A new hierarchical spatial index for area entities based on urban morphology

Urban morphology is closely related with natural conditions, functional organization of land use and road network structure. Its basic skeleton is formed by axes of road and hydrographic network. Based on the principle of urban morphology, urban space can be partitioned into four levels with its components being enclave, block, superblock and neighborhood, which is obtained by some topographic features such as roads and hydrographic networks. Obviously, this way of partition is very natural and consistent with human cognition, and thus the basic components at each level are called natural grids. Indeed, the partition can effectively avoid such case that topographic features cross grids, which often occurs in the partition based on regular grids. A new kind of hierarchical spatial index method is further proposed based upon the natural grids, which can be combined into existing classification of spatial index to obtain a new extended classification framework. Particularly, there is no redundancy of index records based on the new spatial index, so that it can improve the efficiency of spatial query. Several experiments are employed to prove that the redundancy of index recording in grid affects its efficiency greatly. In addition, the approaches of space partition and index building in this paper can be further used for map generalization and map updating, thus provide the basis for them in both theory and technology.