Identification And Analysis Of Urban Functional Areas In Chengdu Based On Multisource Data

In recent years,the urbanization level of China keeps increasing,and the rapid development of economy promotes the rapid expansion of urban space,but it is also accompanied by the aggravation of the contradiction between population increase and land utilization.Modern cities urgently need to transform from incremental land for urban expansion to urban development mode of optimizing land stock and realizing land reduction.As an important step in this process,comprehensive and scientific urbanization planning layout is listed as the main component of land space planning policy in the new era.It is of great significance to define the main functions of different Spaces and divide the urban functional areas for the fine management of land development pattern.The data basis of traditional urban land use function identification research is mostly population census data and land census data.Due to the limitations of these data(high acquisition cost,difficult management,poor timeliness,etc.),such research is still at the macro level.With the popularization of remote sensing technology,the recognition and classification of remote sensing image are gradually applied to the research of urban land function recognition.Along with the further in the information age,a variety of sensors in cities,such as smart phones and on-board GPS devices,produce a lot of real-time and reliable location data,and the data acquisition method is becoming more and more easy.The use of big data for urban function form and structure research has become the hotspot in the field of urban space.However,previous studies have not fully mined the temporal characteristics of location data,and most of them use the direct clustering method to judge the difference of urban land use types,resulting in the low accuracy of the recognition results.In this background,This paper adopts the method of combining supervised learning with unsupervised learning to identify and analyze urban functional areas from the perspective of data mining.Taxi trajectory data in 2016,POI(points of interest)data in 2017 and road network data in 2017 of Chengdu were used to construct a multi-source data-based urban functional area identification,result verification and function evaluation system.The main research contents and results of this paper are as follows:(1)A method for identifying urban functional areas is constructed: First,based on trajectory data from Didi Chuxing within the high-speed road surrounding Chengdu,we generated trajectory time sequence data and used the dynamic time warping(DTW)algorithm to generate a time series similarity matrix.Second,we utilized the K-medoid clustering algorithm to generate preliminary results of land clustering and selected the results with high classification accuracy as the training samples.Then,the k-nearest neighbour(KNN)classification algorithm based on DTW was performed to classify and identify the urban functional areas.Finally,with the help of point-of-interest(POI)auxiliary analysis,the final functional layout in Chengdu was obtained.In order to verify the accuracy of the results,the identified functional areas were compared with Google earth images,Gaode maps,and real photos.The comparison results of typical areas were selected for display.(2)Based on the identification results of urban functional areas,the urban spatial morphology of the study area is analyzed from the perspective of region and function by using the indicators such as land mixing degree,location entropy,Moran's I and average nearest neighbor distance.The main conclusions of this paper are as follows:(1)It is found that this method has high accuracy for the identification of urban functional structure,and also has good effect in the functional areas which are difficult to be distinguished by traditional methods.This method can make up for the disadvantages of traditional data such as land and remote sensing.(2)The distribution pattern of the functional areas is similar to that of the loop lines,while the dominant function of different administrative areas is obviously different.Each loop line and each administrative region land use degree is high,but the degree of land mixing in many areas is not consistent with its urban spatial vitality.The urban expansion of Chengdu is single-center development,which is different from the multi-center and group-type development model of many developed cities.This model is prone to a variety of problems such as unbalanced land allocation,disharmony of layout,limited urban functional radiation range and so on.In view of the above findings,this paper puts forward some suggestions for the optimization of future development: Optimize the function of the downtown area and reduce the development intensity;Implement the new urbanization development strategy based on the integration of circles and give full play to the core advantages of the first circle;Improve the charm of the city,improve the supporting facilities,promote the reform and development of tourism;To develop and construct a multi-center,group-type urban development pattern.