Urban Land Use Evolution Method and System Considering Multi-level Spatial Interaction Heterogeneity

Abstract

The present invention discloses a method and system for urban land use evolution considering multi-level spatial interaction heterogeneity. First, the research area is divided into multiple spatial structure units through an urban spatial structure diagram; then, the urbanization index of different-level zoning units is calculated, and the spatial field intensity between different-level zoning units and spatial structure units is obtained based on the urbanization index using the gravity model, thereby determining the spatial interaction effect of different-level zoning units on spatial structure units; at the local level, the neighborhood effect and environmental effect of plots within the spatial structure units are calculated; then, considering the multi-level spatial interaction effects at different levels and locally, a cellular automaton model is constructed and the weight parameters of the model are determined using the hill climbing method; finally, the constructed cellular automaton model is used to simulate the land use evolution process. The present invention can capture the characteristics of land use changes more accurately and has high accuracy and applicability.

Description

Technical Field

[0001] The present invention relates to the fields of geography and geographic modeling, and particularly relates to a method and system for urban land use evolution considering multi-level spatial interaction heterogeneity. Background Art

[0002] With the continuous acceleration of regional urbanization and urban regionalization, the intensity of regional spatial interaction has become increasingly significant. Due to the theoretical simplicity of the Cellular Automata (CA) modeling and the high efficiency in simulating complex systems, it has been widely adopted in simulating land use dynamics and analyzing the human-land interaction mechanism. Existing CA model studies have increasingly focused on the regional spatial interaction effect in the urban evolution process, constructed diverse spatial interaction models from perspectives such as population flow, information flow, two-way urban flow, and comprehensive urban influence index, and integrated them with the CA model to explore the impact of the interaction between cities (urban flow) on urban expansion.

[0003] Spatial heterogeneity is an endowment of the spatial distribution of urban resources and also a significant feature in the urban evolution process. Urban development can be regarded as a non-stationary process in time and space. Many CA model studies mainly focus on the spatial heterogeneity of the geographical environment and transformation rules. In addition, some scholars have not only confirmed that the urban evolution presents a concentric circle spatial structure pattern, but also revealed that the land use combination and transformation rules within different structural units show differences. Existing studies have revealed the impact of regional spatial interaction (between cities) on spatial evolution, but they have ignored the multi-level characteristics of regional interaction, that is, the land use change is not only caused by the interaction between cities and between local areas (defined as neighborhood effect in the CA model), but is the result of the synergistic effect of multi-level spatial interactions at different levels and local (neighborhood). In addition, although existing studies have found the differences in land use combination and transformation rules within different concentric circle spatial structure units, the problem of heterogeneity of multi-level spatial interaction effects caused by the urban concentric circle spatial structure pattern has not been considered. Summary of the Invention

[0004] Object of the Invention: Aiming at the problems existing in the above-mentioned prior art, the object of the present invention is to provide a land use evolution method and system, by considering the urban spatial interaction structure, measuring the multi-level spatial interaction effects at different levels, and combining them with the cellular automata model, to construct a cellular automata model considering multi-level spatial interaction heterogeneity (MRC-CA) to more scientifically and realistically simulate the land use evolution process.

[0005] Technical Solution: To achieve the above object of the invention, the present invention adopts the following technical solutions:

[0006] A method for urban land use evolution considering multi-level spatial interaction heterogeneity, comprising the following steps:

[0007] The research area is divided into multiple spatial structure units through the urban spatial structure diagram;

[0008] Calculate the urbanization indices of the zoning units at three levels: between cities, within cities, and in local areas. Use the gravity model to obtain the spatial field intensity between different-level zoning units and spatial structure units based on the urbanization indices, and then determine the spatial interaction effect of different-level zoning units on spatial structure units; at the local level, calculate the neighborhood effect and environmental effect of the plots within the spatial structure unit;

[0009] Comprehensively consider the multi-level spatial interaction effects at different levels and locally, and construct a cellular automaton model, where the land use conversion rules are jointly determined by the comprehensive effects weighted by the multi-level spatial interaction effects;

[0010] Use the hill climbing method to determine the weight parameters of the model;

[0011] Use the constructed cellular automaton model to simulate the land use evolution process.

[0012] Preferably, the spatial interaction effect of different-level zoning units on spatial structure units is determined according to the following formula:

[0013]

[0014] where i = 1 or 2 represents the levels of between cities and within cities, SFI i,j,k represents the spatial field intensity between the i-level zoning unit j and the spatial structure unit k, UC i,j is the urbanization index of the i-level zoning unit j, UC 3,k is the urbanization index of the spatial structure unit k in the local area, D(x i,j , y i,j , x k , y k ) represents the Euclidean distance from the center point (x i,j , y i,j ) of the i-level zoning unit j to the center point (x k , y k ) of the spatial structure unit k; RI i,k represents the interaction effect of all i-level zoning units on the spatial structure unit k, m is the number of i-level zoning units; RI′ i,k is the spatial interaction effect of each plot within the spatial structure unit k, S i,k is the area of the i-level zoning spatial structure unit k; is the normalization value.

[0015] Preferably, the calculation method of the urbanization index incorporates indicators such as the night light index, urban population quantity, and construction land area. The specific calculation formula is as follows:

[0016]

[0017] Among them, UC i,j is the urbanization index of the i-th level zoning unit j; light i,j , PoP i,j , built i,j respectively represent the night light index, population quantity, and construction land area of the i-th level zoning unit j.

[0018] Preferably, when calculating the neighborhood effect of a plot, the first-order and second-order adjacent plots are set as the neighborhood range; when calculating the environmental effect of a plot, the altitude of the plot, the distance from the center of different-level zoning units, and the distances from highways, main roads, and secondary roads are considered, and the Logistic regression model is used to obtain the environmental effect of each plot.

[0019] Preferably, the land use conversion rule in the cellular automaton model is expressed as:

[0020]

[0021] Among them, P a,b is the transition probability from land use type a to b, where land use type a is in the spatial structure unit k, w i,k,a,b is the interaction effect weight of the i-th zoning level of, Ω a,b is the neighborhood effect of changing from land use type a to b, E a,b is the environmental effect that promotes the change of land use type, w 3,k,a,b and w 4,k,a,b are the weights of the neighborhood effect and the environmental effect.

[0022] Preferably, the model parameters are determined using the hill climbing method, including:

[0023] Suppose there are N land use types in the research data, and each type is affected by M levels of regional interaction effects. Then the model has M×N weights to be determined;

[0024] First, randomly generate an initial weight combination W, and then increase any one weight element by the step size p while keeping the other weight elements unchanged; by comparing the accuracy kappa i of the land use simulation results under different changed weight combinations, select the weight combination W with the maximum simulation accuracy max(kappa i )max ;

[0025] Then, when max(kappa i ) is greater than the maximum simulation accuracy kappa when none of the weight elements increase the step size p t,m , assign max(kappa i ) and W max to kappa t,m and W m+1 respectively;

[0026] Continue to randomly select a weight element to increase the step size p to simulate land use change, compare the simulation accuracies of different weight combinations, and update the values of kappa t,m and W m+1 until the simulation accuracy max(kappa i ) is less than the maximum simulation accuracy kappa t,m ;

[0027] To avoid local optimal solutions, set multiple iterations to randomly generate different initial weight combinations. In each iteration, compare the maximum simulation accuracy kappa t,m+1 of each iteration with the maximum simulation accuracy kappa t,1:m of the previous iteration result. When kappa t,m+1 is greater than kappa t,1:m , set the weight combination W t,m+1 corresponding to kappa m+1 as the optimal weight combination.

[0028] Based on the same inventive concept, a urban land use evolution system considering multi-level spatial interaction heterogeneity provided by the present invention includes:

[0029] A geographical unit division module, configured to divide a research area into multiple spatial structure units through an urban spatial structure diagram;

[0030] A multi-level spatial interaction effect measurement module, configured to calculate the urbanization indices of zoning units at three levels of between cities, within cities, and local regions, use a gravity model to obtain the spatial field intensity between zoning units at different levels and spatial structure units based on the urbanization indices, and further determine the spatial interaction effects of zoning units at different levels on spatial structure units; and at the local level, calculate the neighborhood effect and environmental effect of plots within spatial structure units;

[0031] CA model construction and simulation module, which is used to comprehensively consider the multi-level spatial interaction effects at different levels and locally, construct a cellular automaton model, where the land use conversion rules are jointly determined by the comprehensive effects weighted by the multi-level spatial interaction effects; use the hill climbing method to determine the weight parameters of the model; and use the constructed cellular automaton model to simulate the land use evolution process.

[0032] Based on the same inventive concept, a computer program product provided by the present invention includes a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of the urban land use evolution method considering multi-level spatial interaction heterogeneity are implemented.

[0033] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0034] 1. By dividing the spatial structure units, the present invention uses the gravity model to measure the spatial field intensity between different-level zoning units and spatial structure units based on the urbanization index, obtains the spatial interaction effects at different levels, and combines the neighborhood effect and environmental effect at the local level of the cellular automaton model to couple multi-level spatial interactions, and finally obtains the comprehensive effect received by each plot, effectively overcoming the problem that the traditional method ignores the influence of multi-level spatial interaction effects and urban spatial structure on land use change, resulting in the limitations and deficiencies of the model. Compared with the traditional model, the MRC-CA model constructed by the present invention considers the urban spatial structure, multi-level spatial interaction effects and environmental factors, and can more accurately capture the characteristics of land use change.

[0035] 2. Combining the basic principle of the cellular automaton model, multi-level spatial interaction effects and urban spatial structure, the present invention designs multi-level land use conversion rules, considering the influence of multi-level spatial interaction effects, neighborhood effects and environmental effects, as well as the differences in their synergistic effects within different structural units (urban area - inner fringe area - outer fringe area - rural hinterland). The research results can effectively reveal the driving mechanisms and characteristics of land use conversion within different spatial structure units.

[0036] 3. Based on the hill climbing method, taking the multi-level spatial interaction effects as input data, the present invention determines the model parameters of different types of spatial structure units. Through multiple iterations, different initial weight combinations are randomly generated, the accuracy of the simulation results under different weight combinations is compared, and the optimal weight combination is selected. The constructed model shows high accuracy and applicability in predicting the evolution pattern. Brief Description of the Drawings

[0037] Figure 1 It is a schematic diagram of the overall method flow of the embodiment of the present invention.

[0038] Figure 2Schematic diagram of the process for determining the model weight parameters in the embodiments of the present invention.

[0039] Figure 3 Example diagram of the land evolution simulation in Jiangyin City in 2017 based on the MRC-CA model. Specific implementation manners

[0040] The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and specific embodiments.

[0041] To solve the problem of multi-level spatial interaction heterogeneity caused by the urban concentric circle spatial structure model, embodiments of the present invention disclose an urban land use evolution method considering multi-level spatial interaction heterogeneity. First, the study area is divided into multiple spatial structure units through the urban spatial structure diagram. For example, each city is a spatial structure unit. Secondly, the urbanization index of different-level zoning units is calculated, and the spatial field intensity between different-level zoning units and spatial structure units is obtained based on the urbanization index using the gravity model, and then the spatial interaction effect of different-level zoning units on spatial structure units is determined. At the local level, the neighborhood effect and environmental effect of the plots within the spatial structure unit are calculated. Thirdly, by comprehensively considering the multi-level spatial interaction effects at different levels and locally (neighborhood effect and environmental effect), an MRC-CA model is constructed (where the land use conversion rules are jointly determined by the comprehensive effect weighted by the multi-level spatial interaction effects), and the weight parameters of the model are determined using the hill climbing method. Finally, the constructed MRC-CA model is used to simulate the land use evolution process. In this embodiment, the evolution simulation of Jiangyin City is taken as an example and compared with the actual data to illustrate the accuracy and applicability of the model. The whole process is as Figure 1 shown, and the detailed steps are as follows:

[0042] S1: Division of the urban spatial structure.

[0043] The present invention uses the Delaunay triangulation method proposed by Yang, J. (2023). Modeling land-use change using partitioned vector cellular automata while considering urban spatial structure. Environment and Planning B: Urban Analytics and City Science, 50(8), 2273-2293. to identify the urban concentric circle spatial structure units (urban area - inner edge area - outer edge area - rural hinterland), and superimpose them on the township-level zoning units of the study area to obtain the spatial structure units for spatial interaction calculation.

[0044] S2: Calculation of urbanization index.

[0045] In this embodiment, taking the night light index, urban population quantity, and construction land area as examples, the urbanization index of different-level zoning units is calculated based on formula (1). This method can reflect the urbanization intensity levels of different spatial units.

[0046]

[0047] Among them, UC i,j is the urbanization index of the j-th zoning unit at the i-th level. The first and second levels (i = 1 or 2) are zoning units within and between cities, while the third level (i = 3) is the zoning unit of a local area; light i,j represents the night light index of the j-th zoning unit at the i-th level; PoP i,j represents the population quantity of the j-th zoning unit at the i-th level; built i,j represents the construction land area of the j-th zoning unit at the i-th level.

[0048] S3: Measurement of spatial interaction effects at different levels.

[0049] The gravity model is a commonly used model to describe the attraction level between two regions. It is directly proportional to the urbanization level and inversely proportional to the spatial distance. The present invention uses the following equation to measure the spatial field intensity between the spatial structure unit and zoning units at different levels, and determine the spatial interaction effects of zoning units at different levels on the spatial structure unit:

[0050]

[0051] Among them, SFI i,j,k represents the spatial field intensity between the j-th zoning unit at the i-th level and the k-th spatial structure unit. D(x i,j ,y i,j ,x k ,y k ) represents the Euclidean distance from the center point (x i,j ,y i,j ) of the j-th zoning unit at the i-th level to the center point (x k ,y k ) of the k-th spatial structure unit. RI i,k represents the interaction effect of the i-th level zoning unit on the k-th spatial structure unit, and m is the number of units at the i-th level. RI′ i,k is the spatial interaction effect of each plot within the k-th spatial structure unit, which is obtained by evenly distributing the spatial interaction effect RI i,k to each plot according to the k-th spatial structure unit. S i,k refers to the area of the k-th spatial structure unit in the i-th level zoning unit. is the normalized value, min(RI′ i,k ), and max(RI′ i,k ) are the minimum and maximum values of the spatial interaction effect RI′ i,k respectively. When i equals 1 or 2, RI′ represents the external spatial interaction effect of all inter-city or intra-city zoning units on the land use plots in the study area.

[0052] At the local level, the traditional cellular automata (CA) model considers neighborhood effect and environmental effect as two important factors to be considered when simulating land use changes. In this embodiment, we set the first-order and second-order adjacent plots as the neighborhood range and apply the research of Yao, Y., Liu, X., Li, X., Liu, P., Hong, Y., Zhang, Y., & Mai, K. (2017). Simulating urban land-use changes at a large scale by integrating dynamic land parcel subdivision and vector-based cellular automata. International Journal of Geographical Information Science, 31(12), 2452 - 2479. to calculate the neighborhood effect (Ω a,b ) of each plot. Meanwhile, based on the data such as elevation, distance from the city center, distance from the county center, distance from the highway, distance from the main road, and distance from the secondary road provided by the research of Shi, F. (2021). Research on vector CA neighborhood calculation and urban growth simulation considering spatial heterogeneity. Nanjing Normal University., the Logistic regression model is used to obtain the environmental effect of each plot:

[0053]

[0054] where E a,b represents the plot conversion probability from land use type a to b under the environmental influence. x l is the independent variable value of the l-th environmental factor of the plot. β l and α are the Logistic regression parameters and regression constant, which are obtained based on the data of seven environmental factors and the land use status data.

[0055] S4: Obtaining land use conversion rules.

[0056] Combining the basic principle of the CA model and the above-mentioned multi-level spatial interaction effects, this study designed an improved land use conversion rule for territorial space:

[0057]

[0058] Among them, P a,b is the conversion probability from land use type a to b, and land use type a is located in the spatial structure unit k. w i,k,a,b is the interaction effect weight of the i-th zoning level, which promotes the change of land use type from a to b. Ω is the neighborhood effect of changing from land use type a to b, and E a,b is the environmental effect that promotes the change of land use type. w a,b and w 3,k,a,b and w 4,k,a,b are the weights of the neighborhood effect and the environmental effect, which promote the change of land use from type a to b. The weight values are all within the range of [0, 1].

[0059] S5: Determination of model parameters.

[0060] Inspired by the hill climbing method, the proposed model uses multi-level spatial interaction effects as input data to determine the model parameters of different types of urban spatial structure units. Taking the urban area as an example, the probabilities of different types of land use types within this area being transformed by the spatial interaction effects at different levels are different, which is determined by the multiple weights of the multi-level interaction effects in the proposed model. Assume that there are n land use types in the research data, and each type is affected by the regional interaction effects at four levels. Therefore, the proposed model has 4 × n weights to be determined. As Figure 2 shown, first, a randomly generated initial weight combination W(4 × n) is generated, and then any one weight element is increased by 0.1 while the other weight elements remain unchanged. By comparing the accuracy (kappa i ) of the land use simulation results under different changed weight combinations, the weight combination (W i ) with the maximum simulation accuracy (max(kappa max )) is selected; then, when max(kappa i ) is greater than the maximum simulation accuracy (kappa t,m ) when none of the weight elements are increased by 0.1, max(kappa i ) and W max are respectively assigned to kappa t,m and W m+1, this is to continue to randomly select a weight element and increase it by 0.1 to simulate land use change, compare the simulation accuracies of different weight combinations, and update kappa t,m and W m+1 values until the simulation accuracy max(kappa i ) is less than the maximum simulation accuracy (kappa t,m ); To avoid the problem that the weight combination found by the hill climbing method is a local optimal solution, we set 10,000 iterations to randomly generate different initial weight combinations W(4×n). In each iteration, the maximum simulation accuracy kappa t,m+1 of each iteration is compared with the maximum simulation accuracy kappa t,1:m of the previous iteration result. When kappa t,m+1 is greater than kappa t,1:m , the weight combination W t,m+1 corresponding to kappa m+1 is set as the optimal weight combination of the land use conversion rule in the urban area.

[0061] S6: Evaluation of model accuracy.

[0062] By combining parameters and the model, set the night light index, urban population, and construction land area of Jiangyin City in 2012 as input data to predict the land use pattern of Jiangyin City in 2017( Figure 3 ), and compare it with the real land use data in 2017 using indicators such as the Kappa coefficient, overall accuracy (OA), and FoM (Figure of Merit) to analyze the accuracy and applicability of the model.

[0063] Table 1 Simulation accuracy of the MRC-CA model

[0064]

[0065] As can be seen from Table 1, the overall accuracy of the simulation results of the MRC-CA model exceeds 0.85 (Kappa and OA indicators), and the FoM accuracy reaches 0.17. Comparatively speaking, the simulation effects in the outer edge area and rural hinterland are better, which further confirms the important driving role of multi-level spatial interaction in the expansion of land use in the national territory space.

[0066] Based on the same inventive concept, an embodiment of the present invention also discloses an urban land use evolution system considering multi-level spatial interaction heterogeneity, including: a geographical unit division module for dividing a research area into multiple spatial structure units through an urban spatial structure diagram; a multi-level spatial interaction effect measurement module for calculating the urbanization indices of zoning units at three levels, namely between cities, within cities, and local regions, obtaining the spatial field intensity between zoning units at different levels and spatial structure units based on the urbanization indices using a gravity model, and further determining the spatial interaction effect of zoning units at different levels on spatial structure units; and at the local level, calculating the neighborhood effect and environmental effect of plots within spatial structure units; a CA model construction and simulation module for comprehensively considering the multi-level spatial interaction effects at different levels and locally, constructing a cellular automaton model, where the land use conversion rules are jointly determined by the comprehensive effect weighted by the multi-level spatial interaction effects; determining the weight parameters of the model using the hill climbing method; and simulating the land use evolution process using the constructed cellular automaton model.

[0067] Based on the same inventive concept, an embodiment of the present invention also discloses a computer program product, including computer programs / instructions, which when executed by a processor implement the steps of the urban land use evolution method considering multi-level spatial interaction heterogeneity as described above.

Claims

1. A method for urban land use evolution taking into account multi-level spatial interactive heterogeneity, characterized in that: The steps include: The study area is divided into multiple spatial structural units through the urban spatial structure map; Calculate the urbanization index of the three levels of zoning units, namely, between cities, within cities, and local areas. The urbanization index is calculated by integrating the indicators of night light index, urban population, and construction land area. Use the gravity model to obtain the spatial field intensity between zoning units at different levels and spatial structure units based on the urbanization index according to the following formula, and then determine the spatial interaction effect of zoning units at different levels on spatial structure units. Among them, i = 1 or 2 represents the level of division between cities and within cities, SFI i,j,k represents the spatial field strength between the i-th level division unit j and the spatial structure unit k, UC i,j is the urbanization index of the i-th level division unit j, UC 3,k is the urbanization index of the spatial structural unit k in the local area, D(x i,j ,y i,j ,x k ,y k ) represents the center point of the i-th level division unit j (x i,j ,y i,j ) to the center point (x k ,y k )’s Euclidean distance; RI i,k represents the interaction effect of all zoning units at level i on spatial structure unit k, m is the number of zoning units at level i; RI′ i,k is the spatial interaction effect of each plot within the spatial structural unit k, S i,k is the area of ​​the spatial structural unit k of the i-th level division; is the normalized value; At the local level, the neighborhood and environmental effects of the blocks within the spatial structural unit are calculated; Taking into account the spatial interaction effects at different levels and the local neighborhood effects and environmental effects, a cellular automaton model is constructed, in which the land use conversion rules are determined by the comprehensive effects obtained by weighting the spatial interaction effects, neighborhood effects and environmental effects. The weight parameters of the model are determined using the hill climbing method; The constructed cellular automaton model is used to simulate the land use evolution process.

2. The urban land use evolution method taking into account multi-level spatial interactive heterogeneity according to claim 1, characterized in that: The specific calculation formula of the urbanization index is as follows: Among them, UC i,j is the urbanization index of the i-th level division unit j; light i,j 、PoP i,j 、built i,j They respectively represent the night light index, population and construction land area of ​​the i-th level zoning unit j.

3. The urban land evolution method taking into account multi-level spatial interactive heterogeneity according to claim 1, characterized in that: When calculating the neighborhood effect of a plot, the first-order and second-order neighboring plots are set as the neighborhood range; when calculating the environmental effect of a plot, the elevation of the plot, the distance from the center of different levels of zoning units, and the distance from highways, main roads, and secondary roads are considered, and the environmental effect of each plot is obtained using a logistic regression model.

4. The urban land use evolution method taking into account multi-level spatial interactive heterogeneity according to claim 1, characterized in that: The land use conversion rule in the cellular automaton model is expressed as: Among them, P a,b is the transition probability from land use type a to land use type b, land use type a is located in spatial structure unit k, w i,k,a,b is the interaction effect at the ith zoning level The weight of a,b is the neighborhood effect of changing from land use type a to b, E a,b is the environmental effect that drives the change in land use type, w 3,k,a,b and w 4,k,a,b is the weight of neighborhood effect and environmental effect.

5. The urban land evolution method taking into account multi-level spatial interactive heterogeneity according to claim 1 is characterized in that: The method of determining the model parameters by using the hill climbing method comprises: Assuming that there are N land use types in the research data, and each type is affected by M levels of regional interaction effects, the model has M×N weights to be determined; First, an initialized weight combination W is randomly generated, and then any weight element is increased by step size p, while other weight elements remain unchanged; the accuracy kappa of land use simulation results under different weight combinations is compared i , select the one with the maximum simulation accuracy max(kappa i ) is the weight combination W max ; Then, when max(kappa i ) is greater than the maximum simulation accuracy kappa when all weight elements do not increase the step size p t,m When max(kappa i ) and W max Assign kappa respectively t,m and W m+1 ; Continue to randomly select a weight element and increase the step size p to simulate land use change, compare the simulation accuracy of different weight combinations, and update kappa t,m and W m+1 The value of , until the simulation accuracy is max(kappa i ) is less than the maximum simulation accuracy kappa t,m ; To avoid local optimal solutions, multiple iterations are set to randomly generate different initialization weight combinations. In each iteration, the maximum simulation accuracy kappa of each iteration is t,m+1 The maximum simulation accuracy kappa compared with the previous iteration result t,1:m When comparing kappa t,m+1 Greater than kappa t,1:m When t,m+1 The weight combination W m+1 Set as the optimal weight combination.

6. An urban land use evolution system that takes into account multi-level spatial interactive heterogeneity, characterized by: include: Geographic unit division module, which is used to divide the study area into multiple spatial structure units through the urban spatial structure map; The multi-level spatial interaction effect measurement module is used to calculate the urbanization index of the three-level zoning units between cities, within cities, and local areas. The urbanization index is calculated by integrating the indicators of night light index, urban population, and construction land area; the gravity model is used to obtain the spatial field intensity between the zoning units of different levels and the spatial structure units based on the urbanization index according to the following formula, and then determine the spatial interaction effect of the zoning units of different levels on the spatial structure units; Among them, i = 1 or 2 represents the level of division between cities and within cities, SFI i,j,k represents the spatial field strength between the i-th level division unit j and the spatial structure unit k, UC i,j is the urbanization index of the i-th level division unit j, UC 3,k is the urbanization index of the spatial structural unit k in the local area, D(x i,j ,y i,j ,x k ,y k ) represents the center point of the i-th level division unit j (x i,j ,y i,j ) to the center point (x k ,y k )’s Euclidean distance; RI i,k represents the interaction effect of all zoning units at level i on spatial structure unit k, m is the number of zoning units at level i; RI′ i,k is the spatial interaction effect of each plot within the spatial structural unit k, S i,k is the area of ​​the spatial structural unit k of the i-th level division; is the normalized value; and, at the local level, the neighborhood and environmental effects of plots within the spatial structural unit are calculated; The CA model construction and simulation module is used to comprehensively consider the spatial interaction effects at different levels and the local neighborhood effects and environmental effects, and to construct a cellular automaton model, in which the land use conversion rules are determined by the comprehensive effects obtained by weighting the spatial interaction effects, neighborhood effects and environmental effects; the weight parameters of the model are determined by the hill climbing method; and the constructed cellular automaton model is used to simulate the land use evolution process.

7. The urban land evolution system taking into account multi-level spatial interactive heterogeneity according to claim 6 is characterized in that: The land use conversion rule in the cellular automaton model is expressed as: Among them, P a,b is the transition probability from land use type a to b, land use type a is located in spatial structure unit k, w i,k,a,b is the interaction effect at the ith zoning level The weight of a,b is the neighborhood effect of changing from land use type a to b, E a,b is the environmental effect that drives the change in land use type, w 3,k,a,b and w 4,k,a,b is the weight of neighborhood effect and environmental effect.

8. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the urban land evolution method taking into account multi-level spatial interactive heterogeneity according to any one of claims 1 to 5 are implemented.

Images