Land planning deduction and evaluation system based on multi-source data driving

Abstract

The application discloses a land planning deduction and evaluation system based on multi-source data driving and relates to the technical field of land planning, which comprises the following steps: collecting multi-source heterogeneous data of basic geographic space, social and economic operation and ecological environment monitoring of a target region, aligning the data through administrative division boundaries and time stamp fields after standardization processing, and constructing a global multi-dimensional space-time cube; superimposing current control rules to identify spatial conflict points, calling a land use evolution simulation model to predict natural growth trends, and generating three deduction scenarios of benchmarks, ecological protection and compact development; starting a parallel computing framework, executing full-factor space evolution simulation on the three scenarios, and outputting planning deduction trajectory dataset with time series attributes. The system realizes deep fusion of multi-source data and parallel deduction of multiple scenarios and is suitable for the field of land space planning.

Description

Technical Field

[0001] This invention belongs to the field of land planning technology, specifically a land planning simulation and evaluation system driven by multi-source data. Background Technology

[0002] Currently, land planning simulation and evaluation often rely on a single data source or simply pieced together multi-source data. Data processing often only achieves coordinate unification and format conversion of basic geospatial data, socio-economic operation data, and ecological environment monitoring data, without deeply exploring the spatiotemporal correlation of multi-source heterogeneous data. Traditional data fusion methods are often used, which can only complete the simple overlay of data and cannot form a data carrier that can integrate the entire region, multiple dimensions, and time sequence.

[0003] Existing technologies often employ single-scenario simulation modes during planning and extrapolation, and their calculation methods are primarily serial. This makes it difficult to simultaneously simulate the evolution of all elements under different planning orientations, and it is impossible to output planning and extrapolation trajectories with complete time-series attributes. Consequently, the identification of spatial overlap and conflict points in ecological protection red line areas, permanent basic farmland protection areas, and urban development boundary areas is not accurate enough, and the prediction of natural land use growth trends lacks comprehensiveness. This makes it difficult to meet the actual needs of multi-scenario comparison, accurate extrapolation, and scientific evaluation in national land planning.

[0004] The lack of deep spatiotemporal fusion of multi-source heterogeneous data makes it impossible to construct a multi-dimensional time-series data carrier covering the entire domain. At the same time, the lack of multi-scenario parallel full-element evolution simulation capabilities makes it impossible to accurately capture the time trajectory of planning evolution, which limits the scientificity and accuracy of land planning simulation and evaluation. Summary of the Invention

[0005] This invention aims to solve at least one of the technical problems existing in the prior art;

[0006] To this end, this invention proposes a land planning simulation and evaluation system based on multi-source data, comprising:

[0007] The data standardization module collects basic geospatial data, socio-economic operation data, and ecological environment monitoring data of the target area. It performs coordinate system unification and format conversion on the collected multi-source heterogeneous data to form a standardized planning base map dataset.

[0008] The spatiotemporal fusion module imports the standardized planning base map dataset into the spatiotemporal data fusion engine, aligns the data based on administrative division boundaries and timestamp fields, and constructs a multi-dimensional spatiotemporal cube covering the entire region.

[0009] The conflict identification module overlays the current management rules of the national land space planning onto the multi-dimensional spatiotemporal cube to identify spatially overlapping conflict points between ecological protection red line areas, permanent basic farmland protection areas, and urban development boundary areas.

[0010] The situation prediction module calls the land use evolution simulation model, reads the historical land use change trajectory in the multi-dimensional spatiotemporal cube, and predicts the natural growth trend under the condition of no new planning intervention.

[0011] The scenario construction module takes the existing control rules and the natural growth trend as input into the scenario construction module to generate three simulation scenarios: the baseline scenario, the ecological protection scenario, and the compact development scenario.

[0012] The parallel simulation module starts a parallel computing framework to perform full-element spatial evolution simulations on three simulation scenarios, and outputs a planning simulation trajectory dataset with time series attributes.

[0013] Furthermore, the process of performing coordinate system and format conversion on the collected multi-source heterogeneous data to form a standardized planning base map dataset specifically includes:

[0014] Read the vector layer files in the basic geospatial data, extract the geometric information and attribute tables, and convert the coordinate system from the local independent coordinate system to the national geodetic coordinate system;

[0015] Analyze statistical yearbook tables in socio-economic operation data, transform unstructured text descriptions into numerical indicators, and convert them according to a unified unit of measurement.

[0016] Radiometric calibration and atmospheric correction are performed on remote sensing images in ecological and environmental monitoring data to generate raster data with uniform resolution.

[0017] Establish globally unique data identifiers for the vector layers, numerical indicators, and raster data, and record their source and acquisition time in the metadata;

[0018] The processed vector layers, numerical indicators, and raster data are packaged and stored to form a structured, standardized planning base map dataset.

[0019] Furthermore, the step of calling the land use evolution simulation model, reading the historical land use change trajectory in the multidimensional spatiotemporal cube, and predicting the natural growth trend under the condition of no new planning intervention specifically includes:

[0020] Land use status maps from several consecutive years in the past are retrieved from the multidimensional spatiotemporal cube and converted into discrete raster cell matrices.

[0021] The frequency of land use type transfers for each grid cell within a historical period is statistically analyzed to generate a Markov transfer probability matrix that reflects the inertia of urban expansion.

[0022] By combining the population density gradient and road accessibility data in the multidimensional spatiotemporal cube, a dynamic suitability weight is assigned to each grid cell;

[0023] The Markov transition probability matrix and the suitability weights are used as input variables and loaded into the land use evolution simulation model for iterative calculation.

[0024] Extract the final land use distribution pattern from the simulation model output and label it as the natural growth trend.

[0025] Furthermore, by combining the population density gradient and road accessibility data in the multidimensional spatiotemporal cube, dynamic suitability weights are assigned to each grid cell, specifically including:

[0026] Extract a population density raster layer with the same spatial range and resolution as the raster unit matrix from the multidimensional spatiotemporal cube, and read the population density value corresponding to the center point of each raster unit.

[0027] Road network vector data is extracted from the multidimensional spatiotemporal cube, the Euclidean distance from the center point of each grid cell to the nearest main road is calculated, and the Euclidean distance is mapped to a road accessibility index.

[0028] The population density value is normalized to eliminate the influence of dimensions and converted into a population density influence factor between 0 and 1.

[0029] The road accessibility index is subjected to inverse distance attenuation processing, where the closer the distance, the higher the accessibility. It is also converted into a road accessibility influence factor between 0 and 1.

[0030] Based on the pre-set land use type conversion driving model, the contribution coefficients of population density influence factor and road accessibility influence factor are respectively assigned to different types of land use conversion behavior;

[0031] Based on the contribution coefficient, the population density influence factor and road accessibility influence factor of each grid cell are weighted and summed to calculate the dynamic suitability weight of the grid cell under specific land use conversion behavior.

[0032] Furthermore, the module that inputs the existing control rules and the natural growth trend into the scenario construction module generates three simulation scenarios: a baseline scenario, an ecological protection scenario, and a compact development scenario. Specifically, these include:

[0033] The current total scale and growth rate of construction land are extracted from the multidimensional spatiotemporal cube and used as the core control parameters of the baseline scenario.

[0034] Based on the existing control rules, the development resistance coefficient of ecologically sensitive areas will be increased, and the land use within the permanent basic farmland area will be locked to construct a parameter configuration set for ecological protection scenarios.

[0035] Based on the existing control rules, the quota for reserved land for construction outside the urban development boundary will be relaxed, and the upper limit of plot ratio around rail transit stations will be increased, so as to construct a parameter configuration set for compact development scenarios.

[0036] The core control parameters of the baseline scenario, the parameter configuration set of the ecological protection scenario, and the parameter configuration set of the compact development scenario are respectively written into the preset scenario configuration file.

[0037] The scenario configuration file is sent to the scenario building module, which parses the constraints in the configuration file and generates three simulation scenarios that can be called by the simulation model.

[0038] Furthermore, the parallel computing framework is activated to perform full-element spatial evolution simulations on the three simulation scenarios, outputting a planning simulation trajectory dataset with time-series attributes, specifically including:

[0039] The data slices in the multidimensional spatiotemporal cube are distributed to each computing node in the parallel computing cluster to ensure that each node has a balanced load.

[0040] Within each computing node, the corresponding simulation scenario is loaded, and the evolution rule set of the cellular automaton is initialized according to the constraints in the script.

[0041] The cellular automata model is driven to evolve cyclically within a set simulation period, and the road network density and public service facility accessibility indicators within the region are updated synchronously after each period ends.

[0042] Collect the intermediate evolution results of all computing nodes in each simulation cycle, and splice and recombine them in the order of the time series;

[0043] Abnormal oscillation data during the evolution process are removed, and the time series data is smoothed by interpolation to finally generate a continuous planning and simulation trajectory dataset.

[0044] Furthermore, it also includes a load-bearing capacity calculation module, used for:

[0045] Read the time-series data on water consumption, energy consumption, and carbon emissions from the planning and simulation trajectory dataset;

[0046] The time-series data is compared with a pre-entered resource environment baseline database to calculate the resource supply and demand difference at different simulation times.

[0047] When the difference between resource supply and demand at a certain simulated moment is detected to exceed the preset safety threshold, the planning simulation trajectory dataset is traced back to locate the dominant land use type that caused the resource overload.

[0048] Based on the spatial distribution of the dominant land use types, high-risk warning areas for resource carrying capacity are marked on the map;

[0049] The marked high-risk warning areas and their corresponding overload factors are written into the intermediate data layer of the assessment report.

[0050] Furthermore, it also includes a connectivity evaluation module for:

[0051] The distribution data of ecological source patches and the topological structure data of corridors are extracted from the planning and simulation trajectory dataset.

[0052] The shortest path distance between the ecological source patches was calculated using graph theory algorithms, and the number of breakpoints in the corridor was counted.

[0053] Substitute the calculated path distance and number of breakpoints into the landscape connectivity index calculation model to obtain the ecological network connectivity score under different simulation scenarios.

[0054] The difference between the ecological network connectivity scores under different simulation scenarios and the scores under the baseline scenario is calculated to obtain the connectivity gain / loss value.

[0055] The connectivity gain / loss value is used as a key evaluation indicator and updated to the intermediate data layer of the evaluation report.

[0056] Furthermore, it also includes a conflict resolution module for:

[0057] By overlaying and analyzing the planning and simulation trajectory datasets output from the three simulation scenarios, areas with spatial ownership disputes between different scenarios can be identified.

[0058] For each area with disputed spatial ownership, the dominant functional positioning of the area in three hypothetical scenarios is extracted, and the reasons for their contradictions are analyzed.

[0059] Based on the cause of the conflict, search the existing management and control rule base for applicable flexible adjustment strategies, such as the replacement of construction land quotas or the staggered arrangement of development time.

[0060] The selected flexible adjustment strategy is fed back to the scenario construction module to correct the parameters of the original simulation scenario;

[0061] Re-execute the spatial evolution simulation under the parallel computing framework until no new spatial ownership disputes appear in the planning and simulation trajectory dataset.

[0062] Furthermore, it also includes a results visualization module for:

[0063] Extract land use snapshots at key time points from the planning and simulation trajectory dataset and render them as a three-dimensional terrain scene;

[0064] In the three-dimensional terrain scene, differentiated texture maps and lighting effects are applied according to different land use properties to enhance the sense of spatial hierarchy;

[0065] The high-risk warning areas obtained from the resource carrying capacity calculation step and the connectivity gain / loss values ​​obtained from the connectivity evaluation step are overlaid on the three-dimensional terrain scene in the form of a heat map.

[0066] Generate an interactive visualization interface that allows operators to switch between different simulation moments using a timeline slider and dynamically observe the evolution of the planning scheme;

[0067] Export the generated 3D visualization scene screenshots and animation sequences as attachments to the final report.

[0068] Compared with the prior art, the beneficial effects of the present invention are:

[0069] Data alignment is performed on standardized multi-source heterogeneous planning base map data based on administrative division boundaries and timestamp fields to construct a multi-dimensional spatiotemporal cube covering the entire region. This breaks through the limitations of conventional multi-source data, which only involves format conversion and simple splicing. It achieves deep spatiotemporal fusion of basic geospatial data, socio-economic operation data, and ecological environment monitoring data. It can integrate data information of different dimensions and time series across the entire region, allowing the spatiotemporal correlation of various types of data to be fully explored. This avoids information disconnect caused by data fragmentation, making the planning base map data more complete and correlated, and overcoming the problem of insufficient data correlation under conventional data processing methods.

[0070] A parallel computing framework was launched to perform full-element spatial evolution simulations on three scenarios: baseline scenario, ecological protection scenario, and compact development scenario. The output is a planning simulation trajectory dataset with time series attributes. This overcomes the limitations of conventional planning simulations that rely on single scenarios and serial computation. It can simultaneously conduct full-element evolution simulations for multiple scenarios, improving the efficiency of planning simulations. At the same time, the complete time series trajectory can clearly present the dynamic process of land space evolution under different scenarios, avoiding the problems of conventional simulations that lack temporal sequence and cannot compare the evolutionary differences between different scenarios. This makes the planning simulation results more targeted and comprehensive, and can accurately reflect the trend of land space change under different planning interventions. This is different from conventional planning simulation methods that rely on single scenarios and lack complete time series simulation trajectories. Attached Figure Description

[0071] Figure 1 This is a timeline diagram of the land planning simulation and evaluation system based on multi-source data as described in this invention.

[0072] Figure 2 A flowchart for predicting the natural growth trend of land use;

[0073] Figure 3 A graph showing the trend of changes in the scale of construction land under three different scenarios;

[0074] Figure 4 A chart showing the changing trend of resource carrying capacity over 20 years under a compact development scenario;

[0075] Figure 5 This is a trend chart showing the changes in land use type area over 20 years. Detailed Implementation

[0076] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0077] See Figure 1The data standardization module collects basic geospatial data, socio-economic operation data, and ecological environment monitoring data of the target area. It performs coordinate system unification and format conversion on the collected multi-source heterogeneous data to form a standardized planning base map dataset. The spatiotemporal fusion module imports the standardized planning base map dataset into the spatiotemporal data fusion engine, aligns the data based on administrative boundaries and timestamp fields, and constructs a multi-dimensional spatiotemporal cube covering the entire region. The conflict identification module overlays the current land use planning control rules onto the multi-dimensional spatiotemporal cube, identifying spatially overlapping conflict points in ecological protection red line areas, permanent basic farmland protection areas, and urban development boundary areas. The situation prediction module calls the land use evolution simulation model, reads the historical land use change trajectory in the multi-dimensional spatiotemporal cube, and predicts the natural growth trend under the condition of no new planning intervention. The scenario construction module inputs the current control rules and the natural growth trend to generate three simulation scenarios: a baseline scenario, an ecological protection scenario, and a compact development scenario. The parallel simulation module starts a parallel computing framework to perform full-element spatial evolution simulations on the three simulation scenarios, outputting a planning simulation trajectory dataset with time-series attributes.

[0078] In one embodiment of the present invention, the data standardization module collects basic geospatial data, socio-economic operation data, and ecological environment monitoring data of the target area. It performs coordinate system unification and format conversion on the collected multi-source heterogeneous data to form a standardized planning base map dataset. Taking a city in East China as an example, the basic geospatial data includes an administrative division vector layer. The administrative division vector layer uses a local independent coordinate system, such as the Beijing 54 coordinate system. The data standardization module reads the vector layer file, extracts the geometric information and attribute tables from the vector layer file, and converts the coordinate system from the Beijing 54 coordinate system to the National Geodetic Coordinate System 2000. After conversion, the coordinate values ​​are consistent with the national benchmark. In some embodiments, the socio-economic operation data includes statistical yearbooks. The statistical yearbook table contains unstructured text descriptions such as "approximately 1000 mu of cultivated land." The data standardization module parses the statistical yearbook table, converting "approximately 1000 mu of cultivated land" into the numerical indicator "cultivated land area," and converts it to "hectares" according to a unified unit of measurement, resulting in "cultivated land area 66.67 hectares." Optionally, the ecological environment monitoring data includes remote sensing imagery, specifically Landsat 8 data. The data standardization module performs radiometric calibration and atmospheric correction on the remote sensing imagery. Radiometric calibration converts digital quantization values ​​into physical quantities, while atmospheric correction eliminates the effects of atmospheric scattering and absorption, generating raster data with a unified resolution of 30 meters. The radiometric calibration process uses the following formula:

[0079]

[0080] in: This represents the radiance of the upper atmosphere, measured in watts per square meter per steradian per micrometer. Indicates the gain coefficient. The digital quantization value representing remote sensing imagery. This represents the bias coefficient; the surface reflectance of the processed raster data ranges from 0 to 1.

[0081] In practical implementation, globally unique data identifiers are established for vector layers, numerical indicators, and raster data, and their source and acquisition time are recorded in the metadata. For example, the identifier for a vector layer is "VEC_2025_001," the source is "Municipal Surveying and Mapping Bureau," and the acquisition time is "June 2025." The identifier for a numerical indicator is "TAB_2025_002," the source is "Municipal Statistics Bureau," and the acquisition time is "May 2025." The identifier for raster data is "RAS_2025_003," the source is "Remote Sensing Data Center," and the acquisition time is "April 2025." In some embodiments, the processed vector layers... Layers, numerical indicators, and raster data are packaged and stored to form a structured, standardized planning base map dataset. The standardized planning base map dataset is stored in GeoPackage format and includes vector layers, attribute tables, and raster layers. The vector layers use the WGS84 geographic coordinate system, the attribute tables contain standardized numerical fields, and the raster layers have a consistent spatial reference and resolution. It can be understood that the data identification code adopts a hierarchical structure encoding, with the encoding rule being "data type_year_serial number" to ensure global uniqueness. Optionally, the data standardization module uses a conversion factor when converting units, for example, 1 hectare equals 15 mu, so the mu value is divided by 15 to get the hectare value.

[0082] In one embodiment of the present invention, the situation prediction module calls a land use evolution simulation model to read the historical land use change trajectory in a multi-dimensional spatiotemporal cube, and predicts the natural growth trend under the condition of no new planning intervention. Taking a city planning assessment as an example, see [reference]. Figure 2 The process involves retrieving land use maps from a multidimensional spatiotemporal cube, covering five consecutive years. These maps include arable land, forest land, construction land, water areas, and unused land. Each year's land use map is converted into a discrete raster cell matrix with a resolution of 30 meters. The value of each raster cell represents its land use type code. The frequency of land use type transfers for each raster cell over the five historical years is calculated. For example, the number of times a raster cell coded as arable land is converted to construction land in the following year is recorded. Based on the transfer frequency counts of all raster cells, a Markov transition probability matrix reflecting the inertia of urban expansion is generated. This Markov transition probability matrix is ​​an n-order square matrix, where n is the number of land use types. The elements in the matrix... This represents the probability of transferring land from type m to type n.

[0083] In practical implementation, combining population density gradient and road accessibility data in a multi-dimensional spatiotemporal cube, dynamic suitability weights are assigned to each raster cell. A population density raster layer with the same spatial range and resolution as the raster cell matrix is ​​extracted from the multi-dimensional spatiotemporal cube. The population density value corresponding to the center point of each raster cell is read, with the unit being people per square kilometer. Road network vector data, including highways, national roads, and urban arterial roads, is extracted from the multi-dimensional spatiotemporal cube. The Euclidean distance from the center point of each raster cell to the nearest major road is calculated, with the unit being meters. This Euclidean distance is then mapped to a road accessibility index using the following mapping function:

[0084]

[0085] in: This represents the road accessibility index of the j-th grid cell. This represents the Euclidean distance from the j-th raster cell to the nearest main road. It can be understood that normalizing the population density value eliminates the influence of dimensions, converting it into a population density influence factor between 0 and 1. The normalization formula is:

[0086]

[0087] in: This represents the population density influence factor of the i-th raster cell. This represents the original population density value of the i-th raster cell. and These represent the minimum and maximum population density values ​​in all grid cells, respectively. In some embodiments, the road accessibility index undergoes inverse distance attenuation processing; the closer the distance, the higher the accessibility. It is also converted into a road accessibility influence factor between 0 and 1. This process is achieved through linear scaling. Optionally, based on a preset land use type conversion driving model, contribution coefficients of the population density influence factor and the road accessibility influence factor are configured for different types of land use conversion behaviors. For example, in the conversion behavior of arable land to construction land, the contribution coefficient of the population density influence factor is set to 0.6, and the contribution coefficient of the road accessibility influence factor is set to 0.4. Based on the contribution coefficients, the population density influence factor and the road accessibility influence factor of each grid cell are weighted and summed to calculate the dynamic suitability weight of the grid cell under a specific land use conversion behavior. The formula for calculating the dynamic suitability weight is:

[0088]

[0089] in: This represents the dynamic suitability weight of grid cell i when a land use type j conversion occurs. This represents the contribution coefficient of the population density influencing factor in the corresponding conversion type. This represents the population density influence factor of raster cell i. This represents the contribution coefficient of the road accessibility influencing factor in the corresponding conversion type. The road accessibility impact factor for grid cell j can be understood as a contribution coefficient. and The sum is 1.

[0090] In specific implementation, the Markov transition probability matrix and the calculated dynamic suitability weights are used as input variables and loaded into the land use evolution simulation model for iterative calculation. The land use evolution simulation model is based on the principle of cellular automata. In each iteration, it provides the global transition probability according to the Markov transition probability matrix and adjusts its neighborhood competition rules according to the dynamic suitability weight of each grid cell. The simulation process iterates a set number of times, which corresponds to the number of predicted future years. The final land use distribution pattern output by the simulation model is extracted. The final land use distribution pattern is represented as a new grid cell matrix and marked as the natural growth trend under the condition of no new planning intervention. In some embodiments, the calculation of dynamic suitability weights considers multiple driving factors, among which population density and road accessibility are core factors. Optionally, the iterative calculation of the land use evolution simulation model is executed in a parallel computing environment to improve efficiency.

[0091] In one embodiment of the present invention, the scenario construction module inputs current control rules and natural growth trends to generate three simulation scenarios: a baseline scenario, an ecological protection scenario, and a compact development scenario. It extracts the current total scale and growth rate of construction land from a multi-dimensional spatiotemporal cube as core control parameters for the baseline scenario. For example, it extracts the current total scale of construction land as 500 square kilometers with an average annual growth rate of 3%. Based on current control rules, it increases the development resistance coefficient in ecologically sensitive areas and locks the land use within the permanent basic farmland area. This constructs the parameter configuration set for the ecological protection scenario. For example, it adjusts the development resistance coefficient of construction land within the ecological red line area from 1.0 to 5.0 and marks the land use attribute field of permanent basic farmland plots as "prohibited from conversion." Based on current control rules, it relaxes the construction reserve land indicators outside the urban development boundary and increases the land use indicators for rail transit stations. The maximum floor area ratio (FAR) of the surrounding area is used to construct the parameter configuration set for the compact development scenario. For example, the reserved land quota for construction within 10 kilometers outside the urban development boundary is increased from 50 hectares per square kilometer to 80 hectares, and the maximum FAR of residential land within 800 meters of rail transit stations is increased from 2.5 to 4.0. The core control parameters of the baseline scenario, the parameter configuration set of the ecological protection scenario, and the parameter configuration set of the compact development scenario are written into a preset JSON format scenario configuration file. The configuration file contains key-value pairs such as "maximum construction land quota", "ecological resistance coefficient", "farmland locking indicator", "reserved land quota", and "maximum floor area ratio". The scenario configuration file is sent to the scenario construction module, which parses the constraints in the configuration file and generates three simulation scenarios that can be called by the simulation model. Each simulation scenario is encapsulated as an object containing all constraint parameters.

[0092] In practical implementation, the parallel simulation module initiates a parallel computing framework to perform full-element spatial evolution simulations on three simulation scenarios, outputting a planning simulation trajectory dataset with time-series attributes. Data slices from the multi-dimensional spatiotemporal cube are distributed to various computing nodes in the parallel computing cluster to ensure balanced load on each node. For example, the study area is divided into 12 spatial blocks, each assigned to a computing node. Each computing node processes all raster data within a block. Within each computing node, the corresponding simulation scenario is loaded, and the evolution rule set of the cellular automaton is initialized according to the constraints in the script. The evolution rule set includes neighborhood transition rules, global constraint rules, and random disturbance factors. This drives the cellular automaton model to evolve cyclically within a set simulation period of 20 years, with one year as a time step. After each period, the road network density and public service facility accessibility indicators within the region are updated synchronously. The road network density indicator is dynamically calculated based on newly generated construction land raster cells, and the public service facility accessibility indicator is calculated based on an improved two-step move search method. The cell state transition probability of the cellular automaton model is also considered. The formula is:

[0093]

[0094] in: Let represent the probability that a cell located at position (u,v) is converted to the target land use type at time t. It is a random number that takes a value between 0 and 1. It is a dispersion coefficient that controls the degree of randomness. This indicates the suitability of development based on neighborhood state. Indicates global transfer potential. This represents the spatial constraint factor derived from the scenario configuration.

[0095] In practice, intermediate evolution results from all computing nodes across each simulation cycle are collected and reassembled in chronological order. Each computing node outputs a land use type raster map for its assigned block at the end of each simulation cycle. All block raster maps are then stitched together according to their spatial index to form a complete land use pattern map of the region at that moment. Abnormal fluctuations in the evolution process are removed, and the time series data is smoothed and interpolated to ultimately generate a continuous planning trajectory dataset. For example, this involves detecting and removing land use types that fall within the "arable land - construction land - arable land" range over three consecutive time steps. The isolated grid cells oscillate rapidly. Cubic spline interpolation is used to smooth the time series curves of various land use areas. Optionally, the removal of abnormal oscillation data is based on a set threshold. When the number of state transitions of a grid cell exceeds the threshold of 3 in a short period of time, the change sequence of the grid cell in that period will be regarded as abnormal and replaced with the mean of its previous and subsequent states. It can be understood that the planning and simulation trajectory dataset is a multidimensional array containing the time dimension. The dimensions of the array include time, spatial location and land use type attributes. Optionally, the parallel computing framework uses a message passing interface to realize data synchronization between computing nodes.

[0096] See Figure 3 This is a chart showing the changing trends of construction land scale under three different land use planning scenarios, visually comparing the differences in construction land scale growth between 2024 and 2044. The baseline scenario starts at 500 square kilometers, with an average annual growth rate of approximately 3%, reaching approximately 875 square kilometers in 2044, a growth of about 75% over 20 years. This represents the natural growth path under current control rules, with a stable growth rate, serving as a benchmark for planning evaluation. The ecological protection scenario starts at 500 square kilometers, with an average annual growth rate of approximately 1.5%, reaching approximately 660 square kilometers in 2044, a growth of about 32% over 20 years. By increasing the ecological resistance coefficient and locking in permanent basic farmland, it significantly curbs the expansion of construction land, representing a development path prioritizing ecology. The compact development scenario starts at 500 square kilometers, with an average annual growth rate of approximately 4.5%, reaching approximately 1150 square kilometers in 2044, a growth of about 130% over 20 years. By relaxing the restrictions on reserved land in the perimeter and increasing the plot ratio around the sites, the construction land expands the fastest, representing a high-intensity development path.

[0097] In one embodiment of the present invention, the carrying capacity calculation module reads time-series data on water consumption, energy consumption, and carbon emissions from the planning simulation trajectory dataset. Taking the compact development scenario simulation results of a river valley city in southern China as an example, the planning simulation trajectory dataset contains the annual water consumption, energy consumption, and carbon emissions corresponding to various types of construction land over the next 20 years. The time-series data on water consumption, energy consumption, and carbon emissions are compared with a pre-entered resource and environmental baseline database. The resource and environmental baseline database contains the total exploitable water resources, the upper limit of energy supply, and the carbon sink capacity threshold for the region. The resource supply and demand difference at different simulation times is calculated. For example, in the 10th simulation year, the water consumption is 350 million cubic meters, the total exploitable water resources are 300 million cubic meters, and the water supply and demand difference is -50 million cubic meters. When a certain... When the resource supply and demand difference at a simulated moment exceeds a preset safety threshold, for example, when the negative difference in water resource supply and demand exceeds 0.2 billion cubic meters, the planning and simulation trajectory dataset is reviewed to locate the dominant land use type causing resource overload. The retrospective analysis reveals that the water resource overload in that year was mainly due to the unexpected expansion of industrial and warehousing land. Based on the spatial distribution of industrial and warehousing land, high-risk warning areas for resource carrying capacity are marked on the map with red polygons. These high-risk warning areas are mainly distributed in industrial parks downstream of river valleys. The marked high-risk warning areas and their corresponding overload factors are written into the intermediate data layer of the assessment report. The intermediate data layer stores the geometric boundaries of the high-risk warning areas and the attribute information "overload factor: water resources, overload amount: 0.5 billion cubic meters" in GeoJSON format. See Table 1 for an example of the carrying capacity calculation module's calculation data for the three types of resource pressure.

[0098] Table 1: Example Table of Resource Supply and Demand Calculations in Simulated Year 10 under the Compact Development Scenario

[0099] ;

[0100] In its implementation, the system includes a connectivity assessment module. This module extracts distribution data of ecological source patches and topological data of corridors from the planning and simulation trajectory dataset. For example, it extracts eight ecological source patches, including forests and large lakes, and 12 potential corridors connecting these patches from the baseline scenario's year 20 data. It uses graph theory algorithms to calculate the shortest path distance between ecological source patches and counts the number of corridor breakpoints. The shortest path distance calculation is based on a cost-distance model. Corridor breakpoints refer to corridor segments blocked by construction land. The calculated path distances and breakpoint numbers are then substituted into the landscape connectivity index calculation model to obtain the ecological network connectivity score under different simulation scenarios. The landscape connectivity index calculation model uses a probabilistic connectivity index, with the following formula:

[0101]

[0102] in: Indicates the landscape connectivity index. This indicates the total number of ecological source patches. and They represent plaques. and plaque area or weight, Indicates plaque With plaque The maximum connectivity probability among all possible paths is calculated, determined by the inter-patch distance and land use resistance surface. The calculated ecological network connectivity score is 0.65 for the baseline scenario, 0.78 for the ecological protection scenario, and 0.52 for the compact development scenario. The difference between the ecological network connectivity scores in different scenarios and the baseline scenario is calculated to obtain the connectivity gain / loss value. For example, the connectivity gain / loss value for the ecological protection scenario relative to the baseline scenario is +0.13, and for the compact development scenario it is -0.13. This connectivity gain / loss value is used as a key evaluation indicator and updated to the intermediate data layer of the assessment report. The intermediate data layer records the record "Scenario Name: Ecological Protection, Connectivity Gain / Loss Value: +0.13". In some embodiments, the number of corridor breakpoints is counted based on raster data. A buffer zone of a preset width is scanned for each potential corridor, and the proportion of construction land raster cells is identified. When the proportion exceeds 50%, the location is determined to be a breakpoint. Optionally, the connectivity probability in the landscape connectivity index calculation model... This was calculated using the minimum cumulative resistance model.

[0103] In its implementation, the system includes a conflict resolution module. This module overlays and analyzes the planning trajectory datasets output from three simulation scenarios to identify areas with spatial ownership disputes between different scenarios. For example, by overlaying the base scenario, ecological protection scenario, and compact development scenario on the land use raster map of year 15, it is found that a raster cell numbered G-1024 is planned as construction land, forest land, and cultivated land in the three scenarios, respectively. This raster cell is thus identified as an area with spatial ownership disputes. For each area with spatial ownership disputes, the dominant functional positioning of the area in the three simulation scenarios is extracted, and the reasons for their contradictions are analyzed. For example, the functional positioning of raster cell G-1024 is extracted as "urban development," "ecological conservation," and "agricultural production." The contradiction lies in the fact that this cell is located in the overlapping area of ​​urban development boundary, ecological red line, and permanent basic farmland. In some embodiments, based on the cause of the contradiction, applicable flexible adjustment strategies are retrieved from the existing management rule base, such as the replacement of construction land indicators or the misalignment of development sequence. In response to the conflict involving unit G-1024, the existing control rule base retrieves a "spatial substitution" strategy. This involves reducing one unit of construction land quota in area A and simultaneously adding one unit of construction land quota in area B, thereby releasing unit G-1024 from urban development. This can be understood as feeding the selected flexible adjustment strategy back into the scenario construction module to correct the parameters of the original simulation scenarios. For example, the parameter configuration file for the area where unit G-1024 is located in the compact development scenario is modified, reducing the original planned construction land quota and correspondingly adding construction land quota for another suitable area. The spatial evolution simulation under the parallel computing framework is then re-executed until no new spatial ownership disputes appear in the planning simulation trajectory dataset. After parameter correction and re-simulation, the results of the three scenarios are overlaid and analyzed again. In the compact development scenario, unit G-1024 has been adjusted to cultivated land or forest land, and the spatial ownership conflict with other scenarios has been eliminated. Optionally, the overlay analysis of the conflict handling module uses a raster calculator tool to identify raster locations with inconsistent attribute values ​​through map algebra methods.

[0104] See Figure 4This is a chart showing the 20-year trend of resource carrying capacity under a compact development scenario, illustrating the normalized supply-demand difference trends of the three core resources from 2025 to 2045. It visually reflects the dynamic evolution of resource overload risk. Water resources will decline continuously from 2025 to 2034, bottoming out in 2034 and gradually recovering, returning to near initial levels by 2045. Energy resources will decline continuously from 2025 to 2037, bottoming out in 2037 and recovering, with overall fluctuations smaller than those of water resources and carbon emissions. Carbon emissions will rise rapidly from 2025 to 2036, peaking in 2036 and then slowly declining, representing the highest supply-demand difference among the three resources. The water resource overload threshold is defined as a water resource overload when the supply-demand difference falls below this threshold. Water resources and carbon emissions are the core overload risk points, requiring priority control of industrial and warehousing land expansion, optimization of the energy structure, and enhancement of carbon sequestration capacity.

[0105] In one embodiment of the present invention, the results visualization module extracts land use snapshots of key time nodes from the planning and simulation trajectory dataset and renders them as a three-dimensional terrain scene. Taking the simulation results of an ecological protection scenario in a certain area as an example, land use raster data of five time nodes—year 0, year 5, year 10, year 15, and year 20—are extracted from the planning and simulation trajectory dataset as land use snapshots of key time nodes. The land use snapshots are overlaid with the corresponding digital elevation model data, and a three-dimensional terrain scene with topographic relief is generated in the three-dimensional graphics engine. The resolution of the digital elevation model data is 10 meters. The rendering of the three-dimensional terrain scene adopts perspective projection. In the three-dimensional terrain scene, differentiated texture maps and lighting effects are assigned according to different land use properties to enhance the sense of spatial hierarchy. For example, green vegetation texture maps are assigned to raster areas of cultivated land, gray artificial surface texture maps are assigned to raster areas of construction land, and blue water texture maps are assigned to raster areas of water bodies. The lighting effect is dynamically calculated based on the solar altitude angle and azimuth angle parameters, generating shadows at the topographic relief. The synthesis of texture maps and lighting effects follows the formula:

[0106]

[0107] in: This represents the final display brightness value of the 3D terrain scene at pixel coordinates (x, y). This indicates the number of light sources that affect the scene display. This represents the intensity of the k-th light source. Indicates the texture map at wavelength The reflectance function is determined by the texture color. Let represent the surface normal vector at pixel (x,y). This represents the direction vector from pixel (x,y) to the k-th light source.

[0108] In practice, the high-risk warning areas obtained from the resource carrying capacity calculation step and the connectivity gain / loss values ​​obtained from the connectivity evaluation step are overlaid on the 3D terrain scene in the form of a heatmap. The high-risk warning areas obtained from the resource carrying capacity calculation step are overlaid with a semi-transparent red gradient layer, and the layer transparency is positively correlated with the severity of resource overload. The connectivity gain / loss values ​​obtained from the connectivity evaluation step are represented by a gradient color band from blue to red, with negative connectivity gain / loss values ​​displayed in blue and positive connectivity gain / loss values ​​displayed in red. A mapping relationship is established between the color band and the specific connectivity gain / loss values. In some embodiments, the overlay of the heatmap layer uses alpha mixing technology, which is integrated with the underlying 3D terrain field. The scene is pixel-level fused to generate an interactive visualization interface, allowing operators to switch between different simulation moments using a timeline slider and dynamically observe the evolution of the planning scheme. The interactive visualization interface is built on WebGL technology and includes a timeline slider control. The slider position corresponds to the simulation year. When the operator drags the slider, the land use snapshot, high-risk warning area heat map, and connectivity gain / loss value heat map displayed in the 3D terrain scene are synchronously updated to the data state of the corresponding year. The operator can rotate and zoom the 3D terrain scene 360 ​​degrees to observe the evolution of the spatial pattern from different perspectives. It can be understood that the movement of the timeline slider triggers data reloading and rendering pipeline updates.

[0109] In practice, the generated 3D visualization scene screenshots and animation sequences are exported as attachments to the final report. The 3D visualization scene screenshots are saved in PNG format with a resolution of 1920x1080 pixels. The screenshots contain static images of the 3D scene from a specific perspective and at a specific simulation moment. The animation sequence is generated by automatically and continuously playing scene changes along the timeline and exported in MP4 video format with a frame rate of 30 frames per second and a duration of 30 seconds, showing the complete evolution from year 0 to year 20. Optionally, the camera flight path can be locked when exporting the animation sequence to give the animation a coherent roaming perspective. In some embodiments, the results visualization module allows adding annotations such as legends, scale bars, and titles to the screenshots and animations before exporting. It is understood that the rendering detail level of the 3D terrain scene can be dynamically adjusted according to the performance of the computer graphics hardware to ensure smooth interaction. Optionally, the exported image and video files are stored according to a preset naming rule, which includes scene name, time node, and file type information.

[0110] See Figure 5This is a 20-year land use type area change trend map, showing the dynamic changes in the area of ​​four land use types—arable land, construction land, water area, and forest land—during the 0–20 year simulation period, intuitively reflecting the evolution of land use structure. Construction land expanded, increasing from 1200 grid cells to 4900, an increase of 308%, making it the only land use type that continued to grow, reflecting the strong demand for construction land from urbanization and industrialization. Arable land and forest land decreased; arable land decreased from 6000 to 3900, a reduction of 35%, and forest land decreased from 1600 to 0, completely replaced by other land use types. The reduction in both is highly matched with the expansion of construction land, reflecting the typical land use conversion pattern of "urban construction occupying agricultural and forestry land." Water area remained stable, with its area remaining constant throughout the 20-year simulation period, indicating that the region has implemented strict protection of its aquatic ecological space.

[0111] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A land planning simulation and evaluation system based on multi-source data, characterized in that: include: The data standardization module collects basic geospatial data, socio-economic operation data, and ecological environment monitoring data of the target area. It performs coordinate system unification and format conversion on the collected multi-source heterogeneous data to form a standardized planning base map dataset. The spatiotemporal fusion module imports the standardized planning base map dataset into the spatiotemporal data fusion engine, aligns the data based on administrative division boundaries and timestamp fields, and constructs a multi-dimensional spatiotemporal cube covering the entire region. The conflict identification module overlays the current management rules of the national land space planning onto the multi-dimensional spatiotemporal cube to identify spatially overlapping conflict points between ecological protection red line areas, permanent basic farmland protection areas, and urban development boundary areas. The situation prediction module calls the land use evolution simulation model, reads the historical land use change trajectory in the multi-dimensional spatiotemporal cube, and predicts the natural growth trend under the condition of no new planning intervention. The scenario construction module takes the existing control rules and the natural growth trend as input into the scenario construction module to generate three simulation scenarios: the baseline scenario, the ecological protection scenario, and the compact development scenario. The parallel simulation module initiates a parallel computing framework to perform full-element spatial evolution simulations on three simulation scenarios, outputting a planning simulation trajectory dataset with time-series attributes, specifically including: Data slices from the multidimensional spatiotemporal cube are distributed to various computing nodes in the parallel computing cluster to ensure balanced load on each node. Within each computing node, the corresponding simulation scenario is loaded, and the evolution rule set of the cellular automaton is initialized according to the constraints in the script. The cellular automaton model is driven to evolve cyclically within a set simulation period, and the road network density and public service facility accessibility indicators in the region are updated synchronously after each period. The intermediate evolution results of all computing nodes in each simulation period are collected and spliced ​​and recombined in the order of the time series. Abnormal oscillation data during the evolution process are removed, and the time series data is smoothed by interpolation to finally generate a continuous planning simulation trajectory dataset. The system also includes a load-bearing capacity calculation module, used for: Read time-series data on water consumption, energy consumption, and carbon emissions from the planning simulation trajectory dataset; compare the time-series data with a pre-entered resource and environmental baseline database to calculate the resource supply-demand difference at different simulation times; when the resource supply-demand difference at a certain simulation time exceeds a preset safety threshold, backtrack the planning simulation trajectory dataset to locate the dominant land use type causing resource overload; based on the spatial distribution of the dominant land use type, mark high-risk warning areas for resource carrying capacity on a map; write the marked high-risk warning areas and their corresponding overload factors into the intermediate data layer of the assessment report; The system also includes a connectivity assessment module for: Distribution data of ecological source patches and topological structure data of corridors are extracted from the planning and simulation trajectory dataset. Graph theory algorithms are used to calculate the shortest path distance between the ecological source patches and to count the number of breakpoints in the corridors. The calculated path distances and breakpoint numbers are substituted into the landscape connectivity index calculation model to obtain ecological network connectivity scores under different simulation scenarios. The difference between the ecological network connectivity scores under different simulation scenarios and the scores under the baseline scenario is calculated to obtain connectivity gain / loss values. These connectivity gain / loss values ​​are used as key evaluation indicators and updated to the intermediate data layer of the evaluation report. The system also includes a conflict resolution module for: The planning and simulation trajectory datasets output from the three simulation scenarios are overlaid and analyzed to identify areas with spatial ownership disputes among different scenarios. For each area with spatial ownership disputes, the dominant functional positioning of the area in the three simulation scenarios is extracted, and the reasons for their contradictions are analyzed. Based on the reasons for the contradictions, applicable flexible adjustment strategies are retrieved from the existing control rule base. The selected flexible adjustment strategies are fed back to the scenario construction module to correct the parameters of the original simulation scenarios. The spatial evolution simulation under the parallel computing framework is re-executed until no new spatial ownership disputes appear in the planning and simulation trajectory dataset.

2. The land planning simulation and evaluation system based on multi-source data as described in claim 1, characterized in that, The process of performing coordinate system and format conversion on the collected multi-source heterogeneous data to form a standardized planning base map dataset specifically includes: Read the vector layer files in the basic geospatial data, extract the geometric information and attribute tables, and convert the coordinate system from the local independent coordinate system to the national geodetic coordinate system; Analyze statistical yearbook tables in socio-economic operation data, transform unstructured text descriptions into numerical indicators, and convert them according to a unified unit of measurement. Radiometric calibration and atmospheric correction are performed on remote sensing images in ecological and environmental monitoring data to generate raster data with uniform resolution. Establish globally unique data identifiers for the vector layers, numerical indicators, and raster data, and record their source and acquisition time in the metadata; The processed vector layers, numerical indicators, and raster data are packaged and stored to form a structured, standardized planning base map dataset.

3. The land planning simulation and evaluation system based on multi-source data as described in claim 2, characterized in that, The process of calling the land use evolution simulation model, reading the historical land use change trajectory in the multi-dimensional spatiotemporal cube, and predicting the natural growth trend without new planning intervention specifically includes: Land use status maps from several consecutive years in the past are retrieved from the multidimensional spatiotemporal cube and converted into discrete raster cell matrices. The frequency of land use type transfers for each grid cell within a historical period is statistically analyzed to generate a Markov transfer probability matrix that reflects the inertia of urban expansion. By combining the population density gradient and road accessibility data in the multidimensional spatiotemporal cube, a dynamic suitability weight is assigned to each grid cell; The Markov transition probability matrix and the suitability weights are used as input variables and loaded into the land use evolution simulation model for iterative calculation. Extract the final land use distribution pattern from the simulation model output and label it as the natural growth trend.

4. The land planning simulation and evaluation system based on multi-source data as described in claim 3, characterized in that, Combining population density gradient and road accessibility data within the multidimensional spatiotemporal cube, dynamic suitability weights are assigned to each grid cell, specifically including: Extract a population density raster layer with the same spatial range and resolution as the raster unit matrix from the multidimensional spatiotemporal cube, and read the population density value corresponding to the center point of each raster unit. Road network vector data is extracted from the multidimensional spatiotemporal cube, the Euclidean distance from the center point of each grid cell to the nearest main road is calculated, and the Euclidean distance is mapped to a road accessibility index. The population density value is normalized to eliminate the influence of dimensions and converted into a population density influence factor between 0 and 1. The road accessibility index is subjected to inverse distance attenuation processing, where the closer the distance, the higher the accessibility. It is also converted into a road accessibility influence factor between 0 and 1. Based on the pre-set land use type conversion driving model, the contribution coefficients of population density influence factor and road accessibility influence factor are respectively assigned to different types of land use conversion behavior; Based on the contribution coefficient, the population density influence factor and road accessibility influence factor of each grid cell are weighted and summed to calculate the dynamic suitability weight of the grid cell under specific land use conversion behavior.

5. The land planning simulation and evaluation system based on multi-source data as described in claim 4, characterized in that, The module that inputs existing management rules and the natural growth trend into a scenario construction module generates three simulation scenarios: a baseline scenario, an ecological protection scenario, and a compact development scenario. Specifically, these include: The current total scale and growth rate of construction land are extracted from the multidimensional spatiotemporal cube and used as the core control parameters of the baseline scenario. Based on the existing control rules, the development resistance coefficient of ecologically sensitive areas will be increased, and the land use within the permanent basic farmland area will be locked to construct a parameter configuration set for ecological protection scenarios. Based on the existing control rules, the quota for reserved land for construction outside the urban development boundary will be relaxed, and the upper limit of plot ratio around rail transit stations will be increased, so as to construct a parameter configuration set for compact development scenarios. The core control parameters of the baseline scenario, the parameter configuration set of the ecological protection scenario, and the parameter configuration set of the compact development scenario are respectively written into the preset scenario configuration file. The scenario configuration file is sent to the scenario building module, which parses the constraints in the configuration file and generates three simulation scenarios that can be called by the simulation model.

6. The land planning simulation and evaluation system based on multi-source data as described in claim 5, characterized in that, It also includes a results visualization module, used for: Extract land use snapshots at key time points from the planning and simulation trajectory dataset and render them as a three-dimensional terrain scene; In the three-dimensional terrain scene, differentiated texture maps and lighting effects are applied according to different land use properties to enhance the sense of spatial hierarchy; The high-risk warning areas obtained from the resource carrying capacity calculation step and the connectivity gain / loss values ​​obtained from the connectivity evaluation step are overlaid on the three-dimensional terrain scene in the form of a heat map. Generate an interactive visualization interface that allows operators to switch between different simulation moments using a timeline slider and dynamically observe the evolution of the planning scheme; Export the generated 3D visualization scene screenshots and animation sequences as attachments to the final report.

Images