中文摘要：

利用影像设计地理超图模型，将地理超图模型用于虚拟环境选线系统。用任意多边形表达综合地形图上各特定区域对应的图像物体，并将该多边形区域设计成广义象图标。把数据结构的有向图表示与集合的图示相结合，建立地理超图的数据模型。地理超图数据模型主要包括类、对象、类的属性、对象的属性、类与类之间的链、对象与对象之间的链六种基本成分。在地理超图的知识模型中，采用分层模型将领域知识划分为单元级、概念级和实体级，各知识单元通过关系链相连；图像的二维字符串表示定性地描述了图像各物体之间的空间关系，从而将图像匹配的问题转化为二维字符串匹配的问题。在地理超图查询中，采用任意多边形广义象检索算法快速定位广义象；基于广义象建立可视查询功能，把电脑与人的眼脑识别模式相结合，将系统查询与人们的视觉紧密联系在一起，可以直观地查询信息和知识。

英文摘要：

This paper applies image to the design of the geo-hypergraph model of virtual environment railway system. In the model of this paper, image objects are expressed with generalized icons in arbitrary polygon. By combining vector graph of data structure with assembly, the paper constructed a data model of geo-hypergraph. The geo-hypergraph of data model includes class, object, attributes of class, attribute of object, chain of class, and chain of object. In the knowledge model of geo-hypergraph, domain knowledge is divided into principle knowledge, idea knowledge and practice knowledge connected with related chains. By using 2-dimension characters to express the 3-dimension relation of objects, an image matching can be changed as a 2-dimension character matching. At last, this paper introduces a method for knowledge digging from the geo-hypergraph model based on arbitrary polygon generalized icons. Based on generalized icons, visual inquiry function has been established, which combines the recognition mode of computer with that of human eyes. System inquiry is closely connected with people vision so that information and knowledge can be visually inquired.