Land resource visualization calculation method based on land utilization pattern spots

Abstract

The invention relates to the technical field of land visualization, in particular to a land resource visualization calculation method based on land utilization pattern spots. The method comprises the following steps: acquiring land resource utilization data and a land remote sensing image; performing land utilization type classification on the land resource utilization data to generate land utilization type data; performing land attribute feature extraction on the land resource utilization data according to the land utilization type data to obtain land attribute feature data; and performing land map spot visualization on the land remote sensing image through the land utilization type data and the land attribute feature data to generate a land utilization resource spatial distribution map. According to the method, through multi-level data processing, attribute extraction, land utilization change prediction and decision visualization, the accuracy and comprehensiveness of land resource visualization calculation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of land visualization, and in particular, to a land resource visualization calculation method based on land use patches. Background Art

[0002] In the initial stage, the Land Use Patch, as the basic unit for land resource analysis, was mainly represented by a simple two-dimensional map, which could not fully reflect the spatial distribution and spatio-temporal changes of land resources. With the progress of remote sensing technology, especially the popularization of satellite images and high-resolution geographical data, the accurate extraction of land use patches has become possible. Image processing and spatial analysis algorithms have developed significantly, and patch analysis methods have been gradually introduced into the visualization calculation of land resources, enabling a more refined description of land use types, change dynamics, and spatial structures. During this period, the visualization of land resources was not limited to simple map display, but also combined with various advanced technologies such as data mining and model calculation to achieve quantitative assessment and trend prediction of land use changes. Entering the 21st century, with the rise of big data technology, artificial intelligence, and deep learning, the calculation methods for land resource visualization have gradually developed towards the direction of intelligence, real-time, and multi-dimensional analysis. However, at present, the traditional processing of land resource spatial distribution maps is mostly simple regional division, lacking comprehensive analysis of land resources at different levels and attributes. At the same time, traditional land use change predictions are mostly limited to static data analysis, lacking dynamic prediction of future changes, thus resulting in relatively low accuracy and comprehensiveness in land resource visualization calculation. Summary of the Invention

[0003] Based on this, it is necessary to provide a land resource visualization calculation method based on land use patches to solve at least one of the above technical problems.

[0004] To achieve the above object, a land resource visualization calculation method based on land use patches includes the following steps:

[0005] Step S1: Obtain land resource utilization data and land remote sensing images; classify the land resource utilization data by land use type to generate land use type data; extract land attribute feature data from the land resource utilization data according to the land use type data; perform land patch visualization on the land remote sensing images through the land use type data and the land attribute feature data to generate a land use resource spatial distribution map;

[0006] Step S2: Divide the land use resource spatial distribution map into layer plots through multi-attribute plots of land use resources to generate single-layer plots and multi-layer plots of land use resources; calculate the surface land indicators for the single-layer plots of land use resources to obtain the summary data of the main surface resources; generate the summary data of the hidden-layer potential resources of the land by calculating the hidden-layer land indicators for the multi-layer plots of land use resources.

[0007] Step S3: Predict the land use change for the summary data of the main surface resources and the summary data of the hidden-layer potential resources of the land to generate the land use change prediction data; reconstruct the distribution map of the land use resource spatial distribution map through the land use change prediction data to generate the predicted map of the land use distribution change.

[0008] Step S4: Analyze the plot changes of the land use resource spatial distribution map based on the predicted map of the land use distribution change to generate the land use change plot data; visualize the plot change decision for the land use change plot data to generate the plot decision report.

[0009] The present invention provides rich basic data for subsequent analysis by obtaining land resource utilization data and land remote sensing images, ensuring the comprehensiveness and accuracy of land resource management. Classifying land resources by type and extracting land attribute features can accurately reflect the spatial distribution and properties of different land types, improving the management accuracy of land resources. Through the visualization processing of remote sensing images, the spatial distribution of land resources can be intuitively presented, providing a clear spatial view for further analysis. By means of the multi-attribute plot division method, the analysis of land resources is no longer limited to a single level, and data at different levels and dimensions can be comprehensively considered, improving the accuracy and comprehensiveness of land resource analysis. By calculating surface and hidden layer land indicators, the explicit resources and potential resources of land can be analyzed respectively, so as to more comprehensively evaluate the utilization value and potential of land, avoiding the situation of ignoring deep-layer resources in traditional methods. The generated resource summary data provides a reliable data basis for further prediction and decision-making, supporting refined land resource management. By predicting land use changes, the future trends of land use changes can be insighted, providing forward-looking decision-making basis for land planning and resource allocation. Reconstructing the land use distribution map based on the prediction data makes the prediction results more operable in space, and can visually display the predicted changes in the geographical space, helping decision-makers to carry out specific spatial planning and resource optimization. By analyzing the land use change plots, the change areas that need attention can be accurately identified, helping decision-makers to understand the specific trends of land use changes, and thus making targeted countermeasures. Through the visualization of plot change decisions, an easy-to-understand decision report is generated, helping decision-makers quickly understand the analysis results and make scientific and reasonable land management and planning decisions, improving decision-making efficiency and accuracy. Therefore, the present invention improves the accuracy and comprehensiveness of land resource visualization calculation through multi-level data processing, attribute extraction, land use change prediction and decision visualization.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Obtain land resource utilization data and land remote sensing images;

[0012] Step S12: Perform data preprocessing on the land resource utilization data to generate standard land resource utilization data, where the data preprocessing includes data cleaning, data denoising, filling of missing data values, and data standardization;

[0013] Step S13: Classify the standard land resource utilization data by land use type to generate land use type data;

[0014] Step S14: Extract land attribute features from the land resource utilization data according to the land use type data to obtain land attribute feature data; visualize land patches on the land remote sensing image through the land use type data and the land attribute feature data, and generate a spatial distribution map of land use resources.

[0015] Through the complete process from data collection to visual display, the present invention ensures the comprehensive evaluation of land resources. By integrating land resource utilization data and remote sensing images, not only can the accuracy of the data be improved, but also the spatial and temporal factors of land use can be fully considered. Data preprocessing (including data cleaning, denoising, missing value filling, and standardization) can effectively improve the quality of the data, making subsequent analysis and decision-making based on a more reliable foundation. This not only enhances the stability of the data but also reduces errors caused by data noise. By classifying the standardized land resource utilization data, different land use types can be accurately identified. This provides a clear basis for land management decisions and can help relevant departments implement precise policy measures on different land types. The extraction of land attribute features provides more detailed information to help understand the specific characteristics of different land use types. For example, extracting attribute features such as soil type, vegetation cover, and terrain helps accurately evaluate the usability and environmental impact of the land. Through the visualization of land patches on the land remote sensing image, combined with land use type and attribute feature data, the spatial distribution of land resources can be clearly displayed. The generation of the spatial distribution map makes complex land resource data intuitive and easy to understand, helping policymakers and managers make scientific decisions quickly.

[0016] Preferably, step S14 includes the following steps:

[0017] Step S141: Extract land surface type features from the land resource utilization data according to the land use type data to obtain land surface type feature data; analyze the land distribution density of the land resource utilization data through the land surface type feature data to generate land distribution density data;

[0018] Step S142: Analyze the land environmental features of the land resource utilization data based on the land distribution density data to generate land environmental feature data; integrate the land surface type feature data and the land environmental feature data to generate land attribute feature data;

[0019] Step S143: Identify the regional boundaries of the land remote sensing image to generate land area boundary identification data; divide the land remote sensing image through the land area boundary identification data to generate land area division data; segment the land remote sensing image according to the land area division data to generate a land parcel segmentation image;

[0020] Step S144: Import the land attribute feature data into the land plot segmentation image for pixel multi-dimensional color coding to generate a land use resource spatial distribution map.

[0021] Through extracting the land surface type features and combining with the land distribution density analysis, the present invention can accurately describe the spatial distribution characteristics of land resources. This analysis can identify different density regions of land use, providing accurate data support for land planning and environmental management. For example, land density data can reveal the over-developed or undeveloped state of certain regions, providing a basis for optimizing land use. By analyzing the land distribution density data, the environmental characteristics of the land are further extracted, providing strong support for the ecological assessment, environmental monitoring and risk assessment of the land. The generation of land environmental characteristic data can help understand the ecological conditions of different land blocks, and then guide the balance of environmental protection and land development. Regional boundary recognition and land plot segmentation can effectively distinguish different land areas and land units. This process not only improves the accuracy of land division, but also provides a clear spatial scope for subsequent land use analysis. The generation of land plot segmentation images enables land managers to clearly identify the specific locations and distributions of each plot, providing a spatial basis for decision-making. Integrate the land surface type feature data with the land environmental characteristic data to form comprehensive land attribute feature data. This multi-dimensional data fusion provides a more comprehensive situation of land resources, helping to conduct more refined land assessment and comprehensive decision-making. By combining the land attribute feature data with the land plot segmentation image and performing pixel multi-dimensional color coding, a land use resource spatial distribution map is generated. This spatial distribution map not only intuitively presents the spatial layout of land resources, but also clearly displays information such as different land types and environmental characteristics through color coding, facilitating quick understanding and decision-making. Each link jointly improves the accuracy of land resource management. Through detailed surface feature analysis, density and environmental feature extraction, regional division and segmentation, the finally formed spatial distribution map provides a scientific data basis for land planning, environmental protection, urban development, etc. Land use decisions can be more accurate and targeted, reducing resource waste and avoiding environmental degradation. Through the comprehensive extraction and visualization of these data, dynamic monitoring of different land types, regional boundaries and environmental characteristics can be achieved. This is of great significance for decision-making in multiple aspects such as land development, urban construction, and agricultural planning, and can ensure the sustainable use and reasonable development of land resources.

[0022] Preferably, importing the land attribute feature data into the land plot segmentation image for pixel multi-dimensional color coding includes:

[0023] Import the land attribute feature data into the land plot segmentation image for data filling to generate the plot filling attribute data; perform land attribute RGB color coding on the plot filling attribute data to generate the land attribute RGB color coding data; perform land use HSB adjustment on the land attribute RGB color coding data to generate the land use HSB color adjustment data;

[0024] Perform data visualization on the land plot segmentation image through the land attribute RGB color coding data and the land use HSB color adjustment data to generate the initial land use resource spatial distribution map; perform multi-attribute plot screening on the initial land use resource spatial distribution map to obtain the land use resource multi-attribute plots;

[0025] Perform plot attribute layer superposition on the land resource multi-attribute plots to generate the land use resource multi-attribute plot layer; perform attribute value calculation on the land use resource multi-attribute plot layer to obtain the land resource attribute value data; perform attribute proportion sorting on the land resource multi-attribute plots based on the land resource attribute value data to generate the attribute proportion sorting data;

[0026] Control the layer transparency of the land use resource multi-attribute plot layer through the attribute proportion sorting data to optimize the land use resource multi-attribute plots; optimize the layer of the initial land use resource spatial distribution map according to the land use resource multi-attribute optimized plots to generate the land use resource spatial distribution map.

[0027] By importing land attribute feature data into the land plot segmentation image for data filling, the present invention can effectively enhance the attribute expression of land plots. The generated plot filling attribute data provides accurate basic data for subsequent color coding and layer processing, enabling different land types and attributes to be clearly reflected in the image. This provides a more intuitive and accurate land resource distribution map for land management personnel. RGB color coding and HSB color adjustment are performed on land attributes, enabling effective differentiation of different land attributes (such as soil type, vegetation coverage, land use type, etc.) through color changes. This multi-dimensional color coding not only improves the visualization effect of land use resources but also makes the comparison of different land attributes more obvious, thereby providing clear visual information for management personnel to facilitate quick judgment of the status of land resources. Through multi-attribute plot screening and plot attribute layer superposition, key areas with different attributes can be clearly displayed in the land use resource spatial distribution map, helping to identify plots that require priority attention. For example, it can accurately identify hotspots of land use resources or plots with greater development potential, contributing to the optimization of land resource utilization and allocation. Sorting the attribute proportion based on land resource attribute value data can help identify the proportion and importance of each attribute in different land resource plots. This provides a more quantitative and systematic basis for land planning and decision-making. For example, if a certain type of land resource attribute has an overly high proportion in a certain area, adjustments or corresponding management measures need to be taken to optimize the land use structure in that area. By controlling the layer transparency, different attributes of land resources and their optimization effects can be more intuitively displayed. For example, transparency adjustment can make the changes in land attributes more prominent while avoiding the image from being too complex, making the hierarchy between different land features clearer and helping decision-makers understand and make decisions. Layer optimization can also clearly display the spatial distribution of optimized land resources, effectively supporting sustainable land planning. Through this series of steps, the current situation of land resources can be comprehensively analyzed, potential problems identified, and effectively optimized. This can not only help relevant departments make more accurate decisions in land planning, urban construction, and agricultural development but also achieve a balance between environmental protection and resource conservation in the use of land resources, providing a scientific basis for sustainable development. By generating an initial land use resource spatial distribution map and gradually optimizing it (such as layer optimization, transparency control, etc.), a clearer and more hierarchical land resource spatial distribution map can be obtained. This optimization process improves the efficiency of land resource management, making resource allocation and development decisions more accurate, and providing strong support for optimizing land use and environmental management.

[0028] Preferably, step S2 includes the following steps:

[0029] Step S21: Divide the land use resource spatial distribution map by multi-attribute plots of land use resources to generate single-layer plots of land use resources and multi-layer plots of land use resources;

[0030] Step S22: Calculate the land use resource fixed indicators for the single-layer plots of land use resources; calculate the surface land use resource indicators for the multi-layer plots of land use resources to obtain the surface land use resource indicator data;

[0031] Step S23: Summarize the data of the land use resource spatial distribution map according to the surface land use resource indicator data and the land use resource fixed indicators to generate the surface land main resource summary data;

[0032] Step S24: Calculate the hidden land use resource indicators for the multi-layer plots of land use resources to obtain the hidden land use resource indicator data; summarize the data of the land use resource spatial distribution map with the hidden land use resource indicator data and the land use resource fixed indicators to generate the hidden land potential resource summary data.

[0033] Through the layer plot division of the land use resource spatial distribution map, the land resources can be divided into single-layer plots and multi-layer plots. This division can provide more refined spatial units for subsequent analysis, helping to accurately identify different levels and characteristics of land resources. The division of single-layer plots and multi-layer plots can better support the multi-dimensional analysis of complex land resources, assisting managers in making differentiated decisions for different land types. By calculating the land indicators of single-layer plots and multi-layer plots, fixed indicator and surface indicator data can be obtained. Fixed indicators (such as land area, soil type, etc.) and surface indicators (such as vegetation coverage, surface soil moisture, etc.) provide a quantitative standard for the status of land resources, which provides basic data for subsequent analysis and ensures the accuracy and comprehensiveness of the data. By summarizing the surface indicator data and fixed indicator data, the summary data of the main surface resources of the land is generated. This summary can reflect the overall status of the land surface resources, providing a macro perspective for the utilization, development, and protection of land resources. Through these data, managers can better understand the resource distribution on the land surface, and then optimize the land use plan and improve the resource utilization efficiency. By calculating the hidden layer land indicators of multi-layer plots, hidden layer indicator data is obtained. Hidden layer indicators usually involve deeper land resource characteristics (such as groundwater level, soil structure, underground mineral resources, etc.), and these data provide important support for land potential assessment. The summary data of hidden layer potential resources can reveal the potential value of land resources, helping decision-makers identify areas with development potential, thus promoting sustainable resource development and utilization. Through the various calculations and data summaries in step S2, the comprehensive evaluation data of land resources can be obtained. Through the multi-level analysis of surface and hidden layer indicators, the advantages and disadvantages of land resources can be more accurately identified, promoting precise resource development, environmental protection, and land planning. For example, areas rich in underground resources can be determined based on the summary data of hidden layer potential resources, while ecological protection or agricultural development can be carried out based on the summary data of surface resources.

[0034] Preferably, step S21 includes the following steps:

[0035] Step S211: rasterize the land use resource spatial distribution map to generate a land use grid, where the land use grid includes a number of raster cells; screen the raster cells through the multi-attribute plots of land use resources to obtain multi-attribute raster cells;

[0036] Step S212: classify the land use grid through the multi-attribute raster cells to generate single-attribute raster grids and multi-attribute raster grids; classify and mark the land use resource spatial distribution map according to the single-attribute raster grids and multi-attribute raster grids to generate the first type of land use resource plots and the second type of land use resource plots;

[0037] Step S213: Conduct a hierarchical analysis of the first type of land use resource plots and the second type of land use resource plots to generate hierarchical data for the land use resource plots; based on the hierarchical data for the land use resource plots, perform layer plot division on the first type of land use resource plots and the second type of land use resource plots to generate single-layer land use resource plots and multi-layer land use resource plots.

[0038] In the present invention, through rasterization processing, the spatial distribution map of land use resources is transformed into raster cells, which makes the spatial distribution of land resource data more detailed. The rasterized grid cells can better handle the land resource characteristics at different scales, enabling land management to be analyzed at a more precise spatial level. Through multi-attribute raster cell screening, more dimensions of land attributes can be identified within each raster cell, thus providing more decision-making bases for subsequent classification and marking. This helps to identify and evaluate the diversity of land use and the spatial distribution of different attributes. The grid type classification in step S212 enables the raster grids of land resources to be distinguished according to single attributes and multi-attributes. Through the division of single-attribute raster grids and multi-attribute raster grids, areas of different land types can be distinguished. Single-attribute grids are suitable for basic land use assessment, while multi-attribute grids are suitable for detailed analysis of complex areas, such as areas considering factors such as soil type and vegetation cover. According to these classification results, the spatial distribution map of land use resources can be marked as plots of different categories, such as the first type of plots and the second type of plots, providing basic data for subsequent land planning and development. The plot layer hierarchical analysis in step S213 helps to further refine the hierarchical structure of land resources. By analyzing the hierarchical relationship of land resource plots, the spatial structure characteristics of land resources can be revealed. For example, some lands are composite plots composed of multiple attribute levels, while some plots belong to a single attribute category. This analysis helps to distinguish the development priorities, utilization potential, etc. of different plots during land development and planning. Based on the hierarchical data, further layer plot division can help clearly define different types or functional areas of land use, thereby optimizing the allocation and use of land resources. These processing procedures in step S21 make the management of land use resources more refined and more operable. Through rasterization and multi-attribute screening, not only can land resources be carefully divided spatially, but also land can be accurately classified at the attribute level.

[0039] Preferably, calculating the surface land indicators for the multi-layer land use resource plots includes:

[0040] Calculate the proportion of each layer's attributes for the multi-layer land use resource plot to obtain the proportion data of each layer's attributes; confirm the layer with the largest pixel attribute for the multi-layer land use resource plot according to the proportion data of each layer's attributes to obtain the layer with the largest attribute proportion;

[0041] Calculate the chromaticity of the plot pixels for the multi-layer land use resource plot based on the layer with the largest attribute proportion, thereby generating the surface layer index data of the land use resources.

[0042] Through the calculation of the proportion of attributes for each layer in the present invention, the proportion of each layer in the multi-layer plot can be quantified. This calculation of the proportion of attributes provides a quantitative basis for subsequent analysis and can reveal the weights of different land use types, soil categories, vegetation cover, etc. in the spatial distribution. This information is crucial for accurately identifying the spatial characteristics and potential of land resources. Confirming the layer with the largest attribute proportion according to the proportion data of each layer's attributes, this process can help determine the dominant land attribute type in the multi-layer plot. By identifying which layer dominates in the multi-layer plot, the core characteristics of land resources can be better understood, and a more focused and accurate basis can be provided for subsequent land planning, development, and protection decisions. The calculation of the chromaticity of plot pixels helps to convert the layer attributes into digital indicators through color coding, thereby providing a visual and quantitative analysis of the surface layer characteristics of land use resources. This not only enhances the visualization effect of the image but also makes the differences between different land resource attributes more obvious, further providing a basis for the analysis and optimization of land resources. Through this chromaticity mapping, different regions can be clearly classified, and plots with specific resource characteristics can be accurately identified. Through the generated surface layer index data of land use resources, a comprehensive evaluation of the surface layer characteristics of land resources can be carried out.

[0043] Preferably, the calculation of the hidden layer land indicators for the multi-layer land use resource plot includes:

[0044] Calculate the proportion of each layer's attributes for the multi-layer land use resource plot to obtain the proportion data of each layer's attributes; confirm the layer with non-maximum pixel attributes for the multi-layer land use resource plot according to the proportion data of each layer's attributes to obtain the layer with non-maximum attribute proportion;

[0045] Calculate the chromaticity of the plot pixels for the multi-layer land use resource plot based on the layer with non-maximum attribute proportion, thereby generating a set of hidden plot pixel chromaticities; calculate the chromaticity similarity for the set of hidden plot pixel chromaticities to generate chromaticity similarity data; conduct attribute correlation analysis on the set of hidden plot pixel chromaticities according to the chromaticity similarity data to generate hidden attribute correlation data;

[0046] Using the implicit attribute to associate data, perform the associated pixel chromaticity mean processing on the implicit plot pixel chromaticity set to generate the associated pixel chromaticity mean; perform data aggregation on the associated pixel chromaticity mean, thereby generating the hidden layer index data of land use resources.

[0047] Through the confirmation of the non-maximum attribute proportion layer, the present invention can identify those land features that are not significant in a single layer but are potentially important in the overall multi-layer analysis. This process helps to discover some imperceptible land attributes, such as potential ecological resources, soil characteristics, etc., and further provides data support for the in-depth development and optimization of land resources. The calculation of plot pixel chromaticity and the generation of the implicit plot pixel chromaticity set enable the spatial distribution of land resources to extend beyond surface features and penetrate deeper into more hidden land feature levels. Through chromaticity mapping, the implicit features of land resources can be presented in a more intuitive and accurate way, making the multi-dimensional features of the spatial distribution more prominent and facilitating subsequent analysis and decision-making. By calculating the chromaticity similarity of the implicit plot pixel chromaticity set, plots with similar features in space can be identified. This analysis not only helps to identify areas with similar attributes but also discovers the potential correlations of land resources, such as the soil type and climate conditions in adjacent areas having similar impacts on land use. The calculation of chromaticity similarity further improves the clustering analysis ability of land resources and helps to find potential optimization areas among diverse land resources. Through the attribute association analysis of the implicit plot pixel chromaticity set, the different features of land resources can be analyzed for their correlations, and the potential connections between existing land resources can be discovered. For example, the correlations among attributes such as soil type, vegetation coverage, and land use pattern reveal new land use optimization opportunities. The attribute association analysis provides a more refined theoretical basis and decision-making support for subsequent land development, protection, and optimization. The associated pixel chromaticity mean processing and data aggregation will further improve the efficiency and accuracy of data processing. By aggregating the associated pixel chromaticity means, the processing process of large-scale land data can be simplified, and on this basis, the hidden layer index data of land use resources can be generated. This aggregation method not only reduces data noise but also improves the usability of data, enabling the hidden layer features to be clearly reflected in the overall evaluation of land resources. The generation of the hidden layer index data helps to reveal the potential value and development potential of land resources. The implicit indicators often involve features such as the ecological function, soil health, and sustainable use of land, and can provide in-depth analysis support for the long-term planning of land. These indicators will help to discover the optimal development path among various land use patterns and support decision-makers in finding a balance between development and protection.

[0048] Preferably, step S3 includes the following steps:

[0049] Step S31: Conduct a temporal variation analysis on the aggregated data of the main surface resources of the land and the aggregated data of the latent potential resources of the land to generate land use change data;

[0050] Step S32: Divide the land use change data into data sets to generate a model training set and a model test set; Use the long short-term memory neural network algorithm to train the model training set to generate a preliminary land use change prediction model; Optimize and iterate the land use change prediction pre-model through the model test set to generate a land use change prediction model;

[0051] Step S33: Import the land use change data into the land use change prediction model to predict the land use change and generate land use change prediction data;

[0052] Step S34: Reconstruct the distribution map of the land use resource spatial distribution map through the land use change prediction data to generate a predicted map of land use distribution changes.

[0053] Through the temporal variation analysis of the aggregated data of the main surface resources of the land and the aggregated data of the latent potential resources of the land, the present invention can reveal the temporal evolution law of land use and resource changes. This analysis provides historical trend data for subsequent land use prediction and planning decisions, helping to identify the regularity of land resource changes. For example, it can identify characteristics such as the seasonality and periodicity of land use changes, and thus provide a basis for long-term planning. Model training and optimization (based on the long short-term memory (LSTM) neural network algorithm) enable the prediction of land use changes to not only consider the current state but also utilize historical data for dynamic prediction on a long time scale. The LSTM network is particularly good at processing temporal data and can provide highly accurate predictions for land use changes by capturing the time dependence in the data. This means that future land use changes can be accurately predicted through this model, providing a reliable reference for land managers. Through the generated land use change prediction model (Step S32), the future land use trend can be accurately predicted. This prediction ability can provide data support for policymakers and land planners, helping them make more scientific and accurate decisions in aspects such as land use, ecological protection, and urban development. By analyzing the future land use change trend, potential problems and opportunities can be identified in advance, reducing decision-making risks. Data set division and model optimization (using the model test set) ensure the generalization ability and accuracy of the land use change prediction model. By using the model training set and the test set separately, the overfitting problem can be effectively avoided, thereby enhancing the stability and reliability of the prediction results. In addition, the optimization and iteration process ensure that the model can continuously adjust and adapt to new data changes, improving the adaptability and prediction accuracy of the model. The prediction of land use changes (based on the trained model) can dynamically predict the future changes in land use in real time.

[0054] Preferably, step S4 includes the following steps:

[0055] Step S41: Perform plot change analysis on the land use resource spatial distribution map according to the land use distribution change prediction map to generate land use change plot data; construct plot change decisions for the land use change plot data to generate plot change decision data;

[0056] Step S42: Visualize plot change decisions for the land use distribution change prediction map according to the plot change decision data to generate a plot decision report.

[0057] Through plot change analysis, the present invention can conduct in-depth plot change assessment on the spatial distribution of land resources based on the land use distribution change prediction map. This analysis not only reveals the change trends of land use in different plots, but also can effectively identify problems such as resource flow, change in utilization mode, or environmental degradation in which plots. Through this analysis, managers can more accurately grasp the changes in various land resources, providing a strong basis for the next decision-making. Plot change decision construction can generate plot change decision data based on land use change data. This process not only relies on the trend analysis of historical data, but also combines the output of prediction models to construct decision models, thus providing intelligent decision-making solutions for different types of land use changes. Through this decision construction, decision-makers can adopt more reasonable and effective strategies in land resource management and land use change, avoiding human decision-making biases and improving the scientificity and accuracy of decision-making. The generation of plot change decision data helps decision-makers quickly respond to the challenges and problems brought about by land use changes. For example, some plots are at risk of ecological degradation due to overdevelopment, or some regions need to change the land use pattern due to policy adjustments. Through plot change decisions, managers can make early warnings and response measures in advance, providing guarantees for the sustainable management of land resources. The plot change decision data is presented to decision-makers through graphical means. Description of the Drawings

[0058] Figure 1 It is a schematic flow chart of the steps of the land resource visualization calculation method based on land use patches;

[0059] Figure 2 It is Figure 1 a schematic detailed implementation step flow chart of step S2 in

[0060] Figure 3 It is Figure 1 a schematic detailed implementation step flow chart of step S3 in

[0061] The realization, functional features and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings. Detailed implementation manners

[0062] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0063] In addition, the accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor means and / or microcontroller means.

[0064] It should be understood that although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly, the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0065] To achieve the above object, please refer to Figures 1 to 3 , a land resource visualization calculation method based on land use patches, the method comprising the following steps:

[0066] Step S1: Obtain land resource utilization data and land remote sensing images; classify the land use types of the land resource utilization data to generate land use type data; extract land attribute feature data from the land resource utilization data according to the land use type data; perform land patch visualization on the land remote sensing images through the land use type data and the land attribute feature data to generate a land use resource spatial distribution map;

[0067] Step S2: Divide the land use resource spatial distribution map into layer plots through multi-attribute plots of land use resources to generate single-layer plots and multi-layer plots of land use resources; calculate the surface land indicators for the single-layer plots of land use resources to obtain the summary data of the main surface resources; generate the summary data of the potential resources in the hidden layer of the land by calculating the hidden layer land indicators for the multi-layer plots of land use resources.

[0068] Step S3: Predict the land use change for the summary data of the main surface resources and the summary data of the potential resources in the hidden layer of the land to generate the land use change prediction data; reconstruct the distribution map of the land use resource spatial distribution map through the land use change prediction data to generate the predicted map of the land use distribution change.

[0069] Step S4: Analyze the plot changes of the land use resource spatial distribution map based on the predicted map of the land use distribution change to generate the land use change plot data; visualize the plot change decision for the land use change plot data to generate the plot decision report.

[0070] The present invention provides rich basic data for subsequent analysis by obtaining land resource utilization data and land remote sensing images, ensuring the comprehensiveness and accuracy of land resource management. Classifying land resources by type and extracting land attribute features can accurately reflect the spatial distribution and properties of different land types, improving the management accuracy of land resources. Through the visualization processing of remote sensing images, the spatial distribution of land resources can be intuitively presented, providing a clear spatial view for further analysis. Through the multi-attribute plot division method, the analysis of land resources is no longer limited to a single level, and data at different levels and dimensions can be comprehensively considered, improving the accuracy and comprehensiveness of land resource analysis. By calculating the surface and hidden layer land indicators, the explicit resources and potential resources of the land can be analyzed respectively, so as to more comprehensively evaluate the utilization value and potential of the land, avoiding the situation of ignoring deep resources in traditional methods. The generated resource summary data provides a reliable data basis for further prediction and decision-making, supporting refined land resource management. By predicting land use changes, the future trends of land use changes can be insighted, providing a forward-looking decision-making basis for land planning and resource allocation. Reconstructing the land use distribution map based on the prediction data makes the prediction results more operable in space, and can visually display the predicted changes in the geographical space, helping decision-makers to carry out specific spatial planning and resource optimization. By analyzing the land use change plots, the change areas that need attention can be accurately identified, helping decision-makers to understand the specific trends of land use changes, and thus making targeted countermeasures. Through the visualization of plot change decisions, an easy-to-understand decision report is generated, helping decision-makers quickly understand the analysis results and make scientific and reasonable land management and planning decisions, improving decision-making efficiency and accuracy. Therefore, the present invention improves the accuracy and comprehensiveness of land resource visualization calculation through multi-level data processing, attribute extraction, land use change prediction, and decision visualization.

[0071] In an embodiment of the present invention, with reference to Figure 1 as shown, it is a schematic diagram of the step flow of a land resource visualization calculation method based on land use patches of the present invention. In this example, a land resource visualization calculation method based on land use patches includes the following steps:

[0072] Step S1: Obtain land resource utilization data and land remote sensing images; classify the land resource utilization data by land use type to generate land use type data; extract land attribute features from the land resource utilization data according to the land use type data to obtain land attribute feature data; perform land patch visualization on the land remote sensing image through the land use type data and the land attribute feature data to generate a land use resource spatial distribution map;

[0073] In the embodiments of the present invention, statistical data on land resource utilization is obtained through a Geographic Information System (GIS) or a relevant data platform, including but not limited to information such as land use type, area, utilization status, etc. Land resource utilization data: includes land use type, plot area, crop information, building use, land development degree, etc. Images taken by remote sensing satellites or drones are used to obtain surface image data of the corresponding area. Data sources: images provided by remote sensing satellites (such as Landsat, Sentinel, etc.), or high-resolution images obtained in real time through drones. Data formats: common formats of remote sensing images include GeoTIFF, JPEG, PNG, etc. Remote sensing image: from the Sentinel-2 satellite, covering an area of 100 square kilometers with a resolution of 10 meters. The land resource utilization data is cleaned and standardized to ensure the accuracy and consistency of the data, removing missing values, outliers, duplicate data, etc., and unifying the unit and time format. Different regions of land are classified according to the land resource utilization data to determine the main land use types (such as agricultural land, urban land, forest land, etc.), and the K-means clustering algorithm is used for unsupervised classification. Classification results: Region A: agricultural land, Region B: urban land, Region C: water area. According to the land use type data, the attribute features of each plot in the land resource data are extracted. Attribute features include: land surface cover type (such as forest land, grassland, water surface, etc.), land quality (such as soil fertility, humidity, etc.), land development status (such as whether it is farmland, construction land, etc.). Attribute features extracted for agricultural land: soil type (such as sandy soil), irrigation situation (such as having an irrigation system), farming type (such as crop rotation system). Attribute features extracted for urban land: building density, greening coverage rate, land price, etc. An image segmentation algorithm (such as K-means or a deep learning algorithm) is used to segment the image to identify different types of land use areas. Techniques such as contrast enhancement and color adjustment are used to improve the visualization effect of the image. A GIS software (such as ArcGIS, QGIS, etc.) is used for map drawing, or a web-based visualization tool (such as Leaflet, Google Earth Engine, etc.). Different regions of land use types are displayed in different colors or icons, and the attribute feature information is displayed on the map as additional information. Through different color markings, agricultural land is green, urban land is gray, and water area is blue. Clicking on each area on the map pops up its detailed attribute information, such as soil type, irrigation status, etc.

[0074] Step S2: Divide the spatial distribution map of land use resources by multi-attribute plots of land use resources to generate single-layer plots of land use resources and multi-layer plots of land use resources; calculate the surface land indicators for the single-layer plots of land use resources to obtain the summary data of the main surface resources of the land; generate the summary data of the potential hidden land resources by calculating the hidden land indicators for the multi-layer plots of land use resources.

[0075] In the embodiments of the present invention, multi-attribute information of each plot is extracted from the land use type data, land attribute characteristic data, and remote sensing image data obtained in step S1. These attributes include, but are not limited to: land use type (agricultural land, urban land, forest land, etc.), land cover type (such as farmland, grassland, buildings, etc.), land quality indicators (such as soil quality, humidity, vegetation cover, etc.). For those areas with a single attribute (such as a single crop planting area, a single land use type), independent division is carried out. For those areas with multiple attribute overlays (such as urban green belts, mixed areas of farmland and forest, etc.), multi-layer division is carried out, and each layer represents an attribute characteristic. In the land use space, area A is a single agricultural land, so it is divided into a single-layer plot; area B is an urban area, including buildings, green spaces, and roads, so it is divided into multiple layer plots, representing the building area, greening area, and traffic area respectively. Surface land indicators refer to the land resources and environmental characteristics that have a direct impact on the surface layer (such as the soil surface, vegetation layer, etc.). Common surface indicators include: soil type (such as sandy soil, loam, clay, etc.), vegetation coverage, water resource situation (such as irrigation conditions, water source distribution, etc.), land use rate (such as cultivation ratio, urbanization degree, etc.). Using soil sample data or remote sensing data, calculate the soil type and soil quality index. Evaluate the vegetation coverage of each plot through the vegetation index (such as NDVI) in the remote sensing image. Evaluate the water resource situation of the plot according to the water source distribution map and irrigation data. Analyze the utilization conditions such as cultivation, construction, and greening according to the land use data. For the single-layer agricultural plot A, calculate that its soil type is loam, the vegetation coverage is 80%, and the irrigation conditions are good. For the single-layer urban plot B, calculate that its land use rate is 50% (including commercial land, residential areas, and green spaces).

[0076] Step S3: Predict the land use change for the summary data of the main surface resources of the land and the summary data of the potential hidden land resources to generate land use change prediction data; reconstruct the spatial distribution map of land use resources through the land use change prediction data to generate a predicted map of land use distribution change.

[0077] In the embodiments of the present invention, appropriate prediction models are selected, such as Support Vector Machine (SVM), Random Forest (RF), Long Short-Term Memory (LSTM) neural network, etc. Input variables: summary data of main resources on the land surface, summary data of potential resources in the land hidden layer, historical land use data, external factor data, etc. Output variable: future land use change, usually the prediction result of the change in the spatial distribution of various land use types (such as farmland, urban areas, forests, etc.). The model is trained using historical data sets, and the prediction accuracy of the model is evaluated by methods such as cross-validation and regression analysis. During the training process, appropriate evaluation metrics are selected, such as Mean Squared Error (MSE), Accuracy, F1 score, etc., to evaluate the performance of the model in different scenarios. Through the land surface resource data and the hidden layer potential resource data, combined with historical land use changes, a land use change prediction model based on random forest is trained. This model can predict the change trend of agricultural land to urban land in a certain area within the next five years. Using the land use change prediction data, the existing spatial distribution map of land use resources is reconstructed. The reconstruction process can be carried out through the following methods: According to the existing plot data and the predicted changes, a spatial interpolation method (such as Kriging interpolation) is used to predict the land use type of the changed area. According to the driving factors of land use change prediction (such as land surface resources, hidden layer resources, etc.), new land use types are assigned to different plots. The land use spatial distribution map and the prediction data are converted into a raster form, and the land use type is updated according to the predicted value of each unit in the raster. By combining the land use change prediction data with the existing land use spatial distribution map, a land use distribution change prediction map is generated. This map can show different types of land use changes (such as newly added urban land, reduced agricultural land, expanded forest land, etc.). If the prediction data involves multiple time periods, the map can display the temporal evolution process of land use change, for example, showing each stage of land use change from 2020 to 2025.

[0078] Step S4: Perform plot change analysis on the land use resource spatial distribution map according to the land use distribution change prediction map to generate land use change plot data; perform visualization of plot change decisions on the land use change plot data to generate a plot decision report.

[0079] In the embodiments of the present invention, based on the land use distribution change prediction map and the existing land use resource spatial distribution map, the image differential analysis method (such as raster calculation, pixel-level change detection, etc.) is used to detect changes in plots, and the areas where the land use types have changed are identified. For each changed plot, calculate the land types before and after the change and the size of the changed area to obtain the change situation of each plot. Extract the plots detected with changes to generate land use change plot data. This data contains detailed information of each changed plot, including: Change type: For example, farmland converted to urban land, forest converted to farmland, etc. Changed area: The changed area of each plot, and the unit can be square kilometers, mu, etc. Time period when the change occurred: The stage when the change occurred within the prediction period. Driving factor analysis: Analyze the main driving factors affecting the change, such as climate change, policy adjustment, market demand, etc. The generated land use change plot data includes information such as the spatial location, change type, and changed area of the changed area, and usually consists of fields such as plot number, land type, change type, area, and time period. Use the GIS platform to display the plots with land use changes, highlight the changed plots in the form of layers, and support user interactive operations, such as clicking to view detailed information. Convert the information such as the area and change type of the changed plots into a heat map to help users identify the areas with the most obvious changes. Display the dynamic process of land use changes, play the history of land use changes through the time axis, and help decision-makers intuitively understand the change process of each plot. On the existing land use resource spatial distribution map, highlight the plots that have changed with different colors or symbols. Different colors represent different change types. For example, red indicates the conversion of agriculture to urban land, and green indicates the conversion of urban areas to nature reserves. Generate a plot decision report according to the land use change plot data.

[0080] Preferably, step S1 includes the following steps:

[0081] Step S11: Obtain land resource utilization data and land remote sensing images;

[0082] Step S12: Perform data preprocessing on the land resource utilization data to generate standard land resource utilization data, where the data preprocessing includes data cleaning, data denoising, filling of missing data values, and data standardization;

[0083] Step S13: Classify the land use types of the standard land resource utilization data to generate land use type data;

[0084] Step S14: Extract the land attribute characteristics of the land resource utilization data according to the land use type data to obtain land attribute characteristic data; perform land patch visualization on the land remote sensing image through the land use type data and the land attribute characteristic data to generate a land use resource spatial distribution map.

[0085] In the embodiments of the present invention, land resource utilization data is obtained through the land management department or a remote sensing data provider (such as a satellite remote sensing platform). These data usually include the land use status (such as arable land, forest land, urban construction land, etc.). Land remote sensing images come from remote sensing devices of satellites or drones, providing image data in different bands and covering the land cover conditions at different time nodes. Common remote sensing data includes Landsat images, Sentinel-2 images, etc. Remote sensing images are usually stored in the GeoTIFF format and contain geographical location information. The land resource utilization data can be in the format of a Geographic Information System (GIS) such as Shapefile, GeoJSON, or CSV. Remote sensing image data is downloaded using public platforms or APIs (such as Google Earth Engine, USGS Earth Explorer). Incomplete or incorrect records are removed, such as data with missing values or illogical data (such as coordinate points outside the geographical range). For remote sensing images, denoising algorithms (such as median filtering, mean filtering, or Gaussian filtering) are used to reduce the influence of factors such as the atmosphere and clouds. Outlier detection methods (such as the Z-Score method) are used for the land resource utilization data to identify and remove noisy data. Appropriate filling methods (such as mean, median imputation, or using machine learning methods for filling) are adopted to fill in the missing values to ensure data integrity. Numerical data (such as land area, population density, etc.) is standardized or normalized to ensure that data with different dimensions can be compared on the same scale. Spectral value normalization is performed on the remote sensing images to improve the comparability between data from different sensors or different times. Based on the standardized land resource data, supervised learning methods (such as Support Vector Machine SVM, decision tree, random forest) or unsupervised learning methods (such as K-means clustering) are used for land use type classification. Different types of land use (such as agricultural land, industrial land, residential areas, etc.) can be automatically identified by analyzing the spatial distribution of land use data and the changes in land cover. Ground survey data or high-resolution images are used to verify the classification results and evaluate the classification accuracy. Classification accuracy metrics are calculated, such as the Kappa coefficient, overall accuracy, and class accuracy. According to the land use type data (such as arable land, forest land, urban land, etc.), land attribute features related to it are extracted, such as land area, land use intensity, economic value of land resources, etc. For each plot in the remote sensing image, its spectral features, texture features, and shape features are extracted to further refine the physical properties of the land. Feature selection and dimensionality reduction (such as Principal Component Analysis PCA, LDA) are performed to ensure that the features most relevant to land use and remote sensing image classification are selected. Clustering or association analysis is used to discover the potential relationship between land use types and their geospatial distribution characteristics. Combining the land use type data and the land attribute feature data, a GIS platform (such as ArcGIS, QGIS) is used for visualization to generate a spatial distribution map of land resources.Intuitively display the distribution and attribute differences of different types of land by using color gradients, heat maps, contour maps, etc.

[0086] Preferably, step S14 includes the following steps:

[0087] Step S141: Extract land surface type features from land resource utilization data according to land use type data to obtain land surface type feature data; perform land distribution density analysis on land resource utilization data through the land surface type feature data to generate land distribution density data;

[0088] Step S142: Perform land environment feature analysis on land resource utilization data based on the land distribution density data to generate land environment feature data; integrate the land surface type feature data and the land environment feature data to generate land attribute feature data;

[0089] Step S143: Identify the regional boundaries of the land remote sensing image to generate land area boundary identification data; divide the land remote sensing image through the land area boundary identification data to generate land area division data; segment the land remote sensing image according to the land area division data to generate a land parcel segmentation image;

[0090] Step S144: Import the land attribute feature data into the land parcel segmentation image for pixel multi-dimensional color coding to generate a land use resource spatial distribution map.

[0091] In the embodiments of the present invention, different land surface types are distinguished by using the spectral characteristics (such as infrared and visible light bands) of remote sensing images. A convolutional neural network (CNN) can be applied for deep learning to extract features. Texture features (such as gray-level co-occurrence matrix, local binary pattern, etc.) are used to further distinguish the surface features of the land. Based on the extracted surface type features, spatial statistical analysis is carried out. For example, calculate the spatial distribution density of different land types, that is, the distribution of different land use types within each plot or region. Spatial autocorrelation analysis (such as Moran's index) or heat map method can be used to calculate the density of land distribution and generate land distribution density data. According to the land distribution density data, analyze the environmental characteristics around the land. For example: slope, altitude, terrain undulation, etc. By combining meteorological data, analyze the land use differences in different climate regions, and analyze the ecological function areas of the land, such as protected areas, wetlands, etc. Spatial analysis tools (such as buffer analysis and overlay analysis in ArcGIS) can be used to evaluate the environmental characteristics of land resources. Integrate the land surface type feature data and land environmental feature data to form comprehensive land attribute feature data. Through methods such as principal component analysis (PCA) or weighted average method, standardize and synthesize different environmental features to obtain a unified land attribute feature dataset. Use edge detection algorithms (such as Canny edge detection, Sobel operator) or deep learning methods (such as boundary extraction based on convolutional neural network) to identify the boundaries of land areas in remote sensing images. Post-process the boundary recognition results to remove noise and refine the boundaries to ensure the accuracy of land areas. Based on the regional boundary recognition data, divide the remote sensing image into multiple land areas. The watershed algorithm or region growing algorithm can be used to achieve precise division of the regions. The divided regional data can be used for subsequent land parcel segmentation and detailed analysis. On the basis of land area division, conduct parcel segmentation. Use image segmentation techniques (such as image thresholding, K-means clustering, superpixel segmentation method) to further divide each land area into independent land parcels. During the segmentation process, consider factors such as the shape, size, and surrounding environment of the land parcels to ensure the rationality of parcel segmentation. Import the land attribute feature data (such as land type, environmental characteristics, etc.) into the segmented land parcel images, and assign a multi-dimensional color coding to each pixel or each parcel. The multi-dimensional color coding can use different color gradients to represent according to factors such as land use type, environmental characteristics, and parcel density. For example: agricultural land: green, construction land: red, water body: blue, ecological protection area: yellow. Use RGB or HSB color space for color coding to ensure good distinguishability of land attribute feature information in terms of color. Generate the final spatial distribution map of land use resources from the processed pixel data. Each region or each parcel in the map will be identified with a specific color according to its attribute characteristics (such as land type, environmental characteristics).The spatial distribution map of land use resources is displayed through a GIS platform or a custom-developed visualization tool (such as Matplotlib, Plotly, ArcGIS, etc.), and the generated spatial distribution map can be used for decision-making support in fields such as spatial planning, land management, and environmental monitoring.

[0092] Preferably, importing the land attribute feature data into the land parcel segmentation image for pixel multi-dimensional color coding includes:

[0093] Importing the land attribute feature data into the land parcel segmentation image for data filling to generate land parcel filled attribute data; performing land attribute RGB color coding on the land parcel filled attribute data to generate land attribute RGB color coding data; performing land use HSB adjustment on the land attribute RGB color coding data to generate land use HSB color adjustment data;

[0094] Visualizing the land parcel segmentation image through the land attribute RGB color coding data and the land use HSB color adjustment data to generate an initial spatial distribution map of land use resources; performing multi-attribute parcel screening on the initial spatial distribution map of land use resources to obtain multi-attribute parcels of land use resources;

[0095] Performing layer superposition of parcel attributes on the multi-attribute parcels of land resources to generate a multi-attribute parcel layer of land use resources; calculating attribute value data for the multi-attribute parcel layer of land use resources to obtain land resource attribute value data; performing attribute proportion ranking on the multi-attribute parcels of land resources based on the land resource attribute value data to generate attribute proportion ranking data;

[0096] Controlling the layer transparency of the multi-attribute parcel layer of land use resources through the attribute proportion ranking data, thereby optimizing the multi-attribute parcels of land use resources; optimizing the layer of the initial spatial distribution map of land use resources according to the multi-attribute optimized parcels of land use resources to generate a spatial distribution map of land use resources.

[0097] In the embodiments of the present invention, relevant information is extracted from the attribute characteristic data of land resources (including land type, environmental characteristics, land density, etc.). According to the spatial position of the land parcel segmentation image, the attribute characteristic data of each parcel is matched. Each parcel is filled to ensure that each area in the image has accurate land attribute data. The data filling can adopt interpolation methods (such as bilinear interpolation) or directly extract the corresponding values from the land attribute data. After filling, the attribute data of each parcel is generated. These data include the land type of the parcel (such as arable land, forest land, urban land, etc.), environmental characteristics (such as vegetation coverage, climate type), land use intensity and other information. In order to visualize the land attribute data, the land attribute data needs to be converted into RGB color coding. For example: arable land: green (representing agricultural land), construction land: red (representing urban construction land), forest land: dark green (representing forest land), water body: blue. Different attribute values (such as land use type, land cover type, etc.) are mapped to different colors in the RGB color space. The attribute value of each parcel will determine its corresponding color. Through a color mapping algorithm (such as based on linear interpolation or custom color scale), RGB color coding data is generated for each parcel according to different land types. In this way, each parcel will be displayed as a specific color in the image. The HSB (hue, saturation, brightness) model is used to further adjust the color performance of the image. By adjusting the hue (H), saturation (S) and brightness (B) values, the colors of the image can be made more vivid and have higher distinguishability. Change the main color tone of the image (for example, change the green of arable land to a brighter green, or adjust the blue of the water body). Control the vividness of the colors to ensure that the land types in the image are clearly visible. Increase or decrease the brightness of the image so that different areas can be clearly displayed under different lighting conditions. The final color values of each land parcel are generated through the adjusted HSB data to ensure that the image is more visually in line with the actual land use characteristics. The RGB color coding data and the color data adjusted by HSB are combined and applied to the land parcel segmentation image. Each parcel is rendered according to its attributes and adjusted colors. GIS or image processing software (such as ArcGIS, QGIS, Matplotlib, etc.) is used for visualization to generate an initial spatial distribution map of land use resources. On the spatial distribution map of land use resources, parcels with specific attributes are selected according to different screening criteria (such as land use type, environmental characteristics, land density, etc.). For example, select all construction land, agricultural land, ecological protection areas, etc. The parcels that meet the conditions can be selected through the attribute query tool of GIS or a custom algorithm. The multi-attribute parcel data after screening is overlaid on the map layer. For example, attributes such as different land types and environmental characteristics can be placed in different layers respectively. Using the layer overlay technology, different land use attributes are displayed on one map to help decision-makers understand the comprehensive characteristics of land resources.By overlaying layers, a multi-attribute layer of land use resources containing various attributes is generated. Each layer contains plot information with different attributes, facilitating subsequent analysis and visualization. According to the different attributes of land resources, specific attribute values of each plot are calculated. For example, calculate the areas, ecological values, resource utilization intensities, etc. of different land use types. Spatial analysis tools (such as buffer analysis, spatial statistics, etc.) can be used to calculate these attribute values. Attribute value data of land resources are generated based on the calculation results. These data include information such as the attribute scores, land utilization rates, land values, etc. of each plot. According to the attribute value data of land resources, the attribute ratios of each land plot are calculated. For example, the ratios of attributes such as land use type, ecological type, development intensity, etc. The land plots are sorted according to the attribute ratios to find the most representative plots, or priority divisions are made according to the high or low attribute ratios. The sorted data of attribute ratios are output for subsequent optimization analysis and decision-making. According to the sorted data of attribute ratios, the transparencies of different plot layers are adjusted. For plots with higher attribute ratios, higher transparencies are set; for plots with lower ratios, lower transparencies are set to highlight important areas. This transparency control can be achieved through GIS software or custom-developed visualization tools. By adjusting the layer transparencies, the visualization effect of the land use resource layer is optimized to ensure that important plots can be prominently displayed on the map. The optimized land use resource layers are synthesized to generate the final spatial distribution map of land use resources. Through layer optimization, the spatial distribution and attribute characteristics of land use are made clearly visible. The final spatial distribution map of land use resources will be presented based on the optimized multi-attribute layer to help decision-makers better conduct land planning, resource allocation, and environmental monitoring.

[0098] As an example of the present invention, refer to Figure 2 shown, in this example, step S2 includes:

[0099] Step S21: Divide the spatial distribution map of land use resources into layer plots through multi-attribute plots of land use resources to generate single-layer plots of land use resources and multi-layer plots of land use resources;

[0100] Step S22: Calculate land indicators for the single-layer plots of land use resources to obtain fixed land use resource indicators; calculate surface land indicators for the multi-layer plots of land use resources to obtain surface land use resource indicator data;

[0101] Step S23: Summarize the data of the spatial distribution map of land use resources according to the surface land use resource indicator data and the fixed land use resource indicators to generate summary data of the main surface resources;

[0102] Step S24: Obtain hidden layer indicator data of land use resources by calculating hidden layer land indicators for multi-layer plots of land use resources; summarize the hidden layer indicator data of land use resources and fixed indicators of land use resources on the spatial distribution map of land use resources to generate summary data of hidden layer potential resources of land.

[0103] In the embodiments of the present invention, for land areas with similar attributes (such as the same land use type or similar environmental characteristics), they are divided into single-layer plots. Each single-layer plot contains a certain specific land attribute and is suitable for simple analysis. For land areas with complex attribute combinations (such as involving multiple land use types, ecological functions, land cover types, etc. at the same time), they are divided into multi-layer plots. Each multi-layer plot contains multiple attribute information and is suitable for comprehensive analysis. For example, all cultivated land, forest land, water bodies, etc. respectively form different single-layer plots. For example, if a certain plot simultaneously contains the functions of urban land and agricultural land, it needs to be divided into multiple layer plots to reflect the overlapping areas of different land functions. For land with a single attribute (such as pure cultivated land or pure forest land), a series of fixed land indicators can be calculated, such as: land area: the area of each single-layer plot, land utilization rate: the development or utilization intensity of the land within the plot, greening coverage rate: such as the greening degree of forest land or the vegetation coverage of cultivated land, soil quality indicators: such as soil pH value, soil organic matter content, etc. For land plots containing multiple attributes, surface layer indicators are calculated to reflect the resource characteristics of the upper layer of the plot. For example: surface land use intensity: combining the area ratios of different use types within the plot to calculate the comprehensive utilization intensity of the plot. Land environmental quality: evaluating the environmental quality of the land by analyzing factors such as air quality, vegetation coverage, water source conditions, etc. within the plot. Ecological function assessment: evaluating the ecological protection function of the plot through data such as vegetation coverage, land slope, humidity, etc., and generating a table containing surface layer indicator data such as land area, utilization intensity, environmental quality, etc. for further analysis. According to the surface layer indicator data and fixed indicator data of each single-layer plot and multi-layer plot, summary statistics are carried out. The contents of the summary can include: total area: the total area of plots with different land use types and attributes, land use intensity: the average value or weighted value of land use intensity in different regions, environmental quality score: the environmental quality score calculated based on surface layer indicators, ecological function value: the ecological function score calculated based on factors such as land use type, soil quality, vegetation coverage, etc., generating a data table summarizing all the main surface layer resources of the land (such as land use intensity, environmental quality, ecological function, etc.), and further providing a basis for land planning and resource management. Hidden layer indicators usually refer to those indicators that are not easily directly observable but have an important impact on land resources. For example: groundwater level: estimating the burial depth of groundwater through a hydrological model or remote sensing technology. Soil potential: evaluating the potential of the soil through data such as the water holding capacity and nutrient content of the soil. Geological conditions: evaluating the geological stability of the plot, the potential risk of landslides or collapses, etc. Land development potential: evaluating the future development potential of the plot based on the geographical location, environmental conditions, etc. of the plot. Through model analysis or data processing, hidden layer indicators of the land are calculated, such as groundwater level changes, soil potential, etc., providing a basis for the long-term planning and resource protection of the land.Summarize the hidden layer index data and fixed index data to form the comprehensive resource data of the land. For example: Total land potential resources: Based on the comprehensive analysis of the hidden layer index and the surface layer index, evaluate the resource potential of each plot. Potential distribution map: Generate a map of the hidden layer potential resources of the land to show the distribution of potential resources in each region. Finally, generate the summary data of the hidden layer potential resources of the land including the hidden layer index and the fixed index. These data not only include the current land use situation, but also reflect the future development and resource potential of the land, providing support for the long-term management and planning of the land.

[0104] Preferably, step S21 includes the following steps:

[0105] Step S211: rasterize the spatial distribution map of land use resources to generate a land use grid, where the land use grid includes a number of raster cells; perform multi-attribute raster cell screening on the raster cells through the multi-attribute land use resource plots to obtain multi-attribute raster cells;

[0106] Step S212: classify the land use grid through the multi-attribute raster cells to generate single-attribute raster grids and multi-attribute raster grids; classify and mark the regions of the spatial distribution map of land use resources according to the single-attribute raster grids and multi-attribute raster grids to generate the first type of land use resource plots and the second type of land use resource plots;

[0107] Step S213: perform plot layer level analysis on the first type of land use resource plots and the second type of land use resource plots to generate plot layer level data; divide the first type of land use resource plots and the second type of land use resource plots into layer plots based on the plot layer level data to generate single-layer land use resource plots and multi-layer land use resource plots.

[0108] In the embodiments of the present invention, by rasterizing the spatial distribution map of land use resources, land use grids are generated. The grid consists of multiple raster cells, and the size and resolution of the raster cells should be set according to application requirements. Rasterization is usually carried out in the form of a regular two-dimensional grid, and a land use type or resource attribute is assigned to each grid cell. According to various attributes of land use resources (such as land use, soil type, vegetation coverage, etc.), the raster cells are screened to generate multi-attribute raster cells. Through multi-attribute analysis (such as weighting, similarity measurement, etc.), each raster cell is weighted and screened to ensure that the attribute information it contains can effectively reflect the actual situation of land use. Based on the screened multi-attribute raster cells, grid type classification is carried out, and the land use grids are divided into single-attribute raster grids and multi-attribute raster grids. Single-attribute raster grids refer to areas with relatively single and easily classifiable attributes, such as pure agricultural land or urban construction land; multi-attribute raster grids contain multiple intertwined attributes, indicating that the land use types in these areas are complex or diverse. According to the classification results of single-attribute and multi-attribute raster grids, regional classification marks are made on the spatial distribution map of land use resources. The classified results divide the spatial distribution map of land use resources into different plot types, generating the first type of plots of land use resources (such as pure agricultural land, industrial land, etc.) and the second type of plots (such as mixed land use areas, development potential areas, etc.). The goal of this process is to provide clear regional boundaries and marks for subsequent land planning and resource management. Layer-level analysis is carried out on the first type of plots and the second type of plots. Layer-level analysis includes hierarchically dividing the plots according to factors such as plot type, depth of land use, and degree of land development, generating plot layer-level data. The core of this step is to organize the plots at different levels through the spatial relationship and attribute characteristics of the plots, so as to provide support for subsequent planning, management, and analysis. Based on the generated plot layer-level data, layer plot division is carried out on the first type of plots and the second type of plots of land use resources. This division will divide the plots into single-layer plots and multi-layer plots according to levels. The specific division methods are as follows: Single-layer plots refer to a certain type of land use resources, and their attributes and utilization characteristics are consistent, and are suitable for single-level planning and management plots. Multi-layer plots refer to lands with multiple uses or attributes. These plots involve multiple levels in planning and use and require complex utilization planning. Use common GIS software (such as ArcGIS, QGIS) to rasterize the land use map, select an appropriate raster size (usually based on project scale and accuracy requirements), and ensure that the raster cells contain sufficient land information. Use multi-attribute decision analysis methods in GIS, such as the Analytic Hierarchy Process (AHP), fuzzy logic model, or weighted average method, combined with expert opinions or historical data to weight and screen the raster cells. Statistical methods such as the Analytic Hierarchy Process (AHP) and cluster analysis method can be applied to hierarchically divide the plots.By comprehensively considering factors such as the land use type, soil characteristics, and ecological functions of the plot, different types of plots are classified into different layer levels.

[0109] Preferably, the calculation of the surface land indicators for the multi-layer plot of land use resources includes:

[0110] Calculating the proportion of each layer attribute of the multi-layer plot of land use resources to obtain the proportion data of each layer attribute; confirming the layer with the largest pixel attribute of the multi-layer plot of land use resources according to the proportion data of each layer attribute to obtain the layer with the largest proportion of attributes;

[0111] Calculating the chromaticity of the plot pixels of the multi-layer plot of land use resources according to the layer with the largest proportion of attributes, thereby generating the surface index data of the land use resources.

[0112] In the embodiments of the present invention, by analyzing multi-layered plots of land use resources, the proportion data of the attributes of each layer is obtained by calculating the spatial proportion of each layer in the plot. The specific calculation method is as follows: The multi-layered plot of land use resources is divided into a number of grid cells (with an appropriate resolution set according to requirements). For each grid cell, the corresponding layer attribute information is extracted. These attributes include land use type, soil type, vegetation cover, etc. The proportion of the attributes of different layers in each grid cell is counted and summarized to obtain the proportion data of each layer in the entire multi-layered plot. For example, for a plot, if the vegetation coverage rate of a certain layer accounts for 30% of the total area, the proportion of this layer is 30%. Through the above calculations, the proportion data of the attributes of each layer is obtained. These data can be used for subsequent layer analysis and confirmation of the layer with the maximum attributes. According to the proportion data of the attributes of each layer, the layer with the maximum attribute proportion is identified. The specific method is as follows: The proportion of the attributes of each layer is sorted to determine the layer with the highest proportion. If a certain layer has the largest proportion in the plot, then this layer is the layer with the maximum attribute proportion. In practical applications, a threshold is set. If the proportion of a certain layer exceeds the threshold, then this layer is the dominant layer and further analysis is carried out. By comparing the proportion data of each layer, the layer with the maximum attribute proportion of the plot is finally determined. This layer represents the main land use type or resource characteristics of the plot and has important guiding significance for subsequent land use decision-making and planning. Based on the attribute data of the layer with the maximum attribute proportion, pixel chromaticity calculation is performed on the multi-layered plot of land use resources. The purpose of this process is to generate a chromaticity index reflecting the state of land use resources through image processing technology and further extract the environmental and utilization characteristics of the surface land. The specific calculation method is as follows: The pixel values of the grid cells are extracted from the layer with the maximum attribute proportion. The pixel value of each grid cell represents the land use attribute corresponding to this area in space. According to the pixel values and proportion data of the maximum attribute layer, a chromaticity mapping algorithm (such as RGB, HSV models, etc.) is used to map different attribute values to specific colors. For example, agricultural land is mapped to green, urban construction land is mapped to gray or red, and water bodies are mapped to blue. The chromaticity distribution of the layer is calculated, and through statistical quantities such as the average and standard deviation of chromaticity values, surface chromaticity data of a plot is generated. These chromaticity data can reflect the spatial distribution and utilization of land resources. Finally, surface index data of land use resources is generated using the chromaticity calculation results. These data can reflect surface characteristics such as the main land use type, soil properties, and vegetation cover degree of the plot. Specific surface indicators include: determined by the layer with the maximum attribute proportion, indicating the distribution of different land use types in the entire plot. Based on the results of chromaticity calculation, environmental characteristic indices are generated, such as ecological greenness index, water body coverage rate, etc. By combining the surface land index data, the land use potential of this area is evaluated, such as agricultural production suitability, urban construction suitability, etc.

[0113] Preferably, the calculation of the hidden land index for the multi-layer land parcels of land use resources includes:

[0114] Calculating the proportion of each layer attribute of the multi-layer land parcels of land use resources to obtain the proportion data of each layer attribute; confirming the non-maximum layer of pixel attributes for the multi-layer land parcels of land use resources according to the proportion data of each layer attribute to obtain the non-maximum attribute proportion layer;

[0115] Calculating the chromaticity of the plot pixels of the multi-layer land parcels of land use resources according to the non-maximum attribute proportion layer to generate a set of hidden plot pixel chromaticities; calculating the chromaticity similarity of the set of hidden plot pixel chromaticities to generate chromaticity similarity data; performing attribute correlation analysis on the set of hidden plot pixel chromaticities according to the chromaticity similarity data to generate hidden attribute correlation data;

[0116] Processing the mean chromaticity of associated pixels of the set of hidden plot pixel chromaticities by using the hidden attribute correlation data to generate the mean chromaticity of associated pixels; aggregating the mean chromaticity of associated pixels to generate the hidden layer index data of land use resources.

[0117] In the embodiments of the present invention, by calculating the proportion of the attributes of each layer for the multi-layer plots of land use resources, the method is similar to the calculation of surface land indicators. The specific operations are as follows: rasterize the spatial distribution of land use resources to generate a land use grid, which contains a number of grid cells. For each grid cell, extract the attribute data of the corresponding layer and calculate the proportion of the attributes of each layer in this grid cell (such as land use type, soil type, vegetation cover, etc.). By statistically calculating the proportion of the layer attributes of all grid cells, the proportion data of the attributes of each layer is obtained, which reflects the distribution proportion of different layers in the plot. By comparing the proportion data of the attributes of each layer, identify the layers with non-maximum attribute proportions. The purpose of this process is to determine the other layers except the layer with the maximum proportion, and these layers have important impacts on the implicit characteristics of land use. The specific method is as follows: First, exclude the layer with the maximum proportion from the proportion data, and the remaining layers are the layers with non-maximum attribute proportions. Screen these non-maximum layers to determine the layers that have significant impacts on the implicit land indicators. Select the layers with greater impacts and relevance for subsequent calculations. For the layers with non-maximum attribute proportions, calculate the pixel chromaticity of the plot. The goal of this process is to further reveal the implicit characteristics of the plot by calculating the pixel chromaticity values of the layers. The specific steps are as follows: Extract the pixel values of the grid cells from the layers with non-maximum attribute proportions. Each pixel value reflects the land use attribute corresponding to this area in space (for example, soil humidity, vegetation coverage, etc.). Map the pixel value of each grid cell to a specific chromaticity model (such as RGB or HSV model). The different attributes of each layer are represented by chromaticity values, thereby revealing the potential characteristics of its land use. Through the above calculations, the pixel chromaticity set of the implicit plot is obtained. This chromaticity set represents the implicit attributes of the non-maximum layers in the land use plot, and through these chromaticity values, some potential or imperceptible characteristics of land use can be revealed. Calculate the chromaticity similarity of the pixel chromaticity set of the implicit plot. The purpose of this process is to evaluate the similarity of their implicit attributes by comparing the chromaticity values of different grid cells. The specific method is as follows: Use the similarity measurement method between chromaticity values (such as cosine similarity, Euclidean distance, etc.) to calculate the chromaticity similarity between each pair of grid cells. According to the calculated chromaticity similarity, generate a chromaticity similarity data set. This data set reflects the implicit attribute similarity between different regions and helps to reveal the potential correlation relationships in land use. Based on the chromaticity similarity data, conduct an attribute correlation analysis on the pixel chromaticity set of the implicit plot. Through this analysis, the implicit attribute correlations in land use can be identified, and further provide support for the management and planning of land resources. The specific analysis method is as follows: Apply association rule mining algorithms (such as Apriori algorithm, FP-growth algorithm, etc.) to analyze the chromaticity similarity data and mine the implicit attribute association rules between grid cells.Through attribute correlation analysis, latent attribute correlation data is generated, which reflects the potential correlations between different layers and different land use types. Based on the latent attribute correlation data, the correlation pixel chromaticity mean processing is performed on the latent plot pixel chromaticity set. This processing aims to aggregate raster cells with similar latent attributes to generate highly representative mean chromaticity data. The specific steps are as follows: According to the latent attribute correlation data, raster cells with similar latent characteristics are selected. The chromaticity means of these selected raster cells are calculated to obtain a comprehensive correlation pixel chromaticity mean, representing the overall characteristics of the region in terms of latent attributes. The calculated correlation pixel chromaticity means are aggregated. Aggregation methods usually include weighted average, normalization, or other statistical methods. Through aggregation, the comprehensive latent attribute characteristics of all raster cells in the region are obtained. Finally, the latent layer index data of land use resources is generated through the aggregation process.

[0118] As an example of the present invention, referring to Figure 3 as shown, in this example, step S3 includes:

[0119] Step S31: Perform temporal variation analysis on the summary data of the main resources on the land surface and the summary data of the latent potential resources of the land to generate land use change data;

[0120] Step S32: Divide the land use change data into data sets to generate a model training set and a model test set; Use the long short-term memory neural network algorithm to train the model training set to generate a land use change prediction pre-model; Optimize and iterate the land use change prediction pre-model through the model test set to generate a land use change prediction model;

[0121] Step S33: Import the land use change data into the land use change prediction model to predict the land use change and generate land use change prediction data;

[0122] Step S34: Reconstruct the spatial distribution map of land use resources through the land use change prediction data to generate a predicted map of land use distribution changes.

[0123] In the embodiments of the present invention, time series analysis is performed on the main resource data of the land surface layer and the potential resource data of the hidden layer. By selecting an appropriate time window, the changing trends of these two types of resources over time are tracked. Time series analysis methods (such as moving average method, exponential smoothing method) are used to identify the changing trends of resource distribution. If the resource distribution is affected by seasonal or periodic factors, techniques such as Fourier transform can be used to extract periodic change characteristics. The rates of change of various resources are calculated to observe the speed of increase or decrease of resources. The time series change results of the summary data of the main resources of the land surface layer and the summary data of the potential resources of the hidden layer of the land are combined to generate comprehensive land use change data. This data set will contain information such as the changing trend, periodic characteristics, and rate of change of land use. The land use change data is divided into a model training set and a model test set according to a certain ratio. A common ratio is 80% for training and 20% for testing, or the ratio can be adjusted according to the data volume and actual requirements. Feature extraction and annotation are performed on the data to ensure that the training data set contains sufficient input features (such as historical resource changes, climate factors, land use types, etc.) and corresponding target outputs (such as resource distribution at a future moment). The LSTM algorithm is used to train the model on the training set. LSTM is a variant of the recurrent neural network (RNN), which is specifically designed to handle the long-term dependence problems of time series data and is suitable for time series prediction of land use changes. The input land use change data is normalized and standardized to improve the model training effect. A specific LSTM network architecture is designed, including hyperparameters such as the number of layers, the number of neurons in each layer, and activation functions. The Backpropagation Through Time (BPTT) algorithm is used to train the model, optimize the model parameters (weights and biases), and minimize the loss function (such as mean squared error MSE). The prediction accuracy of the trained model is evaluated through the model test set, and the model is optimized. For example, adjust the learning rate, increase or decrease the number of LSTM layers, or use regularization techniques such as Dropout to prevent overfitting. The cross-validation method is used to evaluate the model to find the optimal combination of hyperparameters and ensure that the model has strong generalization ability. After training and optimization are completed, a preliminary land use change prediction pre-model is generated, and preliminary prediction and verification can be performed at this stage. The generated land use change data is imported into the trained land use change prediction model. The current resource distribution data and the trained LSTM model are input, and the time series prediction of land use changes is performed using the model. Through the feedforward process of the model, the future land use change situation is predicted using the patterns learned during the training process, and land use change prediction data is generated. The accuracy of the prediction results is evaluated by comparing with the actual data. Evaluation metrics can include error rates (such as mean absolute error MAE), mean squared error MSE, etc. If the model prediction error is large, return to step S32 for further optimization iteration.The land use change data generated by the prediction model is used to predict the future distribution of land resources according to the time dimension. The predicted change data is spatially reconstructed according to the geographic space (such as grid cells), and the spatial distribution of resources at a certain moment in the future is estimated in combination with the existing spatial distribution of land use resources. The land use distribution change prediction map is generated through GIS technology or spatial analysis software (such as ArcGIS, QGIS, etc.) to show the changes in future land use, including different types of land resources, expansion or contraction of land use, and other information. The generated land use distribution change prediction map is visualized, including the distribution changes of different types of land use, to help planners and decision makers understand the trend of land resource changes.

[0124] Preferably, step S4 comprises the following steps:

[0125] Step S41: performing a plot change analysis on the land use resource spatial distribution map according to the land use distribution change prediction map to generate land use change plot data; performing a plot change decision construction on the land use change plot data to generate plot change decision data;

[0126] Step S42: Visualize the plot change decision on the land use distribution change prediction map according to the plot change decision data, and generate a plot decision report.

[0127] In the embodiments of the present invention, by using the predicted map of land use distribution changes generated in the previous step, which includes the spatial distribution of land use resources and the predicted data of future changes, the predicted results of land use changes are analyzed plot by plot to determine the change type of each plot (such as conversion, expansion, contraction, etc. of land use types). The changes of these plots can be analyzed according to time periods (such as annual, quarterly, etc.). The analyzed change types include but are not limited to: newly added plots: newly added land use areas in the prediction, reduced plots: reduction or disappearance of the original land use areas, converted plots: changes in land use types (for example, agricultural land converted to residential land), expanded / contracted plots: expansion or contraction of the existing land use types. Based on the change information of each plot obtained through analysis, land use change plot data is generated. This data includes key information such as the change type, change degree, and time range of change for each plot. According to the land use change plot data, corresponding decision-making schemes are formulated to determine which plot changes meet the planning objectives and which changes require measures to be adjusted or optimized. By comparing with the existing land use planning objectives (such as urban development objectives, agricultural protection objectives, environmental protection objectives, etc.), it is judged whether the land changes meet these objectives. For plots that do not meet the planning objectives, corresponding optimization or adjustment schemes are designed. For example, for plots with a reduction in environmental protection areas, protection needs to be strengthened or land use needs to be adjusted; for areas with high urban construction demands, the land development plan needs to be optimized. The economic, environmental, and social values of each plot are evaluated, and the plots are prioritized according to the evaluation results to provide decision-making support. According to the above analysis, plot change decision data is generated, which includes: specific decision-making suggestions for each plot change (such as changing the use, suspending development, accelerating development, etc.). Improvement measures or optimization schemes proposed for plots that do not meet the plan, and management measures required. The plot change decision data is imported into a Geographic Information System (GIS) or a map drawing tool for visual display. Multiple decision-making layers are created in the GIS platform to distinguish different types of plot decisions: land use type adjustment layer: showing which plots need to adjust or change their land use types. Land protection / development layer: marking plots that need to be strengthened in protection or accelerated in development. Optimization suggestion layer: marking plots that need to optimize or change the land development direction. Support users to switch different decision-making layers according to their needs to interactively view the change situations and decision-making suggestions of plots. The change types and their detailed information (such as area, changed land use, decision-making suggestions, etc.) of plots can be dynamically marked on the map.

Claims

1. A land resource visualization calculation method based on land use patches, characterized in that, It includes the following steps: Step S1: Obtain land resource utilization data and land remote sensing images; Classify the land use types of the land resource utilization data to generate land use type data; Extract the land attribute characteristics from the land resource utilization data according to the land use type data to obtain land attribute characteristic data; Visualize the land patches of the land remote sensing images through the land use type data and the land attribute characteristic data to generate a spatial distribution map of land use resources; Step S2: Divide the layers of the spatial distribution map of land use resources through multi-attribute plots of land use resources to generate single-layer plots of land use resources and multi-layer plots of land use resources; Calculate the surface land index for the single-layer plots of land use resources to obtain the summary data of the main surface resources of the land; Generate the summary data of the potential hidden land resources by calculating the hidden land index for the multi-layer plots of land use resources; Step S3: Predict the land use change based on the summary data of the main surface resources of the land and the summary data of the potential hidden land resources to generate land use change prediction data; Reconstruct the spatial distribution map of land use resources through the land use change prediction data to generate a predicted map of land use distribution change; Step S4: Analyze the plot changes of the spatial distribution map of land use resources according to the predicted map of land use distribution change to generate land use change plot data; Visualize the plot change decision of the land use change plot data to generate a plot decision report.

2. The land resource visualization calculation method based on land use patches according to claim 1, wherein Step S1 includes the following steps: Step S11: Obtain land resource utilization data and land remote sensing images; Step S12: Perform data preprocessing on the land resource utilization data to generate standard land resource utilization data, where the data preprocessing includes data cleaning, data denoising, filling of missing data values, and data standardization; Step S13: Classify the land use types of the standard land resource utilization data to generate land use type data; Step S14: Extract the land attribute characteristics from the land resource utilization data according to the land use type data to obtain land attribute characteristic data; Visualize the land patches of the land remote sensing images through the land use type data and the land attribute characteristic data to generate a spatial distribution map of land use resources.

3. The land resource visualization calculation method based on land use patches according to claim 2, wherein, Step S14 includes the following steps: Step S141: Extract the land surface type characteristics from the land resource utilization data according to the land use type data to obtain land surface type characteristic data; Analyze the land distribution density of the land resource utilization data through the land surface type characteristic data to generate land distribution density data; Step S142: Analyze the land environmental characteristics of the land resource utilization data based on the land distribution density data to generate land environmental characteristic data; Integrate the land surface type characteristic data and the land environmental characteristic data to generate land attribute characteristic data; Step S143: Identify the regional boundaries of the land remote sensing image to generate land area boundary identification data; divide the land remote sensing image based on the land area boundary identification data to generate land area division data; segment the land remote sensing image according to the land area division data to generate a land parcel segmentation image; Step S144: Import the land attribute feature data into the land parcel segmentation image for pixel multi-dimensional color coding to generate a land use resource spatial distribution map.

4. The land resource visualization calculation method based on land use patches according to claim 3, characterized in that, Importing the land attribute feature data into the land parcel segmentation image for pixel multi-dimensional color coding includes: Import the land attribute feature data into the land parcel segmentation image for data filling to generate parcel filling attribute data; perform land attribute RGB color coding on the parcel filling attribute data to generate land attribute RGB color coding data; perform land use HSB adjustment on the land attribute RGB color coding data to generate land use HSB color adjustment data; Visualize the data of the land parcel segmentation image through the land attribute RGB color coding data and the land use HSB color adjustment data to generate an initial land use resource spatial distribution map; perform multi-attribute parcel screening on the initial land use resource spatial distribution map to obtain multi-attribute parcels of land use resources; Overlay the parcel attribute layers of the multi-attribute parcels of land resources to generate a multi-attribute parcel layer of land use resources; calculate the attribute values of the multi-attribute parcel layer of land use resources to obtain land resource attribute value data; sort the multi-attribute parcels of land resources based on the land resource attribute value data to generate attribute ratio sorting data; Control the layer transparency of the multi-attribute parcel layer of land use resources through the attribute ratio sorting data, thereby optimizing the multi-attribute parcels of land use resources; optimize the layer of the initial land use resource spatial distribution map according to the optimized multi-attribute parcels of land use resources to generate a land use resource spatial distribution map.

5. The land resource visualization calculation method based on land use patches according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Divide the layer parcels of the land use resource spatial distribution map through the multi-attribute parcels of land use resources to generate single-layer parcels of land use resources and multi-layer parcels of land use resources; Step S22: Calculate the land indicators for the single-layer parcels of land use resources to obtain fixed land use resource indicators; calculate the surface land indicators for the multi-layer parcels of land use resources to obtain surface land indicator data of land use resources; Step S23: Summarize the data of the land use resource spatial distribution map according to the surface land indicator data of land use resources and the fixed land use resource indicators to generate summary data of the main surface land resources; Step S24: Calculate the hidden layer land indicators for the multi-layer parcels of land use resources to obtain hidden layer land indicator data of land use resources; summarize the data of the land use resource spatial distribution map with the hidden layer land indicator data of land use resources and the fixed land use resource indicators to generate summary data of the potential hidden layer land resources.

6. The land resource visualization calculation method based on land use patches according to claim 5, characterized in that Step S21 includes the following steps: Step S211: Rasterize the land use resource spatial distribution map to generate a land use grid, where the land use grid includes a number of raster cells; Screen multi-attribute raster cells through the land use resource multi-attribute plots to obtain multi-attribute raster cells; Step S212: Classify the land use grid by the multi-attribute raster cells to generate single-attribute raster grids and multi-attribute raster grids; Regionally classify and label the land use resource spatial distribution map according to the single-attribute raster grids and multi-attribute raster grids to generate the first type of land use resource plots and the second type of land use resource plots; Step S213: Conduct a plot layer hierarchy analysis on the first type of land use resource plots and the second type of land use resource plots to generate plot layer hierarchy data; Based on the plot layer hierarchy data, divide the first type of land use resource plots and the second type of land use resource plots into layer plots to generate single-layer land use resource plots and multi-layer land use resource plots.

7. The land resource visualization calculation method based on land use patches according to claim 5, characterized in that Calculating the surface land indicators for the multi-layer land use resource plots includes: Calculate the proportion of each layer attribute for the multi-layer land use resource plots to obtain the proportion data of each layer attribute; Confirm the layer with the largest pixel attribute for the multi-layer land use resource plots according to the proportion data of each layer attribute to obtain the layer with the largest attribute proportion; Calculate the plot pixel chromaticity for the multi-layer land use resource plots according to the layer with the largest attribute proportion, thereby generating the surface layer index data of the land use resources.

8. The land resource visualization calculation method based on land use patches according to claim 5, wherein Calculating the hidden layer land indicators for the multi-layer land use resource plots includes: Calculate the proportion of each layer attribute for the multi-layer land use resource plots to obtain the proportion data of each layer attribute; Confirm the layer with non-maximum pixel attributes for the multi-layer land use resource plots according to the proportion data of each layer attribute to obtain the layer with non-maximum attribute proportion; Calculate the plot pixel chromaticity for the multi-layer land use resource plots according to the layer with non-maximum attribute proportion, thereby generating a set of hidden plot pixel chromaticities; Calculate the chromaticity similarity for the set of hidden plot pixel chromaticities to generate chromaticity similarity data; Conduct an attribute correlation analysis on the set of hidden plot pixel chromaticities according to the chromaticity similarity data to generate hidden attribute correlation data; Use the hidden attribute correlation data to perform an associated pixel chromaticity mean process on the set of hidden plot pixel chromaticities to generate an associated pixel chromaticity mean; Aggregate the associated pixel chromaticity mean data, thereby generating the hidden layer index data of the land use resources.

9. The land resource visualization calculation method based on land use patches according to claim 1, characterized in that Step S3 includes the following steps: Step S31: Conduct a time series change analysis on the land surface main resource summary data and the land hidden potential resource summary data to generate land use change data; Step S32: Divide the land use change data into data sets to generate a model training set and a model test set; Use the long short-term memory neural network algorithm to train the model training set to generate a land use change prediction pre-model; Optimize and iterate the land use change prediction pre-model through the model test set, thereby generating a land use change prediction model; Step S33: Import the land use change data into the land use change prediction model to conduct land use change prediction and generate land use change prediction data; Step S34: Reconstruct the spatial distribution map of land use resources through the land use change prediction data to generate a predicted map of land use distribution changes.

10. The land resource visualization calculation method based on land use patches according to claim 1, wherein, Step S4 includes the following steps: Step S41: Conduct plot change analysis on the spatial distribution map of land use resources based on the predicted map of land use distribution changes to generate land use change plot data; construct plot change decisions for the land use change plot data to generate plot change decision data; Step S42: Visualize the plot change decisions for the predicted map of land use distribution changes based on the plot change decision data to generate a plot decision report.

Images