Study And Application On Spatial Clustering Algorithm Based On Surface Area With Integrated Multi-type Data

Clustering analysis is the most simple human understanding objective things, one of the most commonly used method, with the development of the cluster analysis, and the study of clustering analysis in recent years, data collected more and more various aspects, such as agricultural meteorological data, such as agricultural data, which is produced by the mesoscale earth observation system of large scale, sources, characteristics of diversity, rich resources and cheap data, clustering provides a good data base for the space, but at present study of spatial clustering mainly focused on the spatial point attribute clustering, as for the spatial attributes of the planar area spatial clustering is a lot of research is relatively less, meanwhile, multi-source, multi-dimensional, multi-type data clustering analysis, there are also some problems, such as the spatial data value degree is not high, duality, mixed attribute data type is not unified, attention degree is not high question, spatial data algorithm needs to be updated, space attribute data connectivity is not quite close together and processing software and platform of complicated problems.According to the problems existing in the previous research, this article from the spatial and attribute data of planar, thinning of spatial and attribute data, standardizing of attribute data and weight into the space, the K-as algorithm was improved and updated, puts forward the Touches adjacency, adjacency merging (space) for the first time, seed the search space (second) and so on the new algorithm, and finally developed a spatial clustering software system based on planar area, effectively solve the data processing is simple, old algorithm, attaches great importance to the lack of spatial data, spatial and attribute data connection is not quite close together, problems such as complex processing platform diversity, especially seeds search algorithm is put forward, very good to avoid the planar clustering of clustering, messy scattered blocks, and explain the emergence of single block area (enclave), and eventually generate secondary zoning map the national agricultural regionalization, achieved good effect.