A land space unit division method based on multi-source data fusion and related device

Abstract

The application provides a land space unit division method based on multi-source data fusion and related devices, and belongs to the technical field of land space unit division. In each macro space range, a differentiated chain intelligent modeling method is used to divide the urban space, ecological space and agricultural space, to obtain initial urban units, ecological units and agricultural units; quality inspection is performed on the obtained initial urban units, ecological units and agricultural units, if quality defects or conflict problems are found through quality inspection, a parameter back transmission mechanism and a local recalculation mechanism are triggered, until the obtained urban units, ecological units and agricultural units meet the quality requirements, to obtain final urban units, ecological units and agricultural units. The application solves the problem of low accuracy of land space unit division.

Description

Technical Field

[0001] This invention belongs to the field of land spatial unit division technology, specifically relating to a land spatial unit division method and related apparatus based on multi-source data fusion. Background Technology

[0002] Currently, existing methods for dividing land spatial units based on multi-source data fusion mainly include the following types: (1) Rule overlay and manual interpretation method: Using remote sensing images, current land use, basic topography, etc. as input, spatial type identification and boundary delineation are completed based on GIS (Geographic Information System) overlay analysis, threshold determination, experience correction and other methods; (2) Control line constraint zoning method: The main constraints are ecological protection red line, permanent basic farmland, urban development boundary, etc., and spatial types are trimmed, embedded and classified through priority rules; (3) Indicator evaluation and weighted comprehensive method: Construct an indicator system for ecological sensitivity, arable land quality, development suitability, etc., form spatial types or functional zones through gridded evaluation and hierarchical zoning, and further refine them into units by artificial means; (4) Local intelligent identification method: In some scenarios, classifiers or segmentation models are introduced to identify land features / land use types, but they are mostly used for "map patch identification" and are difficult to directly convert into manageable "land space units".

[0003] The above four methods can achieve basic partitioning in engineering applications, but they still generally have the following defects and shortcomings: (1) Weak fusion of multi-source data, incomplete information expression The input data comes from various sources (remote sensing, DEM (Digital Elevation Model), land use, ecological environment, socio-economic and planning control, etc.), but existing methods mostly use "simple overlay / stitching", lacking a unified data standardization, spatiotemporal alignment, scale conversion and conflict resolution mechanism. This leads to inconsistent attributes of the same plot under different data sources, and spatial discrimination results are prone to problems such as "fractures, voids, overlaps and misalignments".

[0004] (2) Insufficient data staticity and timeliness make it difficult to support rolling updates. Relying on annual change surveys, interim interpretation results, or one-off thematic data results in long update cycles. When changes occur such as ecological restoration, farmland consolidation, and urban expansion, it is difficult to adjust unit boundaries and attributes in a timely manner, causing the results to lag behind management needs and affecting the timeliness and consistency of land use control.

[0005] (3) The empirical rules for dividing logical dependencies are not stable and reusable. Key steps such as threshold setting, priority pruning, and boundary correction rely on human experience, and the rules vary across different regions and compilation teams. Results from the same region show poor repeatability across different batches of updates, making it difficult to establish a standardized process that can be widely adopted, and also failing to meet the requirements for results review and traceability.

[0006] (4) Insufficient hierarchical connection between the “three spaces” and the internal units of the space Existing technologies often remain at the level of macro-level zoning (ecology / agriculture / urban) or land category identification, failing to adequately depict the functional differences, ecological process differences, production suitability differences, and development intensity differences within each space. This makes it difficult to subsequently form a refined national spatial unit system that is manageable, statistically accurate, assessable, and interconnected with projects.

[0007] (5) The toolchain is fragmented and lacks a closed-loop system of acquisition-processing-division-output. Data preprocessing, feature extraction, model discrimination, unit generation, and output are scattered across multiple software programs and processes. The lack of a unified process orchestration, quality control, version management, and result comparison mechanism leads to low efficiency, excessive manual intervention, and unstable quality, making it difficult to meet the needs of large-scale promotion and application.

[0008] The core problem caused by the above-mentioned defects is that, under the conditions of multi-source heterogeneous data and complex management and control constraints, it is difficult to simultaneously achieve consistency of results, refinement of boundaries, traceability of processes, and dynamic updability in the division of national land spatial units.

[0009] Therefore, there is an urgent need for an integrated approach that can provide multi-source data fusion, intelligent processing, and collaborative division for land and space planning and management, so as to first divide the land into three spaces: ecological, agricultural, and urban, and then form land and space unit results that can be implemented within each space based on spatial attributes and control requirements. Summary of the Invention

[0010] The purpose of this invention is to provide a method and related apparatus for dividing land spatial units based on multi-source data fusion, in order to solve the problem of low accuracy in the division of land spatial units in the prior art.

[0011] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for dividing territorial spatial units based on multi-source data fusion, comprising the following steps: Acquire multi-source heterogeneous data, including vector-based land infrastructure data, raster-based natural environment data, and planning and control element data; The multi-source heterogeneous data is cleaned and its spatiotemporal reference is unified to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. Based on vector-based land base data and raster-based natural environment data after cleaning and unified spatiotemporal benchmarking, urban features, ecological features and agricultural features are extracted. Data corresponding to urban, ecological, and agricultural characteristics are mapped to a unified spatial carrier, and multidimensional spatial characteristic data are constructed in the unified spatial carrier. Based on the planning and control element data after cleaning and unified processing of spatiotemporal reference, it is determined whether the land space to be divided belongs to the control area; for the control area, the candidate category of the control area is determined according to the constraints of the control area; for the non-control area, the multi-dimensional spatial feature data is input into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories. By using a constraint mask, the probability vectors of candidate categories and several macro-spaces in the control area are consistently corrected to obtain three macro-space ranges: urban space, ecological space, and agricultural space. Within each macro-space scope, a differentiated chain-like intelligent modeling method is used to divide the urban space, ecological space, and agricultural space to obtain initial urban units, ecological units, and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm process customized according to the spatial type. The initial urban, ecological, and agricultural units are subjected to quality checks. If quality defects or conflicts are found during the quality checks, a parameter feedback mechanism and a local recalculation mechanism are triggered until the urban, ecological, and agricultural units meet the quality requirements, thus obtaining the final urban, ecological, and agricultural units.

[0012] A further improvement of this invention lies in dividing the urban space to obtain initial urban class units, specifically including: Acquire construction land map patch data, extract the centroid or representative point of the construction land map patch from the construction land map patch data, and convert the centroid or representative point of the construction land map patch into point features; The point features are clustered using the DBSCAN density clustering algorithm. During the clustering analysis, core points are determined based on a preset neighborhood radius and minimum number of points to identify areas of concentrated land use. Based on road network data, roads are overlaid with the concentrated land use distribution area as a barrier boundary to determine the physical boundary shape of various functional areas and obtain the initial town unit. The cluster density levels of functional zones within the initial urban unit were calculated using the kernel density estimation method. Based on the cluster density levels of the functional zones and the dominant land use types within the initial urban units, the initial urban units are classified into residential living units, comprehensive service units, commercial and business units, industrial development units, and logistics and warehousing units.

[0013] A further improvement of this invention lies in dividing the ecological space to obtain initial ecological class units, specifically including: Based on the assessment results of the importance of ecosystem service functions, identify ecological source areas; Based on the identified ecological source areas, ecological resistance surfaces are constructed by selecting land use types, elevation and slope factors of digital elevation models (DEMs), and minimum cost paths are extracted using the minimum cumulative resistance model as ecological corridors. In ecological resistance surfaces and ecological corridors, circuit theory is introduced to simulate the migration and diffusion process of species in the landscape surface, and ecological pinch points and obstacle points are identified by calculating current density. An ecological degradation risk index, which includes ecological function indicators and ecological pressure indicators, is constructed based on the identified ecological pinch points and obstacles. Based on the ecological degradation risk index, the ecological space is divided into ecological conservation units, ecological corridor units, water security units, and ecological degradation units.

[0014] A further improvement of this invention is that the formula for calculating the ecological degradation risk index is:

[0015] in, DRI This is an ecological degradation risk index. Pressure u For the first u Ecological stress factors Function v For the first v Ecological function factors, α u and β v These are the weighting coefficients.

[0016] A further improvement of this invention lies in dividing the agricultural space to obtain initial agricultural units, specifically including: Using the field chief system grid or management grid as the basic statistical carrier, calculate the area ratio of cultivated land, orchard, facility agricultural land and village land within the grid; Based on the area ratio of cultivated land, orchards, facility agricultural land and village land within the grid, and in accordance with the dominant and secondary dominant rules, agricultural space is divided into agricultural production units, agricultural supporting units and agricultural mixed units. Based on crop type and farmland contiguousness, agricultural production units, agricultural supporting units, and agricultural mixed units are further subdivided into grain production units, cash crop units, aquaculture units, rural residential units, and multiple mixed units. Farmland contiguousness and irrigation conditions are written into the corresponding unit attribute database.

[0017] A further improvement of this invention lies in the fact that the consistency correction of the candidate categories of the controlled region and the probability vectors of several macroscopic spaces through the constraint mask specifically includes: Based on ecological protection red lines, permanent basic farmland, and urban development boundaries, a constraint mask is constructed; Based on the constraint mask, the spatial suitability probability vector is corrected, and the calculation formula for the correction is as follows:

[0018] in, This is the corrected spatial fitness probability vector. This is the initial spatial fitness probability vector. For constraining the mask.

[0019] A further improvement of the present invention is that the quality check includes topological consistency check, constraint consistency check and statistical consistency check.

[0020] Secondly, the present invention provides a land spatial unit division system based on multi-source data fusion, comprising: The data acquisition module is used to acquire multi-source heterogeneous data, including vector-based land basic data, raster-based natural environment data, and planning and control element data. The preprocessing module is used to clean and unify the spatiotemporal reference of the multi-source heterogeneous data to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. The feature extraction module is used to extract urban features, ecological features, and agricultural features from vector-based land base data and raster-based natural environment data after unified processing based on cleaning and spatiotemporal benchmarks. The multidimensional data construction module is used to map data corresponding to urban characteristics, ecological characteristics and agricultural characteristics to a unified spatial carrier, and to construct multidimensional spatial characteristic data in the unified spatial carrier. The classification and prediction module is used to determine whether the land space to be divided belongs to the control area based on the planning and control element data after cleaning and unified processing of spatiotemporal reference; for the control area, the candidate category of the control area is determined according to the constraints of the control area; for the non-control area, the multi-dimensional spatial feature data is input into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories. The correction module is used to perform consistency correction on the candidate categories and probability vectors of several spatial categories of the control area through a constraint mask, so as to obtain three macro-spatial ranges: urban space, ecological space and agricultural space. The spatial division module is used to divide the urban space, ecological space and agricultural space within various macro-space ranges using a differentiated chain-like intelligent modeling method to obtain initial urban units, ecological units and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm flow customized according to the spatial type. The quality inspection module is used to perform quality checks on the initial urban, ecological, and agricultural units. If quality defects or conflicts are found through the quality inspection, the parameter feedback mechanism and local recalculation mechanism are triggered until the urban, ecological, and agricultural units meet the quality requirements, thus obtaining the final urban, ecological, and agricultural units.

[0021] Thirdly, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the land spatial unit division method based on multi-source data fusion described above.

[0022] Fourthly, the present invention provides a storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the land spatial unit division method based on multi-source data fusion described above.

[0023] Compared with the prior art, the present invention has the following beneficial effects: The proposed method for delineating land spatial units based on multi-source data fusion acquires multi-source heterogeneous data (vector-based land infrastructure data, raster-based natural environment data, and planning and control element data). This demonstrates that the method considers diverse data types when delineating land spatial units, including not only vector-based land infrastructure data but also raster-based natural environment data and planning and control element data, thus improving the accuracy of subsequent land spatial unit delineation. Furthermore, by using a constraint mask to perform consistency correction on the probability vectors of candidate categories and several macro-space categories within the control area, three macro-space ranges—urban space, ecological space, and agricultural space—are obtained. This operation not only avoids automatic category suppression but also eliminates conflicts between neighboring spatial categories, further improving the accuracy of land spatial unit delineation. Moreover, the invention employs a differentiated chain-based intelligent modeling method to delineate urban, ecological, and agricultural spaces. This avoids the problem of single methods failing to adequately characterize different types of spatial features, reducing misclassification, omissions, and boundary misjudgments, thereby improving the accuracy of land spatial unit delineation. Furthermore, after obtaining the initial urban, ecological, and agricultural units, this invention performs a quality check on these units, ensuring that the land space division results meet control requirements. Attached Figure Description

[0024] Figure 1 This is a flowchart of the land spatial unit division method based on multi-source data fusion of the present invention; Figure 2 This is a schematic diagram of the land spatial unit division system based on multi-source data fusion of the present invention; Figure 3 This is an architecture diagram of the land spatial unit division system based on multi-source data fusion in Embodiment 4 of the present invention; Figure 4 This is a flowchart of the land spatial unit division method based on multi-source data fusion in Embodiment 4 of the present invention; Figure 5 This is a schematic diagram of the three spatial internal division units in Embodiment 4 of the present invention; Figure 6 This is a flowchart illustrating the division of urban units in Embodiment 4 of the present invention; Figure 7 This is a flowchart illustrating the division of ecological units in Embodiment 4 of the present invention; Figure 8 This is a flowchart illustrating the division of agricultural units in Embodiment 4 of the present invention; Figure 9 This is a flowchart of the quality control and closed-loop update process in Embodiment 4 of the present invention; Figure 10 This is a schematic diagram of the structure of the electronic device of the present invention. Detailed Implementation

[0025] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0026] Example 1: The flowchart of the land spatial unit division method based on multi-source data fusion of this invention is as follows: Figure 1 As shown, the land spatial unit division method based on multi-source data fusion of the present invention includes the following steps: S1. Acquire multi-source heterogeneous data, which includes vector-based land infrastructure data, raster-based natural environment data, and planning and control element data; S2. The multi-source heterogeneous data is cleaned and its spatiotemporal reference is unified to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. S3. Based on the vector-based land base data and raster-based natural environment data after cleaning and unified spatiotemporal benchmarking, extract urban features, ecological features and agricultural features; S4. Map the data corresponding to urban characteristics, ecological characteristics and agricultural characteristics to a unified spatial carrier, and construct multi-dimensional spatial characteristic data in the unified spatial carrier; S5. Based on the planning and control element data after cleaning and unified processing of spatiotemporal reference, determine whether the land space to be divided belongs to the control area; for the control area, determine the candidate category of the control area according to the constraints of the control area; for the non-control area, input the multi-dimensional spatial feature data into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories. S6. By using a constraint mask, the probability vectors of candidate categories and several macro-spaces in the control area are consistently corrected to obtain three macro-space ranges: urban space, ecological space, and agricultural space. S7. Within each macro-space range, a differentiated chain-like intelligent modeling method is used to divide the urban space, ecological space, and agricultural space to obtain initial urban units, ecological units, and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm flow customized according to the spatial type. S8. Perform quality checks on the initial urban, ecological, and agricultural units obtained. If quality defects or conflicts are found through the quality check, trigger the parameter feedback mechanism and local recalculation mechanism until the urban, ecological, and agricultural units obtained meet the quality requirements, and obtain the final urban, ecological, and agricultural units.

[0027] Example 2: A schematic diagram of the land spatial unit division system based on multi-source data fusion of this invention is shown below. Figure 2 As shown, the land spatial unit division system based on multi-source data fusion of the present invention includes: The data acquisition module is used to acquire multi-source heterogeneous data, including vector-based land basic data, raster-based natural environment data, and planning and control element data. The preprocessing module is used to clean and unify the spatiotemporal reference of the multi-source heterogeneous data to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. The feature extraction module is used to extract urban features, ecological features, and agricultural features from vector-based land base data and raster-based natural environment data after unified processing based on cleaning and spatiotemporal benchmarks. The multidimensional data construction module is used to map data corresponding to urban characteristics, ecological characteristics and agricultural characteristics to a unified spatial carrier, and to construct multidimensional spatial characteristic data in the unified spatial carrier. The classification and prediction module is used to determine whether the land space to be divided belongs to the control area based on the planning and control element data after cleaning and unified processing of spatiotemporal reference; for the control area, the candidate category of the control area is determined according to the constraints of the control area; for the non-control area, the multi-dimensional spatial feature data is input into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories. The correction module is used to perform consistency correction on the candidate categories and probability vectors of several spatial categories of the control area through a constraint mask, so as to obtain three macro-spatial ranges: urban space, ecological space and agricultural space. The spatial division module is used to divide the urban space, ecological space and agricultural space within various macro-space ranges using a differentiated chain-like intelligent modeling method to obtain initial urban units, ecological units and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm flow customized according to the spatial type. The quality inspection module is used to perform quality checks on the initial urban, ecological, and agricultural units. If quality defects or conflicts are found through the quality inspection, the parameter feedback mechanism and local recalculation mechanism are triggered until the urban, ecological, and agricultural units meet the quality requirements, thus obtaining the final urban, ecological, and agricultural units.

[0028] Example 3: The method for dividing territorial spatial units based on multi-source data fusion of this invention includes the following steps: S1. Acquire multi-source heterogeneous data, which includes vector-based land basic data, raster-based natural environment data, and planning and control element data.

[0029] S2. The multi-source heterogeneous data is cleaned and its spatiotemporal reference is unified to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification.

[0030] S3. Based on the vector-based land base data and raster-based natural environment data after cleaning and unified spatiotemporal benchmarking, extract urban features, ecological features and agricultural features.

[0031] S4. Map the data corresponding to urban characteristics, ecological characteristics and agricultural characteristics to a unified spatial carrier, and construct multi-dimensional spatial characteristic data in the unified spatial carrier.

[0032] S5. Based on the planning and control element data after cleaning and unified processing of spatiotemporal reference, determine whether the land space to be divided belongs to the control area; for the control area, determine the candidate category of the control area according to the constraints of the control area; for the non-control area, input the multi-dimensional spatial feature data into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories.

[0033] S6. By using a constraint mask, the probability vectors of candidate categories and several macro-spaces in the control area are consistently corrected to obtain three macro-space ranges: urban space, ecological space, and agricultural space.

[0034] The consistency correction of the candidate categories and probability vectors of several macroscopic spaces in the controlled area through constraint masks specifically includes: Based on ecological protection red lines, permanent basic farmland, and urban development boundaries, a constraint mask is constructed; Based on the constraint mask, the spatial suitability probability vector is corrected, and the calculation formula for the correction is as follows:

[0035] in, This is the corrected spatial fitness probability vector. This is the initial spatial fitness probability vector. For constraining the mask.

[0036] S7. Within each macro-space scope, a differentiated chain-based intelligent modeling method is used to divide urban space, ecological space, and agricultural space to obtain initial urban units, ecological units, and agricultural units.

[0037] In this step, chain-based intelligent modeling refers to a sequentially linked algorithmic process customized according to the spatial type.

[0038] This step involves dividing the urban space to obtain initial urban class units, specifically including: Element-to-point conversion: Obtain construction map patch data, extract the centroid or representative point of the construction map patch from the construction map patch data, and convert the centroid or representative point of the construction map patch into point features; Density clustering processing: The point features are clustered using the DBSCAN density clustering algorithm. During the clustering analysis, core points are determined based on the preset neighborhood radius and minimum number of points to identify areas of concentrated land use distribution. Road network boundary processing: Based on road network data, roads are superimposed on the concentrated land distribution areas as barrier boundaries to determine the physical boundary shapes of various functional areas and obtain the initial urban units; Kernel density grading: The functional zone cluster density level within the initial urban unit is calculated using the kernel density estimation method; Unit classification processing: Based on the cluster density level of the functional area and the dominant land use type within the initial urban unit, the initial urban unit is classified into residential living units, comprehensive service units, commercial and business units, industrial development units, and logistics and warehousing units (the resulting initial urban unit categories).

[0039] This step involves dividing the ecological space to obtain initial ecological units, specifically including: Source area identification and processing: Based on the assessment results of the importance of ecosystem service functions, identify the ecological source areas; Resistance surface construction and corridor extraction: Based on the identified ecological source areas, land use type, elevation and slope factors of digital elevation model (DEM) are selected to construct ecological resistance surfaces, and the minimum cumulative resistance model is used to extract the minimum cost path as ecological corridor. Circuit theory connectivity analysis: In ecological resistance surfaces and ecological corridors, circuit theory is introduced to simulate the migration and diffusion process of species in the landscape surface, and ecological pinch points and obstacle points are identified by calculating current density. Degradation zoning: Based on the identified ecological pinch points and obstacle points, an ecological degradation risk index containing ecological function indicators and ecological pressure indicators is constructed. Based on the ecological degradation risk index, the ecological space is divided into ecological conservation units, ecological corridor units, water security units, and ecological degradation units (the initial ecological units obtained).

[0040] The formula for calculating the ecological degradation risk index is:

[0041] in, DRI This is an ecological degradation risk index. Pressure u For the first u Ecological stress factors Function v For the firstv Ecological function factors, α u and β v These are the weighting coefficients.

[0042] This step involves dividing the agricultural space to obtain initial agricultural units, specifically including: Grid feature statistical processing: Using the field chief system grid or management grid as the basic statistical carrier, calculate the area ratio of cultivated land, orchard, facility agricultural land and village land within the grid; Rule-based discrimination and processing: Based on the area ratio of cultivated land, orchards, facility agricultural land and village land within the grid, agricultural space is divided into agricultural production units, agricultural supporting units and mixed agricultural units according to the dominant and secondary dominant rules. Subdivision processing: Based on crop type and farmland contiguousness, agricultural production units, agricultural supporting units, and agricultural mixed units are subdivided into grain production units, cash crop units, aquaculture units, rural residential units, and multiple mixed units (the initial agricultural units), and farmland contiguousness and irrigation conditions are written into the corresponding unit attribute library.

[0043] S8. Perform quality checks on the initial urban, ecological, and agricultural units obtained. If quality defects or conflicts are found through the quality check, trigger the parameter feedback mechanism and local recalculation mechanism until the urban, ecological, and agricultural units obtained meet the quality requirements, and obtain the final urban, ecological, and agricultural units.

[0044] The parameter return mechanism and the local recalculation mechanism are explained below: The parameter feedback mechanism and the local recalculation mechanism refer to the process of sending adjustment instructions back to the corresponding link of the chain-based intelligent modeling according to the defect type or conflict type to update the model parameters, and only using the updated parameters to re-divide the units in the local area where there are defects or conflicts.

[0045] The quality checks in this step include topological consistency checks, constraint consistency checks, and statistical consistency checks.

[0046] Example 4: The land spatial unit delineation method based on multi-source data fusion of this invention is implemented through a land spatial unit delineation system based on multi-source data fusion. The architecture diagram of the land spatial unit delineation system based on multi-source data fusion is shown below. Figure 3 As shown below, Figure 3 Detailed explanation: This invention's land spatial unit division system based on multi-source data fusion adopts a three-layer architecture: data input layer, intelligent processing layer, and decision output layer. Each layer's modules form a closed loop through data flow and feedback mechanisms. The data input layer is equipped with multiple interfaces to access multi-source data, which includes at least vector-based land infrastructure data (including land use patches, administrative boundaries, water systems, and roads), raster-based natural environment data (including remote sensing imagery, DEMs and derived slope and aspect ratios, and vegetation indices), and planning and control element data (ecological protection red lines, permanent basic farmland, and urban development boundaries, etc.).

[0047] Multi-source data can also include human activity data such as population raster, nighttime light, and POIs, used to enhance urban spatial identification and unit attribute labeling. Before being stored, the data input layer performs format parsing and initial screening, generates metadata (source, time phase, precision, coordinate system, resolution, and version number), and marks missing fields, obvious misalignments, and topological anomalies.

[0048] The intelligent processing layer comprises a spatial calibration and standardization module, a multi-source fusion module, three spatial discrimination modules, a unit generation module, and a quality control and feedback module. The spatial calibration and standardization module is responsible for coordinate system one, geometric registration, scale unification, raster / vector conversion, topology repair, and anomaly handling. The multi-source fusion module is responsible for feature construction, evidence fusion, and conflict resolution. The three spatial discrimination modules output three types of spatial masks (ecological, agricultural, and urban) and their confidence levels. The unit generation module generates and optimizes urban, ecological, and agricultural units within the masks, respectively. The quality control and feedback module performs topological consistency, constraint consistency, and statistical consistency checks on the results. If a conflict exceeding the threshold or an abnormal fragmentation rate is found, the problem location and cause are fed back to the fusion or unit generation stage to trigger a local recalculation, thus forming a closed loop of computation—quality control—rollback—recalculation.

[0049] The decision output layer outputs three types of results: three spatial partition results, three types of unit boundary and attribute library results, and quality control results (conflict list, statistical report, version and parameter log), and supports querying, tracing and service publishing.

[0050] The flowchart of the land spatial unit division method based on multi-source data fusion of this invention is as follows: Figure 4 As shown, the method of the present invention will be described in detail below. The method for dividing territorial spatial units based on multi-source data fusion of this invention includes the following steps: S1. Obtain vector-based land baseline data, raster-based natural environment data, and planning and control element data.

[0051] S2. Based on vector-based land base data and raster-based natural environment data, extract urban features, ecological features, and agricultural features.

[0052] Before extracting urban, ecological, and agricultural features from vector-based land use data and raster-based natural environment data, preprocessing is performed on both datasets. Preprocessing includes data cleaning, spatial calibration, scale unification, and missing data imputation. The preprocessing of vector-based land use data and raster-based natural environment data is described in detail below: During the data cleaning and spatial calibration phase, multi-source data (vector-based land use data and raster-based natural environment data) are unified to the target coordinate system and spatial reference. Coordinate transformation employs a two-dimensional affine model or a seven-parameter model. Taking the two-dimensional affine model as an example, the coordinate transformation formula is as follows:

[0053]

[0054] The parameters a0, a1, a2, b0, b1, and b2 are obtained through least-squares fitting of control points. Control points can be extracted from high-precision base maps, survey results, or stable features. Topology repair (self-intersection, overlap, gaps, and hanging lines) is performed on vector polygons, and fragmentation is handled according to the smallest cartographic unit; outlier removal and noise suppression are performed on raster data.

[0055] During the scale unification and missing interpolation stage, rasters of different resolutions are unified to the target resolution and aligned with vector boundaries. Raster resampling preferably uses bilinear interpolation, calculated using the following formula:

[0056] For missing items in tables or rasters, spatiotemporal interpolation methods are preferred, such as the weighting form of spatiotemporal kriging:

[0057] The interpolation identifier and confidence level are written into the metadata to ensure subsequent traceability.

[0058] S3. Map the data corresponding to ecological characteristics, agricultural characteristics and urban characteristics to a unified spatial carrier, and construct multi-dimensional spatial characteristic data in the unified spatial carrier.

[0059] S4. Based on the planning and control element data, determine whether the land space to be divided belongs to the control area; for the control area, determine the candidate category of the control area according to the constraints of the control area; for the non-control area, input the multi-dimensional spatial feature data into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories.

[0060] Based on planning and control element data, determine whether the land space to be delineated belongs to the control area; for the control area (the area falling within the ecological protection red line, permanent basic farmland and urban development boundary), determine the candidate category of the control area according to the constraints of the control area.

[0061] For non-controlled areas, multi-dimensional spatial feature data are input into a pre-trained land space classification model to obtain probability vectors of which the non-controlled areas belong to several macro-space categories.

[0062] S5. By using a constraint mask, the probability vectors of candidate categories and several macro-spaces in the control area are consistently corrected to obtain three macro-space ranges: urban space, ecological space, and agricultural space.

[0063] The formula for calculating consistency correction is:

[0064]

[0065] in, To constrain the mask, categories are prevented from being suppressed to 0 or a minimum value within the constrained region, thus preventing the results from exceeding rigid boundaries. Then, connectivity correction, hole removal, and boundary regularization are performed on the three spatial results to form a manageable spatial extent.

[0066] S6. Based on urban space, ecological space, and agricultural space, divide urban units, ecological units, and agricultural units.

[0067] Based on urban space, ecological space, and agricultural space, the city is divided into urban units, ecological units, and agricultural units. The diagram illustrating this division is as follows: Figure 5 As shown, specifically, a differentiated chain-based intelligent modeling method is adopted, which divides urban space, ecological space and agricultural space into urban units, ecological units and agricultural units. Chain-based intelligent modeling refers to a sequentially connected algorithm process customized according to the spatial type.

[0068] The following is a detailed explanation of the division of urban space, ecological space, and agricultural space: Urban spatial planning employs a chain-like process: element point conversion processing—density clustering processing—road network edge construction processing—kernel density hierarchical processing—unit classification processing, such as... Figure 6 As shown. The construction map patches are converted into point features (centroids or representative points), and DBSCAN clustering is performed to identify concentrated distribution areas. The neighborhood and core point conditions for DBSCAN are:

[0069] After clustering, the clustering results are combined with the road network to construct boundaries, using a road barrier + concave hull / construction zone approach to generate manageable boundaries; subsequently, kernel density estimation is performed on functional points or functional areas to form intensity levels, calculated using the following formula:

[0070] Where h is the bandwidth and K(·) is the kernel function. Finally, based on the dominant land use composition and intensity level, urban units are classified into residential living, comprehensive services, commercial and business, industrial development, logistics and warehousing, etc., and the dominant land use proportion, intensity level, road network accessibility and other attribute fields are written into them.

[0071] Ecological space management employs a chain-like process: source area identification and processing — resistance surface construction and corridor extraction processing — circuit theory connectivity analysis processing — degradation zoning processing, such as... Figure 7 As shown. The selection of ecological source areas is based on the comprehensive importance index of ecosystem services, and the calculation formula is as follows:

[0072] The contiguousness threshold is then superimposed to form a source area set. The resistance surface is preferably determined by a weighted superposition of factors such as land use type, elevation, and slope.

[0073] The minimum cumulative resistance (MCR) is calculated as follows:

[0074] in, This represents the grid movement distance. The resistance value is used. Based on the minimum cost path output by MCR and the extraction of ecological corridors, ecological conservation units and important ecological corridor units are formed; water security units are formed for large water bodies and wetlands according to area thresholds; at the same time, the degradation risk index is superimposed to delineate degradation units. The index is preferably constructed by ecological function factors (negative) + ecological pressure factors (positive), and the calculation formula is as follows:

[0075] The degradation unit range is output according to the graded threshold, and the rest are classified as general ecological units.

[0076] Agricultural spatial processing employs a chain-like process of grid feature statistical processing—rule-based discrimination processing—sub-detailing processing, such as... Figure 8As shown, the system uses field-chief grids or rule-managed grids as statistical carriers to calculate the area ratio of cultivated land, orchards, facility agricultural land, and village land within the grid. Based on dominant thresholds, agricultural space is divided into three categories: agricultural production, agricultural support, and mixed agriculture. Agricultural production is further subdivided into units such as grain production, cash crops, and aquaculture; agricultural support is subdivided into units such as agricultural support and rural residence; and mixed space is subdivided into multiple mixed units according to a dominant + secondary dominant combination rule. Fields such as dominant type, proportion, cultivated land contiguousness, slope grade, and irrigation conditions are written into the attribute database to ensure direct integration with supervision and assessment.

[0077] After obtaining the urban, ecological, and agricultural units, a quality check is conducted on the urban, ecological, and agricultural units to obtain a quality control report and a conflict list.

[0078] Quality checks include topological consistency checks, constraint consistency checks, and statistical consistency checks.

[0079] The following is a detailed explanation of the quality inspection process for urban units, ecological units, and agricultural units: The system performs topological consistency checks (overlap, gaps, breaks, and fragmentation rate), constraint consistency checks (area and location conflicting with red lines / basic farmland / development boundaries), and statistical consistency checks (area summary closure and type proportion rationality) on three types of units, generating quality control reports and conflict lists. Repairable issues are automatically fixed by the system, while issues exceeding thresholds trigger local rollback and recalculation. Finally, the system solidifies data versions, parameter versions, and processing logs, outputting unit result layers, attribute libraries, and reports, and supports service publishing, such as... Figure 9 As shown.

[0080] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention solves the problem of difficulty in overlaying multi-source data with a reference in traditional land spatial unit division, overcoming boundary misalignment and attribute conflicts caused by inconsistencies in coordinates, scales, and temporal phases. Through spatial calibration and standardization chains (2D affine / parametric coordinate transformation, raster bilinear resampling, vector topology repair) and missing information interpolation chains (spatiotemporal kriging interpolation), it achieves unified expression and consistent database entry of vector-raster-tabular data. For example, by aligning remote sensing images, DEM slope rasters, and map patches under the same coordinate system and resolution reference, "patch boundaries—topographic constraints—ecological features" can be consistently calculated within the same unit; missing items in statistical or monitoring fields are interpolated using the following formula to ensure the integrity and continuity of the feature stack:

[0081] In summary, the method of this invention achieves full-element fusion with the same caliber, significantly improving the input reliability and result consistency of unit division.

[0082] 2. This invention solves the problem that traditional unit division can only achieve macro-level partitioning and is difficult to refine to manageable units, overcoming the shortcomings of simply dividing into three zones without forming units or having unit boundaries that fluctuate with personnel experience. Through a hierarchical collaborative mechanism of three spatial discriminations and the generation of units within the space, three spatial masks—urban, ecological, and agricultural—are first output, and then urban, ecological, and agricultural units are formed respectively. The unit type, dominant factor, confidence level, and constraint compliance are written into an attribute database. For example, within the urban space, a chain algorithm of construction land transfer point—density clustering—road network constraint boundary construction—kernel density grading is used to automatically identify functional clusters and form manageable boundaries; within the agricultural space, area proportion is determined using regulatory grids / field length grids, achieving automatic database entry of production / supporting / mixed and subdivided units. In summary, the method of this invention achieves feasible integration from partitioning to units, enhancing the direct support capability of the results for land use control and regulatory assessment.

[0083] 3. This invention solves the problems of rigid external control constraints and post-event conflict discovery in traditional methods, overcoming the limitations of overlapping conflicts between the results and bottom-line elements such as ecological red lines, permanent basic farmland, and urban development boundaries, leading to repeated rework. It uses a constraint mask endogenous mechanism to perform consistency correction on the model output, avoiding automatic category suppression, and simultaneously performs constraint consistency verification and conflict localization during the unit generation stage. For example, when a candidate unit crosses a control boundary, the system automatically marks the conflict type and conflict area and triggers local backtracking and recalculation, ensuring that the results meet the bottom-line constraints during the generation stage. Its consistency correction form can be expressed as:

[0084] In summary, the method of this invention achieves consistent output of constraints that are compliant from the moment of generation, significantly reducing rework rate and approval risk.

[0085] 4. The method of this invention solves the problems of traditional ecological unit division lacking calculable channels and degradation identification mechanisms, and overcomes the shortcomings of coarse internal boundaries, weak functions, and uninterpretable processes in ecological spaces. Through an ecological algorithm chain of source area—resistance surface—minimum cumulative resistance—corridor extraction—degradation zoning, ecological conservation, corridor connectivity, and degradation risk are incorporated into the same computational framework. The resistance surface construction, minimum cumulative resistance solution, and degradation risk index can be expressed as follows:

[0086]

[0087]

[0088] In summary, the method of this invention enables ecological units to be calculable, interpretable, and hierarchical, thereby improving the professionalism and regulatory capacity of the division of internal ecological spaces.

[0089] 5. The method of this invention solves the problems of fragmentation, untraceable processes, and difficult updates in traditional toolchains, overcoming the shortcomings of manual multi-software interconnection, non-reproducible parameters, and incomparable versions. Through a closed-loop mechanism of quality control—recalculation feedback—version solidification, the system automatically performs quality control on topological consistency, constraint consistency, and statistical consistency, and feeds back conflict locations and causes to the fusion or unit generation stage to trigger local recalculation; at the same time, it solidifies data versions, parameter versions, and processing logs, achieving reproducible, auditable, and continuously updated results. In summary, the method of this invention achieves engineered closed-loop production and traceable delivery, significantly improving the ability for large-scale deployment and long-term operation and maintenance.

[0090] Example 5: Please see Figure 10 As shown, the present invention also provides an electronic device 100 for a land spatial unit division method based on multi-source data fusion; the electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on the at least one processor 102, and at least one communication bus 104.

[0091] The memory 101 can be used to store the computer program 103. The processor 102 implements the steps of the land spatial unit division method based on multi-source data fusion described in Embodiment 1 by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101. The memory 101 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0092] The at least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 102 may be a microprocessor or any conventional processor. The processor 102 is the control center of the electronic device 100, connecting various parts of the electronic device 100 via various interfaces and lines.

[0093] The memory 101 in the electronic device 100 stores multiple instructions to implement a land spatial unit division method based on multi-source data fusion, and the processor 102 can execute the multiple instructions to achieve the following: Acquire multi-source heterogeneous data, including vector-based land infrastructure data, raster-based natural environment data, and planning and control element data; The multi-source heterogeneous data is cleaned and its spatiotemporal reference is unified to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. Based on vector-based land base data and raster-based natural environment data after cleaning and unified spatiotemporal benchmarking, urban features, ecological features and agricultural features are extracted. Data corresponding to urban, ecological, and agricultural characteristics are mapped to a unified spatial carrier, and multidimensional spatial characteristic data are constructed in the unified spatial carrier. Based on the planning and control element data after cleaning and unified processing of spatiotemporal reference, it is determined whether the land space to be divided belongs to the control area; for the control area, the candidate category of the control area is determined according to the constraints of the control area; for the non-control area, the multi-dimensional spatial feature data is input into the pre-trained land space classification model to obtain the probability vector of the non-control area belonging to several macro-space categories. By using a constraint mask, the probability vectors of candidate categories and several macro-spaces in the control area are consistently corrected to obtain three macro-space ranges: urban space, ecological space, and agricultural space. Within each macro-space scope, a differentiated chain-like intelligent modeling method is used to divide the urban space, ecological space, and agricultural space to obtain initial urban units, ecological units, and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm process customized according to the spatial type. The initial urban, ecological, and agricultural units are subjected to quality checks. If quality defects or conflicts are found during the quality checks, a parameter feedback mechanism and a local recalculation mechanism are triggered until the urban, ecological, and agricultural units meet the quality requirements, thus obtaining the final urban, ecological, and agricultural units.

[0094] Example 6: If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0095] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0097] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0098] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for dividing territorial spatial units based on multi-source data fusion, characterized in that, Includes the following steps: Acquire multi-source heterogeneous data, including vector-based land infrastructure data, raster-based natural environment data, and planning and control element data; The multi-source heterogeneous data is cleaned and its spatiotemporal reference is unified to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. Based on vector-based land base data and raster-based natural environment data after cleaning and unified spatiotemporal benchmarking, urban features, ecological features and agricultural features are extracted. Data corresponding to urban, ecological, and agricultural characteristics are mapped to a unified spatial carrier, and multidimensional spatial characteristic data are constructed in the unified spatial carrier. Based on the planning and control element data after cleaning and unified processing of spatiotemporal benchmarks, it is determined whether the land space to be delineated belongs to the control area. For controlled areas, candidate categories of controlled areas are determined based on the constraints of controlled areas; for non-controlled areas, the multi-dimensional spatial feature data is input into a pre-trained land space classification model to obtain probability vectors of non-controlled areas belonging to several macro-space categories. By using a constraint mask, the probability vectors of candidate categories and several macro-spaces in the control area are consistently corrected to obtain three macro-space ranges: urban space, ecological space, and agricultural space. Within each macro-space scope, a differentiated chain-like intelligent modeling method is used to divide the urban space, ecological space, and agricultural space to obtain initial urban units, ecological units, and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm process customized according to the spatial type. The initial urban, ecological, and agricultural units are subjected to quality checks. If quality defects or conflicts are found during the quality checks, a parameter feedback mechanism and a local recalculation mechanism are triggered until the urban, ecological, and agricultural units meet the quality requirements, thus obtaining the final urban, ecological, and agricultural units.

2. The method for dividing territorial spatial units based on multi-source data fusion according to claim 1, characterized in that, The urban space is divided to obtain initial urban class units, specifically including: Acquire construction land map patch data, extract the centroid or representative point of the construction land map patch from the construction land map patch data, and convert the centroid or representative point of the construction land map patch into point features; The point features are clustered using the DBSCAN density clustering algorithm. During the clustering analysis, core points are determined based on a preset neighborhood radius and minimum number of points to identify areas of concentrated land use. Based on road network data, roads are overlaid with the concentrated land use distribution area as a barrier boundary to determine the physical boundary shape of various functional areas and obtain the initial town unit. The cluster density levels of functional zones within the initial urban unit were calculated using the kernel density estimation method. Based on the cluster density levels of the functional zones and the dominant land use types within the initial urban units, the initial urban units are classified into residential living units, comprehensive service units, commercial and business units, industrial development units, and logistics and warehousing units.

3. The method for dividing territorial spatial units based on multi-source data fusion according to claim 1, characterized in that, The ecological space is divided to obtain initial ecological class units, specifically including: Based on the assessment results of the importance of ecosystem service functions, identify ecological source areas; Based on the identified ecological source areas, ecological resistance surfaces are constructed by selecting land use types, elevation and slope factors of digital elevation models (DEMs), and minimum cost paths are extracted using the minimum cumulative resistance model as ecological corridors. In ecological resistance surfaces and ecological corridors, circuit theory is introduced to simulate the migration and diffusion process of species in the landscape surface, and ecological pinch points and obstacle points are identified by calculating current density. An ecological degradation risk index, which includes ecological function indicators and ecological pressure indicators, is constructed based on the identified ecological pinch points and obstacles. Based on the ecological degradation risk index, the ecological space is divided into ecological conservation units, ecological corridor units, water security units, and ecological degradation units.

4. The method for dividing territorial spatial units based on multi-source data fusion according to claim 3, characterized in that, The formula for calculating the ecological degradation risk index is as follows: in, DRI This is an ecological degradation risk index. Pressure u For the first u Ecological stress factors Function v For the first v Ecological function factors, α u and β v These are the weighting coefficients.

5. The method for dividing territorial spatial units based on multi-source data fusion according to claim 1, characterized in that, The agricultural space is divided to obtain initial agricultural units, specifically including: Using the field chief system grid or management grid as the basic statistical carrier, calculate the area ratio of cultivated land, orchard, facility agricultural land and village land within the grid; Based on the area ratio of cultivated land, orchards, facility agricultural land and village land within the grid, and in accordance with the dominant and secondary dominant rules, agricultural space is divided into agricultural production units, agricultural supporting units and agricultural mixed units. Based on crop type and farmland contiguousness, agricultural production units, agricultural supporting units, and agricultural mixed units are further subdivided into grain production units, cash crop units, aquaculture units, rural residential units, and multiple mixed units. Farmland contiguousness and irrigation conditions are written into the corresponding unit attribute database.

6. The method for dividing territorial spatial units based on multi-source data fusion according to claim 1, characterized in that, The consistency correction of the candidate categories and probability vectors of several macroscopic spaces in the controlled area through constraint masks specifically includes: Based on ecological protection red lines, permanent basic farmland, and urban development boundaries, a constraint mask is constructed; Based on the constraint mask, the spatial suitability probability vector is corrected, and the calculation formula for the correction is as follows: in, This is the corrected spatial fitness probability vector. This is the initial spatial fitness probability vector. For constraining the mask.

7. The method for dividing territorial spatial units based on multi-source data fusion according to claim 1, characterized in that, The quality checks include topological consistency checks, constraint consistency checks, and statistical consistency checks.

8. A land spatial unit division system based on multi-source data fusion, characterized in that, include: The data acquisition module is used to acquire multi-source heterogeneous data, including vector-based land basic data, raster-based natural environment data, and planning and control element data. The preprocessing module is used to clean and unify the spatiotemporal reference of the multi-source heterogeneous data to obtain multi-source heterogeneous data after cleaning and spatiotemporal reference unification. The feature extraction module is used to extract urban features, ecological features, and agricultural features from vector-based land base data and raster-based natural environment data after unified processing based on cleaning and spatiotemporal benchmarks. The multidimensional data construction module is used to map data corresponding to urban characteristics, ecological characteristics and agricultural characteristics to a unified spatial carrier, and to construct multidimensional spatial characteristic data in the unified spatial carrier. The classification and prediction module is used to determine whether the land space to be delineated belongs to the control area based on the planning and control element data after cleaning and unified processing of spatiotemporal benchmarks. For controlled areas, candidate categories of controlled areas are determined based on the constraints of controlled areas; for non-controlled areas, the multi-dimensional spatial feature data is input into a pre-trained land space classification model to obtain probability vectors of non-controlled areas belonging to several macro-space categories. The correction module is used to perform consistency correction on the candidate categories and probability vectors of several spatial categories of the control area through a constraint mask, so as to obtain three macro-spatial ranges: urban space, ecological space and agricultural space. The spatial division module is used to divide the urban space, ecological space and agricultural space within various macro-space ranges using a differentiated chain-like intelligent modeling method to obtain initial urban units, ecological units and agricultural units. The chain-like intelligent modeling refers to a sequentially connected algorithm flow customized according to the spatial type. The quality inspection module is used to perform quality checks on the initial urban, ecological, and agricultural units. If quality defects or conflicts are found through the quality inspection, the parameter feedback mechanism and local recalculation mechanism are triggered until the urban, ecological, and agricultural units meet the quality requirements, thus obtaining the final urban, ecological, and agricultural units.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the land spatial unit division method based on multi-source data fusion as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the land spatial unit division method based on multi-source data fusion as described in any one of claims 1 to 7.

Images