A town development boundary optimization method based on an artificial intelligence model

Abstract

This application provides a method for optimizing urban development boundaries based on an artificial intelligence model, belonging to the field of geographic information science and technology. The method includes: acquiring urban land use data, urban expansion data, geological disaster risk data, and POI data for various industries; obtaining existing industrial land distribution data based on driving factor data, POI data for various industries, and urban land use data; using the existing industrial land distribution data as independent variables and driving factor data as dependent variables, using the XGBoost ensemble learning method for feature sampling learning, constructing a hotspot land suitability evaluation model, and outputting the development probability of various types of industrial land; obtaining hotspot plots for urban development boundary optimization through a hierarchical threshold identification method based on data statistical distribution; and superimposing hotspot plots for spatial registration and comparative verification to complete technical verification and accuracy evaluation. This invention overcomes the problems of traditional methods relying on human experience and subjective weighting.

Description

Technical Field

[0001] This application relates to the field of geographic information science and technology, and in particular to a method for optimizing urban development boundaries based on an artificial intelligence model. Background Technology

[0002] With the accelerating urbanization process in my country, urban land area has expanded rapidly, and extensive development has led to problems such as low land use efficiency and environmental damage. The accurate delineation and dynamic optimization of urban development boundaries is a key technical task in geospatial information processing. Its core requirement is to achieve the scientific definition of land use space through intelligent technology, thereby improving the integration efficiency of multi-source spatial data, enhancing the quantitative accuracy of land suitability assessment, and ensuring the dynamic adaptability of boundary adjustments. However, existing urban development boundary optimization technologies have significant technical limitations in their implementation: (1) Insufficient dynamic response and automation capabilities: Existing methods mostly rely on static data overlay or manual experience adjustment, lacking the ability to dynamically capture the rapid evolution of urban land use space. During the boundary generation process, the automation link from data input to result output is broken, resulting in slow response speed and low accuracy of boundary adjustment, which cannot adapt to the real-time update characteristics of multi-source spatial data.

[0003] (2) Weak quantitative analysis and multi-factor integration capabilities: There is a lack of systematic quantitative models for urban development boundary optimization. Traditional methods are unable to effectively model the nonlinear relationships of multiple driving factors such as topography, transportation, infrastructure, and disaster risk, resulting in highly subjective and inaccurate land use suitability assessment results. At the same time, existing technologies lack the ability to quantitatively calculate key indicators such as land development probability, and cannot provide objective numerical basis for boundary optimization.

[0004] In summary, there is currently no comprehensive intelligent technical solution based on multi-source spatial data to address the technical challenges of strong subjectivity and insufficient dynamic adaptation in urban development boundary optimization. Therefore, there is an urgent need to develop a technical approach that integrates artificial intelligence and geospatial analysis to achieve quantitative assessment of the suitability of industrial land and intelligent optimization of development boundaries, providing objective and efficient technical support for boundary adjustments. Summary of the Invention

[0005] The purpose of this invention is to provide an urban development boundary optimization method based on an artificial intelligence model, in order to address the problems of insufficient integration of multiple factors, lagging dynamic response, and strong subjectivity in quantitative evaluation in existing urban development boundary optimization methods.

[0006] The above-mentioned objective of this application is achieved through the following technical solution: S1: Acquire initial land use classification data, geological disaster data, point of interest data, and industry project database, and preprocess them to obtain urban land use data, urban expansion data, geological disaster risk data, and POI data for each industry; S2: Collect data on five types of driving factors: transportation, topography, infrastructure, urban form, and industrial layout; based on the driving factor data, POI data of each industry, and urban land use data, obtain the existing industrial land distribution data; S3: Using existing industrial land distribution data as independent variables and driving factor data as dependent variables, the XGBoost ensemble learning method is used to perform feature sampling learning, construct a hotspot land suitability evaluation model, and output the development probability of various types of industrial land. S4: Based on the development probability of various industrial land uses, a hierarchical threshold identification method based on data statistical distribution is used to obtain hotspot plots for urban development boundary optimization; S5: Obtain existing urban development boundary optimization data, overlay hotspot plots for spatial registration and comparative verification, and complete technical verification and accuracy assessment.

[0007] Optionally, step S1 includes: S11: Reclassify and overlay the initial land use classification data to obtain urban land use data and urban expansion data; S12: Based on the intensity level of geological disasters and geological disaster data, a gradient buffer is generated using the Euclidean distance algorithm to obtain geological disaster risk data; S13: Based on the project list file, perform keyword matching and classification on the industry project database, and extract POI data for each industry in the secondary industry, tertiary industry, and transportation land.

[0008] Optionally, step S2 includes: S21: Obtain data on five types of driving factors affecting the construction of industrial land and construct a suitability evaluation system for hot land use. S22: Based on the hotspot land use suitability evaluation system, kernel density analysis is performed on the POI data of each industry to determine the service impact range threshold; based on the determined service impact range threshold, buffer analysis is performed on the POI data of each industry to obtain the service impact range data of industrial land use. S23: Overlay and analyze the data on the service impact range of industrial land with urban land data to obtain the existing industrial land distribution data.

[0009] Optionally, step S3 includes: S31: Random point sampling is performed on existing industrial land data and driving factor data using uniform sampling or proportional sampling strategies; S32: Normalize or Z-score standardize the sampled data; S33: Using the XGBoost algorithm, a suitability evaluation model for hotspot land use is constructed and trained using processed sampled data. The objective function of the model is:

[0010] in Indicates the number of samples. Represents the objective function and the loss / error function. Description of industrial land In suitability level The difference between the predicted target value and the actual target value of the final score; regularization term Used to control the complexity of the tree and prevent overfitting. It is the number of leaf nodes in the decision tree. Indicates the weight of the leaf node. yes The penalty coefficient, This represents the coefficient of the regularization penalty term; S34: Industrial land use is determined by the softmax function. In suitability level The final score is converted into the probability of industrial land development, as follows:

[0011] in, The total number of suitability levels; S35: Input all driving factor data into the trained model to obtain the development probabilities of various land uses.

[0012] Optionally, step S4 includes: S41: Based on geological disaster risk data and development probability data, select land parcels with development probability below the threshold as candidate land parcels to be removed from the list of land parcels affected by disasters; S42: Overlay disaster-affected candidate land parcels with urban expansion data to obtain disaster-affected recommended land parcels to be removed; S43: Based on urban land use data, the morphological dilatation and erosion method is used to generate urban development boundaries; S44: Based on urban development boundary data and the development probability of various industrial land, select land parcels with a development probability higher than the threshold outside the development boundary as candidate land parcels to be transferred in; S45: Based on statistical distribution, set a tiered threshold to divide candidate land parcels into first-level and second-level priority land parcels, thus obtaining tiered priority land parcels. Land parcels affected by disasters are recommended to be transferred out and land parcels to be transferred in with priority based on their classification, namely, hot parcels for optimizing urban development boundaries.

[0013] Optionally, step S43 includes: The morphological expansion corrosion method includes: A set of points Structural element Inflation is defined as follows:

[0014] A set of points Structural element The definition of corrosion is as follows:

[0015] in It is a binary simulation result that includes only urban and non-urban land; structuring element It is A sliding window, excluding the four pixels at the corner of the square; This indicates a variable that iterates through all pixels in X; This variable represents the offsets used to traverse all elements in B. Represents a candidate point in space; The closing and opening operations are performed sequentially to smooth the boundaries and integrate the patches.

[0016] An electronic device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to enable the electronic device to perform an urban development boundary optimization method based on an artificial intelligence model.

[0017] A computer-readable storage medium storing instructions that, when executed, perform a method for optimizing urban development boundaries based on an artificial intelligence model.

[0018] The beneficial effects of the technical solution provided in this application are: To construct a comprehensive suitability evaluation system for hotspot land use adjustments, encompassing multiple dimensions such as land suitability, ecological sustainability, transportation accessibility, infrastructure carrying capacity, and socio-economic needs, the aim is to provide a quantitative and systematic basis for the scientific adjustment of urban development boundaries. An ensemble learning algorithm (XGBoost) was employed to establish nonlinear relationships between spatial data of various urban functional land uses and multidimensional evaluation factors, yielding development suitability evaluation results for land in different urban functional land uses. This significantly improved the scientific rigor and accuracy of land use suitability evaluation. Furthermore, by combining the development characteristics of different regions with influencing factors such as geological hazards and policies, a scientific basis was provided for the precise inclusion and exclusion of hotspot plots. Comparison and verification between the hotspot plots identified by the model and the actual urban development boundary optimization public announcement schemes demonstrated good scientific rigor, practicality, and promotional value, showing potential for widespread application in various urban spatial governance practices. Attached Figure Description

[0019] The present application will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a step diagram in an embodiment of the present invention; Figure 2 This is a schematic diagram of the driving factors of the hotspot land suitability evaluation system in an embodiment of the present invention; Figure 3 This is a probability map of secondary industry land development output by the hotspot land suitability evaluation model in this embodiment of the invention; Figure 4 This is a probability map of tertiary industry land development output by the hotspot land suitability evaluation model in this embodiment of the invention; Figure 5 This is a probability map of transportation land development in Hubei Province in an embodiment of the present invention; Figure 6 This is a probability space map of secondary industry land development output by the hotspot land suitability evaluation model in this embodiment of the invention; Figure 7 This is the probability space map of tertiary industry land development output by the hotspot land suitability evaluation model in this embodiment of the invention; Figure 8 These are candidate land parcels for transfer into or out of the second industry sector in this embodiment of the invention. Figure 9 These are candidate land parcels for transfer into or out of the tertiary industry in this embodiment of the invention; Figure 10 This is a comparison chart of the actual land parcels transferred into Caidian District and the simulation results in an embodiment of the present invention; Figure 11 This is a schematic diagram of the electronic device structure in the embodiments of this application. Detailed Implementation

[0020] To provide a clearer understanding of the technical features, objectives, and effects of this application, the specific embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0021] The embodiments of this application provide a method for optimizing urban development boundaries based on an artificial intelligence model.

[0022] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of an urban development boundary optimization method based on an artificial intelligence model in an embodiment of this application, including: S1: Acquire initial land use classification data, geological disaster data, point of interest data, and industry project database, and preprocess them to obtain urban land use data, urban expansion data, geological disaster risk data, and POI data for each industry; S2: Collect data on five types of driving factors: transportation, topography, infrastructure, urban form, and industrial layout; based on the driving factor data, POI data of each industry, and urban land use data, obtain the existing industrial land distribution data; S3: Using existing industrial land distribution data as independent variables and driving factor data as dependent variables, the XGBoost ensemble learning method is used to perform feature sampling learning, construct a hotspot land suitability evaluation model, and output the development probability of various types of industrial land. S4: Based on the development probability of various industrial land uses, a hierarchical threshold identification method based on data statistical distribution is used to obtain hotspot plots for urban development boundary optimization; S5: Obtain existing urban development boundary optimization data, overlay hotspot plots for spatial registration and comparative verification, and complete technical verification and accuracy assessment.

[0023] As one example, by constructing a multi-dimensional comprehensive hotspot land suitability evaluation system, the problems of single factor consideration, strong subjectivity in evaluation results, and insufficient spatial expression in traditional methods are solved, significantly improving the scientificity and accuracy of industrial land suitability analysis. By introducing an integrated learning model (XGBoost) to efficiently model driving factor data and existing industrial land distribution data, the problems of strong subjectivity in setting multi-factor weights and weak model generalization ability in traditional methods are solved. Through a hotspot plot identification mechanism with graded thresholds, differentiated and priority-defined boundary optimization suggestions are generated. By registering and comparing the model identification results with the government's publicized plans, the applicability of the optimization method and its ability to support real-world decision-making are further enhanced, comprehensively improving the scientificity, intelligence, and dynamic adjustment capabilities of urban development boundary optimization.

[0024] The technical effects of this invention are as follows: Constructing an intelligent evaluation link of "driving factors - XGBoost - development probability": Integrating multi-source geographic data and machine learning, for the first time, multi-dimensional driving factors such as industrial POI, geological disasters, and transportation are nonlinearly modeled through the XGBoost model to achieve automated and high-precision quantitative assessment of land use suitability, overcoming the problems of traditional methods relying on human experience and subjective weighting.

[0025] A boundary dynamic adjustment mechanism of "hierarchical threshold + morphological optimization" is proposed: the priority level is automatically divided by combining data statistical distribution, and the morphological dilation and erosion algorithm is introduced to intelligently smooth and integrate urban patches, so as to realize the adaptive optimization of the boundary and enhance the spatial continuity, thereby improving the scientificity and manageability of the boundary delineation.

[0026] Establish a two-way verification system of "model output - actual solution": By spatially overlaying the hot spots identified by the model with the government's publicized solutions and conducting accuracy assessments, a closed-loop verification mechanism is formed, which significantly improves the reliability of the optimization results and the decision support capability.

[0027] Step S1 includes: S11: Reclassify and overlay the initial land use classification data to obtain urban land use data and urban expansion data; S12: Based on the intensity level of geological disasters and geological disaster data, a gradient buffer is generated using the Euclidean distance algorithm to obtain geological disaster risk data; S13: Based on the project list file, perform keyword matching and classification on the industry project database, and extract POI data for each industry in the secondary industry, tertiary industry, and transportation land.

[0028] Step S2 includes: S21: Obtain data on five types of driving factors affecting the construction of industrial land and construct a suitability evaluation system for hot land use. S22: Based on the hotspot land use suitability evaluation system, kernel density analysis is performed on the POI data of each industry to determine the service impact range threshold; based on the determined service impact range threshold, buffer analysis is performed on the POI data of each industry to obtain the service impact range data of industrial land use. S23: Overlay and analyze the data on the service impact range of industrial land with urban land data to obtain the existing industrial land distribution data.

[0029] Step S3 includes: S31: Random point sampling is performed on existing industrial land data and driving factor data using uniform sampling or proportional sampling strategies; As one example, a uniform sampling strategy or a proportional sampling strategy is used to randomly sample existing industrial land use data and driving factor data to ensure the sample has objective representativeness and obtain sampled data. The number of sampling points can be automatically determined through cross-validation to balance computational efficiency and representativeness.

[0030] S32: Normalize or Z-score standardize the sampled data; As one example, the sampled data is normalized to eliminate the dimensional differences between different driving factors and improve the stability of model training. Outliers in the data can be processed using Z-score standardization to further enhance the robustness of the model.

[0031] S33: Using the XGBoost algorithm, a suitability evaluation model for hotspot land use is constructed and trained using processed sampled data. The objective function of the model is:

[0032] in Indicates the number of samples. Represents the objective function and the loss / error function. Description of industrial land In suitability level The difference between the predicted target value and the actual target value of the final score; regularization term Used to control the complexity of the tree and prevent overfitting. It is the number of leaf nodes in the decision tree. Indicates the weight of the leaf node. yes The penalty coefficient, This represents the coefficient of the regularization penalty term; S34: Industrial land use is determined by the softmax function. In suitability level The final score is converted into the probability of industrial land development, as follows:

[0033] in, The total number of suitability levels; As an example, in order to convert the final score of industrial land i on suitability level c into the probability of industrial land development, this invention uses the softmax function, which is a standard function for converting scores into probabilities in commonly used multi-class classification problems. It ensures that the sum of the probabilities of all categories is 1, provides an interpretable distribution, and enhances the model's ability to handle uncertainty.

[0034] S35: Input all driving factor data into the trained model to obtain the development probabilities of various land uses.

[0035] As one example, using existing industrial land use data as independent variables and driving factor data as dependent variables, an ensemble learning method (XGBoost) is employed for feature sampling learning to construct a suitability evaluation model for hotspot land use. This model, based on the Gradient Boosting Decision Tree (GBDT) framework, achieves nonlinear relationship modeling and automated feature selection for high-dimensional geographic data, supporting both column and row subsampling, significantly reducing computational complexity and improving generalization ability. Parameters are automatically optimized through cross-validation, avoiding subjective intervention and ensuring the model's objectivity and robustness.

[0036] Step S4 includes: S41: Based on geological disaster risk data and development probability data, select land parcels with development probability below the threshold as candidate land parcels to be removed from the list of land parcels affected by disasters; S42: Overlay disaster-affected candidate land parcels with urban expansion data to obtain disaster-affected recommended land parcels to be removed; S43: Based on urban land use data, the morphological dilatation and erosion method is used to generate urban development boundaries; S44: Based on urban development boundary data and the development probability of various industrial land, select land parcels with a development probability higher than the threshold outside the development boundary as candidate land parcels to be transferred in; S45: Based on statistical distribution, set a tiered threshold to divide candidate land parcels into first-level and second-level priority land parcels, thus obtaining tiered priority land parcels. Land parcels affected by disasters are recommended to be transferred out and land parcels to be transferred in with priority based on their classification, namely, hot parcels for optimizing urban development boundaries.

[0037] Step S43 includes: The morphological expansion corrosion method includes: A set of points Structural element Inflation is defined as follows:

[0038] A set of points Structural element The definition of corrosion is as follows:

[0039] in It is a binary simulation result that includes only urban and non-urban land; structuring element It is A sliding window, excluding the four pixels at the corner of the square; This indicates a variable that iterates through all pixels in X; This variable represents the offsets used to traverse all elements in B. Represents a candidate point in space; The closing and opening operations are performed sequentially to smooth the boundaries and integrate the patches.

[0040] As one embodiment, this structured element shape prevents the boundary contour of expansion or erosion from approaching a rectangle, and more closely approximates the boundary line of urban plots. During erosion, the origin of the structured element scans urban pixels. If not all pixels in the structured element are urban pixels, the urban pixels at the origin of the structured element are removed during erosion. Conversely, as the origin of the structured element moves around urban pixels, the expansion process converts all non-urban pixels in the structured element into urban pixels. The erosion process can remove small and scattered urban patches with low compactness, as very small urban patches are not easy to delineate and manage urban development boundaries. Furthermore, the expansion process can connect urban plots suitable for inclusion in BUD.

[0041] The planning of BUD (Block Layout and Unification) should meet two requirements: removing isolated patches and filling the gaps between city clusters. Therefore, we need to combine dilation and erosion methods. Applying dilation after erosion using the same structuring element is called the opening operation. The erosion step in the opening operation removes isolated city patches and the boundaries of city blocks, while the dilation step restores most of the boundary cells without restoring noise. The opening operation tends to "open" small gaps or spaces between contacting blocks in the image. The opening operation is defined by the following formula:

[0042] Closing is similar to opening, but the difference is that closing first performs dilation and then erosion using the same structuring element. Closing is more efficient at filling small gaps in an image and "closing" them. Closing is defined by the following formula:

[0043] Opening tends to shrink city pixels because it essentially represents the intersection of X and B. Conversely, closing creates... and The union of these elements thus increases the number of city pixels. Both operations can smooth out plot boundaries in the image.

[0044] Because opening and closing operations possess the aforementioned specific functions, this study first applies closing operations to connect adjacent urban plots, and then uses opening operations to delete isolated small urban plots that are unsuitable for planning as BUD (Building Unified Urban Areas). Theoretically, combining opening and closing operations can ensure that the final BUD area does not deviate significantly from the planned quantity target.

[0045] As one example, a morphological dilation and erosion method is used to process urban land use data, removing isolated urban patches and integrating patch clusters with the potential to develop into contiguous urban areas, generating easily manageable urban development boundaries. The specific steps are: dilation -> erosion -> erosion -> dilation. The process of dilation followed by erosion is a closing operation, and the subsequent erosion and dilation is an opening operation. By performing one closing operation and one opening operation on the urban-non-urban binary data, a raster-structured urban development boundary can be obtained. Then, the raster is converted to a vector in GIS software to obtain the final urban development boundary result. This method is based on set theory and achieves automated boundary smoothing and patch integration without the need for manual threshold setting.

[0046] The present invention provides an embodiment as follows: The study focuses on Hubei Province, using land use data from 2010 and 2020 as initial data. Based on experience in land use change simulation and combining historical and existing data from Wuhan City, this case study selects sixteen driving factors. These factors are: transportation factors (including primary roads, secondary roads, tertiary roads, arterial roads, expressways, and railways); topographic factors (including elevation and slope); infrastructure factors (including government buildings, high-speed rail stations, bus stops, and educational facilities); urban morphology factors (including city centers and important towns); and industrial layout factors (including existing large factories and large commercial areas). Urban development boundary optimization data from Caidian District, Yangtze River New Area, and Jiangxia District of Wuhan City are used as supporting data for the analysis.

[0047] The first step involves reclassifying and overlaying the initial land use classification data to obtain urban land use data and urban expansion data; based on the geological hazard intensity level setting, buffer analysis is performed using the Euclidean distance algorithm to obtain geological hazard risk data; POI data is obtained, and POI data for secondary industry, tertiary industry, and transportation land are selected based on key projects.

[0048] The second step involves collecting driving factor data based on the impact on industrial land development to construct a suitability evaluation system for hotspot land use; conducting buffer analysis on POI data for secondary, tertiary, and transportation land to obtain data on the service impact range of industrial land; and overlaying the service impact range data of industrial land with urban land data to obtain data on the distribution of existing industrial land.

[0049] Step 3: Perform random point sampling on the driving factor data and existing industrial land distribution data. The model provides two sampling methods: 1. Uniform sampling strategy; 2. Proportional sampling strategy. This example uses the uniform sampling strategy to extract existing industrial land distribution data and... Figure 2 The driving factor data in the data.

[0050] Step 4: Use the ensemble learning algorithm (XGBoost) to train the sampled data, set the training parameters and sampling rate, and then train the regression tree.

[0051] Step 5: Input all driving factor data into the trained hotspot land suitability evaluation model, and output the development probability of various land uses, such as... Figure 3 , Figure 4 , Figure 5 As shown.

[0052] Step 6: Using urban development boundary and geological disaster risk data as development constraints, overlay them with the development probability results of various industrial land uses to identify hotspots that should be prioritized for inclusion and recommended for exclusion from urban development boundaries. This reveals the spatial differentiation patterns of industrial land use at the provincial and county levels. Figure 6 , Figure 7 , Figure 8 and Figure 9 As shown.

[0053] Step 7: Obtain urban development boundary optimization data for Caidian District, Yangtze River New Area, and Jiangxia District of Wuhan City, overlay the model to identify hotspots for urban development boundary adjustments, perform spatial registration and comparative verification, complete technical verification and accuracy assessment, and confirm the accuracy and feasibility of the model output results. Figure 10 As shown in the figure. The comparative verification results show that the proposed model can effectively identify areas with high suitability for urban development, and the unidentified plots are all located near the simulation results, demonstrating good scientific validity, practicality, and promotional value, and has the potential to be widely applied in different urban spatial governance practices.

[0054] This application also discloses an electronic device. (See reference...) Figure 11 , Figure 11 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application. The electronic device 500 may include: at least one processor 501, at least one network interface 504, a user interface 503, a memory 505, and at least one communication bus 502.

[0055] The communication bus 502 is used to enable communication between these components.

[0056] The user interface 503 may include a display screen, and optionally, the user interface 503 may also include a standard wired interface or a wireless interface.

[0057] The network interface 504 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0058] This application also discloses a computer-readable storage medium storing multiple instructions adapted for loading by a processor to execute the aforementioned method for optimizing urban development boundaries based on an artificial intelligence model.

[0059] The above are merely exemplary embodiments of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure.

[0060] This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for optimizing urban development boundaries based on an artificial intelligence model, characterized in that, The method includes the following steps: S1: Acquire initial land use classification data, geological disaster data, point of interest data, and industry project database, and preprocess them to obtain urban land use data, urban expansion data, geological disaster risk data, and POI data for each industry; S2: Collect data on five types of driving factors: transportation, topography, infrastructure, urban form, and industrial layout; based on the driving factor data, POI data of each industry, and urban land use data, obtain the existing industrial land distribution data; S3: Using existing industrial land distribution data as independent variables and driving factor data as dependent variables, the XGBoost ensemble learning method is used to perform feature sampling learning, construct a hotspot land suitability evaluation model, and output the development probability of various types of industrial land. S4: Based on the development probability of various industrial land uses, a hierarchical threshold identification method based on data statistical distribution is used to obtain hotspot plots for urban development boundary optimization; S5: Obtain existing urban development boundary optimization data, overlay hotspot plots for spatial registration and comparative verification, and complete technical verification and accuracy assessment.

2. The method for optimizing urban development boundaries based on an artificial intelligence model as described in claim 1, characterized in that, Step S1 includes: S11: Reclassify and overlay the initial land use classification data to obtain urban land use data and urban expansion data; S12: Based on the intensity level of geological disasters and geological disaster data, a gradient buffer is generated using the Euclidean distance algorithm to obtain geological disaster risk data; S13: Based on the project list file, perform keyword matching and classification on the industry project database, and extract POI data for each industry in the secondary industry, tertiary industry, and transportation land.

3. The method for optimizing urban development boundaries based on an artificial intelligence model as described in claim 1, characterized in that, Step S2 includes: S21: Obtain data on five types of driving factors affecting the construction of industrial land and construct a suitability evaluation system for hot land use. S22: Based on the hotspot land use suitability evaluation system, kernel density analysis is performed on the POI data of each industry to determine the service impact range threshold; based on the determined service impact range threshold, buffer analysis is performed on the POI data of each industry to obtain the service impact range data of industrial land use. S23: Overlay and analyze the data on the service impact range of industrial land with urban land data to obtain the existing industrial land distribution data.

4. The method for optimizing urban development boundaries based on an artificial intelligence model as described in claim 1, characterized in that, Step S3 includes: S31: Random point sampling is performed on existing industrial land data and driving factor data using uniform sampling or proportional sampling strategies; S32: Normalize or Z-score standardize the sampled data; S33: Using the XGBoost algorithm, a suitability evaluation model for hotspot land use is constructed and trained using processed sampled data. The objective function of the model is: in Indicates the number of samples. Represents the objective function and the loss / error function. Description of industrial land In suitability level The difference between the predicted target value and the actual target value of the final score; regularization term Used to control the complexity of the tree and prevent overfitting. It is the number of leaf nodes in the decision tree. Indicates the weight of the leaf node. yes The penalty coefficient, This represents the coefficient of the regularization penalty term; S34: Industrial land use is determined by the softmax function. In suitability level The final score is converted into the probability of industrial land development, as follows: in, The total number of suitability levels; S35: Input all driving factor data into the trained model to obtain the development probabilities of various land uses.

5. The method for optimizing urban development boundaries based on an artificial intelligence model as described in claim 1, characterized in that, Step S4 includes: S41: Based on geological disaster risk data and development probability data, select land parcels with development probability below the threshold as candidate land parcels to be removed from the list of land parcels affected by disasters; S42: Overlay disaster-affected candidate land parcels with urban expansion data to obtain disaster-affected recommended land parcels to be removed; S43: Based on urban land use data, the morphological dilatation and erosion method is used to generate urban development boundaries; S44: Based on urban development boundary data and the development probability of various industrial land, select land parcels with a development probability higher than the threshold outside the development boundary as candidate land parcels to be transferred in; S45: Based on statistical distribution, set a tiered threshold to divide candidate land parcels into first-level and second-level priority land parcels, thus obtaining tiered priority land parcels. Land parcels affected by disasters are recommended to be transferred out and land parcels to be transferred in with priority based on their classification, namely, hot parcels for optimizing urban development boundaries.

6. The method for optimizing urban development boundaries based on an artificial intelligence model as described in claim 5, characterized in that, Step S43 includes: The morphological expansion corrosion method includes: A set of points Structural element Inflation is defined as follows: A set of points Structural element The definition of corrosion is as follows: in It is a binary simulation result that includes only urban and non-urban land; structuring element It is A sliding window, excluding the four pixels at the corner of the square; This indicates a variable that iterates through all pixels in X; This variable represents the offsets used to traverse all elements in B. Represents a candidate point in space; The closing and opening operations are performed sequentially to smooth the boundaries and integrate the patches.

7. An electronic device, characterized in that, The device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to enable the electronic device to perform the urban development boundary optimization method based on an artificial intelligence model as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a computer, perform the urban development boundary optimization method based on an artificial intelligence model as described in any one of claims 1-6.

Images