization idea of different algorithms in different parts are systematically analyzed,the applicability,advatages and disadvatages of each algorithm are analyzed,after that content extraction method of graph is briefly described.(2)Content importance based edge bundling algorithm is proposed: relation calculate model is proposed,and force directed algorithm is improved to use relation information.And a specific bundling plan is designed to ensure the important content is independent of each other after bundling.(3)Expectation based 3D force-directed edge bundling algorithm is proposed: expectation model is proposed,with this model the force directed system is improved,which enhanced the stability of the force system,and ensures the validity of the algorithm with large scale of date in 3D space.With this model expansion of the rules is supported,which makes it possible to accomplish the graph visualization edge bundling for geographic information data in 3D space,combined with 3D environment.