High and cold pasturing area territorial space function classification data processing method and system

Abstract

The invention provides a high and cold pasturing area territorial space function classification data processing method and system, and relates to the technical field of data processing, and the method comprises the steps: 1, collecting and processing multi-source space data of a high and cold pasturing area, and constructing a territorial space element database; step 2, based on the territorial space element database, evaluating the grazing degree of each space unit, and generating a spatial distribution data set of the grazing degree of the alpine pasturing area; and step 3, performing spatial overlay analysis on the spatial distribution data set of the grazing degree of the high and cold pasturing area and the range of the prohibited development area in the main body functional area planning, and performing geometric fusion on an intersection area generated by overlay to obtain a preliminary territorial space functional partition data set. The ecological protection and agricultural development requirements are coordinated, and the intelligence of territorial space function classification is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for processing land spatial functional classification data in high-altitude pastoral areas. Background Technology

[0002] Current methods for classifying the spatial functions of high-altitude pastoral areas have many limitations and cannot meet the actual needs of refined management and sustainable utilization of land space. Specifically, most existing technologies rely on single evaluation indicators (such as vegetation cover or administrative area as the basis for division) or static administrative boundaries for regional division. They have not formed a systematic multi-source spatial data processing and fusion mechanism, and cannot comprehensively integrate key land space elements such as topography, climate conditions, grassland type, soil texture, hydrological characteristics, and intensity of human activities in high-altitude pastoral areas. This results in significant defects in the constructed basic data system, making it difficult to support subsequent accurate functional classification analysis.

[0003] Crucially, existing methods lack consideration for the attribution and association characteristics of spatial units. They fail to optimize classification results by calculating the attribution and association parameters between spatial assessment units and their neighboring units, making it difficult to adapt and adjust the classification data to the actual spatial distribution characteristics. Furthermore, in the feature extraction stage, advanced intelligent algorithms are not fully utilized to mine the spatial structural characteristics of grid units, and traditional statistical analysis methods are still relied upon. This results in low accuracy of the final land space function classification results, making it impossible to accurately distinguish the boundaries and ranges of different functional types such as ecological space and agricultural space (pasture). These deficiencies are particularly prominent in ecologically fragile high-altitude pastoral areas, which can easily lead to an imbalance between ecological protection and grassland utilization. This could result in the destruction of fragile ecosystems due to overgrazing, or the waste of grassland resources due to unreasonable delineation of protection areas. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide a method and system for processing land spatial function classification data in high-altitude pastoral areas, to coordinate the needs of ecological protection and agricultural development, and to improve the intelligence of land spatial function classification.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: Firstly, a method for processing land spatial function classification data in high-altitude pastoral areas, the method comprising: Step 1: Collect and process multi-source spatial data from high-altitude pastoral areas to construct a national land spatial element database; Step 2: Based on the national land spatial element database, assess the grazing suitability of each spatial unit and generate a spatial distribution dataset of grazing suitability in high-altitude pastoral areas; Step 3: Spatial overlay analysis is performed on the dataset of the spatial distribution of grazing suitability in high-altitude pastoral areas and the scope of prohibited development zones in the main functional area planning. The intersecting areas generated by the overlay are geometrically fused to obtain a preliminary dataset of national land spatial functional zoning. Step 4: Perform regular gridding processing on the preliminary land space functional zoning dataset, transforming the original irregular spatial unit functional zoning results into regular grid units of preset size, forming a gridded land space functional classification dataset; Step 5: Select spatial assessment units based on the gridded land spatial function classification dataset, analyze the spatial distribution characteristics of the spatial assessment units, and calculate the attribution association parameters between each spatial assessment unit and its surrounding adjacent units; use the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset. Step 6: Based on the updated gridded land spatial function classification dataset, a pre-trained graph neural network model is used to extract the spatial structure features of the grid units, and finally generate the land spatial function classification result dataset for the high-altitude pastoral area.

[0006] Secondly, the data processing system for the functional classification of land use in high-altitude pastoral areas includes: The processing module is used to collect and process multi-source spatial data from high-altitude pastoral areas to build a national land spatial element database. The assessment module is used to assess the grazing suitability of each spatial unit based on the national land spatial element database, and generate a spatial distribution dataset of grazing suitability in high-altitude pastoral areas; The fusion module is used to perform spatial overlay analysis on the spatial distribution dataset of grazing suitability in high-altitude pastoral areas and the prohibited development zone scope in the main functional area planning. The intersecting areas generated by the overlay are geometrically fused to obtain a preliminary dataset of national land spatial functional zoning. The conversion module is used to perform regular gridding processing on the preliminary land spatial functional zoning dataset, converting the original irregular spatial unit functional zoning results into regular grid units of preset size to form a gridded land spatial functional classification dataset. The adaptation module is used to select spatial assessment units based on the gridded land spatial function classification dataset, analyze the spatial distribution characteristics of the spatial assessment units, calculate the attribution association parameters between each spatial assessment unit and its surrounding adjacent units, and use the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset. The classification module is used to extract the spatial structure features of grid cells based on the updated gridded land spatial function classification dataset, and finally generate the land spatial function classification result dataset for high-altitude pastoral areas.

[0007] Thirdly, a computing device, comprising: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the method.

[0008] Fourthly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0009] The above-described solution of the present invention has at least the following beneficial effects: By integrating multi-source spatial data from high-altitude pastoral areas and constructing a standardized national land spatial element database, the one-sidedness of single data sources can be effectively avoided. The grazing suitability assessment based on the national land spatial element database can reflect the grazing suitability characteristics of different spatial units in high-altitude pastoral areas, avoiding subjectivity and blindness in functional zoning. By spatially overlaying and geometrically fusing the grazing suitability spatial distribution dataset with the scope of prohibited development zones, the needs of ecological protection and grazing suitability characteristics can be organically combined. This ensures that the preliminary national land spatial functional zoning not only meets the regional ecological protection requirements but also takes into account the actual situation of grazing development, guaranteeing the rationality and feasibility of functional zoning. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the method for processing land spatial function classification data in high-altitude pastoral areas provided in an embodiment of the present invention.

[0011] Figure 2 This is a schematic diagram of the land spatial function classification data processing system for high-altitude pastoral areas provided in an embodiment of the present invention. Detailed Implementation

[0012] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0013] like Figure 1 As shown, an embodiment of the present invention proposes a method for processing land spatial function classification data in high-altitude pastoral areas, the method comprising the following steps: Step 1: Collect and process multi-source spatial data from high-altitude pastoral areas to construct a national land spatial element database; Step 2: Based on the national land spatial element database, assess the grazing suitability of each spatial unit and generate a spatial distribution dataset of grazing suitability in high-altitude pastoral areas; Step 3: Spatial overlay analysis is performed on the dataset of the spatial distribution of grazing suitability in high-altitude pastoral areas and the scope of prohibited development zones in the main functional area planning. The intersecting areas generated by the overlay are geometrically fused to obtain a preliminary dataset of national land spatial functional zoning. Step 4: Perform regular gridding processing on the preliminary land space functional zoning dataset, transforming the original irregular spatial unit functional zoning results into regular grid units of preset size, forming a gridded land space functional classification dataset; Step 5: Select spatial assessment units based on the gridded land spatial function classification dataset, analyze the spatial distribution characteristics of the spatial assessment units, and calculate the attribution association parameters between each spatial assessment unit and its surrounding adjacent units; use the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset. Step 6: Based on the updated gridded land spatial function classification dataset, a pre-trained graph neural network model is used to extract the spatial structure features of the grid units, and finally generate the land spatial function classification result dataset for the high-altitude pastoral area.

[0014] In this embodiment of the invention, multi-source spatial data from high-altitude pastoral areas are integrated. By constructing a standardized national land spatial element database, the one-sidedness of a single data source is effectively avoided. The grazing suitability assessment based on the national land spatial element database can reflect the grazing suitability characteristics of different spatial units in high-altitude pastoral areas, avoiding subjectivity and blindness in functional zoning. By spatially overlaying and geometrically fusing the grazing suitability spatial distribution dataset with the prohibited development zone, the organic combination of ecological protection needs and grazing suitability characteristics can be achieved. This ensures that the initially delineated national land spatial functional zoning not only meets regional ecological protection requirements but also takes into account the actual development of grazing, guaranteeing the rationality and feasibility of the functional zoning.

[0015] In a preferred embodiment of the present invention, step 1 above, which involves collecting and processing multi-source spatial data from high-altitude pastoral areas to construct a national spatial element database, may include: In this embodiment of the invention, the geographical boundaries of the target high-altitude pastoral area are first defined. Based on latitude and longitude coordinates, administrative boundaries, or natural geographical boundaries, the spatial scope of data collection is delineated to ensure coverage of the entire study area without redundant extensions. Subsequently, the core data types to be included in the database are determined, focusing on key elements affecting the classification of national land space functions. Specifically, this includes the following types of multi-source spatial data: basic geospatial data, including digital elevation model (DEM) data, used to extract topographic information such as elevation and slope, topographic vector data such as boundaries of landform types such as mountains, plains, and valleys, and administrative division data; grassland resource-related data, including grassland type distribution maps, grassland carrying capacity survey data, and grassland utilization status data; climate and environmental data, including spatial interpolation data of climate elements such as multi-year average temperature, precipitation, and accumulated temperature, and climate zoning data; disaster risk data, including vector data or raster data of the historical frequency and impact range of major pastoral disasters such as snow disasters and droughts; national land space planning data, including main functional area planning data and ecological protection red line data; and other auxiliary data, including transportation network data and water source distribution data.

[0016] Digital elevation model (DEM) data was downloaded from platforms such as the Geospatial Data Cloud and the National Bureau of Surveying and Mapping; climate data was obtained from multi-year observation data and spatial interpolation products from the National Meteorological Science Data Center and local meteorological bureaus; grassland type and vegetation distribution data were referenced from the survey results of local grassland authorities; data on functional zoning and ecological protection red lines were extracted from official documents issued by the Ministry of Natural Resources and the National Development and Reform Commission or from spatial data service platforms; disaster data was obtained from disaster survey results of emergency management departments and research institutes; for some areas lacking authoritative data sources or with insufficient data accuracy, field surveys were conducted to supplement data collection, and survey teams were organized to enter pastoral areas to record the location of settlements, grazing paths, and actual grassland utilization status through GPS positioning; quadrat surveys were used to measure the grass yield and vegetation coverage of grasslands in different areas to verify and supplement existing grassland resource data; descriptions of disaster occurrence and grazing habits by local herders were recorded, and disaster risk and grazing-related data were improved by combining field observation data; during the collection process, metadata information such as the original format, coordinate system, and data accuracy of various types of data were recorded simultaneously.

[0017] The raw data in different formats were uniformly converted into a standard format that facilitates subsequent analysis and storage. For example, raster data was converted to TIFF format, vector data to SHP format, and tabular attribute data to CSV format, and these were linked with the corresponding spatial data. To address potential coordinate system differences between different data sources, the authoritative coordinate system commonly used in the study area was used as a benchmark. Spatial data processing was used to transform the coordinates of all data to ensure that all types of spatial data were within the same coordinate framework, avoiding spatial location deviations. Based on the pre-defined study area boundaries, the collected global data was cropped, retaining valid data within the study area and eliminating redundant parts. For fragmented data of the same type, data splicing was performed to form a complete dataset covering the entire study area, ensuring spatial continuity. For missing data areas, spatial interpolation or analogy based on the similarity of adjacent areas was used to supplement the missing data, ensuring data integrity. Statistical analysis and field verification were used to identify and remove outlier data. A topological check was performed on the vector data to correct topological errors such as overlaps, gaps, and dangling nodes, ensuring the spatial logical consistency of the vector data.

[0018] The attribute information of various types of data is standardized and unified. Based on data type and application requirements, a hierarchical database structure is designed, divided into raster data layer, vector data layer, and attribute data layer. Further subdivisions are made based on feature themes; for example, the raster data layer includes sub-layers such as elevation raster, climate element raster, and disaster risk raster; the vector data layer includes sub-layers such as grassland type vector, administrative division vector, and prohibited development zone vector; and the attribute data layer stores detailed attribute information for each spatial feature. Simultaneously, data association rules are designed to ensure that spatial data and attribute data can be quickly linked and queried using unique identifiers. The preprocessed and standardized data are sequentially imported into the database management system according to the designed database structure. During the import process, batch verification of the data is performed to ensure that the data format, coordinate system, and attribute information conform to the import standards; for data that does not conform... Data on the same topic is categorized, stored, and indexed to optimize data query and retrieval efficiency. For example, spatial indexes are created for vector data such as commonly used grassland types and prohibited development zones, and keyword indexes are created for attribute data. A complete metadata archive is established for each type of data in the database, recording key information such as data source, collection time, processing method, coordinate system, data precision, and attribute description, facilitating data traceability and use. Database usage documentation is compiled, clarifying the database structure, data query methods, data update process, and other content. After data is entered into the database, comprehensive database testing is conducted, including data integrity testing, spatial consistency testing, and query performance testing. Based on the test results, the database structure and index design are optimized to ensure stable database operation and efficient data retrieval, meeting the data usage requirements of the land spatial function classification in high-altitude pastoral areas.

[0019] In a preferred embodiment of the present invention, step 2 above, which assesses the grazing suitability of each spatial unit based on the national land spatial element database and generates a spatial distribution dataset of grazing suitability in high-altitude pastoral areas, may include: In this embodiment of the invention, step 220 involves selecting multiple spatial attribute data affecting the suitability of animal husbandry from the national land spatial element database and performing normalization processing to form a standardized set of spatial attribute data. Specifically, this includes: determining the data selection criteria and specific data items; combining the core logic of animal husbandry production in high-altitude pastoral areas; selecting six direct and crucial categories of spatial attribute data from the constructed national land spatial element database; each category of data having clearly defined specific indicators; grassland type data including grassland level and carrying capacity level, directly reflecting the basic production potential of grassland; and climate data including multi-year average temperature, annual precipitation, and accumulated temperature ≥0℃. The core indicators directly affect the growth cycle, growth rate, and yield stability of pasture; elevation data, i.e., the actual altitude values ​​of each spatial unit within the study area, directly relates to the ease with which humans and livestock can enter the area; topographic slope data, i.e., the actual slope angle of each spatial unit, directly affects the convenience and safety of livestock grazing, as well as the growth and distribution of pasture; snow disaster risk data, i.e., the annual frequency of snow disasters in each spatial unit in history, directly relates to the risk of disaster losses in livestock production; grazing radius data, i.e., the straight-line distance of each spatial unit from the nearest settlement, directly affects the efficiency of herders in managing livestock and the convenience of livestock drinking water and foraging.

[0020] Data normalization is necessary because the original measurement standards of the six major data categories are completely different. Directly using them for comprehensive calculations would lead to distorted results due to dimensional differences. Therefore, a complete normalization operation needs to be performed on each data category separately. The first step is to statistically analyze the global range of values ​​for each category of data. For each category of spatial attribute data, all spatial units within the entire high-altitude pastoral area research area are traversed, and all original values ​​for that category are recorded. From these, the minimum and maximum values ​​are selected to define the value range for that category of data. The second step is to perform normalization calculations unit by unit. For the original value of each spatial unit in that category of data, the normalization is calculated as (original value of the spatial unit - minimum value of the category) ÷ (original value of the category of data). The calculation is performed using the logic of (maximum value - minimum value of this type of data). Through this operation, the original values ​​of all spatial units in this type of data are uniformly converted into dimensionless values ​​between 0 and 1. The third step is to integrate and form a standardized dataset. After all spatial units in the six categories of spatial attribute data have been normalized, these six categories of dimensionless data that have been converted to the 0 to 1 range are summarized and organized according to the structure of spatial unit ID + grassland type standardized value + climate standardized value + elevation standardized value + terrain slope standardized value + snow disaster risk standardized value + grazing radius standardized value to form a standardized set of spatial attribute data, ensuring that all types of data are comparable and superimposed on the same dimension.

[0021] Step 221 involves assigning importance to each data item in the standardized spatial attribute dataset, forming an importance allocation scheme. This includes determining the three bases for weight allocation: the actual production logic, and considering the specific scenarios of pastoral production in high-altitude pastoral areas. Grassland type directly determines forage yield and carrying capacity, serving as the core foundation of pastoral production and having the highest impact. Climate conditions determine whether forage can grow stably, acting as a key support for pastoral production, with the next highest impact. Topographic slope and grazing radius directly relate to the feasibility of grazing operations and management efficiency, having a moderate impact. Elevation indirectly affects pastoral production by influencing climate and accessibility for humans and livestock. The impact on production is relatively low; although snow disaster risk will cause losses, its occurrence is accidental and its impact is relatively concentrated, resulting in the lowest impact. Literature references include authoritative research findings in grassland ecological protection, livestock planning, and land space suitability assessment, referencing the weighting ratios of various influencing factors in similar assessments of high-altitude cold regions to ensure consistency with industry research consensus. Expert consultation is also provided, with a consultation team composed of grassland ecologists, senior livestock production managers, and land space planning experts conducting three rounds of specialized discussions. In the first round, experts independently proposed weighting allocation suggestions; the second round combined field research in high-altitude pastoral areas... Data, such as pasture yield, grazing frequency, livestock survival rate, and disaster loss records in different attribute regions, were analyzed one by one to demonstrate the impact strength of each type of data. In the third round, opinions were summarized and a unified plan was formed to ensure that the weight allocation reflected the actual situation. The specific weight allocation results were clarified, and based on the above three criteria, fixed weights were assigned to the six data items in the standardized spatial attribute data set, with the total weight of all data being 100%. The specific allocation is as follows: grassland type data weight 30%, as it directly determines the basic potential of livestock production and is a core influencing factor; climate data weight 20%, as it determines the stability of pasture growth and is a key supporting factor; topography... Slope data has a weight of 15% because it directly affects the convenience and safety of grazing operations, making it an important guarantee factor; grazing radius data has a weight of 15% because it is related to management efficiency and the satisfaction of livestock's basic needs, making it an important guarantee factor; elevation data has a weight of 10% because it indirectly affects livestock production, making it a secondary influencing factor; snow disaster risk data has a weight of 10% because its impact is accidental and localized, making it a secondary influencing factor. The above weight allocation results, the basis for the three-fold allocation, expert demonstration records, and supporting explanations from field survey data are compiled into a booklet, clarifying the origin and rationality of each data weight, forming a formal importance allocation scheme.

[0022] Step 222: Based on the importance allocation scheme, comprehensively calculate the spatial attribute data of each spatial unit to obtain the comprehensive evaluation value of pastoral suitability for each spatial unit. Specifically, this includes: determining the calculation unit and data extraction rules; taking a single spatial unit within the research area of ​​the high-altitude pastoral region as an independent calculation object; spatial units can be regular grids or vector plots; each unit has clear boundaries and does not overlap, ensuring coverage of the entire research area without omissions; for each spatial unit, extracting the six standardized data items corresponding to that unit from the standardized spatial attribute data set according to the principle of precise matching of spatial unit IDs: standardized value of grassland type, standardized value of climate, standardized value of terrain slope, standardized value of grazing radius, standardized value of elevation, and standardized value of snow disaster risk; ensuring that the extracted data corresponds one-to-one with the spatial unit, without mismatches or omissions; performing weight multiplication calculations item by item; and calculating the standardized data value × corresponding weight for each standardized data item of the spatial unit according to the weight ratio in the importance allocation scheme; obtaining the grassland type weight score; and determining the grassland type label for that spatial unit. The following weights are calculated: Standardized value × 30%; Climate weight score (climate standardized value of the spatial unit × 20%); Topographic slope weight score (topographic slope standardized value of the spatial unit × 15%); Grazing radius weight score (grazing radius standardized value of the spatial unit × 15%); Elevation weight score (elevation standardized value of the spatial unit × 10%); Snow disaster risk weight score (snow disaster risk standardized value of the spatial unit × 10%). These six weight scores are added sequentially: grassland type weight score + climate weight score + topographic slope weight score + grazing radius weight score + elevation weight score + snow disaster risk weight score. The sum is the comprehensive evaluation value of pastoral suitability for the spatial unit, which is between 0 and 1. Following a unified process of data extraction, weight multiplication, and summation, calculations are performed on all spatial units within the research area of ​​the high-altitude pastoral region, ensuring that the calculation steps for each unit are completely consistent without any differential operations. Finally, the comprehensive evaluation value of pastoral suitability for each spatial unit is obtained, forming a list of spatial unit IDs and corresponding comprehensive evaluation values.

[0023] Step 223: Based on the pre-set rules for classifying pastoral suitability levels, classify the comprehensive evaluation values ​​of pastoral suitability into levels to determine the pastoral suitability level of each spatial unit. Specifically, this includes: developing detailed criteria and standards for the level classification rules; combining the actual utilization scenarios of pastoral production in high-altitude pastoral areas, ecological protection red line requirements, historical suitability evaluation practice cases, and expert opinions; pre-setting five levels of pastoral suitability levels; clarifying the core characteristics of each level and the corresponding comprehensive evaluation value range; ensuring that the level classification is both realistic and clearly defined; the core characteristics of the extremely high suitability level are high forage yield, strong livestock carrying capacity, good climate adaptability, gentle terrain, proximity to settlements, and no snow. High suitability level: High suitability is characterized by high forage yield and strong carrying capacity, favorable climate, relatively flat terrain, proximity to settlements, low snow disaster risk, and stable grazing requiring only routine management. Medium suitability level: Medium forage yield and moderate carrying capacity, moderate climate, some slope in the terrain but not affecting grazing, moderate distance from settlements, moderate snow disaster risk, requiring targeted grazing management. Low suitability level: Low suitability is characterized by... Low grass yield and weak carrying capacity, average climate conditions, steep terrain, far from settlements, and high risk of snow disasters make grazing limited to specific seasons or require intensive management. The corresponding comprehensive evaluation value range is 0.2 to 0.4. The core characteristics of the extremely low suitability level are extremely low grass yield and weak carrying capacity, harsh climate conditions, steep terrain, or extremely high risk of snow disasters. It is unsuitable for grazing and should be primarily used for ecological protection. The corresponding comprehensive evaluation value range is 0 to 0.2. A unit-by-unit classification operation is performed, extracting the comprehensive evaluation value of pastoral suitability for each spatial unit. The value is then compared to the preset classification rules to determine its range. If the comprehensive evaluation value is 0.83, it falls within the range specified above. A score between 0.8 and 1.0 is considered extremely suitable; a score of 0.72 falls between 0.6 and 0.8 and is considered high suitable; a score of 0.55 falls between 0.4 and 0.6 and is considered moderately suitable; a score of 0.34 falls between 0.2 and 0.4 and is considered low suitable; a score of 0.18 falls between 0 and 0.2 and is considered extremely low suitable. During the classification process, if the comprehensive evaluation value is exactly equal to the interval boundary value, it is classified into the next higher level according to the principle of choosing the higher value, ensuring that all comprehensive evaluation values ​​can uniquely correspond to one suitability level, without omissions or overlaps.

[0024] Step 224 involves matching and integrating the pastoral suitability level of each spatial unit with its geographical location information to generate a spatial distribution dataset of pastoral suitability in high-altitude pastoral areas. Specifically, this includes: extracting complete geographical location information; for each spatial unit with a determined pastoral suitability level, extracting comprehensive and accurate geographical location information from the national land spatial elements database, including latitude and longitude coordinates accurate to the second for the unit's center point, and the set of latitude and longitude coordinates for all inflection points of the unit boundary to ensure precise spatial location; spatial boundary data presented as unit boundary information in vector coordinate string form, clearly defining the shape and extent of the unit; administrative affiliation information including the names and corresponding administrative codes of the provincial, municipal, county, and township-level administrative regions to which the unit belongs, facilitating statistical analysis by administrative region; adjacent unit association information including a list of spatial unit IDs directly adjacent to the unit, providing support for spatial analysis; and establishing precise matching relationships. By using unique spatial unit IDs, the pastoral suitability level of each unit is bound one-to-one with the extracted geographic location information, forming a complete data record consisting of unit ID + center point latitude and longitude + boundary inflection point coordinates + administrative affiliation + adjacent unit IDs + suitability level name + suitability level code. This ensures that each geographic location uniquely corresponds to a suitability level, without any association errors or chaotic many-to-one or one-to-many relationships. All complete data records of spatial units bound to suitability levels are summarized and organized according to industry-standard spatial data formats. If it is a vector format, it includes a vector boundary file and an attribute data table, with the attribute data table linked to the vector boundary through the unit ID. If it is a raster format, the suitability level code corresponding to each pixel is precisely mapped to the latitude and longitude coordinates of that pixel, forming a raster matrix. Finally, a spatial distribution dataset of pastoral suitability covering the entire research area of ​​the high-altitude pastoral region is generated.

[0025] By screening key spatial attribute data that affect the suitability of animal husbandry and normalizing them, the comparability of the data and the rationality of the calculation are ensured; the clear five-level classification makes the suitability results intuitive and easy to understand; the matching and integration of geographical location and suitability level forms a structured and usable spatial distribution dataset.

[0026] In a preferred embodiment of the present invention, step 3 above involves spatially overlaying the dataset of the spatial distribution of grazing suitability in high-altitude pastoral areas with the prohibited development zones in the main functional zoning plan, and geometrically fusing the intersecting areas generated by the overlay to obtain a preliminary dataset of national land spatial functional zoning, which may include: In this embodiment of the invention, step 330 involves forming an input data layer to be analyzed based on the spatial distribution dataset of grazing suitability in high-altitude pastoral areas and the data on prohibited development zones in the main functional area planning scheme. Specifically, this includes: first, determining the detailed composition of the two core datasets; the spatial distribution dataset of grazing suitability in high-altitude pastoral areas must cover complete information on every independent spatial unit within the area, including but not limited to grazing suitability level, precise latitude and longitude coordinate range, set of inflection point coordinates of unit boundaries, comprehensive evaluation value of grazing suitability, name of the sub-region to which the unit belongs, and grassland type; the data on prohibited development zones in the main functional area planning scheme must contain all key information on legally prohibited development zones, specifically including the unique administrative number of the prohibited development zone, set of inflection point coordinates of the legal boundaries, area type, description of control requirements, total area of ​​the area, and administrative jurisdiction. Next, rigorous data preprocessing is carried out: first, unifying the data format by converting both datasets into a common vector data format, ensuring that each dataset contains complete components such as geometric information files, attribute information files, and projection information files, where the geometric information files record boundary inflection point coordinates, and the attribute information files correspond to the aforementioned attribute data; second, unifying the coordinate system. Both types of data use the WGS-84 geographic coordinate system and the UTM projected coordinate system. Coordinate transformation tools were used to transform the original data one by one, ensuring that the coordinate values ​​of the same location are completely consistent in both types of data, eliminating spatial location deviations caused by coordinate system differences. Thirdly, comprehensive data cleaning was performed, verifying each spatial unit in both types of data. For units with abnormal boundaries, corrections were made by consulting the original data source; for units with incorrect attributes, corrections were made by referring to relevant planning documents and assessment reports; for units with duplicate entries, the most complete and up-to-date record was retained, and redundant data was deleted, ensuring the accuracy and uniqueness of both datasets. Finally, an input data layer for analysis was constructed. The preprocessed spatial distribution dataset of grazing suitability in high-altitude pastoral areas was used as one independent spatial layer, and the data on prohibited development zones in the main functional area planning was used as another independent spatial layer. Both layers were loaded into the same spatial analysis platform and spatially aligned based on a unified coordinate reference frame, ensuring that the two layers correspond accurately both visually and logically. This ensures that information from the same geographic location can be accurately correlated during subsequent overlay analysis, ultimately forming an input data layer containing the two correlated layers.

[0027] Step 331: Perform spatial overlay calculation on the two datasets in the input data layer to generate overlay analysis results describing the spatial relationship between them. Specifically, this includes: using a unit-by-unit, full-coverage spatial overlay calculation method, taking each independent spatial unit in the dataset of grazing suitability of high-altitude pastoral areas as the smallest analysis unit, and comprehensively determining the spatial relationship between each unit and each prohibited development area in the dataset of prohibited development areas of the main functional area in ascending order of unit number. The determination process follows clear logical rules and is specifically divided into three categories for verification. The first category is to determine whether the grazing suitability spatial unit is completely located within a certain prohibited development area. The verification method is to confirm that the coordinates of all boundary inflection points of the grazing suitability unit fall within the corresponding prohibited development area. The first category is to determine whether a suitable grazing unit is within the area enclosed by the coordinates of the boundary inflection points, and whether any point of the suitable grazing unit exceeds the boundary of the prohibited development area. The second category is to determine whether a suitable grazing unit partially overlaps with a prohibited development area. The verification method is to check whether at least one boundary inflection point of the suitable grazing unit falls within the prohibited development area, or whether the boundary line of the suitable grazing unit intersects with the boundary line of the prohibited development area at least once, or whether the graphic ranges of the two partially overlap. The third category is to determine whether a suitable grazing unit has no spatial overlap with any prohibited development area. The verification method is to confirm that all boundary inflection points of the suitable grazing unit do not fall within any prohibited development area, and that its boundary line does not intersect with the boundary lines of any prohibited development area, and that the graphic ranges do not intersect.

[0028] During the comparison of each suitable grazing unit with the prohibited development area, detailed comparison results are recorded simultaneously, including spatial relationship type identifiers. If there is complete overlap, the administrative number, area type, and control requirements of the corresponding prohibited development area are recorded, as well as the grazing suitability level, comprehensive evaluation value, and unit number of the suitable grazing unit. If there is partial overlap, in addition to recording the basic attributes of the suitable grazing unit and the prohibited development area, the set of boundary inflection point coordinates of the overlapping area, the shape description of the overlapping area, the proportion of the overlapping area to the total area of ​​the suitable grazing unit, and the proportion of the overlapping area to the total area of ​​the corresponding prohibited development area are also recorded. At the same time, the set of boundary inflection point coordinates of the non-overlapping part of the suitable grazing unit, the shape and range description of the non-overlapping part are also recorded. If there is no overlap, only the complete attribute information of the suitable grazing unit and the non-overlap identifier are recorded. All comparison records are named and sorted in a unified format of suitable grazing unit number - prohibited development area number, and compiled to form a complete overlay analysis result, ensuring that the spatial relationship between each suitable grazing unit and each prohibited development area is clearly recorded without omissions or mismatches.

[0029] Step 332: Based on the overlay analysis results, extract all areas in the spatial distribution data of grazing suitability in high-altitude pastoral areas that spatially overlap with the prohibited development zone, generating a set of intersecting areas. Specifically, this includes: first, establishing clear screening rules; from the overlay analysis results generated in Step 331, selecting all comparison records marked as completely encompassing or partially overlapping, and directly excluding records marked as non-overlapping, ensuring that only all associated data with spatial overlap are retained; for each selected completely encompassing comparison record, extracting the complete spatial information of the grazing suitability spatial unit, including the set of all boundary inflection point coordinates of the unit, the complete graphic range of the unit, grazing suitability level, comprehensive evaluation value, unit number, and the name of the sub-region to which it belongs; associating this information with the corresponding administrative number and region type of the prohibited development zone, treating the complete graphic of the grazing suitability unit as an independent geometric feature, and clarifying that the feature's attribute is completely located within the prohibited development zone; for each selected partially overlapping comparison record, extracting detailed spatial information of the overlapping areas; first, by comparing the boundary inflection point coordinates of the grazing suitability unit and the prohibited development zone, identifying all intersections that simultaneously belong to the boundaries of both the grazing suitability unit and the prohibited development zone. Point coordinates, as well as inflection point coordinates belonging only to suitable pasture units and falling within prohibited development areas, and inflection point coordinates belonging only to prohibited development areas and falling within suitable pasture units, are sequentially connected in a clockwise or counterclockwise order to form a complete set of boundary inflection point coordinates for the overlapping area, thus determining the graphic range of the overlapping area. Then, the graphic shape description, boundary inflection point coordinate set, suitable pasture unit number and suitable pasture level, and corresponding prohibited development area administrative number and area type of the overlapping area are recorded. This independent overlapping area is treated as a separate geometric graphic element, and its attribute is clearly defined as partially located within the prohibited development area. After extracting all independent geometric graphic elements, a unique identification number is assigned to each element to ensure that each element's identifier is unique. Simultaneously, all extracted geometric graphic elements are checked to eliminate duplicate or incomplete extractions due to data errors. For duplicate elements, the most complete record is retained; for incomplete elements, the original coordinate data is supplemented to complete the data. Finally, all checked, numbered, and attribute-completed independent geometric graphic elements are collected and sorted by identification number to form a set of intersecting areas with a clear structure and complete information.

[0030] Step 333: Perform boundary fusion processing on all independent geometric figures within the intersecting region set to generate one or more seamless merged polygonal features. Specifically, this includes: firstly, conducting a comprehensive spatial position verification on all independent geometric figures within the intersecting region set; selecting each geometric figure as the target figure and comparing it pairwise with all other geometric figures in the set to check their spatial relationships; specifically, checking for adjacency (the minimum distance between the boundary lines of two figures is less than 1 meter and there is no overlap); nesting (all boundary inflection points of one figure fall within the range of another figure, forming an inclusion relationship); secondary overlap (both figures are partially overlapping features with additional overlapping areas); and lack of association (the boundary distance between two figures is greater than 1 meter). Within 1 meter, with no inclusion or overlap, during the verification process, the spatial relationship type and related details of each group of graphics are recorded, such as the specific line segments of adjacent boundaries, the numbers of nested inner and outer layer graphics, and the range of areas with secondary overlap. Subsequently, boundary fusion processing is carried out according to the spatial relationship type. For graphics with adjacent relationships, the boundary line segments of the two graphics that are adjacent to each other are extracted, and the inflection point coordinates of the adjacent line segments are checked to see if they can be accurately connected. For example, whether the boundary endpoint coordinates of the previous graphic and the boundary start coordinates of the next graphic are consistent. If there is a slight deviation, fine-tuning is performed based on the coordinates of the closer inflection point to ensure that the two boundaries can be seamlessly connected. Then, the adjacent boundary line segments that overlap in the two graphics are deleted, and the remaining boundary line segments are connected in a unified clockwise order to form a continuous and complete boundary line, which is then merged into a unified geometric figure.

[0031] For graphics with a double overlap, the overlapping regions of the two graphics are extracted. The boundary inflection point coordinates of these regions are then integrated sequentially. The union of the two graphics, i.e., the smallest graphic containing the entire range of both graphics, is taken as the range of the merged graphic. Duplicate boundary segments and inflection point coordinates are removed from both graphics to ensure that there are no overlapping regions within the merged graphic, while preserving the original attribute information of both graphics. For graphics with a nested relationship, the outermost graphic in the nested structure is first identified, i.e., the graphic containing the most other graphics and having the largest boundary range. Then, the boundary ranges of all inner graphics are included within the boundary range of the outermost graphic. Duplicate boundary segments between inner and outer graphics are removed, while the complete boundary of the outermost graphic is preserved, forming a single boundary containing all graphics. The inner region features a unified geometry; the merged graphic attributes record the identification number and corresponding attributes of all included inner graphics to ensure no attribute information is lost; after all fusion operations are completed, the generated merged polygon features are finally verified to check whether the boundaries of each merged polygon are continuous, whether there are blank or overlapping areas inside, and whether the graphic range accurately covers all independent graphic ranges in the original intersecting area set; for those with broken boundaries, the inflection point coordinates are added for repair; for those with internal blanks or repetitions, the boundary range is readjusted; ensuring that each merged polygon feature is a geometric shape with complete boundaries, seamless interior, and accurate range, ultimately forming one or more seamless merged polygon features that meet the requirements.

[0032] Step 334 involves integrating the merged polygonal features with the non-overlapping areas in the spatial distribution data of grazing suitability in high-altitude pastoral areas to generate a preliminary national land spatial functional zoning dataset including prohibited development zone markers. Specifically, this includes: first, extracting the non-overlapping areas from the spatial distribution data of grazing suitability in high-altitude pastoral areas; based on the overlay analysis results of Step 331, selecting all spatial units marked as non-overlapping grazing suitability; verifying the boundary inflection point coordinates of each unit to confirm that it does not intersect with the boundary of any prohibited development zone or fall within the scope of any prohibited development zone; extracting the complete information of these units, including the set of boundary inflection point coordinates, grazing suitability level, comprehensive evaluation value, unit number, name of the sub-region to which it belongs, grassland type, etc., ensuring that the spatial extent and attribute information of each non-overlapping unit are complete, while also providing... Non-overlapping areas are assigned a unified category identifier. Next, spatial integration of the two types of areas is performed. The merged polygonal elements generated in step 333 are loaded onto the extracted non-overlapping areas within the same spatial reference frame, and the entire high-altitude pastoral area is spatially stitched according to its latitude and longitude range. During the stitching process, a zonal verification and full-area coverage method is adopted. The high-altitude pastoral area is divided into several 1 km × 1 km small grids according to latitude and longitude. The regional coverage within each small grid is checked one by one to ensure that each grid has a corresponding area without any spatial omissions. Simultaneously, it is checked whether there is any overlapping coverage of areas, i.e., the same grid contains both merged polygonal elements and non-overlapping areas. If so, the attribution of the overlapping areas is confirmed by comparing with the original overlay analysis results. The correct area type is retained, and the overlapping parts are deleted.

[0033] Then, functional zoning identifiers are added to all integrated areas. For merged polygonal elements, the functional identifier is uniformly set as a prohibited development area, and the corresponding administrative number, area type, and control requirements of the prohibited development area are supplemented in the attribute information. For non-overlapping areas, preliminary functional type identifiers are set according to their original grazing suitability levels. Non-overlapping units with grazing suitability levels of very high, high, and medium are tentatively designated as agricultural space candidate areas, indicating that their core function is inclined towards pastoral production. Non-overlapping units with grazing suitability levels of low and very low are tentatively designated as ecological space candidate areas, indicating that their core function is inclined towards ecological protection. Finally, the attribute information and format of the dataset are improved, and basic information such as the total area, central latitude and longitude coordinates, and number of boundary inflection point coordinates are supplemented for each area. The geometric information and attribute information of all areas are organized according to the standard vector data format to ensure that the geometric information file, attribute information file, and projection information file match each other. The dataset undergoes an overall quality check to eliminate problems such as missing attributes, incorrect coordinates, and chaotic identifiers. Finally, a preliminary national land space functional zoning dataset is generated that fully covers the high-altitude pastoral area, has clear functional identifiers, complete attribute information, and a standardized format.

[0034] The data's usability and operability are improved. The integrated preliminary dataset has a standardized format and complete information, and can be used for grid processing, adaptation and adjustment without the need for a lot of additional data sorting. This reduces the difficulty of subsequent processes, while the detailed attribute records also facilitate data traceability and the implementation of control measures.

[0035] In a preferred embodiment of the present invention, step 4 above, which involves performing regular gridding processing on the preliminary land spatial functional zoning dataset to transform the original irregular spatial unit functional zoning results into regular grid units of a preset size, forming a gridded land spatial functional classification dataset, may include: In this embodiment of the invention, step 440 involves generating a spatial grid array covering the entire study area, composed of multiple regular grids of identical size and shape, based on the preset regular grid unit size and the entire area of ​​the high-altitude pastoral region. Specifically, this includes: firstly, determining the core basic parameters: one is the preset size of the regular grid units, requiring a unified side length standard to ensure that all grid units have identical length and width, and are all standard squares; the other is defining the entire area of ​​the high-altitude pastoral region by collecting data on the region's legal administrative boundaries and remote sensing image coverage data to determine the extreme latitude and longitude values ​​of the entire region, i.e., the easternmost east longitude coordinate, the westernmost east longitude coordinate, the northernmost north latitude coordinate, and the southernmost north latitude coordinate, forming... The rectangular spatial bounding box of the entire area clearly defines the coverage boundary of the spatial grid array. Next, the core parameters of the grid array are calculated: First, the number of horizontal grids is calculated by subtracting the easternmost east longitude coordinate from the westernmost east longitude coordinate of the entire area to obtain the east-west longitude distance. This longitude distance is then divided by the longitude distance corresponding to the side length of a single grid cell. If the result is an integer, the number of horizontal grids is that integer; if the result has a decimal, it is rounded up to ensure that the easternmost region is completely covered. Second, the number of vertical grids is calculated by subtracting the southernmost north latitude coordinate from the northernmost north latitude coordinate of the entire area to obtain the north-south latitude distance. This latitude distance is then divided by the latitude distance corresponding to the side length of a single grid cell, again following the same principle of rounding up. The number of vertical grid cells is determined by rounding up. Then, each regular grid cell is generated sequentially, starting from the top left corner of the rectangular area encompassing the entire high-altitude pastoral region. The first row of grid cells is generated along the east-west direction. The top left corner coordinates of the first grid cell are the starting point coordinates, and the bottom right corner coordinates are the westernmost longitude coordinates plus the longitude distance corresponding to the single grid side length, and the northernmost latitude coordinates minus the latitude distance corresponding to the single grid side length. The top left corner coordinates of the second grid cell are the top right corner coordinates of the first grid cell, and the bottom right corner coordinates are the first grid cell's bottom right longitude coordinates plus the longitude distance corresponding to the single grid side length, plus the first grid cell's bottom right latitude coordinates, and so on, until all horizontal grid cells in the first row are generated, covering the entire east-west region. Then, starting from the first... The coordinates of the bottom left corner of the leftmost grid in the first row become the coordinates of the top left corner of the first grid in the second row. Following the same logic, the second row of grids is generated along the east-west direction, and this process continues downwards until all vertical grids are generated, covering the entire north-south area. Finally, a spatial grid array is verified and formed, checking the coordinate accuracy of each grid to ensure seamless connection between adjacent grids. Even if edge grids exceed the legal administrative boundaries of the high-altitude pastoral area, they retain their complete grid shape without being cut or split. Each grid is assigned a unique identification number, recording the coordinates of its complete boundary inflection points, side length, area, and other information. This ultimately forms a spatial grid array covering the entire high-altitude pastoral area, with uniform grid size and shape, unique numbers, and accurate coordinates.

[0036] Step 441 involves calculating the spatial location relationship between the spatial grid array and the preliminary national land spatial functional zoning dataset to obtain spatial association results describing the coverage relationship between each regular grid and different functional zoning units. Specifically, this includes: first, determining the basis for comparison between the two types of data; the spatial grid array is a set of grids of uniform specifications, with each grid being an independent comparison unit; the preliminary national land spatial functional zoning dataset contains three types of functional zoning units: prohibited development areas, agricultural space candidate areas, and ecological space candidate areas, each of which is an independent irregular spatial graphic. The two types of data are loaded into the same spatial analysis environment, using the previously unified WGS-84 geographic coordinate system and UTM projection coordinate system to ensure spatial alignment without deviation. Next, a grid-by-grid, full-traversal comparison method is used to calculate the spatial location relationship. Each regular grid is selected in ascending order of grid identification number. The target grid is used to determine its spatial relationship with each functional partition unit in the initial partitioning data. The determination logic is divided into four core cases, which are checked one by one: 1. The target grid is completely covered by a functional partition unit, that is, the coordinates of the four vertices and all internal points of the target grid are within the boundary range of the functional partition unit, with no part exceeding it; 2. The target grid partially overlaps with a functional partition unit, that is, part of the target grid is within the range of the functional partition unit, and another part is within other partition units or areas not covered by any partition unit; 3. The target grid overlaps with multiple functional partition units simultaneously, that is, different areas of the target grid are within the range of two or more different functional partition units; 4. The target grid does not overlap with any functional partition units, that is, the four vertices and all internal points of the target grid are not within the boundary range of any functional partition unit.

[0037] During the comparison of each set of target grids and functional partition units, detailed information is recorded simultaneously: first, the coverage relationship type identifier; second, the associated functional partition unit information, including the name, function type, and partition unit number of the partition unit; and third, the specific information of the overlapping area. If it is a complete coverage, the confirmation identifier that the partition unit completely covers the target grid is recorded; if it is a partial overlap or multiple unit overlap, the coordinate set of the boundary inflection points of the overlapping area and the location description of the overlapping area within the target grid are recorded; if there is no overlap, only the target grid number and the no-overlap identifier are recorded. Finally, the spatial association results are compiled, and all comparison records are summarized according to the identification number of each target grid. Each grid corresponds to a complete association record, and the recorded content includes the grid number, the number and specific number of associated functional partition units, the function type of each associated partition unit, the coverage relationship type, and the coordinates and location description of the overlapping area. This ensures that the spatial relationship between each grid and all functional partition units is completely recorded without omissions or mismatches, ultimately forming a structured spatial association result.

[0038] Step 442: Based on the spatial correlation results, calculate the area of ​​each type of national land space functional zoning unit covered within each regular grid, and generate a statistical table recording the area ratio of each type of functional zoning unit within each grid. Specifically, this includes: First, calculating the total area of ​​each regular grid. Since all grids have the same size, the total area of ​​a single grid = grid side length × grid side length. This is calculated and recorded in the grid basic information beforehand, eliminating the need for repeated calculations. Next, calculate the coverage area of ​​each type of functional zoning unit according to the grid number. For fully covered grids, the area of ​​a certain type of functional zoning unit covered within the grid = the total area of ​​the grid; the coverage area of ​​other types of functional zoning units = 0. For partially overlapping grids, extract the coordinates of the boundary inflection points of the overlapping area between the grid and the corresponding functional zoning unit. Use the geometric segmentation and summation method to calculate the overlapping area, dividing the irregular overlapping area into multiple regular small shapes. Calculate the area of ​​each small shape separately (triangle area = base × height ÷ 2, rectangle area = length × width), and then sum the areas of all small shapes to obtain the coverage area of ​​the functional zoning unit within the grid. For other types of functional zoning units... Coverage area = 0; For multi-unit overlapping grids, for each functional partition unit overlapping with the grid, calculate their respective overlapping area using the geometric segmentation and summation method described above, ensuring that the coverage area of ​​each partition unit is counted separately and without confusion; For non-overlapping grids, the coverage area of ​​all functional partition units is 0; Then, calculate the area proportion of each type of functional partition. For each grid, divide the coverage area of ​​a certain type of functional partition unit by the total area of ​​the grid to obtain the area proportion of that type of functional partition within the grid; if the coverage area of ​​a certain type of functional partition is 0, then the proportion is 0%; Finally, generate a standardized statistical table. The statistical table uses the grid identification number as the core index, with each row corresponding to a regular grid. Columns include grid number, total grid area, coverage area of ​​prohibited development areas, area proportion of prohibited development areas, coverage area of ​​agricultural space candidate areas, area proportion of agricultural space candidate areas, coverage area of ​​ecological space candidate areas, area proportion of ecological space candidate areas, coverage area of ​​no partition, and area proportion of no partition, ensuring that the area and proportion data of each type of grid are clearly traceable, the statistical table format is uniform, the data is accurate, and there are no calculation errors.

[0039] Step 443: Based on the area percentage data recorded in the statistical table and following the preset functional type determination rules, determine and assign a dominant national land space functional type to each rule grid, while retaining the internal functional type composition information, forming a set of grid units with dominant functional type and internal functional type composition information; specifically including: first, determining the preset functional type determination rules to ensure consistent determination logic; priority rules, with functional type priority from high to low as follows: prohibited development areas > agricultural space candidate areas > ecological space candidate areas; percentage priority rules, if the area percentage of a certain type of functional zone is the highest in the grid and no other type of functional zone has the same percentage, then the type of functional zone is directly determined as the dominant functional type of the grid; parallel percentage rules, if the area percentage of two or more types of functional zones is the same and both are the highest, then according to the above priority rules, the type of functional zone with the highest priority is selected as the dominant functional type; no coverage rule, if the coverage area percentage of all functional zones in the grid is 0%, then its dominant functional type is temporarily set as an area to be further determined; then, determine the dominant functional type for each rule grid one by one. The process involves determining the type of functional area. Following grid number order, the statistical table is retrieved to determine the area percentage of each functional area within that grid. First, it's determined if a single, highest-performing functional area type exists. If a single highest-performing type exists, it's directly assigned as the dominant function. If multiple highest-performing types exist, the dominant function is selected based on priority rules. If all percentages are 0%, the area awaiting further determination is assigned as the dominant function. Then, the internal functional type composition information for each grid is retained. For each grid, detailed information on all internal functional areas is recorded, including the name, coverage area, and area percentage of each type. Even if a functional area has a 0% percentage, the coverage area (0 square meters) and percentage (0%) must be recorded to ensure complete and traceable composition information and the actual distribution of functional types within the grid. Finally, a grid unit set is formed. For each grid, the dominant function type identifier is added, integrating the grid's core information: grid identification number, boundary inflection point coordinates, side length, total area, dominant function type, and the coverage area and percentage of each internal functional area. This information is then organized and summarized in numerical order to form a grid unit set where each unit has complete information, a clear dominant function, and traceable composition information.

[0040] Step 444 integrates the spatial geometric information, dominant functional type, and internal functional type composition information of the grid unit set to generate a gridded land space function classification dataset with unified geometric form. Specifically, this includes: first, sorting out and integrating all necessary information to ensure no omissions; this includes: 1) spatial geometric information, namely the complete boundary inflection point coordinates, grid shape, side length, and area of ​​each grid unit; 2) functional attribute information, namely the dominant functional type of each grid unit, and the coverage area and area ratio of various internal functional zones; and 3) basic identification information, namely the unique identification number and data source of each grid unit. Next, unifying the data format and standards, all integrated information is organized according to the standard vector data format, divided into three categories: geometric information files, attribute information files, and projection information files. The geometric information files record the boundary inflection point coordinates and graphic type of each grid; the attribute information files are associated with grid numbers and sequentially enter attribute data such as the dominant functional type, the coverage area and ratio of various functional zones, and the grid area. The projection information files retain unified WGS-84 geographic coordinate system and UTM projection coordinate system parameters to ensure that the coordinate system of the dataset is consistent with all previous data, allowing direct integration with subsequent analysis processes. Then, data association and verification are performed, checking the precise correlation between the geometric and attribute information of each grid cell, i.e., ensuring no mismatch between the boundary coordinates corresponding to the grid number and the functional attributes; verifying the accuracy of the area calculation for each grid (grid area = side length × side length, consistent with the area recorded in the attribute information); verifying that the determination of the dominant functional type conforms to preset rules; checking the data format for standardization, and promptly correcting any errors found; finally, generating the final dataset by packaging and integrating the verified geometric information files, attribute information files, and projection information files, clearly defining the dataset's name, version, generation time, applicable scope, and other metadata information to ensure the dataset's integrity and identifiability; ultimately forming a gridded national spatial function classification dataset with unified geometric form, clear functional attributes, accurate coordinates, and standardized format.

[0041] By standardizing and unifying spatial units, the original irregular functional partition units are transformed into regular grid units, improving the universality of data and preserving the core information of functional partitions. This not only clarifies the core functional orientation of each grid through the dominant functional type, but also fully preserves the composition and proportion of internal functional types, thereby improving the coherence and efficiency of data processing.

[0042] In a preferred embodiment of the present invention, step 5 above, which involves selecting spatial assessment units based on a gridded land spatial function classification dataset, analyzing the spatial distribution characteristics of the spatial assessment units, calculating the attribution association parameters between each spatial assessment unit and its neighboring units, and using the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset, may include: In this embodiment of the invention, step 550 involves selecting regular grids as spatial assessment units based on the gridded land space function classification dataset to obtain a set of spatial assessment units. Specifically, this includes using the previously generated gridded land space function classification dataset as the core basis. The size of the regular grid in this dataset needs to be determined by considering the actual management needs of high-altitude pastoral areas, the resolution of remote sensing image data, and the complexity of the terrain. For example, it can be uniformly set to a 1km × 1km square grid to ensure that the grid size accurately reflects local functional differences while also considering the efficiency of overall data processing. Each regular grid contains clear spatial attribute information, including latitude and longitude boundary coordinates, planar coordinates in the projected coordinate system, the determined dominant land space function type, and various internal functions. Detailed information such as the specific distribution range and actual occupied area of ​​each type is included. When selecting spatial assessment units, no additional grid units are added, filtered, or removed. Each regular grid in the dataset is directly identified as an independent spatial assessment unit, and each assessment unit is assigned a unique identifier. The numbering rule adopts the regional administrative division code + grid row and column number to ensure that each unit can be accurately located and traced through the number. Then, in the order of grid horizontal row and column number from left to right and vertical row and column number from top to bottom, all regular grid assessment units with unique identifiers are fully integrated to form a set of spatial assessment units that cover the entire high-altitude pastoral area without omissions or duplications. The spatial range of each unit in the set is seamlessly connected, and together they constitute a complete grid-based coverage system for the high-altitude pastoral area.

[0043] Step 551: Process the gridded land space function classification dataset using a set of spatial assessment units. By statistically analyzing the area proportion of functional types within each unit, the internal composition and spatial arrangement patterns of unit functions are analyzed to obtain the spatial distribution characteristics of each spatial assessment unit. Specifically, this includes: calling the established set of spatial assessment units, and for each spatial assessment unit with a unique identifier in the set, performing a complete process of area calculation, proportion statistics, and feature analysis. First, calculate the total area of ​​the unit. Since the grid uses a projected coordinate system, it can be directly calculated using the grid side length (total area = horizontal side length × vertical side length, i.e., 1km × 1km = 1km). 2If there are discrepancies in the actual surface area due to topographic relief, corrections must be made using Digital Elevation Model (DEM) data to ensure the accuracy of area calculations. Next, the functional types and corresponding areas within each unit are identified and extracted. Based on the spatial boundary coordinates of each functional type within the unit, the actual occupied areas of ecological space, agricultural space, and prohibited development space are calculated using the area calculation function of geographic information tools. Then, the area percentage of each function is calculated by dividing the actual occupied area of ​​each functional type by the corrected total area of ​​the unit. The percentage results are rounded to the nearest integer to ensure data consistency. Based on the area percentage data, the functional composition characteristics are analyzed to identify the dominant functional type with the highest percentage. The spatial arrangement pattern is analyzed by combining the least prevalent secondary functional types, the balance of the proportion of various functions, and the location of functional distribution. By comparing the relative positions of the boundary coordinates of various functions with the unit boundary, the functional distribution pattern is determined. If a certain type of function is concentrated in the central area of ​​the unit and continuously occupies more than 70% of the area, it is judged as a centralized distribution; if a certain type of function is divided into three or more discontinuous small blocks distributed in the unit, it is judged as a dispersed distribution; if a certain type of function is distributed in a strip along the edge of the unit, it is judged as an edge distribution. Combining the functional composition characteristics with the spatial arrangement pattern forms the unique spatial distribution characteristics of the unit. Finally, the characteristic results of all units are summarized by unique identifier number to form a complete spatial distribution characteristic dataset.

[0044] Step 552: Based on the spatial distribution feature results, for each target spatial evaluation unit, identify all adjacent units within the preset spatial range and calculate the similarity in functional type composition between the target unit and each adjacent unit. Specifically, this includes: first, setting a unified preset spatial range, based on the core characteristics of high-altitude pastoral areas, referring to the radius of daily livestock grazing activities, the spatial influence range of herders' migration routes, and the size of a 1km×1km regular grid. The range is determined by extending one grid of the same size in each of the four directions (east, west, south, and north) from the geometric center coordinates of the target unit, forming a 3×3 grid matrix. This means the target unit is located at the center, surrounded by eight adjacent grids, and forms a rectangular area. This range can comprehensively cover potential adjacent units with actual functional connections to the target unit, avoiding omission of key related areas. Based on the spatial distribution feature dataset generated in step 551, for each target spatial evaluation unit... For each spatial evaluation unit, the geometric center coordinates of all evaluation units are compared. If the center coordinates of a unit are within the boundary of a 3×3 grid matrix, it is considered an adjacent unit of the target unit. Units that directly overlap with the geometric boundary of the target unit are directly adjacent units, while those whose boundaries do not overlap but whose center coordinates are within the matrix are not directly adjacent units. Subsequently, the similarity of functional types is calculated. The area ratio data of the three functional types between the target unit and each adjacent unit are extracted. The similarity is determined by summing the absolute values ​​of the differences in the ratios of each functional type. The absolute values ​​of the differences in the ratios of the corresponding functional types are calculated one by one, and the three absolute values ​​are added together to obtain the total difference value (2%+2%+0%=4%). The smaller the total difference value, the more similar the functional composition of the two units is, and the higher the similarity. The larger the total difference value, the lower the similarity. A one-to-one similarity record is established for each target unit and its adjacent units to ensure that there is a clear basis for calculating the correlation parameters.

[0045] Step 553: Based on similarity, and considering the spatial adjacency or spatial distance between the target unit and its adjacent units, and integrating both functional and spatial factors, calculate the attribution association parameters between the target unit and each adjacent unit. Specifically, this includes: first, determining the rules for setting the spatial relationship weight coefficient. These rules reference industry practices regarding the functional impact of adjacent areas in the spatial planning of high-altitude pastoral areas. For directly adjacent units, due to their close spatial connection, the functional impact is more direct and significant, so a fixed weight coefficient of 0.8 is set. For non-directly adjacent units, the weight coefficient is calculated according to the spatial attenuation law where the closer the spatial distance, the larger the weight coefficient. The center-to-center straight-line distance between the target unit and its adjacent unit is calculated. The maximum center-to-center distance within the 3×3 matrix is ​​preset to be 1414m. Therefore, the distance weight coefficient = 1 - (actual center-to-center distance ÷ maximum center-to-center distance). The weighting coefficient is kept between 0 and 0.8 to objectively reflect the impact of spatial distance on the degree of association. Then, the preliminary association value is calculated by multiplying the functional similarity between the target unit and a certain adjacent unit by the spatial relationship weighting coefficient corresponding to that adjacent unit to obtain the preliminary association value. Then, normalization processing is performed by summing the preliminary association values ​​of all adjacent units of the target unit. The preliminary association value of each adjacent unit is divided by the sum to obtain the attribution association parameter. The value of the normalized parameter is between 0 and 1, which is convenient for unified comparison with the threshold. The larger the parameter value, the closer the comprehensive association between the target unit and the corresponding adjacent unit in terms of functional composition and spatial relationship. Finally, the attribution association parameter is bound and stored with the identification number of the corresponding adjacent unit to form a correspondence table of target unit, adjacent unit, and attribution association parameter to ensure data traceability.

[0046] Step 554: Based on the attribution association parameters, compare them with the preset association strength threshold, and select spatial assessment units whose attribution association parameters are lower than the preset association strength threshold to obtain the set of target units to be adjusted; specifically, this includes: firstly, determining the preset association strength threshold, the process is as follows: select typical sub-regions, covering different ecological types in high-altitude pastoral areas, with each sub-region having an area of ​​not less than 100 km². 2To ensure the representativeness of the sample, the distribution of statistical parameters was analyzed. The attribution correlation parameters of all spatial assessment units within each typical sub-region were statistically analyzed, and the mean, median, and standard deviation were calculated. Considering the ecological vulnerability of high-altitude pastoral areas, it was necessary to strengthen the correlation of land use functions to avoid functional fragmentation. Therefore, an adjustment of 0.05 was added to the average value as a threshold adjustment range, ultimately determining the correlation strength threshold to be 0.4 (0.35 + 0.05). This threshold setting conforms to the data distribution pattern and meets the needs of ecological protection and refined management. Subsequently, a comparison was conducted, comparing the attribution correlation parameters of each spatial assessment unit calculated in step 553 with the preset threshold of 0.4. If the unit's attribution correlation parameter is lower than 0.4, it indicates that the unit... If a unit has a weak overall connection with its neighboring units in terms of function and spatial relationship, and its functional distribution does not conform to the overall spatial connection pattern of the region, it may have problems with functional fragmentation or incoordination with the surrounding environment, requiring optimization and adjustment. If the unit's association parameter is higher than or equal to 0.4, it indicates that it is closely connected with its surroundings and its functional distribution is reasonable, requiring no adjustment. If the parameter is exactly equal to the threshold, it is considered to meet the requirements and is not included in the set to be adjusted to avoid ambiguity. Finally, the set is integrated, and all the selected units with substandard association strength are sorted and integrated by unique identifier number to form a set of target units to be adjusted. The association parameter of each unit and the reason for not meeting the standard are marked to ensure that the set only contains the spatial evaluation units that truly need optimization.

[0047] Step 555: For each target unit in the set of target units to be adjusted, call the corresponding attribution association parameters of the target unit, and select one or more adjacent units with more prominent parameter values ​​to generate an adaptation reference unit for the target unit. Specifically, this includes: for each target unit in the set of target units to be adjusted, using its unique identifier, calling the attribution association parameters of the target unit and all adjacent units, as well as the identifiers and functional characteristics of the adjacent units, from the target unit, adjacent unit, and attribution association parameter correspondence table; sorting all the attribution association parameters of the target unit's adjacent units from largest to smallest; setting selection rules for adaptation reference units, balancing adjustment accuracy and representativeness; if the first parameter after sorting is much higher than the others... For parameters, only the first one is selected as the adaptation reference unit to avoid ambiguity in the adjustment direction due to multiple reference units. If the difference between the first two parameters is small, the first two are selected as adaptation reference units to improve the comprehensiveness of the reference. If the first three parameters are all similar, the first three are selected at most, and no more than three, to avoid confusion in the adjustment logic due to too many reference units. The selection rules are set based on the continuity characteristics of the functional distribution in high-altitude pastoral areas. Two to three reference units can ensure that the adjustment direction conforms to the regional pattern without affecting the adjustment efficiency due to too many units. The selected adjacent units are identified as the adaptation reference units of the target unit. A correspondence table between the target unit and the adaptation reference units is established, and the dominant functional type and associated parameter values ​​of each reference unit are marked.

[0048] Step 556: Based on the adapting reference units of the target unit, extract the dominant land space function type within the adapting reference unit, reclassify or adjust the proportion of the function type composition within the corresponding target unit, and update the dominant function type to obtain the set of target units that have been adjusted. Specifically, this includes: for each target unit to be adjusted, extracting detailed information of the reference unit from the gridded land space function classification dataset using the unique identifier of the adapting reference unit, focusing on clarifying the dominant land space function type of each adapting reference unit, first determining the adjustment benchmark, and if the dominant function type of all adapting reference units is consistent, then directly using that function type as the benchmark. The adjustment benchmark for the target unit is as follows: If multiple reference units have different dominant functional types, the average area ratio of each functional type in these reference units is calculated, and the functional type with the highest average value is taken as the adjustment benchmark. If the number of dominant functional types in the reference units is the same and the average values ​​are close, the functional type that conforms to the planning positioning is selected as the benchmark in conjunction with the main functional area planning of the area where the target unit is located. Subsequently, targeted adjustments are carried out. The adjustment process fully considers the terrain conditions and ecological constraints of the high-altitude pastoral area. In the first case, if the original dominant function of the target unit differs significantly from the adjustment benchmark, functional reclassification is carried out, and the distribution of the original agricultural space within the target unit is located by spatial coordinate positioning. In the first scenario, excluding areas with slopes greater than 25° (especially high-altitude pastoral areas unsuitable for cultivation or grazing) and less than 500m from water sources), the remaining agricultural space adjacent to the reference unit's baseline functional area is selected and reclassified as ecological space. This ensures the adjustment does not damage ecologically sensitive areas and aligns with topographic suitability. In the second scenario, if the target unit's original dominant function matches the adjustment baseline, but the baseline function's proportion is significantly higher than the target unit's, a proportional adjustment is made. The distribution area of ​​secondary functional types within the target unit is selected, excluding agricultural spaces already designated as basic grazing areas. Agricultural spaces accounting for 10% of the target unit's total area are selected and adjusted to ecological space. To ensure that the adjustment does not affect the reasonable needs of agricultural production, after the adjustment is completed, the area and proportion of various functions within the target unit are recalculated. The actual area occupied by each function after the adjustment is divided by the total area of ​​the unit to obtain the new area proportion. It is ensured that the sum of the proportions of the three types of functions is 100% and there are no logical errors. Finally, the dominant function type is updated, and the function type with the highest area proportion after the adjustment is determined as the new dominant function type. The functional adjustment of the target unit is completed. All target units that have completed the functional composition adjustment and dominant function update are integrated by unique identification number, and the changes in the functional proportions before and after the adjustment and the basis for the adjustment are marked to form a set of target units that have completed the adjustment.

[0049] Step 557 involves fusing the adjusted target unit set with the unadjusted units in the original gridded land spatial function classification dataset to complete a global update of the dataset, ultimately generating the updated gridded land spatial function classification dataset. Specifically, this includes: first, extracting unadjusted units from the original gridded land spatial function classification dataset, selecting all units not included in the target unit set to be adjusted. These units have a correlation parameter greater than or equal to 0.4, and their functional distribution conforms to the overall spatial correlation pattern of the region. No adjustments are needed; their original unique identifier, spatial geometric boundary coordinates, dominant function type, distribution location, area, and proportion of various internal functions remain unchanged, serving only as the basis for fusion. Next, an association matching mechanism is established, linking the adjusted target unit set with the original dataset using the unique identifier. This quickly locates the corresponding position of each adjusted unit in the original dataset, replacing the old information of the corresponding unit in the original dataset with the latest information of the adjusted unit. During the replacement process, the spatial geometric boundary coordinates and unique identifier of the unit are strictly maintained; only function-related information is updated to avoid spatial location shifts. Finally, data fusion and verification are carried out, following the guidelines for high-altitude and cold regions. The horizontal row and column order of the entire pastoral area was used to stitch and integrate the adjusted and unadjusted units into a preliminary integrated dataset. Subsequently, multi-dimensional verification was performed to ensure data integrity and logical consistency. Functional type verification ensured that the dominant functional type of each unit must be one of three categories: ecological, agricultural, or prohibited development, with no other invalid types. Percentage verification ensured that the area percentage of each of the three functional categories within each unit was positive and its sum was 100%. Spatial boundary verification ensured that the functional boundaries of adjacent units did not overlap, and the offset error between the unit's spatial coordinates and the original dataset did not exceed 1 meter. Logical verification ensured that no other functions existed within the prohibited development space. His functional type adjustment records show that the dominant function of the units within the ecological protection red line is ecological space or prohibited development space. For the problems found in the verification, the adjustment process was traced and corrected one by one to ensure that there are no logical contradictions in the dataset. Finally, the updated dataset was generated, and the verified fused data was output in GeoJSON format, including fields such as the unit's unique identifier number, center latitude and longitude, boundary coordinate set, dominant function type, ecological space area and proportion, agricultural space area and proportion, prohibited development space area and proportion, and adjustment status. This dataset fully covers the entire high-altitude pastoral area, and the functional distribution is more in line with the regional pattern.

[0050] Taking the most closely related adjacent units as a reference, and combining the topographical conditions, ecological constraints and planning positioning of the high-altitude pastoral areas, the adjustment direction is determined to ensure that the functional type of each unit is consistent with the overall spatial distribution pattern of the region, thereby improving the consistency and coordination of the classification of national land space functions.

[0051] In a preferred embodiment of the present invention, step 6 above, which involves extracting the spatial structure features of grid cells using a pre-trained graph neural network model based on the updated gridded land spatial function classification dataset, and finally generating a land spatial function classification result dataset for high-altitude pastoral areas, may include: In this embodiment of the invention, step 660 involves defining each grid unit as a graph node based on the updated gridded land spatial function classification dataset, and establishing connecting edges between spatially adjacent grid units to generate a graph structure representing the spatial relationships across the entire region. Specifically, this includes: using the updated gridded land spatial function classification dataset as the core foundation, which already contains complete information such as the unique identifier number, center latitude and longitude, boundary coordinates, area and proportion of the three types of functions, dominant function type, and adjustment status of each grid unit, firstly defining the graph nodes, mapping each grid unit in the dataset to a separate graph node, and assigning each node a unique identifier number consistent with the grid unit to ensure a one-to-one correspondence between nodes and original grid units; simultaneously, extracting key information from the dataset as attribute features of the nodes, specifically including the center longitude and center latitude of the grid unit, the proportion of ecological space area, and agricultural... The spatial area ratio, the proportion of prohibited development area, the coding of dominant functional types, and the adjustment of status indicators are used to integrate these attribute information into the initial feature set of each node, ensuring that the node features can comprehensively depict the spatial and functional characteristics of the grid unit. Next, the adjacency relationship determination and connection edge establishment are carried out. First, the adjacency relationship determination criteria are clarified. Combining the spatial distribution characteristics of grid units in high-altitude pastoral areas, a 3×3 grid matrix adjacency rule is adopted. Taking the grid unit corresponding to the target node as the center, the nodes corresponding to the eight grid units of the same size around it are all determined as the adjacent nodes of the target node, including the four nodes that are directly adjacent to each other in the top, bottom, left and right and the four nodes that are indirectly adjacent to each other in the diagonal, ensuring full coverage of adjacent nodes that have a spatial relationship with the target node. Then, an undirected connection edge is established for each pair of adjacent nodes. That is, when node A is adjacent to node B, an edge from A to B is established, and an edge from B to A is also established, reflecting the reciprocity of the adjacency relationship.

[0052] Finally, edge weights are calculated to quantify the degree of connection between adjacent nodes. The calculation method comprehensively considers both functional similarity and spatial distance. For functional similarity calculation, referring to the method of total difference in functional proportion in step 552, the sum of the absolute values ​​of the differences in the area proportions of the three functional categories of two adjacent nodes is calculated. The smaller the total difference value, the higher the functional similarity. The total difference value is converted into a similarity coefficient (similarity coefficient = 1 - total difference value ÷ 100). For example, if the total difference value is 5%, the similarity coefficient is 0.95. For spatial distance weight calculation, the straight-line distance between two adjacent nodes is calculated using the latitude and longitude of the center of the corresponding grid cell. The maximum adjacent distance is set to 1.414 km. Distance weights are then used. =1 - (actual center distance ÷ maximum adjacent distance), for example, the distance weight of directly adjacent nodes is 1 - (1 ÷ 1.414) ≈ 0.3, and the distance weight of diagonally adjacent nodes is 0; the edge weight is finally calculated as edge weight = functional similarity coefficient × distance weight. For example, for a pair of directly adjacent nodes with a functional similarity coefficient of 0.95, the edge weight = 0.95 × 0.3 ≈ 0.285; for a pair of diagonally adjacent nodes with a functional similarity coefficient of 0.9, the edge weight = 0.9 × 0 = 0; all nodes, adjacent node pairs and their corresponding edge weights are integrated to form a complete graph structure. This structure contains a set of nodes and a set of edges, which comprehensively represents the spatial relationships and functional associations of the entire territory of the high-altitude pastoral area.

[0053] Step 661 involves inputting the graph structure into a pre-trained graph neural network model. The graph convolutional layers of the model aggregate the neighboring node information of each node, extracting and obtaining the high-dimensional spatial structure feature vector corresponding to each grid cell. Specifically, this includes: firstly, constructing a pre-trained graph neural network model to ensure its ability to aggregate spatial correlation information and extract high-dimensional features. The construction process is as follows: Pre-training data preparation involves selecting 3-5 regions with geographical characteristics similar to the target high-altitude pastoral area, collecting gridded land spatial function classification datasets from these regions, and converting the datasets of each region into a gridded feature vector according to the method in step 660. The corresponding graph structure forms a pre-training sample set; simultaneously, each node in the pre-training sample set is labeled with a real land spatial function classification label to ensure that the sample set has both diversity and accuracy; the graph neural network model structure is designed, which includes an input layer, three graph convolutional layers, activation layers, and fully connected layers. The input layer receives the initial feature set of graph nodes and converts the initial features into a 128-dimensional initial feature vector, with each feature term corresponding to a specific dimension in the vector, ensuring the structured representation of the features; the core function of each graph convolutional layer is to aggregate neighborhood node information. The first graph convolutional layer receives the 128-dimensional initial feature vector, and then... The influence of neighborhood information is adjusted by edge weights. The specific calculation method is: current node's initial feature vector × its own weight coefficient + each neighboring node's initial feature vector × corresponding edge weight × neighborhood weight coefficient. The own weight coefficient and the neighborhood weight coefficient are adjustable parameters of the graph neural network model. The calculation results of all neighboring nodes are added to the current node's own calculation result to obtain the first layer's aggregated 256-dimensional feature vector. The second layer, a graph convolutional layer, receives the 256-dimensional feature vector and uses the same aggregation logic to sum the 256-dimensional features of the current node and the 256-dimensional features of the neighboring nodes according to the edge weights, resulting in a 512-dimensional feature vector. The third layer... The graph convolutional layer receives a 512-dimensional feature vector and further aggregates feature information from a wider neighborhood (neighborhood of neighborhoods), outputting a 1024-dimensional intermediate feature vector. The activation layer adds activation processing after each graph convolutional layer. Specifically, it converts all values ​​less than 0 in the aggregated feature vector to 0, while keeping the values ​​greater than 0 unchanged. This enhances the non-linear expressive power of the features and avoids linear redundancy of feature information. The fully connected layer receives the 1024-dimensional intermediate feature vector output from the third graph convolutional layer and maps it to a 256-dimensional high-dimensional feature vector through a weight matrix. This vector will serve as the core spatial structure feature vector of the node.

[0054] The pre-training process of a graph neural network (GNN) model involves inputting graph structures from the pre-training sample set into the designed GNN model in batches of 10 samples. For each sample, the GNN model outputs a 256-dimensional feature vector for each node and its corresponding predicted classification label. The sum of the differences between the predicted and true classification labels is calculated, i.e., the loss value; the larger the difference, the higher the loss value. Based on the loss value, the weight coefficients of the GNN model itself, the weight coefficients of its neighborhood, and the weight matrix of the fully connected layers are adjusted to gradually reduce the loss value. This process of inputting samples, predicting, calculating loss, and adjusting parameters is repeated iteratively for 1000 iterations until the loss value stabilizes at a low level. Training is then stopped, resulting in a pre-trained GNN model. After the GNN model pre-training is complete, the model is ready for training. The target high-altitude pastoral area map structure generated in step 660 is completely input into the graph neural network model. First, the initial feature vectors of the nodes in the graph structure are converted into 128-dimensional vectors through the input layer. Then, through the gradual aggregation of three layers of graph convolutional layers, the first layer aggregates the initial features of the direct neighboring nodes, the second layer aggregates the features of the neighboring nodes, and the third layer deepens the spatial correlation features of the entire domain. Each layer precisely adjusts the influence of different neighboring nodes through edge weights, and the activation layer simultaneously optimizes the feature expression. Finally, the 1024-dimensional intermediate feature vector is mapped to a 256-dimensional high-dimensional spatial structure feature vector through a fully connected layer. Each graph node corresponding to each grid cell outputs a unique 256-dimensional feature vector, which fully integrates the functional features of the node itself, the correlation features of the neighboring nodes, and the spatial distribution features of the entire domain.

[0055] Step 662: Based on the high-dimensional spatial structure feature vector corresponding to each grid unit, perform spatial context analysis and correction on the classification of each grid unit in the updated gridded land spatial function classification dataset, and finally generate the land spatial function classification result dataset for the high-altitude pastoral area. Specifically, this includes: firstly, parsing the key information of the 256-dimensional high-dimensional spatial structure feature vector corresponding to each grid unit, extracting three core feature dimensions from the vector: neighborhood function consistency dimension; spatial location adaptability dimension; and functional type association strength dimension. Through the feature representation of these three core dimensions, the spatial context adaptability state of the nodes is comprehensively characterized. Then, spatial context analysis is carried out for each… For each graph node corresponding to a grid cell, high-dimensional spatial structure feature vectors of all its neighboring nodes are extracted. The similarity of the core feature dimensions between the target node and each neighboring node is calculated one by one. For each core dimension, the absolute difference between the corresponding dimension values ​​is calculated. The sum of the absolute differences of all dimensions is then averaged to obtain the similarity of that dimension. The similarities of the three core dimensions are then weighted and summed in a 3:3:4 ratio to obtain the comprehensive similarity between the target node and that neighboring node. The comprehensive similarity between the target node and all neighboring nodes is calculated and averaged to serve as the spatial context fit of the target node. The higher the fit, the more harmonious the current classification of the node is with the surrounding spatial environment.

[0056] Classification correction rules are formulated based on spatial context fit. If the spatial context fit of the target node is higher than the preset fit threshold, and the current dominant function type is consistent with more than 60% of the dominant function types of neighboring nodes, the current classification is deemed reasonable and no correction is needed. If the spatial context fit of the target node is lower than the preset threshold, or the current dominant function type is inconsistent with more than 60% of the dominant function types of neighboring nodes, the current classification is deemed to have a possible bias and needs correction. If the grid cell corresponding to the target node is located within the prohibited development space or the ecological space area accounts for more than 80%, the original dominant function type is retained regardless of the fit, without correction, to ensure the rigid constraint of the ecological protection red line. Targeted correction operations are performed on nodes that need correction, selecting the top 3 nodes with the highest comprehensive similarity among all neighboring nodes of the target node as correction reference nodes. The dominant function types of the reference nodes are extracted, the number of reference nodes for each dominant function is counted, and the dominant function type with the largest number is selected as the correction candidate type. The spatial location adaptability dimension in the high-dimensional feature vector of the target node is compared with the adaptability features of the candidate types. If the adaptability features meet the standard, the grid cell corresponding to the target node is... The dominant function type of the unit is adjusted to the candidate type. If the matching features do not meet the standards, the dominant function type with the second largest number among the reference nodes is selected for re-evaluation until a suitable correction type is determined. After correcting the dominant function type, the area ratio of the three functions of the grid unit is readjusted. Based on the average area ratio of the corresponding correction type in the reference nodes, the area of ​​the three functions is adjusted proportionally according to its own spatial range. After the correction is completed, the dataset is integrated and verified. The information of all grid units is integrated, including the unique identifier number, center latitude and longitude, boundary coordinates, final dominant function type, final area and ratio of the three functions, core dimension value of high-dimensional feature vector, correction status, and spatial context fit. Logical verification is performed to ensure that the dominant function type of each unit is only one of the three categories: ecology, agriculture, and prohibited development, and the total area ratio of the three functions is 100%. There are no agricultural spatial classification records in the prohibited development space. Spatial coherence verification is performed to check the final classification results of adjacent grid units to ensure that there are no large-area isolated classification patches. If there are, a second correction is performed by combining the high-dimensional feature vector. After the verification is passed, the final dataset of land spatial function classification results for high-altitude pastoral areas is generated.

[0057] The graph structure construction fully captures the overall spatial relationships of the country. The pre-trained graph neural network model is trained with multiple samples to improve the depth and reliability of classification. The spatial context analysis and correction process combines high-dimensional feature vectors and neighborhood association features to ensure that the classification results are highly consistent with the overall spatial distribution of the region, reducing functional fragmentation and classification bias.

[0058] like Figure 2As shown, embodiments of the present invention also provide a land spatial function classification data processing system for high-altitude pastoral areas, including: The processing module is used to collect and process multi-source spatial data from high-altitude pastoral areas to build a national land spatial element database. The assessment module is used to assess the grazing suitability of each spatial unit based on the national land spatial element database, and generate a spatial distribution dataset of grazing suitability in high-altitude pastoral areas; The fusion module is used to perform spatial overlay analysis on the spatial distribution dataset of grazing suitability in high-altitude pastoral areas and the prohibited development zone scope in the main functional area planning. The intersecting areas generated by the overlay are geometrically fused to obtain a preliminary dataset of national land spatial functional zoning. The conversion module is used to perform regular gridding processing on the preliminary land spatial functional zoning dataset, converting the original irregular spatial unit functional zoning results into regular grid units of preset size to form a gridded land spatial functional classification dataset. The adaptation module is used to select spatial assessment units based on the gridded land spatial function classification dataset, analyze the spatial distribution characteristics of the spatial assessment units, calculate the attribution association parameters between each spatial assessment unit and its surrounding adjacent units, and use the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset. The classification module is used to extract the spatial structure features of grid cells based on the updated gridded land spatial function classification dataset, and finally generate the land spatial function classification result dataset for high-altitude pastoral areas.

[0059] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0060] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0061] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0062] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing land spatial function classification data in high-altitude pastoral areas, characterized in that, The method includes: Step 1: Collect and process multi-source spatial data from high-altitude pastoral areas to construct a national land spatial element database; Step 2: Based on the national land spatial element database, assess the grazing suitability of each spatial unit and generate a spatial distribution dataset of grazing suitability in high-altitude pastoral areas; Step 3: Spatial overlay analysis is performed on the dataset of the spatial distribution of grazing suitability in high-altitude pastoral areas and the scope of prohibited development zones in the main functional area planning. The intersecting areas generated by the overlay are geometrically fused to obtain a preliminary dataset of national land spatial functional zoning. Step 4: Perform regular gridding processing on the preliminary land space functional zoning dataset, transforming the original irregular spatial unit functional zoning results into regular grid units of preset size to form a gridded land space functional classification dataset. Step 5: Select spatial assessment units based on the gridded land spatial function classification dataset, analyze the spatial distribution characteristics of the spatial assessment units, and calculate the attribution association parameters between each spatial assessment unit and its surrounding adjacent units; use the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset. Step 6: Based on the updated gridded land spatial function classification dataset, a pre-trained graph neural network model is used to extract the spatial structure features of the grid units, and finally generate the land spatial function classification result dataset for the high-altitude pastoral area.

2. The method for processing spatial functional classification data in high-altitude pastoral areas according to claim 1, characterized in that, Based on the national land spatial element database, the suitability for grazing in each spatial unit is assessed, generating a spatial distribution dataset of grazing suitability in high-altitude pastoral areas, including: From the national land space element database, select multiple spatial attribute data that affect the suitability of animal husbandry, and perform normalization processing to form a standardized set of spatial attribute data; Assign importance to each data item in the standardized spatial attribute dataset to form an importance allocation scheme; Based on the importance allocation scheme, the spatial attribute data of each spatial unit are comprehensively calculated to obtain the comprehensive evaluation value of the suitability for animal husbandry of each spatial unit; Based on the pre-defined rules for classifying the suitability of animal husbandry, the comprehensive evaluation value of animal husbandry suitability is classified into levels to determine the animal husbandry suitability level of each spatial unit. The pastoral suitability level of each spatial unit is matched and integrated with the geographical location information of the spatial unit to generate a spatial distribution dataset of pastoral suitability in high-altitude pastoral areas.

3. The method for processing spatial functional classification data in high-altitude pastoral areas according to claim 2, characterized in that, Spatial overlay analysis was performed on the spatial distribution dataset of grazing suitability in high-altitude pastoral areas and the prohibited development zones in the main functional zoning plan. Geometric fusion was then performed on the intersecting areas generated by the overlay to obtain a preliminary dataset of national land spatial functional zoning, including: Based on the spatial distribution dataset of grazing suitability in high-altitude pastoral areas and the data on prohibited development zones in the main functional area planning scheme, an input data layer to be analyzed is formed. Spatial overlay calculations are performed on two datasets in the input data layer to generate overlay analysis results describing the spatial relationship between the two datasets. Based on the overlay analysis results, all areas that spatially overlap with the prohibited development zone in the spatial distribution data of grazing suitability of high-altitude pastoral areas are extracted, and a set of intersecting areas is generated. Perform boundary fusion processing on all independent geometric shapes within the set of intersecting regions to generate one or more seamless merged polygonal features; By integrating the merged polygonal features with the non-overlapping areas in the spatial distribution data of grazing suitability in high-altitude pastoral areas, a preliminary national land spatial functional zoning dataset is generated, which includes the identification of prohibited development areas.

4. The method for processing spatial functional classification data in high-altitude pastoral areas according to claim 3, characterized in that, The preliminary land spatial functional zoning dataset is processed into regular grids, transforming the original irregular spatial units into regular grid units of a preset size, thus forming a gridded land spatial functional classification dataset, including: Based on the preset regular grid cell size and the entire area of ​​the high-altitude pastoral region, a spatial grid array covering the entire study area is generated, consisting of multiple regular grids of identical size and shape. Spatial location relationship calculations were performed between the spatial grid array and the preliminary national spatial functional zoning dataset to obtain spatial association results describing the coverage relationship between each regular grid and different functional zoning units; Based on the spatial correlation results, the area of ​​various types of land space functional zoning units covered within each regular grid is counted, and a statistical table recording the area ratio of various functional zoning units within each grid is generated. Based on the area percentage data recorded in the statistical table and following the preset functional type determination rules, a dominant national spatial functional type is determined and assigned to each rule grid, while retaining the internal functional type composition information, forming a set of grid units with dominant functional type and internal functional type composition information. By integrating the spatial geometric information, dominant functional type, and internal functional type composition information of the grid unit set, a gridded national spatial function classification dataset with unified geometric form is generated.

5. The method for processing spatial functional classification data in high-altitude pastoral areas according to claim 4, characterized in that, Based on a gridded national spatial function classification dataset, spatial assessment units were selected. The spatial distribution characteristics of these units were analyzed, and the attribution association parameters between each spatial assessment unit and its neighboring units were calculated, including: Based on the gridded land spatial function classification dataset, regular grids are selected as spatial assessment units to obtain a set of spatial assessment units; The spatial assessment unit set is used to process the gridded land space function classification dataset. By statistically analyzing the area proportion of the functional type composition information within each unit, the internal composition and spatial arrangement pattern of the unit functions are analyzed to obtain the spatial distribution characteristics of each spatial assessment unit. Based on the spatial distribution characteristics, for each target spatial evaluation unit, all adjacent units within the preset spatial range are identified, and the similarity in functional type composition between the target unit and each adjacent unit is calculated. Based on similarity, and combining the spatial adjacency relationship or spatial distance between the target unit and its neighboring units, the affiliation and association parameters between the target unit and each neighboring unit are calculated by integrating both functional and spatial factors.

6. The method for processing spatial functional classification data in high-altitude pastoral areas according to claim 5, characterized in that, The spatial distribution characteristics of the gridded land spatial function classification dataset are adapted and adjusted using attribution association parameters to generate an updated gridded land spatial function classification dataset, including: Based on the attribution association parameter, and compared with the preset association strength threshold, spatial evaluation units with attribution association parameters lower than the preset association strength threshold are selected to obtain the set of target units to be adjusted. For each target unit in the set of target units to be adjusted, call the corresponding attribution association parameter of the target unit, select one or more adjacent units with more prominent parameter values, and generate the adaptation reference unit of the target unit. Based on the adaptation reference unit of the target unit, the dominant land space function type within the adaptation reference unit is extracted, the function type composition within the corresponding target unit is reclassified or the proportion is adjusted, and the dominant function type is updated to obtain the set of target units that have been adjusted. The adjusted target unit set is merged with the unadjusted units in the original gridded land spatial function classification dataset to complete the global update of the dataset, and finally generate the updated gridded land spatial function classification dataset.

7. The method for processing land spatial function classification data in high-altitude pastoral areas according to claim 6, characterized in that, Based on the updated gridded land spatial function classification dataset, a pre-trained graph neural network model is used to extract the spatial structure features of grid cells, ultimately generating a dataset of land spatial function classification results for high-altitude pastoral areas, including: Based on the updated gridded land space function classification dataset, each grid cell is defined as a graph node, and connecting edges are established between spatially adjacent grid cells to generate a graph structure representing the spatial relationships of the entire domain. The graph structure is input into a pre-trained graph neural network model. The graph convolutional layer of the graph neural network model aggregates the information of the neighboring nodes of each node, and extracts and obtains the high-dimensional spatial structure feature vector corresponding to each grid unit. Based on the high-dimensional spatial structure feature vector corresponding to each grid unit, spatial context analysis and correction are performed on the classification of each grid unit in the updated gridded land spatial function classification dataset, and finally, a land spatial function classification result dataset for alpine pastoral areas is generated.

8. A data processing system for the classification of territorial spatial functions in high-altitude pastoral areas, wherein the system implements the method as described in any one of claims 1 to 7, characterized in that, include: The processing module is used to collect and process multi-source spatial data from high-altitude pastoral areas to build a national land spatial element database. The assessment module is used to assess the grazing suitability of each spatial unit based on the national land spatial element database, and generate a spatial distribution dataset of grazing suitability in high-altitude pastoral areas; The fusion module is used to perform spatial overlay analysis on the spatial distribution dataset of grazing suitability in high-altitude pastoral areas and the prohibited development zone scope in the main functional area planning. The intersecting areas generated by the overlay are geometrically fused to obtain a preliminary dataset of national land spatial functional zoning. The conversion module is used to perform regular gridding processing on the preliminary land spatial functional zoning dataset, converting the original irregular spatial unit functional zoning results into regular grid units of preset size to form a gridded land spatial functional classification dataset. The adaptation module is used to select spatial assessment units based on the gridded land spatial function classification dataset, analyze the spatial distribution characteristics of the spatial assessment units, calculate the attribution association parameters between each spatial assessment unit and its surrounding adjacent units, and use the attribution association parameters to adapt and adjust the spatial distribution characteristics of the gridded land spatial function classification dataset to generate an updated gridded land spatial function classification dataset. The classification module is used to extract the spatial structure features of grid cells based on the updated gridded land spatial function classification dataset, and finally generate a dataset of land spatial function classification results for high-altitude pastoral areas.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images