Abstract: Graphical applications and interactive systems,especially applied in large scene simulation,are often required to display large quantities of objects synchronously．But not much attention has been focused on the organization of the scene structure which plays an important role in determining the speed of displaying the whole scene each frame respectively while much has been spent on surface simplification and other coherency within each object’s representation．Though the coherency within the scene is so simple to understand,this thesis will still present some also simple methodology which can be used to factually improve the efficiency in displaying the customized scene via applying geometric data analysis or related method．We presented a new method for representing a hierarchy of scene objects which partitions the scene into groups of objects termed as clouds of objects．The method has several variants,some of which applies only geometric data analysis,some of which just applies scene’s inherent geometrical property and others which applies both．Finally,we presented some statistics information to verify improvement by applying our method in graphical interactive systems.

Key words: Geometrical data analysis,AHC,Cloud,Culling