Method Of Urban Land Function Classification Based On GF-1 Imagery And POI Data

With the rapid development of social economy and.the continuous improvement of land use in depth and breadth,the direction of development of land use is gradually increased from single to plural.Urban land function classification is of great significance in the analysis of land resource consumption and virtual trade,and the study of the dynamic relationship between social and economic development and land resource pressure.Therefore,the construction of land functional classification system and the identification of urban functional parcels has become a hot topic for scholars at home and abroad.For the problem that we can not build the corresponding relationship between the GF-1 multi-spectral remote sensing images and the land function types,first of all,according to the spatial distribution characteristics of national economic industrial lands,we construct a complete set of land function classification system adapted to the rules of industrial land.Secondly,we use the object-oriented classification method to classify the GF-1 multi-spectral remote sensing images and extract the land cover information from them.Meanwhile,we use the short text classifition method to POI data for national economic industrial classification.Finally,the POI datas classified of the national economic industries are calculated using kernel density estimation,and then we calculate the distance between the normalized kernel density values of the parcels’ area after the multi-scale segmentation of the construction land and the typical industrial parcels’.At last,we form a classification method of land function assisted by POI data.The experimental results show that we can build the the corresponding relationship between the remote sensing images and the land function.The automatic classification method of land function proposed in this paper has high classification accuracy of land function,and can classify urban land well.