Preliminary study on cluster methods for geometric shape based on correlation graph

On study for geographic information science, description and measurement for shape of spatial objects is a very important research subject. But current methods used in this field all have some limitations. In this paper the authors discuss a new method to solve the problem better. The authors also put forward a new association matrix — Angular Matrix, which can represent geometric shape in a better way. Preliminary experiment shows that this method can express characteristics of geometric shape better with desirable calculation efficiency.