Research On Feature Recognition And Pedigree Construction Of Typical Rural Settlement Spatial Form

To meet the needs of transformation and reconstruction villages and towns,rural-settlementrelated research has recently become a popular issue.The extraction and utilization of various characteristics of villages and towns to realize effective classification of rural settlements,known as rural settlement pedigree construction,is capable of providing vital assistance for rural planning and management.However,due to the short history of rural research,there are few studies on the construction of village and town pedigrees,where related technologies and methodology are not yet well-developed.In the meantime,the spatial characteristics of existing rural settlements generally lack comprehensiveness,and most current-existing settlement pedigrees have the problems of onesidedness and poor universality as well.In this study,a 5-layer and 3-dimensional spatial pedigree of rural settlements was constructed based on statistical analysis and urban-rural planning theory,where 115 typical villages and towns in 7 counties(districts)in China are taken as targets.The pedigree is capable of reasonably classifying rural settlements while showing advantages of strong interpretability and universality.The specific work includes two aspects:a)Construction of spatial morphological feature database,which is used to describe the multi-dimensional characteristics of the targets;b)Structural analysis of spatial characteristics of the settlements,and construction of a reasonable pedigree.In the feature database construction,the study introduces landscape analysis methods to quantify the spatial features based on the characteristics of the dataset,while extracting the neighboring relationship of the plots through grid cutting method to enrich the existing data expression.24 spatial features were successfully summarized,and a more comprehensive describing tool of target characteristics was thus established.To avoid the one-sidedness of classification,the whole process of pedigree construction is based on the multi-dimensional spatial feature database in this study.First,the idea of top-down division is adopted.Data analytical techniques such as Factor Analysis and Clustering are used to summarize the spatial morphology of the targets into 5 general features,that is,the 5 levels of the pedigree.Theories of urban and rural planning are also employed to analyze the hierarchical sequence of the pedigree levels.In the last step,different Clustering methods are introduced to divide hierarchical subsets of the pedigree and to solve the problem of excessive differences with regard to the value and number of each type.As the result,a 5-layer,3-dimensional pedigree structure is constructed,the reasonability of which is afterwards demonstrated by the combination of its topographic,economic and transportation factors.