CN116310853A - Multi-source data-based extraction method for edge regions of medium and small cities - Google Patents

Abstract

The application discloses a method for extracting a border region of a medium and small city based on multi-source data, which comprises the following steps: acquiring remote sensing images, POI (PointOfInterest) data, lamplight images and population data of cities; analyzing the remote sensing image to obtain landscape disorder degree; evaluating the nuclear density according to the POI data to obtain the POI nuclear density; determining night light intensity according to the light image; determining a single interpretation effort degree of the view disorder degree, the POI nuclear density and the night light intensity on population data; determining the weight of the landscape disorder degree, the POI nuclear density and the night light intensity; determining the comprehensive index of each region according to the weight; the regions are divided into a plurality of categories according to the comprehensive index, including urban edge areas. The model provided by the application is obviously stronger than a landscape disorder degree threshold method and a POI nuclear density breakpoint analysis method in the aspects of precision, detail and integrity, has higher universality and can meet the research requirements of the border areas of small and medium cities.

Description

A method of extracting the fringe area of small and medium cities based on multi-source data

technical field

The present application relates to the technical field of image processing, in particular to a method for extracting fringe areas of small and medium-sized cities based on multi-source data.

Background technique

As the link between the city and the countryside, the urban fringe area is the area with the fastest changes in land use and spatial structure in the process of urban expansion, and it has the characteristics of diversity, dynamics, and transition. Rapid and accurate identification of urban fringe areas has important practical significance for optimizing urban spatial layout, controlling urban expansion and protecting land resources.

At present, the identification methods of urban fringe areas mainly include urban-rural gradient view method, threshold method, mutation point/breaking point analysis method, etc. The urban-rural gradient view method mainly identifies urban fringe areas based on the spatial gradient changes of factors such as regional land use, socioeconomics, and population density. However, the urban-rural gradient view method is difficult to overcome the subjectivity of determining the dividing point in areas with scattered landscape structures. The threshold method is to determine the urban fringe area according to the threshold range of the distance from the built-up area, population density, building proportion, information entropy and other indicators. The threshold method is simple and practical, but usually the determination of the threshold requires repeated experiments, which has problems such as low efficiency, discontinuous results, and poor versatility. The abrupt change point/break point analysis method is to determine the urban fringe area by calculating the sudden change/break value of single or comprehensive indicators such as nighttime light intensity, impermeable surface index, and landscape disorder in different directions through the model, which is currently the mainstream method.

However, the research on the identification of urban fringes is mainly concentrated in large cities, and there is no extraction model suitable for small and medium-sized urban fringes. Different from large cities, the fringe areas of small and medium-sized cities are often small, and the available data related to urban development (economy, population, lighting images) have low spatial resolution, which increases the difficulty of accurately identifying urban fringe areas.

Contents of the invention

The embodiment of the present application provides a method for extracting fringe areas of small and medium-sized cities based on multi-source data, which is used to solve the problems of high difficulty, low accuracy, poor versatility, and discontinuous results in existing methods for extracting fringe areas of small and medium-sized cities.

On the one hand, the embodiment of the present application provides a method for extracting fringe areas of small and medium-sized cities based on multi-source data, including:

Obtain remote sensing images, POI data, lighting images and population data of the city;

Carry out land use classification on remote sensing images to obtain urban landscape disorder;

Evaluate the city's kernel density based on POI data to obtain the city's POI kernel density;

Determine the nighttime light intensity of the city based on the light image;

Determine the degree of single explanatory power of landscape disorder, POI kernel density, and nighttime light intensity to population data;

Determine the weight of landscape disorder, POI kernel density and night light intensity according to the ratio of single explanatory power to the total explanatory power;

Determine the comprehensive index of each area in the city according to the weight;

Areas of the city are divided into categories based on a composite index, which includes the urban fringe.

A method for extracting fringe areas of small and medium cities based on multi-source data in this application has the following advantages:

Based on GF-2 images, POI data, NPP/VIIRS images, and Woldpop multi-source data, the extraction model of small and medium-sized urban fringe areas was constructed with landscape disorder, POI kernel density, and nighttime light intensity as urban characteristic factors. Through the test of three small and medium-sized cities in Hantai District, Shangzhou District and Hanbin District, it is found that the model proposed in this application is significantly better than the landscape disorder threshold method and POI kernel density breakpoint analysis method in terms of accuracy, detail and integrity. , and has high versatility, which can meet the research needs of small and medium-sized urban fringe areas.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present application. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a flow chart of a method for extracting fringe areas of small and medium-sized cities based on multi-source data provided by the embodiment of the present application;

Fig. 2 is the schematic diagram of landscape disorder in Hantai District provided by the embodiment of the present application;

FIG. 3 is a schematic diagram of kernel density under different bandwidths in the Hantai District provided by the embodiment of the present application;

FIG. 4 is a schematic diagram of the extraction results of the urban fringe area in Hantai District provided by the embodiment of the present application;

Fig. 5 is a schematic diagram of the extraction results of urban fringe areas in Shangzhou District and Hanbin District provided by the embodiment of the present application.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the application with reference to the drawings in the embodiments of the application. Apparently, the described embodiments are only some of the embodiments of the application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

FIG. 1 is a flow chart of a method for extracting fringe areas of small and medium-sized cities based on multi-source data provided by an embodiment of the present application. The embodiment of the present application provides a method for extracting fringe areas of small and medium-sized cities based on multi-source data, including:

S100, acquiring remote sensing images, POI data, light images and population data of the city.

Exemplarily, the data used in this application include five types of data: GF-2, POI, Woldpop, NPP/VIIRS, and administrative boundaries. Table 1 shows the specific information of the data. Among them, the remote sensing images acquired by the GF-2 satellite have achieved sub-meter spatial resolution and multi-spectral comprehensive remote sensing data acquisition, and have the characteristics of high positioning accuracy, high spatial resolution and high temporal resolution. This application uses the GF-2 remote sensing image of 4 scenes covering the study area in July 2020.

POI is a geographic object that can be abstracted as a point, especially some geographic entities that are closely related to people's lives. The POI data in this application comes from Amap (https://lbs.amap.com/).

The light image uses NPP/VIIRS data. The NPP/VIIRS data comes from the National Geophysical Data Center (NGDC) in the United States and is detected by the visible infrared imaging radiation instrument carried by the Suomi NPP satellite. This application selects the monthly data lighting products provided by NGDC in July 2020, with a spatial resolution of 500m.

Population data use Woldpop data. This application selects Woldpop data with a resolution of 100m in July 2020.

Table 1 Data specific information

S110, performing land use classification on remote sensing images to obtain urban landscape disorder.

Exemplarily, S110 specifically includes: using an object-oriented SVM (Supported Vector Machine) classification method to perform land use classification on remote sensing images; and determining the landscape disorder degree according to the land use type.

This application adopts the object-oriented SVM classification method. The object-oriented classification method breaks through the limitations of traditional classification methods that use a single pixel as the basic classification and processing unit, classifies images from the object level, and reduces the loss rate of semantic information contained in traditional pixel-based classification methods. SVM is a classification algorithm based on the VC dimension theory of statistical learning theory and the principle of structural risk minimization. Compared with neural networks or traditional statistics-based classification methods, SVM controls the complexity of the model through the number of vectors, and does not need to reduce the feature variables through dimensionality reduction to control the complexity of the model. Therefore, in the process of classification, The SVM classifier will not lose the characteristic information of the ground object, which reduces the occurrence of some over-fitting phenomena.

The object-oriented SVM classification method takes into account the advantages of object-oriented multi-scale segmentation and SVM. First, multi-scale segmentation is performed according to the properties of the object area on the image, not only considering the spectral information of the image, but also adding texture, geometric shape and spatial topology. and other features, and then use the training samples for support vector machine classification. The object-oriented SVM classification method has obvious advantages in accuracy, generalization and high-dimensional data processing, and has been widely used in remote sensing image classification applications.

Furthermore, before the object-oriented SVM classification method is used to classify remote sensing images for land use, the remote sensing images are also preprocessed. The preprocessing includes atmospheric correction, fusion, mosaic and cropping operations on remote sensing images.

After the preprocessing of the remote sensing image is completed, the land use classification can be carried out on the remote sensing image, and then the degree of landscape disorder can be determined. Specifically, the degree of landscape disorder can represent the degree of fragmentation and dispersion of urban landscape, reflecting the heterogeneity and homogeneity of landscape space. The higher the heterogeneity of land use patches per unit area, the greater the degree of landscape disorder. Urban and rural areas usually have a single type of land use, mostly contiguous construction land or agricultural land, with a low degree of landscape disorder. The urban fringe area is an active expansion zone between the urban landscape and the agricultural hinterland, with various land use types and high landscape disorder. Therefore, the extent of the urban fringe can be judged by the difference in landscape disorder. The formula of landscape disorder degree is as follows:

In the formula, W is the value of landscape disorder, X _n represents the ratio of the nth type of land in the unit area to the unit area; N represents the total number of land use types in the unit area.

S120. Evaluate the city's kernel density according to the POI data to obtain the city's POI kernel density.

Exemplarily, S120 specifically includes: using a kernel density evaluation tool to evaluate the POI data to obtain the POI kernel density.

Kernel density analysis is often used to evaluate the density value of the neighborhood of point or line features, and to simulate the spatial distribution of features, which is widely used in geospatial analysis research. The main principle is that within a certain bandwidth range, the estimated density of elements decreases with the increase of distance, the kernel density is the highest at the center of the element, and 0 at the edge of the bandwidth. Kernel density analysis follows the law of spatial correlation, and the closer the distance is, the greater the correlation is, and the POI data also conforms to this law. The formula for POI kernel density is as follows:

为距离的权重。In the formula, λ _(s) is the calculation of the POI kernel density at the sth area, r is the bandwidth set by the kernel density function, n' is the total amount of all elements involved in the calculation, d _ls is the POI point l and s distance value,

is the distance weight.

Further, before using the kernel density evaluation tool to evaluate the POI data, the POI data is also preprocessed. This preprocessing includes filtering and reprojecting the POI data.

S130. Determine the night light intensity of the city according to the light image.

Exemplarily, S130 specifically includes: determining a DN (Digital Number) value of each image in the light image to obtain the light intensity at night.

Likewise, before determining the DN value of each image in the light image, the light image is also preprocessed, which includes reprojection, denoising and cropping of the NPP/VIIRS data.

S140, determining the degree of single explanatory power of the landscape disorder degree, POI kernel density and night light intensity to the population data respectively.

Exemplarily, S140 specifically includes: using the differentiation and factor detection of the geographical detector to determine the degree of single explanatory power of the landscape disorder degree, POI kernel density and nighttime light intensity to the population data.

Geographic detectors are a set of statistical methods to detect spatial variability and explain the driving forces behind it, including differentiation and factor detection, interaction detection, risk area detection and ecological detection. The main principle of the geographic detector is to assume that the research area is divided into several sub-regions. If the sum of the variances of the sub-regions is less than the total variance of the region, there is spatial differentiation; if the spatial distribution of the two variables tends to be consistent, there is a statistical correlation between the two . Geographic detectors can assess spatial variability, detect explanatory factors, and analyze the interaction between variables, and have been widely used in fields such as nature, environmental science, and human health. Differentiation and factor detection is to detect the spatial differentiation of attribute Y and the explanatory power of a certain factor X to attribute Y, which is measured by q value. Population density in population data is closely related to urban development. This application determines the weight of each factor according to the explanatory power of landscape disorder, POI kernel density, and night light intensity to population data. The formula for q is:

In the formula: L is the stratification of population data Y or landscape disorder, POI kernel density, and night light intensity X, that is, classification or partition; N _h and N' are the number of units in the h-th layer and the whole area, respectively, σ _h ² and σ ² are the variance of the population data Y of the h-th layer and the whole area, respectively, and SSW and SST are the sum of the variance within the layer and the sum of the variance of the whole area, respectively.

S150. Determine the weights of landscape disorder, POI kernel density, and night light intensity according to the ratio of a single explanatory power degree to the total explanatory power degree.

S160. Determine the comprehensive index of each region in the city according to the weight.

Exemplarily, the weighted sum of landscape disorder degree, POI kernel density, night light intensity and corresponding weights may be used as the comprehensive index.

S170, according to the comprehensive index, the various areas of the city are divided into categories, including the urban fringe area.

Exemplarily, the natural break point method can be used to divide various areas of the city into multiple categories, and this application includes three categories, namely urban core area, urban fringe area and rural hinterland.

Natural Breaks is a map binning algorithm that divides the data into two groups, binary natural breaks. The natural breakpoint method believes that there are some natural turning points and breakpoints between any sequence, and the research objects can be divided into groups with similar properties through these turning points and breakpoints. The main principle of the algorithm is to cluster each value, and the condition for the end of clustering is that the variance between groups is the largest and the variance within the group is the smallest. The process of realizing the natural breakpoint method is divided into three steps:

(1) Calculate the sum of squared deviations SDAM of the array mean:

为数组平均值。SDAM is the sum of squared deviations from the mean of the array,

is the average of the array.

(2) For each region within the range, compute the sum of squared deviations SDCM from the class mean and find the smallest one.

(3) Calculate the variance goodness-of-fit GVF:

GVF=(SDAM-SDCM)/SDAM

The value of GVF is between 0 and 1, with 1 indicating an excellent fit and 0 indicating an extremely poor fit. The variance goodness of fit is to verify whether the classification results achieve the grouping purpose of the smallest intra-class difference and the largest inter-class difference.

After dividing each area of the city into multiple categories according to the comprehensive index, this application also performs raster conversion and smoothing on the raster image obtained after the category division to obtain the range of the urban fringe area.

Experiment Description

D. Hantai District is located in the center of the Hanzhong Basin in the southwest of Shaanxi Province, with the Han River in the south and the Qinling Mountains in the north. The terrain of Hantai District is higher in the north and lower in the south. The northern part belongs to the southern slope of Qinling Mountains, with an altitude of 700-2000 meters, accounting for 34% of the total area; the central part is hilly area, with an altitude of 541-700 meters, accounting for 28% of the total area; the south is the alluvial plain of the Han River , accounting for 38% of the total area. Hantai District is the largest commodity distribution center in southern Shaanxi, and is the core area of Qinling and Bashan, with important economic and ecological value. Since the urban construction of Hantai District is mainly concentrated in the south, this application uses the administrative boundaries of 8 towns in the south to form the research area.

results and analysis. The results of landscape disorder threshold method: Based on the preprocessed GF-2 remote sensing images in the study area, the object-oriented SVM classification method is used to divide the land use into vegetation (cultivated land, woodland, grassland), construction land, water body and unused land. class, Figure 2(a). It can be seen from the figure that the construction land is mainly distributed in the middle and south of the study area, the water body is distributed along the southern boundary, the unused land is mainly distributed in the north, and the vegetation is mainly distributed in the northwest, northeast and southeast. According to statistics, the vegetation area in the study area is 80.34km ² , the construction land area is 60.83km ² , the water body area is 6.95km ² , and the unused land area is 3.31km ² .

In order to highlight the circle structure of landscape disorder in the study area, a 100m×100m grid was constructed as the spatial calculation scale unit, and ArcGIS 10.3 software was used to calculate the area ratio of vegetation, construction land, water body and unused land in the grid, and finally Calculate the landscape disorder degree in the study area, Fig. 2 (b). It can be seen from the figure that the landscape structure features of the urban core area are prominent, the landscape disorder is low, and there are concentrated and contiguous low-value areas. After repeated experiments, the threshold value less than 0.46 is used as a sign to identify the urban core area. However, when there is a large area of green space in the urban core area, the landscape disorder is high, and the landscape disorder threshold method cannot identify the complete urban core area. There is no obvious difference between the urban fringe area and the rural landscape disorder, and the landscape disorder is relatively high. Different from large cities, small and medium-sized cities have a small population and small villages, and the study area is located in the Hanzhong Plain, where the villages are relatively concentrated and the cultivated land is relatively scattered, resulting in a high degree of landscape disorder between the urban fringe area and the countryside. Although the landscape disorder threshold method is widely used in the identification of the fringe areas of large cities, in the process of identifying the fringe areas of the study area, the single-factor threshold method generally has the disadvantages of discontinuous results, insufficient details, and poor versatility. was significantly enlarged. Therefore, the landscape disorder threshold method is not suitable for the identification of small and medium urban fringe areas.

Results of POI kernel density breakpoint analysis method: Based on the preprocessed POI data in the study area, the Kernel Density tool of ArcGIS10.3 was used to calculate the POI kernel density. The bandwidth in kernel density analysis has a crucial impact on the results. After referring to the existing research results, 500m, 1000m and 1500m were respectively selected for kernel density analysis. The results are shown in (a)-(c) in Figure 3 Show. It can be seen from the figure that when the bandwidth is 500m, the results of kernel density analysis are fragmented and discontinuous, and the overall distribution of urban POIs is not obvious; When the bandwidth is 1000m, the kernel density analysis results have good stability, and the overall distribution trend is obvious, which can meet the analysis needs of the urban fringe area in the study area.

Figure 3(d) is the result of dividing the POI kernel density with a bandwidth of 1000m in the study area into three categories by using the natural breakpoint method. Combined with (b) and (d) in Figure 3, it can be seen that the study area presents an obvious distribution of circle structure, the urban core area has a large area of continuous high-density areas, the density of urban fringe areas is low, and most rural areas The density is close to 0. Compared with the landscape disorder threshold method, the analysis results of POI kernel density can clearly and completely identify the urban core area, but the identification results of the two methods for the urban fringe area are not ideal. The POI kernel density breakpoint analysis method will identify some well-developed villages with a large number of POIs as urban fringe areas. At the same time, the urban fringe area is in the development stage, with few POI data and slow update speed, resulting in results that are different from the actual urban fringe area. There are large errors. Different from large cities, POI data in the fringe areas of small and medium-sized cities have low integrity and slow update, and the single-factor POI kernel density breakpoint analysis method is difficult to accurately extract the fringe areas of the study area.

The results of the extraction model for the fringe areas of small and medium-sized cities: based on the extraction model for the fringe areas of small and medium-sized cities, the landscape disorder degree, POI kernel density, and night light intensity of the study area are first calculated; then the weight of each factor (landscape disorder degree: 0.10, POI kernel density: 0.51, nighttime light intensity: 0.39) to build a comprehensive index; finally use the natural breakpoint method to identify the urban fringe area, and post-process the results. Figure 4 (a) is the result of dividing the comprehensive index into three categories by using the natural breakpoint method; Figure 4 (b) is the final result of the extraction model of the fringe area of small and medium cities. It can be seen from the figure that the extraction model of small and medium urban fringe areas can accurately and completely identify the urban fringe boundary of the study area. The urban fringes of the study area are mainly concentrated in the north and east, with an area of about 39km ² . Compared with the single-factor landscape disorder threshold method and POI kernel density breakpoint analysis method, the performance of the small and medium-sized urban fringe extraction model has been greatly improved, especially the results of the outer boundary of the urban fringe. The model in this application is relatively consistent with the overall pattern of the inner boundary of the urban fringe (urban core area) extracted by the POI kernel density breakpoint analysis method, but the former is more detailed in comparison. The difference between the two results is mainly concentrated in the southwest of the study area, mainly because the southwest is the Hanjiang New Area, which is still in the stage of rapid development, the landscape pattern changes rapidly, and the update of POI data is slow. Therefore, the POI kernel density breakpoint analysis method extracts There is a discrepancy between the inner boundary of the urban fringe area and the actual boundary. The model of this application focuses on the comprehensive performance of regional landscape disorder, POI kernel density, and night light intensity, and is less dependent on the performance of a single factor. Therefore, it can more accurately identify the range of urban fringe areas.

Evaluation of model accuracy: Since it is difficult to identify the range of urban fringe areas in the study area by the landscape disorder threshold method, this application only evaluates the accuracy of the POI kernel density breakpoint analysis method and the extraction model of small and medium urban fringe areas. Through detailed visual analysis, it is found that the extraction model of the fringe area of small and medium-sized cities is obviously stronger than the POI kernel density breakpoint analysis method in terms of detail and integrity. In order to further evaluate the extraction accuracy of different methods, this application uses two methods of field inspection and landscape pattern index evaluation. Field verification is to uniformly select 100 sample points along the roadside around the urban fringe area. In Figure 4 (b), the accuracy of the extraction results is analyzed through field verification, as shown in Table 2. It can be seen from the table that the overall accuracy of the extraction model for small and medium urban fringe areas is significantly higher than that of the POI kernel density breakpoint analysis method, reaching 98%. The POI kernel density breakpoint analysis method has a large number of false extractions, and the overall accuracy is only 67%.

Table 2 Extraction accuracy of different methods

The landscape pattern index is often used to evaluate the extraction accuracy of urban fringe areas, and the patch density (PD) and Shannon diversity index (SHDI) are selected to evaluate the accuracy of the two methods from the level and landscape level. PD represents the degree of fragmentation of the landscape, and SHDI represents the degree of richness and complexity of the landscape type. PD and SHDI are generally higher in urban fringe areas and lower in urban centers and rural areas. Table 3 is the PD and SHDI values of the two methods in different regions calculated based on Fragstats 4.2 software. It can be seen from the table that in the urban fringe area, the PD and SHDI values of the small and medium-sized urban fringe area extraction model are significantly higher than those of the POI kernel density break point analysis method, which indicates that the landscape in the urban fringe area extracted by the former is fragmented and complex. The degree of sex and diversity is higher than the latter, which can better reflect the characteristics of the urban fringe area. In the rural hinterland, the PD and SHDI values of the small and medium-sized urban fringe area extraction model are significantly lower than those of the POI kernel density breakpoint analysis method, while in the urban core area, the PD and SHDI values are similar due to the small difference in the boundaries extracted by the two methods. In general, the model proposed in this application has high precision and can realize the precise extraction of urban fringe areas in the research area.

Table 3 PD and SHDI values in different regions

Model versatility analysis: In order to further verify the applicability of the model proposed in this application in different regions and different types of small and medium-sized cities, urban fringe areas were identified in Shangzhou District of Shangluo City and Hanbin District of Ankang City. Shangzhou District belongs to a typical belt-shaped urban structure, which is obviously restricted by resource conditions. It can be seen from (a) in Figure 5 that the POI kernel density breakpoint analysis method is incomplete for the extraction results of the urban fringe area of Shangzhou District, mainly concentrated in the southeast. The main reason is that the southeast is an industrial park, and the number of POIs is small and scattered, which is not enough to support the identification of urban fringe areas. Hanbin District is separated from the middle by the Han River, with the old city in the southeast and the new city in the northwest, belonging to a polycentric urban structure. It can be seen from (b) in Figure 5 that the POI kernel density breakpoint analysis method has poor extraction results for the urban fringe area of Hanbin District, especially for the old city in the southeast. The main reason is that the single-factor method requires extremely high data quality when extracting the fringe areas of polycentric cities, while the development of the old urban areas is relatively backward, the population distribution is concentrated, and the POI data is not complete. In contrast, the small and medium urban fringe area extraction model can accurately and completely extract the urban fringe areas of the two regions. The model identifies urban fringe areas based on the comprehensive performance differences of regional landscape disorder, POI kernel density, and night light intensity. It is less dependent on the performance of a single factor and can adapt to different regions and different types of small and medium-sized cities.

While preferred embodiments of the present application have been described, additional changes and modifications to these embodiments can be made by those skilled in the art once the basic inventive concept is appreciated. Therefore, the appended claims are intended to be construed to cover the preferred embodiment and all changes and modifications which fall within the scope of the application.

Obviously, those skilled in the art can make various changes and modifications to the application without departing from the spirit and scope of the application. In this way, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims (8)

1. The method for extracting the edge region of the medium and small cities based on the multi-source data is characterized by comprising the following steps of:

acquiring remote sensing images, POI data, lamplight images and population data of cities;

land utilization classification is carried out on the remote sensing images, and the landscape disorder degree of the city is obtained;

evaluating the nuclear density of the city according to the POI data to obtain the POI nuclear density of the city;

determining the night light intensity of the city according to the light image;

determining single explanatory degrees of the view disorder degree, the POI nuclear density and the night light intensity on the population data respectively;

determining the weight of the landscape disorder degree, the POI nuclear density and the night light intensity according to the proportion of the single interpretation effort degree to the sum of the interpretation effort degrees;

determining the comprehensive index of each region in the city according to the weight;

and dividing each region of the city into a plurality of categories according to the comprehensive index, wherein the categories comprise city edge regions.

2. The method for extracting the border region of the medium and small cities based on the multi-source data according to claim 1, wherein the step of classifying the land utilization of the remote sensing image to obtain the landscape disorder of the cities comprises the following steps:

performing land utilization classification on the remote sensing images by using an object-oriented SVM classification method;

and determining the landscape disorder degree according to the land utilization type.

3. The method for extracting border areas of small and medium cities based on multi-source data according to claim 2, wherein before land utilization classification is performed on the remote sensing images by using an object-oriented SVM classification method, the method further comprises:

and preprocessing the remote sensing image.

4. The method for extracting border areas of small and medium cities based on multi-source data according to claim 1, wherein the step of evaluating the nuclear density of the cities according to the POI data to obtain the POI nuclear density of the cities comprises the following steps:

and evaluating the POI data by using a nuclear density evaluation tool to obtain the POI nuclear density.

5. The method for extracting a border region of a medium and small city based on multi-source data according to claim 4, further comprising, prior to evaluating the POI data using a nuclear density evaluation tool:

and preprocessing the POI data.

6. The method for extracting border areas of small and medium cities based on multi-source data as set forth in claim 1, wherein the determining the night light intensity of the cities according to the light images includes:

and determining DN value of each image in the lamplight images to obtain the night lamplight intensity.

7. The method for extracting border areas of small and medium cities based on multi-source data as set forth in claim 1, wherein said determining a single degree of interpretation of said demographic data by said degree of landscape turbulence, POI kernel density and night light intensity, respectively, comprises:

and determining the single explanatory degree of the landscape disorder degree, the POI nuclear density and the night light intensity on the population data respectively by utilizing the difference and factor detection of the geographic detector.

8. The method for extracting border areas of small and medium cities based on multi-source data as set forth in claim 1, further comprising, after dividing each area of the cities into a plurality of categories according to the composite index:

and carrying out grid conversion vector and smoothing on the grid images obtained after category division to obtain the range of the urban border area.

Images