An artificial intelligence-based land element optimal allocation decision system and method

Abstract

The application relates to the field of land management, in particular to a land element optimized allocation decision system and method based on artificial intelligence, which comprises a land remote sensing module, a region segmentation module, a development prediction module, a virtual merging module and an allocation decision module; the land remote sensing module is used for rasterizing land and inputting land element information; the region segmentation module is used for generating development regions and calculating shape coefficients; the development prediction module is used for calculating the output efficiency of land grids; the virtual merging module is used for merging adjacent development regions; and the allocation decision module is used for planning the development type of conflict grids; the application can reduce the waste of land resources, avoid the low-efficiency utilization caused by land dispersion, improve the land utilization efficiency, reduce the land development and management cost, ensure the scientificity and accuracy of land allocation, and realize the efficient utilization of land resources and the coordinated development of ecological environment.

Description

Technical Field

[0001] This invention relates to the field of land management, specifically to an artificial intelligence-based decision-making system and method for optimizing the allocation of land resources. Background Technology

[0002] Land elements refer to the basic production elements of land resources, including natural attributes such as area, location, fertility, and geological conditions, as well as socio-economic attributes such as use, benefits, and development level. Optimizing the allocation of land elements requires scientific and rational planning and management to achieve the optimal combination of land resources in space, time, and use, thereby realizing the efficient utilization of land resources.

[0003] During land development, due to factors such as ecological environment, land type, and development history, development areas often exhibit a fragmented landscape of small, interspersed land parcels, a phenomenon known as land fragmentation. Fragmented land can negatively impact industrial agglomeration, leading to a decrease in total land output. Furthermore, the differences in land resources across different regions make effective integration of fragmented land difficult, thus affecting land use efficiency.

[0004] Furthermore, the land planning process involves changes in land type, requiring a complex forecasting process before making land allocation decisions. Most forecasting models are time-consuming and the forecast results are not accurate enough, failing to allocate land development types in the most favorable way, and the level of intelligence in decision-making is not high. Summary of the Invention

[0005] The purpose of this invention is to provide a land resource optimization allocation decision system and method based on artificial intelligence to solve the problems mentioned in the background art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a land element optimization allocation decision system based on artificial intelligence, comprising: a land remote sensing module, a region segmentation module, a development prediction module, a virtual merging module, and an allocation decision module;

[0007] The land remote sensing module is used to input regional statistical data into the GIS geographic information system, establish a coordinate system in the satellite remote sensing image of the land to be developed, rasterize the land according to the coordinates, remove the raster of undevelopable land, and input the land element information of each land raster. The land elements include: land development type, land area, land topography, vegetation distribution, population composition, land input and land output.

[0008] The region segmentation module is used to merge rasters of the same land development type to form a development region. The land development type of all adjacent rasters in the development region is different from that of the development region. Using intelligent image segmentation technology, the development region is segmented from the remote sensing image and the region edges are connected. The shape coefficient of each development region is calculated based on the perimeter of the edge and the area of ​​the development region.

[0009] The development prediction module is used to obtain the regional agglomeration effect based on the shape coefficient of the development area and the agglomeration effect coefficient of the industry. It calculates the output efficiency of each land grid based on the input-output ratio, agglomeration effect, land adaptability, development cost and environmental loss of each development type. The output efficiency of all land grids is accumulated to obtain the overall output efficiency, and a mapping relationship between the overall output efficiency and the output efficiency of land grids is established.

[0010] The virtual merging module is used to mark development areas with domain homogeneity below a threshold. The development area that is closest to the connected area and has the same development type is used as the merging area. The shortest path between the connected area and the merging area is established using an artificial intelligence model. The grids passed through by the shortest path are merged into the connected area to connect the connected area and the merging area. Grids around the shortest path are selected to be merged into the connected area to increase the overall output efficiency of the region until the total land efficiency index decreases after any adjacent grid is added to the merging area. All newly merged land grids are marked.

[0011] The configuration decision module is used to identify land graticles with more than one marker as conflict graticles. Using development type as a variable, the conflict graticles are weighted by the output efficiency of each development type. An output influence function is constructed, and planning is carried out with the maximum output efficiency of the entire region as the planning condition and land and environmental constraints as constraints. The planning status of each conflict graticle is determined, and the decision results of the development type of each land graticle are output.

[0012] Furthermore, the land remote sensing module includes: a GIS unit and an element statistics unit;

[0013] The GIS unit is used to acquire satellite remote sensing images of the area to be developed, establish regional coordinates, and rasterize the land according to the coordinates;

[0014] The element statistics unit is used to determine the land element information of each grid from satellite imagery, development logs, and land statistics data.

[0015] Furthermore, the region segmentation module includes: a rasterization unit, a boundary recognition unit, and a shape evaluation unit;

[0016] The gridded unit is used to merge grids according to land development type to form a development area. When there is more than one development type in a grid, the development type with the largest area is taken as the development type of the grid.

[0017] The boundary recognition unit is used to segment the boundary of the development area from satellite remote sensing images using an image recognition and segmentation model;

[0018] The shape evaluation unit is used to calculate the shape coefficient of the development area based on the perimeter and area of ​​the development area boundary.

[0019] Furthermore, the development prediction module includes: a data decision-making unit and an efficiency evaluation unit;

[0020] The data decision unit is used to calculate the agglomeration effect of each development area and to make decisions on changes in development type;

[0021] The efficiency evaluation unit is used to calculate the output efficiency of each land grid and all undeveloped areas.

[0022] Furthermore, the virtual merging module includes: an intelligent addressing unit and a grid marking unit;

[0023] The intelligent addressing unit is used to access the artificial intelligence service interface and perform shortest path planning for regional connectivity between the two development areas.

[0024] The grid marking unit is used to determine the range of the merged area to maximize the overall output efficiency and to mark the newly merged grid.

[0025] Furthermore, the configuration decision module includes: a conflict grid unit, an exponential planning unit, and a grid configuration unit;

[0026] The conflict grid cell is used to determine the output efficiency of each development type in the conflict grid configuration and generate an output impact function.

[0027] The index planning unit is used to plan the development type of conflict grids to maximize the overall output efficiency.

[0028] The grid configuration unit is used to update the remote sensing map, output the new development type for each grid, and provide the total development cost.

[0029] An artificial intelligence-based method for optimizing land resource allocation includes the following steps:

[0030] Step S1. Generate satellite remote sensing images of the area to be developed, establish a coordinate system in the images, rasterize the land according to the coordinates, remove the raster of undevelopable land, and determine the land features of each raster from satellite images, development logs, and land statistics data.

[0031] Step S2. Merge rasters with the same land development type to form development areas. The development areas are connected regions and there are no rasters of the same development type around the development areas. Use an image segmentation model to segment the boundaries of each development area from the remote sensing image. Calculate the shape coefficient of the development area based on the perimeter and area of ​​the boundary.

[0032] Step S3. Multiply the shape coefficient of the development area and the agglomeration effect coefficient of the industry to obtain the regional agglomeration effect. For each type of development, the development parameters and the regional agglomeration effect are weighted and summed to obtain the output efficiency of each land grid and the overall output efficiency.

[0033] Step S4. Select two development areas with the same development type and domain homogeneity higher than the threshold. Use an artificial intelligence model to plan the minimum cost connection path between the two development areas. Merge the grids that the path passes through into the two development areas to form a connected area. Under the premise of increasing the overall output efficiency, merge the adjacent grids of the connected area and mark all merged land grids.

[0034] Step S5. Land rasters with more than one marker are designated as conflict rasters. Planning is carried out with the maximum overall output efficiency as the planning condition and land and environmental constraints as the constraints. The development type change decision for each conflict raster is then output.

[0035] Furthermore, step S1 includes:

[0036] Step S11. Use a GIS geographic information system to acquire satellite remote sensing images of the area to be developed, establish a coordinate system for the remote sensing images, and divide the remote sensing images into square grids with a side length of a, where a is the preset grid width. The number of grids divided is greater than one, and remove grids representing undevelopable land.

[0037] Step S12. Obtain the land area, land topography and vegetation distribution type of the region from the remote sensing image, and obtain the land development type, population composition, land input and land output from the development log and land statistics. When there is more than one development type in the raster, the development type with the largest area is taken as the development type of the raster.

[0038] Each grid cell is labeled with land features, including: land development type, land area, land topography, vegetation distribution, population composition, land input, and land output.

[0039] Furthermore, step S2 includes:

[0040] Step S21. Merge grids according to land development type. Perform the merging method once for each grid to obtain a development area. The merging method is as follows: when there are adjacent grids of the same development type in the current grid, merge the current grid with the adjacent grids to form a region core. If there are adjacent grids of the same development type around the region core, merge the adjacent grids into the region core. If there are no adjacent grids of the same development type around the region core, the region core is transformed into a development area.

[0041] Step S22. Using an image segmentation model, segment the boundaries of each development area from the remote sensing image, obtain the perimeter and area of ​​the boundaries, and calculate the shape coefficient of the development area:

[0042]

[0043] Where, δ k Y represents the shape coefficient of the development area. k A represents the perimeter of the development area. k represents the area of ​​the development region, r represents the raster scaling factor, and c represents the number of rasters contained in the development region.

[0044] Furthermore, step S3 includes:

[0045] Step S31. Calculate the land agglomeration effect of each development area, satisfying P = u·δ k Where P represents the agglomeration effect, and u is the agglomeration effect coefficient of the development type. The development types include: cultivated land, forest land, construction land, transportation land, industrial and mining land and water use land. The agglomeration effect coefficient is determined by the ratio of the output increment to the area increment of the corresponding development type.

[0046] Step S32. Obtain development parameters for each development type, including: input-output ratio, agglomeration effect, land adaptability, development cost, and environmental loss, and determine the output efficiency of the land grid.

[0047]

[0048] Where μ is the output efficiency of the land raster, N is the number of development parameters, and F i For the i-th development parameter value, R i The weight of the i-th development parameter;

[0049] The global output efficiency is obtained by summing the output efficiencies of all grids.

[0050] Furthermore, step S4 includes:

[0051] Step S41. Mark two development regions with the same development type and domain identity below a threshold. The domain identity represents the minimum boundary distance between the two regions. Input the change cost of each grid development type into the artificial intelligence model, plan the connection path between the two regions, and minimize the cumulative change cost of all grids traversed by the connection path.

[0052] Step S42. Merge all the grids traversed by the connected path into the two development regions to obtain the connected region. Simulate each grid around the connected region. If the global output efficiency increases after merging the grid, then merge the grid into the connected region. When there are no grids around the connected region that can be merged, update the connected region and mark all the grids merged in steps S41 to S42.

[0053] Furthermore, step S5 includes:

[0054] Step S51. Take land grids with more than one label as conflict grids, determine the development type of the conflict based on the label of the conflict grid, calculate the output efficiency increment of the conflict grid as a development type, and obtain the output efficiency factor μ(x), where x represents the type of conflict development type, and μ(x) represents the increment of the global output efficiency when the conflict grid adopts the xth development type.

[0055] Step S52. Using the maximization of overall output efficiency as the planning condition, and land and environmental constraints as the constraints, plan for each conflict grid:

[0056]

[0057] Where S represents the overall output efficiency, M1 and M2 represent the number of developed regions and the number of conflict grids, respectively, and Z... j μ represents the output efficiency of the non-collision grid portion in the j-th development region. k (x) represents the output efficiency factor of the k-th conflicting grid, c x and S x c1 and c2 represent the number of land grids and output efficiency of all land using the x-th development type, respectively. c1 and c2 are the upper and lower limits of the number of grids, respectively. S0 is the lower limit of output efficiency. H(S) is a mandatory constraint condition determined by the development policy. H0 is a mandatory constraint parameter.

[0058] Step S53. Determine the development type of each conflict grid to maximize the overall output efficiency while satisfying the constraints, and output the decision results for all land grids.

[0059] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0060] 1. This invention can utilize GIS remote sensing technology and statistical data to rasterize land according to development type, merge raster areas with the same development type, and calculate the shape coefficient and neighborhood identity of the area. By integrating fragmented land, it can reduce the waste of land resources, avoid inefficient use caused by land dispersion, and improve land use efficiency.

[0061] 2. This invention can calculate the efficiency index of each land grid based on the land elements of each land grid, and take land areas with a domain homogeneity higher than the threshold as merged areas. It selects grids around the merged areas to add, thereby increasing the overall efficiency index of the region. It can intelligently select various types of development areas to maximize the benefits of the land allocation process, exchange the same amount of land input for greater resource output, and reduce land development and management costs.

[0062] 3. This invention constructs a global efficiency index function by using conflict areas as variables and the efficiency index increment of each conflict area in the conflict grid as weights. Planning is carried out with the maximum of the global efficiency index function as the planning condition and land and environmental constraints as constraints. Iterative decision-making is performed on the planning results to ensure the scientific and accurate allocation of land resources and achieve efficient use of land resources and coordinated development of the ecological environment. Attached Figure Description

[0063] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0064] Figure 1 This is a schematic diagram of the structure of a land element optimization allocation decision system based on artificial intelligence according to the present invention;

[0065] Figure 2 This is a schematic diagram illustrating the steps of a land element optimization allocation decision-making method based on artificial intelligence according to the present invention. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0067] Please see Figure 1 The present invention provides a technical solution: a land element optimization allocation decision system based on artificial intelligence, comprising: a land remote sensing module, a region segmentation module, a development prediction module, a virtual merging module, and an allocation decision module;

[0068] The land remote sensing module is used to input regional statistical data into the GIS geographic information system, establish a coordinate system in the satellite remote sensing image of the land to be developed, rasterize the land according to the coordinates, remove the raster of undevelopable land, and input the land element information of each land raster. The land elements include: land development type, land area, land topography, vegetation distribution, population composition, land input and land output.

[0069] The land remote sensing module includes: GIS unit and feature statistics unit;

[0070] The GIS unit is used to acquire satellite remote sensing images of the area to be developed, establish regional coordinates, and rasterize the land according to the coordinates;

[0071] The element statistics unit is used to determine the land element information of each grid from satellite imagery, development logs, and land statistics data.

[0072] The region segmentation module is used to merge rasters of the same land development type to form a development region. The land development type of all adjacent rasters in the development region is different from that of the development region. Using intelligent image segmentation technology, the development region is segmented from the remote sensing image and the region edges are connected. The shape coefficient of each development region is calculated based on the perimeter of the edge and the area of ​​the development region.

[0073] The region segmentation module includes: a rasterization unit, a boundary recognition unit, and a shape evaluation unit;

[0074] The gridded unit is used to merge grids according to land development type to form a development area. When there is more than one development type in a grid, the development type with the largest area is taken as the development type of the grid.

[0075] The boundary recognition unit is used to segment the boundary of the development area from satellite remote sensing images using an image recognition and segmentation model;

[0076] The shape evaluation unit is used to calculate the shape coefficient of the development area based on the perimeter and area of ​​the development area boundary.

[0077] The development prediction module is used to obtain the regional agglomeration effect based on the shape coefficient of the development area and the agglomeration effect coefficient of the industry. It calculates the output efficiency of each land grid based on the input-output ratio, agglomeration effect, land adaptability, development cost and environmental loss of each development type. The output efficiency of all land grids is accumulated to obtain the overall output efficiency, and a mapping relationship between the overall output efficiency and the output efficiency of land grids is established.

[0078] The development prediction module includes: a data decision-making unit and an efficiency evaluation unit;

[0079] The data decision unit is used to calculate the agglomeration effect of each development area and to make decisions on changes in development type;

[0080] The efficiency evaluation unit is used to calculate the output efficiency of each land grid and all undeveloped areas.

[0081] The virtual merging module is used to mark development areas with domain homogeneity below a threshold. The development area that is closest to the connected area and has the same development type is used as the merging area. The shortest path between the connected area and the merging area is established using an artificial intelligence model. The grids passed through by the shortest path are merged into the connected area to connect the connected area and the merging area. Grids around the shortest path are selected to be merged into the connected area to increase the overall output efficiency of the region until the total land efficiency index decreases after any adjacent grid is added to the merging area. All newly merged land grids are marked.

[0082] The virtual merging module includes: an intelligent addressing unit and a grid marking unit;

[0083] The intelligent addressing unit is used to access the artificial intelligence service interface and perform shortest path planning for regional connectivity between the two development areas.

[0084] The grid marking unit is used to determine the range of the merged area to maximize the overall output efficiency and to mark the newly merged grid.

[0085] The configuration decision module is used to identify land graticles with more than one marker as conflict graticles. Using development type as a variable, the conflict graticles are weighted by the output efficiency of each development type. An output influence function is constructed, and planning is carried out with the maximum output efficiency of the entire region as the planning condition and land and environmental constraints as constraints. The planning status of each conflict graticle is determined, and the decision results of the development type of each land graticle are output.

[0086] The configuration decision module includes: conflict grid cells, exponential planning cells, and grid configuration cells;

[0087] The conflict grid cell is used to determine the output efficiency of each development type in the conflict grid configuration and generate an output impact function.

[0088] The index planning unit is used to plan the development type of conflict grids to maximize the overall output efficiency.

[0089] The grid configuration unit is used to update the remote sensing map, output the new development type for each grid, and provide the total development cost.

[0090] like Figure 2 As shown, an artificial intelligence-based land resource optimization decision-making method includes the following steps:

[0091] Step S1. Generate satellite remote sensing images of the area to be developed, establish a coordinate system in the images, rasterize the land according to the coordinates, remove the raster of undevelopable land, and determine the land features of each raster from satellite images, development logs, and land statistics data.

[0092] Step S1 includes:

[0093] Step S11. Use a GIS geographic information system to acquire satellite remote sensing images of the area to be developed, establish a coordinate system for the remote sensing images, and divide the remote sensing images into square grids with a side length of a, where a is the preset grid width. The number of grids divided is greater than one, and remove grids representing undevelopable land.

[0094] Step S12. Obtain the land area, land topography and vegetation distribution type of the region from the remote sensing image, and obtain the land development type, population composition, land input and land output from the development log and land statistics. When there is more than one development type in the raster, the development type with the largest area is taken as the development type of the raster.

[0095] Each grid cell is labeled with land features, including: land development type, land area, land topography, vegetation distribution, population composition, land input, and land output.

[0096] Step S2. Merge rasters with the same land development type to form development areas. The development areas are connected regions and there are no rasters of the same development type around the development areas. Use an image segmentation model to segment the boundaries of each development area from the remote sensing image. Calculate the shape coefficient of the development area based on the perimeter and area of ​​the boundary.

[0097] Step S2 includes:

[0098] Step S21. Merge grids according to land development type. Perform the merging method once for each grid to obtain a development area. The merging method is as follows: when there are adjacent grids of the same development type in the current grid, merge the current grid with the adjacent grids to form a region core. If there are adjacent grids of the same development type around the region core, merge the adjacent grids into the region core. If there are no adjacent grids of the same development type around the region core, the region core is transformed into a development area.

[0099] Step S22. Using an image segmentation model, segment the boundaries of each development area from the remote sensing image, obtain the perimeter and area of ​​the boundaries, and calculate the shape coefficient of the development area:

[0100]

[0101] Where, δ k Y represents the shape coefficient of the development area. kA represents the perimeter of the development area. k represents the area of ​​the development region, r represents the raster scaling factor, and c represents the number of rasters contained in the development region.

[0102] Step S3. Multiply the shape coefficient of the development area and the agglomeration effect coefficient of the industry to obtain the regional agglomeration effect. For each type of development, the development parameters and the regional agglomeration effect are weighted and summed to obtain the output efficiency of each land grid and the overall output efficiency.

[0103] Step S3 includes:

[0104] Step S31. Calculate the land agglomeration effect of each development area, satisfying P = u·δ k Where P represents the agglomeration effect, and u is the agglomeration effect coefficient of the development type. The development types include: cultivated land, forest land, construction land, transportation land, industrial and mining land and water use land. The agglomeration effect coefficient is determined by the ratio of the output increment to the area increment of the corresponding development type.

[0105] Step S32. Obtain development parameters for each development type, including: input-output ratio, agglomeration effect, land adaptability, development cost, and environmental loss, and determine the output efficiency of the land grid.

[0106]

[0107] Where μ is the output efficiency of the land raster, N is the number of development parameters, and F i For the i-th development parameter value, R i The weight of the i-th development parameter;

[0108] The global output efficiency is obtained by summing the output efficiencies of all grids.

[0109] Step S4. Select two development areas with the same development type and domain homogeneity higher than the threshold. Use an artificial intelligence model to plan the minimum cost connection path between the two development areas. Merge the grids that the path passes through into the two development areas to form a connected area. Under the premise of increasing the overall output efficiency, merge the adjacent grids of the connected area and mark all merged land grids.

[0110] Step S4 includes:

[0111] Step S41. Mark two development regions with the same development type and domain identity below a threshold. The domain identity represents the minimum boundary distance between the two regions. Input the change cost of each grid development type into the artificial intelligence model, plan the connection path between the two regions, and minimize the cumulative change cost of all grids traversed by the connection path.

[0112] Step S42. Merge all the grids traversed by the connected path into the two development regions to obtain the connected region. Simulate each grid around the connected region. If the global output efficiency increases after merging the grid, then merge the grid into the connected region. When there are no grids around the connected region that can be merged, update the connected region and mark all the grids merged in steps S41 to S42.

[0113] Step S5. Land rasters with more than one marker are designated as conflict rasters. Planning is carried out with the maximum overall output efficiency as the planning condition and land and environmental constraints as the constraints. The development type change decision for each conflict raster is then output.

[0114] Step S5 includes:

[0115] Step S51. Take land grids with more than one label as conflict grids, determine the development type of the conflict based on the label of the conflict grid, calculate the output efficiency increment of the conflict grid as a development type, and obtain the output efficiency factor μ(x), where x represents the type of conflict development type, and μ(x) represents the increment of the global output efficiency when the conflict grid adopts the xth development type.

[0116] Step S52. Using the maximization of overall output efficiency as the planning condition, and land and environmental constraints as the constraints, plan for each conflict grid:

[0117]

[0118] Where S represents the overall output efficiency, M1 and M2 represent the number of developed regions and the number of conflict grids, respectively, and Z... j μ represents the output efficiency of the non-collision grid portion in the j-th development region. k (x) represents the output efficiency factor of the k-th conflicting grid, c x and S x c1 and c2 represent the number of land grids and output efficiency of all land using the x-th development type, respectively. c1 and c2 are the upper and lower limits of the number of grids, respectively. S0 is the lower limit of output efficiency. H(S) is a mandatory constraint condition determined by the development policy. H0 is a mandatory constraint parameter.

[0119] Step S53. Determine the development type of each conflict grid to maximize the overall output efficiency while satisfying the constraints, and output the decision results for all land grids.

[0120] Example: A development area contains 10 grids, including 2 arable land, 3 forest land, 3 industrial and mining land, 1 building land, and 1 water use land. The regulations stipulate that there must be at least 2 arable land grids and at least 1 water use land grid. The output efficiencies of arable land, forest land, industrial and mining land, building land, and water use land are 0.2, 0.1, 0.3, 0.5, and 0.1, respectively. Therefore, the overall output efficiency is 2.2. There is one conflicting grid in the forest land. If the grid is arable land, the overall output efficiency becomes 2.3; if it is building land, the overall output efficiency becomes 2.5. Therefore, the development type of the conflicting grid is changed from forest land to building land.

[0121] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0122] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A land resource optimization allocation decision-making method based on artificial intelligence, characterized in that, The method includes the following steps: Step S1. Generate satellite remote sensing images of the area to be developed, establish a coordinate system in the images, rasterize the land according to the coordinates, remove the raster of undevelopable land, and determine the land elements of each raster from satellite images, development logs and land statistics data. Step S2. Merge rasters with the same land development type to form development areas. The development areas are connected regions and there are no rasters of the same development type around the development areas. Use an image segmentation model to segment the boundaries of each development area from the remote sensing image. Calculate the shape coefficient of the development area based on the perimeter and area of ​​the boundary. Step S3. Multiply the shape coefficient of the development area and the agglomeration effect coefficient of the industry to obtain the regional agglomeration effect. For each type of development, the development parameters and the regional agglomeration effect are weighted and summed to obtain the output efficiency of each land grid and the overall output efficiency. Step S4. Select two development areas with the same development type and neighborhood identity higher than the threshold. Use an artificial intelligence model to plan the minimum cost connection path between the two development areas. Merge the grids that the path passes through into the two development areas to form a connected area. Under the premise of increasing the overall output efficiency, merge the adjacent grids of the connected area and mark all merged land grids. Step S5. Land rasters with more than one marker are designated as conflict rasters. Planning is carried out with the maximum overall output efficiency as the planning condition and land and environmental constraints as the constraints. The development type change decision for each conflict raster is then output.

2. The land element optimization allocation decision-making method based on artificial intelligence according to claim 1, characterized in that: Step S1 includes: Step S11. Use a GIS geographic information system to acquire satellite remote sensing images of the area to be developed, establish a coordinate system for the remote sensing images, and divide the remote sensing images into square grids with a side length of a, where a is the preset grid width. The number of grids divided is greater than one, and remove grids representing undevelopable land. Step S12. Obtain the land area, land topography and vegetation distribution type of the region from the remote sensing image, and obtain the land development type, population composition, land input and land output from the development log and land statistics. When there is more than one development type in the raster, the development type with the largest area is taken as the development type of the raster. Each grid cell is labeled with land features, including: land development type, land area, land topography, vegetation distribution, population composition, land input, and land output.

3. The land element optimization allocation decision-making method based on artificial intelligence according to claim 2, characterized in that: Step S2 includes: Step S21. Merge grids according to land development type. Perform the merging method once for each grid to obtain a development area. The merging method is as follows: when there are adjacent grids of the same development type in the current grid, merge the current grid with the adjacent grids to form a region core. If there are adjacent grids of the same development type around the region core, merge the adjacent grids into the region core. If there are no adjacent grids of the same development type around the region core, the region core is transformed into a development area. Step S22. Using an image segmentation model, segment the boundaries of each development area from the remote sensing image, obtain the perimeter and area of ​​the boundaries, and calculate the shape coefficient of the development area: ； Where, δ k Y represents the shape coefficient of the development area. k A represents the perimeter of the development area. k represents the area of ​​the development region, r represents the raster scaling factor, and c represents the number of rasters contained in the development region.

4. The land element optimization allocation decision method based on artificial intelligence according to claim 3, characterized in that: Step S3 include: Step S31. Calculate the land agglomeration effect of each development area, satisfying P=u·δ k Where P represents the agglomeration effect, and u is the agglomeration effect coefficient of the development type. The development types include: cultivated land, forest land, construction land, transportation land, industrial and mining land and water use land. The agglomeration effect coefficient is determined by the ratio of the output increment to the area increment of the corresponding development type. Step S32. Obtain development parameters for each development type, including: input-output ratio, agglomeration effect, land adaptability, development cost, and environmental loss, and determine the output efficiency of the land grid. ； Where μ is the output efficiency of the land raster, N is the number of development parameters, and F i For the i-th development parameter value, R i The weight of the i-th development parameter; The global output efficiency is obtained by summing the output efficiencies of all grids. Step S4 includes: Step S41. Mark two development areas with the same development type and neighborhood identity below a threshold. The neighborhood identity represents the minimum boundary distance between the two areas. Input the change cost of each grid development type into the artificial intelligence model, plan the connection path between the two areas, and minimize the cumulative change cost of all grids passed through the connection path. Step S42. Merge all the grids traversed by the connected path into the two development regions to obtain the connected region. Simulate each grid around the connected region. If the global output efficiency increases after merging the grid, then merge the grid into the connected region. When there are no grids around the connected region that can be merged, update the connected region and mark all the grids merged in steps S41 to S42.

5. The land element optimization allocation decision-making method based on artificial intelligence according to claim 4, characterized in that: Step S5 includes: Step S51. Take land grids with more than one label as conflict grids, determine the development type of the conflict based on the label of the conflict grid, calculate the output efficiency increment of the conflict grid as a development type, and obtain the output efficiency factor μ(x), where x represents the type of conflict development type, and μ(x) represents the increment of the global output efficiency when the conflict grid adopts the xth development type. Step S52. Using the maximization of overall output efficiency as the planning condition, and land and environmental constraints as the constraints, plan for each conflict grid: ； Where S represents the overall output efficiency, M1 and M2 represent the number of developed regions and the number of conflict grids, respectively, and Z... j μ represents the output efficiency of the non-collision grid portion in the j-th development region. k (x) represents the output efficiency factor of the k-th conflicting grid, c x and S x c1 and c2 represent the number of land grids and output efficiency of all land using the x-th development type, respectively. c1 and c2 are the upper and lower limits of the number of grids, respectively. S0 is the lower limit of output efficiency. H(S) is a mandatory constraint condition determined by the development policy. H0 is a mandatory constraint parameter. Step S53. Determine the development type of each conflict grid to maximize the overall output efficiency while satisfying the constraints, and output the decision results for all land grids.

6. A land resource optimization allocation decision system based on artificial intelligence, characterized in that, The system includes the following modules: land remote sensing module, region segmentation module, development prediction module, virtual merging module, and configuration decision module; The land remote sensing module is used to input regional statistical data into the GIS geographic information system, establish a coordinate system in the satellite remote sensing image of the land to be developed, rasterize the land according to the coordinates, remove the raster of undevelopable land, and input the land element information of each land raster. The land elements include: land development type, land area, land topography, vegetation distribution, population composition, land input and land output. The region segmentation module is used to merge rasters of the same land development type to form a development region. The land development type of all adjacent rasters in the development region is different from that of the development region. Using intelligent image segmentation technology, the development region is segmented from the remote sensing image and the region edges are connected. The shape coefficient of each development region is calculated based on the perimeter of the edge and the area of ​​the development region. The development prediction module is used to obtain the regional agglomeration effect based on the shape coefficient of the development area and the agglomeration effect coefficient of the industry. It calculates the output efficiency of each land grid based on the input-output ratio, agglomeration effect, land adaptability, development cost and environmental loss of each development type. The output efficiency of all land grids is accumulated to obtain the overall output efficiency, and a mapping relationship between the overall output efficiency and the output efficiency of land grids is established. The virtual merging module is used to mark development areas with neighborhood identity below a threshold. The development area that is closest to the connected area and has the same development type is used as the merging area. The shortest path between the connected area and the merging area is established using an artificial intelligence model. The grids passed through by the shortest path are merged into the connected area to connect the connected area and the merging area. Grids around the shortest path are selected to be merged into the connected area to increase the overall output efficiency of the region until the total land efficiency index decreases after any adjacent grid is added to the merging area. All newly merged land grids are marked. The configuration decision module is used to identify land graticles with more than one marker as conflict graticles. Using development type as a variable, the conflict graticles are weighted by the output efficiency of each development type. An output influence function is constructed, and planning is carried out with the maximum output efficiency of the entire region as the planning condition and land and environmental constraints as constraints. The planning status of each conflict graticle is determined, and the decision results of the development type of each land graticle are output.

7. The land element optimization allocation decision system based on artificial intelligence according to claim 6, characterized in that: The land remote sensing module includes: GIS unit and feature statistics unit; The GIS unit is used to acquire satellite remote sensing images of the area to be developed, establish regional coordinates, and rasterize the land according to the coordinates; The element statistics unit is used to determine the land element information of each grid from satellite imagery, development logs, and land statistics data.

8. The land element optimization allocation decision system based on artificial intelligence according to claim 7, characterized in that: The region segmentation module includes: a rasterization unit, a boundary recognition unit, and a shape evaluation unit; The gridded unit is used to merge grids according to land development type to form a development area. When there is more than one development type in a grid, the development type with the largest area is taken as the development type of the grid. The boundary recognition unit is used to segment the boundary of the development area from satellite remote sensing images using an image recognition and segmentation model; The shape evaluation unit is used to calculate the shape coefficient of the development area based on the perimeter and area of ​​the development area boundary.

9. The land element optimization allocation decision system based on artificial intelligence according to claim 8, characterized in that: The development prediction module includes: a data decision-making unit and an efficiency evaluation unit; The data decision unit is used to calculate the agglomeration effect of each development area and to make decisions on changes in development type; The efficiency evaluation unit is used to calculate the output efficiency of each land grid and all undeveloped areas. The virtual merging module includes: an intelligent addressing unit and a grid marking unit; The intelligent addressing unit is used to access the artificial intelligence service interface and perform shortest path planning for regional connectivity between the two development areas. The grid marking unit is used to determine the range of the merged area to maximize the overall output efficiency and to mark the newly merged grid.

10. A land resource optimization allocation decision system based on artificial intelligence according to claim 9, characterized in that: The configuration decision module includes: conflict grid cells, exponential planning cells, and grid configuration cells; The conflict grid cell is used to determine the output efficiency of each development type in the conflict grid configuration and generate an output impact function. The index planning unit is used to plan the development type of conflict grids to maximize the overall output efficiency. The grid configuration unit is used to update the remote sensing map, output the new development type for each grid, and provide the total development cost.

Images