Urban land evolution vector cellular automaton simulation method based on geographic process partition

Abstract

The invention discloses an urban land evolution vector cellular automaton simulation method based on geographic process partitioning. The method comprises the following steps: acquiring a vector neighborhood index, expressing a land interaction rule based on a distance attenuation function, and constructing a distance attenuation sequence curve by taking a neighborhood order as an independent variable and the vector neighborhood index as a dependent variable; calculating a similarity measurement value of the distance attenuation sequence curve by using the improved dynamic time warping distance function; constructing a minimum spanning tree, and performing hierarchical division on the distance attenuation sequence curve by using the minimum spanning tree to complete clustering partition of the urban land; and constructing a vector cellular automaton model based on geographic process partition, and realizing urban land evolution by using the model. According to the method, the dependence of traditional geographic partition on space static attribute data is broken through, the consistency of the evolution process of all space variables in each region is ensured, the dynamic evolution of the land use pattern is reflected more truly, and a more accurate simulation tool is provided for urban space evolution research.

Description

Technical Field

[0001] The present invention relates to the field of earth science technology, and in particular to a vector cellular automaton simulation method for urban land evolution based on geographical process zoning. Background Art

[0002] CA (Cellular Automata) has become an important tool for urban spatial development modeling due to its simplicity, flexibility, and ability to intuitively reflect the spatiotemporal evolution of land use. However, traditional CA models use a spatial evolution mechanism with unified transformation rules, which makes it difficult to depict the inherent spatial heterogeneity of geographical phenomena, resulting in limited simulation accuracy and possible deviations. This heterogeneity is essentially manifested in the agglomeration or exclusion distribution characteristics of geographical elements in space. In order to effectively characterize such spatial differentiation laws, scholars divide the study area into homogeneous sub-areas through geographic zoning methods and construct a zoning CA model based on this. That is, differentiated transformation rules are formulated for different sub-areas and independent simulation systems are established. This spatial differentiation modeling strategy significantly improves the accuracy of urban expansion simulation.

[0003] As a practical manifestation of the laws of spatial dependence and heterogeneity in geography, geographic zoning is a core method for revealing the spatial differentiation patterns of geographical phenomena and delineating geographic units. By aggregating similar adjacent units to form homogeneous regions and separating heterogeneous units to construct boundaries, it effectively simplifies spatial complexity and enhances interpretability. Traditional zoning methods rely on multivariate similarity analysis to identify regional structure, but ignore the dynamic nature of geographic variables. Existing zoning methods based on static boundaries, such as administrative divisions, are suitable for slowly changing phenomena such as landforms but struggle to address dynamic processes such as urban expansion and population migration. Current research often employs spatially constrained clustering algorithms, such as K-means and SOM (Self-Organizing Maps). While these algorithms can reveal regional differences, they fail to fully consider the time-varying characteristics of attribute variables, resulting in zoning results limited to static spatial differentiation and difficulty reflecting the dynamic differences in land use evolution. Summary of the Invention

[0004] The purpose of the present invention is to provide a vector cellular automaton simulation method for urban land use evolution based on geographic process zoning. Based on the theoretical foundations of urban geography and spatial measurement methods, it comprehensively applies relevant technologies such as spatial interaction laws and spatial clustering zoning to aggregate geographic zoning that takes into account the characteristics of spatial distribution differences and change trend similarities. It breaks through the traditional geographic indicator zoning's reliance on spatial static attribute data and realizes the simulation and analysis of urban land use changes from the perspective of dynamic geographic processes.

[0005] The present invention adopts the following technical solution: a vector cellular automaton simulation method for urban land evolution based on geographical process zoning, comprising the following steps:

[0006] S1. Obtain the vector neighborhood index, express the land use interaction law based on the distance decay function, and construct a distance decay sequence curve with the neighborhood order as the independent variable and the vector neighborhood index as the dependent variable.

[0007] S2. Using the improved dynamic time warping distance function, calculate the similarity measure of the distance decay sequence curve.

[0008] S3. Construct a minimum spanning tree and use it to hierarchically divide the distance decay sequence curve to complete the clustering and partitioning of urban land.

[0009] S4. Construct a vector cellular automaton model based on geographic process zoning and use this model to realize urban land evolution.

[0010] Furthermore, in step S1, the vector neighborhood index is obtained by counting the area of ​​the irregular vector blocks. The specific expression is:

[0011]

[0012] Among them, vNI i,k,r The vector neighborhood index of the kth land use type within the i-th plot with a radius of r, a i,k,r represents the total area of ​​the kth land use type within the i-th plot with a radius of r, a i,r It represents the total area of ​​all land use types within the i-th plot with a radius of r.

[0013] Furthermore, in step S2, a constraint window is added to the dynamic time warping distance function to obtain an improved dynamic time warping distance function.

[0014] The calculation formula for the similarity measure of the distance decay sequence curve is:

[0015]

[0016] Among them, DTW() represents the similarity measure of two distance decay sequence curves, dist() represents the Euclidean distance, r c Indicates the data of the cth point in the rth distance decay sequence curve, u d Represents the d-th point data in the u-th distance decay sequence curve, and α represents the size of the constraint window.

[0017] Furthermore, in step S3, the clustering and partitioning of urban land includes the following:

[0018] The distance decay sequence curves are used as tree nodes, and the spatial relationship between the distance decay sequence curves and the similarity measurement value are used to define the edge weights, resulting in a connected graph G, G = (v, l, w), where v represents the tree node set, l represents the edge set, and w represents the edge weight.

[0019] Select the first spanning tree with only one tree node v1 and compare the two edges l connected to the tree node a and l b The similarity measure of l a The similarity measure is less than l b Similarity measure value, then retain l a , l a Connect with two tree nodes v1 and v2 to get the second spanning tree, complete the first iterative edge selection operation, and compare the three edges l connected to v1 and v2 in the second spanning tree. b 、l c and l d The similarity measure of l d The similarity measure value of is the smallest, then retain l d , l d Connect with the three tree nodes v1, v2 and v4 to obtain the third spanning tree, complete the second iterative edge selection operation, and continue the iterative edge selection operation based on the third spanning tree until the set number of iterations is reached; the edges retained in each iterative edge selection operation and the tree nodes connected to the edges constitute the minimum spanning tree.

[0020] Compare the similarity metrics of all edges in the minimum spanning tree, select the edge with the smallest similarity metric value, cut off the edge, and obtain two new spanning trees T1 and T2, completing the first iterative pruning operation. Compare the similarity metrics of all edges in T1 and T2, select the edge with the smallest similarity metric value, cut off the edge, and obtain three new spanning trees T1, T2, and T3, completing the second iterative pruning operation. Continue the iterative pruning operation based on the three new spanning trees T1, T2, and T3 until the set number of pruning times is reached, and the clustering partitioning of urban land is completed.

[0021] Furthermore, the set pruning times = the number of urban land divisions - 1.

[0022] Furthermore, in step S4, the urban land evolution includes the following:

[0023] The vector cellular automaton model based on geographic process partitioning includes cellular units, and the cellular units include irregular plots in vector land use data.

[0024] The transition probability P of the cell unit includes the overall development probability Pg, the constraint factor Pc, the neighborhood effect Ω and the random factor RA; the transition probability P of the i-th plot changing to the k-th land use type at the t-th time isi k,t for:

[0025]

[0026] in, represents the overall development probability of the i-th plot being transformed into the k-th land use type at the t-th time; represents the neighborhood effect of the jth plot on the ith plot at the tth time; Represents the constraint factor of the i-th plot at the t-th time; RA=1+(-lnγ) α , γ represents a random number, and α represents a parameter that controls randomness.

[0027] The random forest regression model is used to obtain the overall development probability. The specific expression is:

[0028]

[0029] Where I() represents the indicator function of the decision tree set, M represents the total number of decision trees, x represents the high-dimensional vector composed of auxiliary spatial variables in the plot, and h n (x) represents the prediction type of the n-th decision tree of x, Y k Represents the label of the k-th land use type.

[0030] The expression of the neighborhood effect of the jth plot on the ith plot at the tth time is:

[0031]

[0032] Where, e represents the exponential constant, d i,j represents the center distance between the i-th plot and the j-th plot, d represents the buffer radius centered on the i-th plot, S i represents the area of ​​the i-th plot, S j represents the area of ​​the jth plot, S max represents the maximum area of ​​the plot in the study area, S min Indicates the minimum area of ​​a parcel in the study area.

[0033] Neighborhood effect of the kth land use type on the ith plot at time t The expression is:

[0034]

[0035] in, represents the neighborhood effect of the jth plot on the ith plot in the kth land use type at the tth time, and buffer_d and No River between i and j means that the buffer radius is d and there is no river crossing between the ith plot and the jth plot.

[0036] If the restricted development zone is set, the constraint factor of the i-th plot at the t-th time is expressed as:

[0037]

[0038] Among them, H i Indicates the development suitability status of the i-th plot.

[0039] The transition probability of each plot of land into different land use types is obtained, and the conversion direction with the highest probability and exceeding the development threshold is selected to complete the urban land evolution.

[0040] Furthermore, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, it implements the steps of the vector cellular automaton simulation method for urban land evolution based on geographic process zoning.

[0041] Furthermore, the present invention also proposes a computer-readable storage medium, which stores a computer program. When the computer program is run by a processor, it executes the urban land evolution vector cellular automaton simulation method based on geographic process zoning.

[0042] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0043] The present invention incorporates the similarity characteristics of the changing trends in the process of land use expansion into geographic zoning, combines the difference and similarity characteristics, and expands the similarity measurement from the static state space to the dynamic process space. It breaks through the traditional geographic zoning's reliance on spatial static attribute data, successfully divides the geographical boundaries of heterogeneous changing processes, ensures the consistency of the evolution process of all spatial variables in each region, and more realistically reflects the dynamic evolution of land use patterns.

[0044] The present invention realizes the dynamic characterization of land use evolution trend, which not only ensures the consistency of the evolution process of spatial variables in each partition, but also effectively overcomes the limitation of traditional methods in reflecting the dynamic characteristics of land use evolution.

[0045] The present invention significantly improves the accuracy of urban expansion simulation in multiple dimensions such as attribute status, expansion form and spatial distribution, providing a more accurate simulation tool for urban spatial evolution research. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is an overall implementation flow chart of the present invention.

[0047] Figure 2 It is a schematic diagram of constructing a minimum spanning tree according to the present invention.

[0048] Figure 3 It is a schematic diagram of the urban land clustering zoning according to the present invention.

[0049] Figure 4 It is a schematic diagram of realizing the evolution of urban land use in an embodiment of the present invention.

[0050] Figure 5 This is a result diagram of the distance decay sequence curve constructed in an embodiment of the present invention.

[0051] Figure 6 This is a clustering and zoning result diagram of urban land in an embodiment of the present invention.

[0052] Figure 7 This is a result diagram of urban land evolution in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0054] Geographic zoning is a fundamental method for revealing the proximal similarities and regional heterogeneity of geographical phenomena. By comparing geographic state zoning based on similarity in indicator characteristics with geographic process zoning based on similarity in change processes, we can tease out the differences and commonalities between the two, revealing the differences in spatial distribution and similarities in change trends during land use expansion. Based on characterizing the differences in land use interaction intensity across regions, we construct similarity metrics to quantify the similarities in change trends during land use evolution, laying the foundation for achieving geographic process zoning aggregation.

[0055] To achieve the above objectives, the present invention proposes a vector cellular automaton simulation method for urban land evolution based on geographic process zoning. By designing similarity measurement indicators for land evolution processes, it quantitatively analyzes local mutations and global evolution laws in urban expansion, and realizes the coordinated expression of spatial distribution differences and similarity of change trends. A VCA (Vector Cellular Automata) model based on geographic process zoning is established, providing a new dynamic perspective for urban land evolution simulation. Figure 1 The specific steps are as follows:

[0056] S1. Taking Jiangyin City as an example, the vector neighborhood index is obtained, and the land use interaction law is expressed based on the distance decay function. A distance decay sequence curve is constructed with the neighborhood order as the independent variable and the vector neighborhood index as the dependent variable. This curve reflects the effect of land use units within a certain range, that is, the vector neighborhood index will gradually decay with the change of distance, corresponding to the dynamic change process in the geographical process zoning, and verifies the existence of spatial heterogeneity.

[0057] The vector neighborhood index is obtained by counting the area of ​​irregular vector plots. The specific expression is:

[0058]

[0059] Among them, vNI i,k,r The vector neighborhood index of the kth land use type within the i-th plot with a radius of r, a i,k,r represents the total area of ​​the kth land use type within the i-th plot with a radius of r, a i,r It represents the total area of ​​all land use types within the i-th plot with a radius of r.

[0060] S2. Using the improved dynamic time warping distance function, we calculated the similarity metric of the distance decay sequence curve and verified the similarity characteristics of the changing trend during the land expansion process. The specific contents are as follows:

[0061] The distance dimension in a distance decay sequence curve is actually represented by the order of the neighborhood rather than the absolute distance. While a certain degree of stretching or displacement has little impact on model analysis, excessive stretching or displacement can lead to semantic errors in the final matching result. Therefore, by adding a constraint window to the dynamic time warping distance function, we obtain an improved dynamic time warping distance function. This not only addresses the curve deformation issue, but also reduces computational effort and improves algorithm efficiency.

[0062] The calculation formula for the similarity measure of the distance decay sequence curve is:

[0063]

[0064] Among them, DTW() represents the similarity measure of two distance decay sequence curves, dist() represents the Euclidean distance, r c Indicates the data of the cth point in the rth distance decay sequence curve, u d Represents the d-th point data in the u-th distance decay sequence curve, and α represents the size of the constraint window.

[0065] The constraint window can be understood as the subscript constraint of the matching point, that is, the subscript distance between the current point and the matching point cannot exceed α.

[0066] S3. Combining the spatial heterogeneity and similarity of the change trends in the above land expansion process, we construct an MST (Minimum Spanning Tree) and use it to hierarchically divide the distance decay sequence curve from top to bottom to complete the clustering and zoning of urban land. The specific contents are as follows:

[0067] like Figure 2 As shown in the figure, distance decay sequence curves are used as tree nodes (i.e., blue circles in the figure). The spatial relationship between distance decay sequence curves and similarity metrics are used to define edge weights, resulting in a connected graph G, G = (v, l, w), where v represents the set of tree nodes, l represents the set of edges, and w represents the edge weights. The MST is a minimum cost spanning tree, where the cost is measured by summing the similarity metrics of all edges. The dashed lines in the figure represent edges in the connected graph, and the solid lines represent edges in the MST.

[0068] Select the first spanning tree with only one tree node v1 and compare the two edges l connected to the tree node a and l b The similarity measure of l a The similarity measure is less than l b Similarity measure value, then retain l a , l a Connect with two tree nodes v1 and v2 to get the second spanning tree, complete the first iterative edge selection operation, and compare the three edges l connected to v1 and v2 in the second spanning tree. b 、l c and l d The similarity measure of l d The similarity measure value of is the smallest, then retain l d , l d Connect with three tree nodes v1, v2 and v4 to get the third spanning tree, complete the second iterative edge selection operation, and compare the three edges l connected to v1, v2 and v4 in the third spanning tree. b 、l c and l e The similarity measure of l c The similarity measure value of is the smallest, then retain l c , l c Connecting to the four tree nodes v1, v2, v3, and v4, we obtain the fourth spanning tree, completing the third iterative edge selection operation. We continue iterating edge selection based on the fourth spanning tree until the set number of iterations is reached. The edges retained in each iterative edge selection operation and the tree nodes connected to them form the MST, which serves as the basis for region partitioning.

[0069] like Figure 3As shown in the figure, after constructing the MST, it is necessary to decompose the MST into a set of continuous subgraphs through pruning. This is equivalent to converting the region partitioning problem into a graph partitioning problem. To partition the p distance decay sequence curves into q regions, (n-1) edges are removed from the MST. Each resulting subgraph will be a tree with all vertices connected and no loops. Initially, all distance decay sequence curves belong to the same tree. When edges are removed from the original MST, a set of disconnected trees will emerge, and each tree in this set has a unique corresponding relationship with a region.

[0070] Compare the similarity metrics of all edges in the MST, select the edge with the smallest similarity metric value, cut off the edge, and obtain two new spanning trees T1 and T2, completing the first iterative pruning operation. Compare the similarity metrics of all edges in T1 and T2, select the edge with the smallest similarity metric value, cut off the edge, and obtain three new spanning trees T1, T2, and T3, completing the second iterative pruning operation. Continue the iterative pruning operation based on the three new spanning trees T1, T2, and T3 until the set number of pruning times is reached, and the clustering partitioning of urban land is completed.

[0071] The set number of pruning times = the number of urban land zones - 1.

[0072] Figure 3 The objective function in is the similarity measure. The smaller the objective function, the lower the similarity. The weakest point represents the most obvious heterogeneity in the geographical attributes of the two regions it connects. The weakest connection point is disconnected to achieve regional division.

[0073] S4. After obtaining dynamic geographic zoning that takes into account the differences in spatial distribution and the similarities in change trends, a VCA model based on geographic process zoning is constructed in each area, and the model is used to realize the evolution of urban land use. The specific contents are:

[0074] like Figure 4 As shown, the vector cellular automaton model based on geographic process partitioning includes cellular units, and the cellular units include irregular plots in vector land use data.

[0075] The transition probability P of the cell unit includes the overall development probability Pg, the constraint factor Pc, the neighborhood effect Ω and the random factor RA; the transition probability P of the i-th plot changing to the k-th land use type at the t-th time is i k,t for:

[0076]

[0077] in, represents the overall development probability of the i-th plot being transformed into the k-th land use type at the t-th time; represents the neighborhood effect of the jth plot on the ith plot at the tth time; Represents the constraint factor of the i-th plot at the t-th time; the factors affecting land use change are extremely complex and highly random, RA=1+(-lnγ) α , γ represents a random number between 0 and 1, and α represents a parameter that controls randomness, with a value between 1 and 10.

[0078] The random forest regression model is used to obtain the overall development probability. The specific expression is:

[0079]

[0080] Where I() represents the indicator function of the decision tree set; M represents the total number of decision trees; x represents the high-dimensional vector composed of auxiliary spatial variables in the plot; h n (x) represents the prediction type of the n-th decision tree of x; Y k represents the label of the k-th land use type, which is the land use type result of each decision tree for the i-th plot conversion.

[0081] The expression of the neighborhood effect of the jth plot on the ith plot at the tth time is:

[0082]

[0083] Where, e represents the exponential constant, d i,j represents the center distance between the i-th plot and the j-th plot, d represents the buffer radius centered on the i-th plot, S i represents the area of ​​the i-th plot, S j represents the area of ​​the jth plot, S max represents the maximum area of ​​the plot in the study area, S min Indicates the minimum area of ​​a parcel in the study area.

[0084] Neighborhood effect of the kth land use type on the ith plot at time t The expression is:

[0085]

[0086] in, represents the neighborhood effect of the jth plot on the ith plot in the kth land use type at the tth time, and buffer_d and No River between i and j means that the buffer radius is d and there is no river crossing between the ith plot and the jth plot.

[0087] Constraints refer to specific land use types that cannot be transformed into other land use types during the simulation process. In this example, water factors, traffic factors, and ecological redline areas are used as restricted development areas. The constraint factor of the i-th plot at time t is expressed as:

[0088]

[0089] Among them, H i Indicates the development suitability status of the i-th plot.

[0090] The transition probability of each plot of land into different land use types is obtained, and the conversion direction with the highest probability and exceeding the development threshold is selected to complete the urban land evolution.

[0091] To further explore the role of trend similarity in modeling geographic process zoning, we used land use vector data from two periods of the study area. The land use data covers 2012 and 2017. The 2012 land use data served as training data for the model calibration phase, used to obtain model parameters, while the 2017 land use data served as reference data for evaluating model accuracy during the application phase. Both data are core and fundamental to model operation. Land use driver data, including slope and aspect data, road classification data, urban development pattern data, vertical grain data, GDP (Gross Domestic Product), and POI (Point of Interest) data, are crucial for model operation. Some of these data serve as the basis for regional differentiation, while others represent drivers of urban growth, playing an irreplaceable role in model operation. By adjusting these driver data and using a random forest algorithm to train the overall adaptability probability of the study area, we constructed a VCA simulation model that accounts for geographic process zoning. The simulation results were systematically analyzed from the perspectives of accuracy assessment, landscape pattern, and urban expansion patterns.

[0092] Combining the individual morphology and system structure of urban plots, we evaluated the accuracy of the simulation results of a VCA model based on geographic process zoning from the perspectives of Figure of Merit (FoM), urban expansion patterns, and spatial distribution. We also compared the model with the classic VCA model in terms of zoning results, determination of the influence range of distance decay, and conversion rules. The evaluation metrics are shown in Table 1. A FoM above 0.2 is considered a good simulation result.

[0093] Table 1 Evaluation index system

[0094]

[0095] Figure 5(a) is the clustering result diagram of the three distance decay sequence curves. The gray lines represent all the distance decay sequence curves in the study area. The blue, orange and green curves represent the clustering results of the three distance decay sequence curves, respectively, indicating that the distance decay sequence curves in the study area can be clustered into the neighborhood exponential decay law represented by these three curves.

[0096] Figure 5 (b) is the clustering result diagram of the four distance decay sequence curves. The gray lines represent all the distance decay sequence curves in the study area, and the blue, orange, green and red curves represent the clustering results of the four distance decay sequence curves, respectively, indicating that the distance decay sequence curves in the study area can be clustered into the neighborhood exponential decay law represented by these four curves.

[0097] Figure 5 (c) is the clustering result diagram of the six distance decay sequence curves. The gray lines represent all the distance decay sequence curves in the study area. The blue, orange, green, red, purple and brown curves represent the clustering results of the six distance decay sequence curves, respectively, indicating that the distance decay sequence in the study area can be clustered into the neighborhood exponential decay law represented by these six curves.

[0098] Figure 6 (a) corresponds to Figure 5 The clustering results of (a) generate the clustering zoning result diagram of three urban lands.

[0099] Figure 6 (b) corresponds to Figure 5 (b) The clustering results of the four urban land use clustering zoning results map.

[0100] Figure 6 (c) corresponds to Figure 5 (c) The clustering results of the six urban land use clustering zoning results map.

[0101] Figure 7 (a) is the optimal partition simulation result diagram obtained by comparing the FoM values ​​of different partition results. Through comparison, it is found that when four geographical partitions are obtained, the simulation accuracy of each partition is relatively high. The specific results are shown in Table 2.

[0102] Table 2 Simulation accuracy of partitions

[0103]

[0104]

[0105] Although the Kappa coefficient of the four geographical divisions is not as high as that of the six geographical divisions, the FoM value of each district is relatively average and is higher than the FoM values ​​of other Kappa coefficients. Therefore, the simulation results of the four geographical divisions are selected as the final simulation results. Figure 7 (b) to (e) are the simulation results of each partition.

[0106] Land use evolution itself is a typical spatiotemporal dynamic process, and exhibits significant spatial heterogeneity. The change patterns in different regions may be completely different. Using the PGR (Process-oriented geographical regionalization) model, the evolution trends of dynamically changing geographical processes can be measured by similarity and regional division can be carried out. This geographical division, which takes into account regional differences, driving factors and dynamic geographical processes, can capture the dynamic spatial differentiation laws of land use conversion while taking into account the spatial heterogeneity of land use distribution, facilitate more accurate expression of land use interaction mechanisms, obtain differentiated conversion rules, and achieve land use change simulation with higher accuracy and applicability.

[0107] An embodiment of the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable by the processor. It should be noted that when the processor executes the computer program, it corresponds to the specific steps of the method provided in the embodiment of the present invention and has the corresponding functional modules and beneficial effects of the method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present invention.

[0108] The present invention also provides a computer-readable storage medium storing a computer program. It should be noted that when executed by a processor, the computer program corresponds to the specific steps of the method provided in the present invention and has the corresponding functional modules and beneficial effects. For technical details not fully described in this embodiment, please refer to the method provided in the present invention.

[0109] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A vector cellular automaton simulation method for urban land evolution based on geographical process zoning, characterized by: include: S1. Obtain the vector neighborhood index, express the land use interaction law based on the distance decay function, and construct a distance decay sequence curve with the neighborhood order as the independent variable and the vector neighborhood index as the dependent variable; S2. Calculate the similarity metric of the distance decay sequence curve using the improved dynamic time warping distance function; S3. Construct a minimum spanning tree and use it to perform hierarchical division on the distance decay sequence curve to complete the clustering and partitioning of urban land. S4. Construct a vector cellular automaton model based on geographic process zoning and use this model to realize urban land evolution.

2. The urban land evolution vector cellular automaton simulation method based on geographic process zoning according to claim 1 is characterized in that: In step S1, the vector neighborhood index is obtained by counting the area of ​​the irregular vector blocks. The specific expression is: Among them, vNI i,k,r The vector neighborhood index of the kth land use type within the i-th plot with a radius of r, a i,k,r represents the total area of ​​the kth land use type within the i-th plot with a radius of r, a i,r It represents the total area of ​​all land use types within the i-th plot with a radius of r.

3. The method for simulating urban land use evolution based on geographic process zoning using vector cellular automata according to claim 1 is characterized in that: In step S2, a constraint window is added to the dynamic time warping distance function to obtain an improved dynamic time warping distance function; The calculation formula for the similarity measure of the distance decay sequence curve is: Among them, DTW() represents the similarity measure of two distance decay sequence curves, dist() represents the Euclidean distance, r c Indicates the data of the cth point in the rth distance decay sequence curve, u d Represents the d-th point data in the u-th distance decay sequence curve, and α represents the size of the constraint window.

4. The method for simulating urban land use evolution based on geographic process zoning using vector cellular automata according to claim 1 is characterized in that: In step S3, the clustering and partitioning of urban land includes the following: The distance decay sequence curves are used as tree nodes, and the spatial relationship between the distance decay sequence curves and the similarity metric are used to define the edge weights. The connected graph G is obtained, G = (v, l, w), where v represents the tree node set, l represents the edge set, and w represents the edge weight. Select the first spanning tree with only one tree node v1 and compare the two edges l connected to the tree node a and l b The similarity measure of l a The similarity measure is less than l b Similarity measure value, then retain l a , l a Connect with two tree nodes v1 and v2 to get the second spanning tree, complete the first iterative edge selection operation, and compare the three edges l connected to v1 and v2 in the second spanning tree. b 、l c and l d The similarity measure of l d The similarity measure value of is the smallest, then retain l d , l d Connect with the three tree nodes v1, v2, and v4 to obtain the third spanning tree, complete the second iterative edge selection operation, and continue the iterative edge selection operation based on the third spanning tree until the set number of iterations is reached; The edges retained in each iterative edge selection operation and the tree nodes connected to the edges constitute the minimum spanning tree; Compare the similarity metrics of all edges in the minimum spanning tree, select the edge with the smallest similarity metric value, cut off the edge, and obtain two new spanning trees T1 and T2, completing the first iterative pruning operation. Compare the similarity metrics of all edges in T1 and T2, select the edge with the smallest similarity metric value, cut off the edge, and obtain three new spanning trees T1, T2, and T3, completing the second iterative pruning operation. Continue the iterative pruning operation based on the three new spanning trees T1, T2, and T3 until the set number of pruning times is reached, and the clustering partitioning of urban land is completed.

5. The method for simulating urban land use evolution based on geographic process zoning using vector cellular automata according to claim 4 is characterized in that: The set number of pruning times = the number of urban land zones - 1.

6. The method for simulating urban land use evolution based on geographic process zoning using vector cellular automata according to claim 1 is characterized in that: In step S4, urban land evolution includes the following: The vector cellular automaton model based on geographic process partitioning includes cellular units, and the cellular units include irregular plots in vector land use data; The transition probability P of the cell unit includes the overall development probability Pg, the constraint factor Pc, the neighborhood effect Ω and the random factor RA; the transition probability of the i-th plot changing to the k-th land use type at the t-th time is for: in, represents the overall development probability of the i-th plot being transformed into the k-th land use type at the t-th time; represents the neighborhood effect of the jth plot on the ith plot at the tth time; Represents the constraint factor of the i-th plot at the t-th time; RA=1+(-lnγ) α ,γ represents a random number, α represents a parameter that controls randomness; The random forest regression model is used to obtain the overall development probability. The specific expression is: Where I() represents the indicator function of the decision tree set, M represents the total number of decision trees, x represents the high-dimensional vector composed of auxiliary spatial variables in the plot, and h n (x) represents the prediction type of the n-th decision tree of x, Y k represents the label of the k-th land use type; The expression of the neighborhood effect of the jth plot on the ith plot at the tth time is: Where, e represents the exponential constant, d i,j represents the center distance between the i-th plot and the j-th plot, d represents the buffer radius centered on the i-th plot, S i represents the area of ​​the i-th plot, S j represents the area of ​​the jth plot, S max represents the maximum area of ​​the plot in the study area, S min Indicates the minimum area of ​​a plot in the study area; Neighborhood effect of the kth land use type on the ith plot at time t The expression is: in, represents the neighborhood effect of the jth plot on the ith plot in the kth land use type at the tth time, buffer_d and No River between i and j represents that the buffer radius is d and there is no river crossing between the ith plot and the jth plot; If the restricted development zone is set, the constraint factor of the i-th plot at the t-th time is expressed as: Among them, H i represents the development suitability status of the i-th plot; The transition probability of each plot of land into different land use types is obtained, and the conversion direction with the highest probability and exceeding the development threshold is selected to complete the urban land evolution.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the urban land evolution vector cellular automaton simulation method based on geographic process zoning as described in any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the urban land evolution vector cellular automaton simulation method based on geographic process zoning according to any one of claims 1 to 6 is executed.

Images