A method and system for delineating mountain protection lines based on a multi-scale hybrid model

Abstract

A method and system for delineating mountain protection lines based on a multi-scale hybrid model, relating to the field of ecological conservation technology. The method includes: collecting multi-source data of a study area according to a preset multi-dimensional index system to obtain a multi-source dataset; dividing the study area into multiple regular grids according to a preset resolution, with each grid serving as a spatial unit; constructing a multi-scale hybrid model; inputting the multi-source dataset into the multi-scale hybrid model for training; and outputting a preliminary protection line; inputting the preliminary protection line into a dynamic optimization algorithm model for spatial layout optimization; and outputting the final mountain protection line. By adopting this application, standardized processing and spatial unit management of multi-source data are achieved, enabling accurate delineation of mountain protection lines and effectively improving the accuracy of the delineation results and their adaptability to different terrain regions.

Description

Technical Field

[0001] This application relates to the field of ecological protection technology, specifically to a method and system for delineating mountain protection lines based on a multi-scale hybrid model. Background Technology

[0002] As a core component of regional ecosystems, the protection of mountains is a crucial aspect of ecological civilization construction, land spatial planning and control, and sustainable urban development. The scientific delineation of mountain protection lines is a fundamental prerequisite for coordinating mountain ecological protection, geological disaster prevention, efficient land resource utilization, and rational landscape development. It directly determines the effective functioning of mountain ecosystems, the success of regional geological disaster risk control, and the achievement of coordinated goals between ecological protection and economic development.

[0003] The current mainstream implementation methods for delineating mountain protection lines are mainly divided into two categories. One category is a quantitative delineation method based on core indicators such as slope and elevation, including factor analysis and multi-factor unified clustering classification. The other category is a qualitative delineation method based on human sensory experience, including visual corridor control and viewing control.

[0004] The existing methods have several problems in practical applications: First, quantitative methods are difficult to select and quantify indicators. Significant differences in natural mountain conditions across different regions make it difficult to construct a universally applicable indicator system and accurately quantify the impact of each indicator on the delineation of protection lines. Second, qualitative methods are too subjective. Different stakeholders have different standards for judging landscape value, which can easily lead to significant deviations in the delineation results. Third, the delineation results are difficult to implement effectively, as they are hard to connect with relevant plans such as territorial spatial planning and ecological protection red line delineation. Furthermore, they involve complex land use and ownership issues, resulting in insufficient operability. Fourth, they cannot effectively balance the relationship between landscape development and ecological protection. Existing methods often focus on the utilization and management of mountain landscapes, neglecting core ecological service functions such as the integrity of mountain ecosystems, water conservation, soil retention, and biodiversity protection. This makes it difficult to accurately identify key ecological function zones and scientifically delineate areas of different protection levels. Summary of the Invention

[0005] This application provides a method and system for delineating mountain protection lines based on a multi-scale hybrid model, which addresses the problems of low accuracy and poor adaptability in existing mountain protection line delineation methods.

[0006] Firstly, this application provides a method for delineating mountain protection lines based on a multi-scale hybrid model, the method comprising: Based on the preset multi-dimensional indicator system, multi-source data of the study area are collected to obtain a multi-source dataset, and the study area is divided into multiple regular grids according to the preset resolution, with each grid serving as a spatial unit; Construct a multi-scale hybrid model, input multi-source datasets into the multi-scale hybrid model for training, and output an initial protection line; The initial protection line is input into the dynamic optimization algorithm model for spatial layout optimization, and the final mountain protection line is output.

[0007] By adopting the above technical solutions, standardized processing and spatial unit management of multi-source data can be achieved, enabling accurate delineation of mountain protection lines and effectively improving the accuracy of the delineation results and their adaptability to different terrain regions.

[0008] In a specific feasible implementation, a multi-scale mixture model is constructed, multi-source datasets are input into the multi-scale mixture model for training, and an initial protection line is output, specifically including: The multi-source datasets were respectively input into the geographic weighted regression model, the random forest model, and the deep learning network; The spatial relationship between mountain protection-related variables and protection needs at different spatial scales was analyzed using a geographically weighted regression model, and spatial heterogeneity parameters were output. The importance of variables in a multi-source dataset is evaluated using a random forest model, key indicators are selected, and the importance weights of the variables are output. Semantic segmentation of remote sensing data from multiple sources is performed using deep learning networks to extract features of mountain boundaries and land cover categories. By combining spatial heterogeneity parameters, variable importance weights, and features of mountain boundaries and land cover categories, a fused feature vector is obtained, and a preliminary protection line is output.

[0009] By adopting the above technical solutions, the computational limitations of a single model can be overcome, and the objectivity, comprehensiveness, and accuracy of the preliminary protection line delineation can be improved.

[0010] In a specific feasible implementation plan, a geographically weighted regression model is used to analyze the spatial relationship between mountain protection-related variables and protection needs at different spatial scales, outputting spatial heterogeneity parameters, specifically including: An adaptive kernel function is used to determine the bandwidth of the geographic weighted regression model. The adaptive kernel function dynamically adjusts the bandwidth according to the local density of the sample points. Spatial heterogeneity at multiple scales, including local topographic scale and regional ecological function scale, was analyzed using a geographically weighted regression model. Output the regression coefficients for each spatial unit as spatial heterogeneity parameters.

[0011] By adopting the above technical solutions, accurate analysis of multi-scale spatial heterogeneity can be achieved, improving the model's adaptability to complex mountain terrain and computational accuracy.

[0012] In a specific feasible implementation, the preliminary protection line is input into a dynamic optimization algorithm model for spatial layout optimization, and the final mountain protection line is output, specifically including: The initial protection line is used as the initial particle swarm and input into the particle swarm optimization algorithm model. Define a fitness function and search for the guard space location that maximizes the fitness function; During the optimization process, constraints are introduced, and the protection line that satisfies the constraints and has the largest fitness function value is output as the final mountain protection line.

[0013] By adopting the above technical solutions, multi-objective optimization balance can be achieved, improving the spatial rationality and feasibility of the protection line.

[0014] In a specific feasible implementation plan, before collecting multi-source data of the study area according to a pre-set multi-dimensional indicator system, the method also includes: Based on the preset ecological prevention and control goals, a multi-dimensional indicator system is constructed; The indicators in the multi-dimensional indicator system are graded and assigned values ​​to determine the weight of each indicator.

[0015] By adopting the above technical solution, the problems of difficulty in selecting indicators and poor universality of traditional methods are solved, providing a unified calculation basis for the delineation of mountain protection lines.

[0016] In a specific feasible implementation, after outputting the final mountain protection line, the method further includes: Continuously access real-time mountain data, which includes at least remote sensing image data, meteorological data, and human activity data; Real-time mountain data is input into a spatiotemporal geographic weighted regression model for dynamic monitoring, and the risk index of the area inside and outside the protection line is output. If the risk index of the area within the protection line exceeds the preset risk threshold, the protection line optimization process is triggered, and the process returns to the step of building a multi-scale hybrid model.

[0017] By adopting the above technical solutions and setting up an automatic optimization process for risk triggering, the long-term effectiveness of the protection line can be guaranteed.

[0018] In one specific feasible implementation, the method further includes: Select typical locations for field reconnaissance to obtain field survey data and verify the accessibility and rationality of the protection line boundary; By comparing changes in land features inside and outside the protection zone using real-time mountain data, the effectiveness of the protection can be assessed.

[0019] By adopting the above technical solutions, the boundary of the protection line was accurately verified, ensuring the feasibility of the delineation results and providing reliable data support for the long-term management and continuous optimization of mountain protection.

[0020] A second aspect of this application provides a mountain protection line delineation system based on a multi-scale hybrid model, the system comprising: The data acquisition module is used to collect multi-source data from the study area; The training module is used to build a multi-scale hybrid model. It takes multi-source datasets as input to train the multi-scale hybrid model and outputs an initial protection line. The dynamic optimization module is used to input the preliminary protection line into the dynamic optimization algorithm model for spatial layout optimization and output the final mountain protection line.

[0021] A third aspect of this application provides an electronic device, comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to execute the above-described method steps.

[0022] A fourth aspect of this application provides a computer storage medium storing a plurality of instructions adapted for loading by a processor and executing the method steps described above. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating a method for delineating mountain protection lines based on a multi-scale hybrid model, as provided in an embodiment of this application. Detailed Implementation

[0024] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0025] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0026] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0027] Please refer to Figure 1 This paper presents a flowchart illustrating a method for delineating mountain protection lines based on a multi-scale hybrid model. This method can be implemented using a computer program, a microcontroller, or run on a mountain protection line delineation system based on a multi-scale hybrid model. The computer program can be integrated into a computer device or run as a standalone application. Specifically, the method includes steps S100 to S300, as follows: S100. Based on the preset multi-dimensional indicator system, collect multi-source data of the study area to obtain a multi-source dataset, and divide the study area into multiple regular grids according to a preset resolution, with each grid serving as a spatial unit. Based on a multi-dimensional indicator system, the types and scope of multi-source data collection are determined, with the collection scope covering the entire study area. The collected multi-source data may include high-resolution remote sensing imagery, digital elevation model (DEM) data, ecological sensitivity thematic data, soil type and physicochemical property thematic data, land use status data, road network vector data, vegetation cover thematic data, meteorological monitoring data, geological disaster survey data, land spatial planning and control data, and ownership and human resource data obtained through field surveys.

[0028] All collected multi-source data are preprocessed to obtain a standardized multi-source dataset. Vector boundary data of the study area are acquired, and the entire study area is divided into multiple continuous and non-overlapping regular grids according to the preset spatial resolution. Each regular grid corresponds to an independent spatial cell. All data in the multi-source dataset are matched to the corresponding spatial cells according to their spatial location, so that each spatial cell corresponds to a complete set of standardized index data.

[0029] During the mesh generation process, multi-resolution adaptation rules are set. For highly heterogeneous areas with terrain undulation exceeding a preset threshold, a higher spatial resolution is used to generate the mesh. For areas with flat terrain and gradual changes in indicators, a relatively lower spatial resolution is used to generate the mesh. This reduces the overall computational load and improves processing efficiency while ensuring computational accuracy.

[0030] In some embodiments, the preset basic spatial resolution can be 10 meters × 10 meters, the spatial resolution of highly heterogeneous areas can be 5 meters × 5 meters, and the spatial resolution of flat areas can be 30 meters × 30 meters.

[0031] Based on the above embodiments, as another optional embodiment, before collecting multi-source data of the study area according to a preset multi-dimensional indicator system, the method further includes: S101. Based on the preset ecological prevention and control goals, construct a multi-dimensional indicator system; In the embodiments of this application, the preset ecological prevention and control objectives may include five core objectives: mountain ecological protection, geological disaster prevention and control, land use optimization, mountain landscape development management and control, and protection of traditional villages and cultural resources.

[0032] To achieve the pre-set ecological protection goals, a multi-dimensional indicator system is constructed, encompassing natural, ecological, and socio-economic factors. The natural factor dimension includes indicators such as mountain elevation, slope, aspect, topographic relief, surface roughness, soil type, soil thickness, and bedrock lithology. The ecological factor dimension includes indicators such as vegetation cover, ecological sensitivity, water conservation capacity, soil retention capacity, biodiversity abundance, geological hazard susceptibility, and soil erosion intensity. The socio-economic factor dimension includes indicators such as land use type, road network density, population density, intensity of human activity disturbance, building distribution density, ownership boundary information, regional territorial spatial planning and control requirements, and ecological protection red line connection requirements.

[0033] In some embodiments, corresponding admission verification rules are set for each dimension of indicators. The admission verification rules correspond one-to-one with the preset ecological prevention and control targets, ensuring that each indicator included in the indicator system can directly respond to at least one ecological prevention and control target, and avoiding the introduction of invalid indicators that would reduce the accuracy and efficiency of subsequent model calculations.

[0034] S102. Assign values ​​to each indicator in the multi-dimensional indicator system according to its level and determine the weight of each indicator.

[0035] For each indicator in the multi-dimensional indicator system, the quantitative processing of the indicator is completed by adopting a hierarchical value assignment method, taking into account the indicator's own attribute characteristics and its impact on mountain protection.

[0036] In some embodiments, corresponding hierarchical assignment rules are set for positive and negative indicators respectively. Positive indicators are those whose larger values ​​have a stronger positive effect on mountain protection. Positive indicators include vegetation coverage, water conservation capacity, soil retention capacity, and biodiversity abundance. Negative indicators are those whose larger values ​​have a stronger negative effect on mountain protection. Negative indicators include geological disaster susceptibility, soil erosion intensity, human activity disturbance intensity, and building distribution density.

[0037] The numerical range of each indicator is divided into multiple consecutive hierarchical intervals, and a corresponding value is assigned to each interval. When determining the weight of each indicator, the analytic hierarchy process (AHP) combined with the entropy weighting method is used to complete the combined weighting calculation of each indicator.

[0038] In some embodiments, the number of grading intervals can be 5, with a value range of 1 to 5, where a higher value for a positive indicator corresponds to a higher value, and a higher value for a negative indicator corresponds to a lower value.

[0039] S200. Construct a multi-scale hybrid model, input the multi-source dataset into the multi-scale hybrid model for training, and output the initial protection line; A multi-scale hybrid model is constructed, which is a hybrid coupled model that integrates the geographically weighted regression (GWR) model, the random forest (RF) model, and the deep learning network. The multi-scale hybrid model sets up parallel computing branches, with three parallel computing branches corresponding to the geographically weighted regression model, the random forest model, and the deep learning network, respectively. The three parallel computing branches synchronously receive the input multi-source datasets and complete independent calculations, and then the calculation results of the three branches are fused.

[0040] The standardized multi-source dataset obtained in step S100 is input into the constructed multi-scale hybrid model. The multi-scale hybrid model trains and processes the input multi-source dataset and outputs the preliminary protection line of the mountain in the study area. The preliminary protection line is a closed vector boundary line that divides the study area into a mountain protection and control area and a non-protection area.

[0041] Based on the above embodiments, as another optional embodiment, a multi-scale hybrid model is constructed. Multi-source datasets are input into the multi-scale hybrid model for training, and an initial protection line is output. Specifically, this includes: S201. Input the multi-source datasets into the geographic weighted regression model, the random forest model, and the deep learning network respectively; analyze the spatial relationship between mountain protection-related variables and protection needs at different spatial scales using the geographic weighted regression model, and output spatial heterogeneity parameters; evaluate the importance of variables in the multi-source datasets using the random forest model, select key indicators, and output variable importance weights; perform semantic segmentation on the remote sensing data of the multi-source datasets using the deep learning network, and extract mountain boundary and land cover category features; The standardized multi-source dataset obtained in step S100 is synchronously input into the three parallel computation branches of the multi-scale hybrid model. The three parallel computation branches are the geographic weighted regression model branch, the random forest model branch, and the deep learning network branch.

[0042] This study analyzes the spatial relationship between mountain protection-related variables and mountain protection needs at different spatial scales using a geographically weighted regression model. The mountain protection-related variables are all quantitative indicators in a multi-dimensional indicator system, and the mountain protection needs are the mountain protection level requirements for the corresponding spatial unit. After calculation using the geographically weighted regression model, the spatial heterogeneity parameters for each spatial unit are output.

[0043] The random forest model branch evaluates the importance of variables in the input multi-source dataset. Through parallel training and voting operations of multiple decision trees, the random forest model branch calculates the importance score of each indicator in the multi-dimensional indicator system for the delineation of the mountain protection line. All indicators are ranked from highest to lowest importance score, and those exceeding a preset importance threshold are selected as key indicators. The variable importance weight corresponding to each key indicator is also output.

[0044] Semantic segmentation of high-resolution remote sensing image data from multiple source datasets is performed using a deep learning network branch. The deep learning network adopts a U-shaped fully convolutional neural network structure. First, the deep learning network is pre-trained using a set of labeled mountain remote sensing image samples. After pre-training, high-resolution remote sensing image data of the study area is input into the deep learning network. The deep learning network performs pixel-by-pixel semantic segmentation on the input remote sensing image data, extracting the initial boundary features of the mountains and the category features of various land features in the study area. The land feature categories include vegetation, buildings, water bodies, roads, bare land, and cultivated land.

[0045] In some embodiments, the number of decision trees in the random forest model branch can be 100, and the preset importance threshold can be 0.05. Indicators with an importance score lower than 0.05 will be removed and will not participate in subsequent fusion operations.

[0046] Based on the above embodiments, as another optional embodiment, a geographically weighted regression model is used to analyze the spatial relationship between mountain protection-related variables and protection needs at different spatial scales, outputting spatial heterogeneity parameters, specifically including: S2011. An adaptive kernel function is used to determine the bandwidth of the geographic weighted regression model. The adaptive kernel function dynamically adjusts the bandwidth according to the local density of the sample points. In the operation of the geographic weighted regression model, an adaptive kernel function is used to determine the bandwidth parameter of the model. The bandwidth parameter is used to control the decay rate of spatial weights in the geographic weighted regression model and determine the spatial influence range of the model.

[0047] In some embodiments, the adaptive kernel function is a bisquare adaptive kernel function. The kernel bandwidth corresponding to each sample point is dynamically adjusted according to the local density of sample points in the study area. The sample point is the center point of each spatial unit obtained by step S100.

[0048] For areas with high local density of sample points, the corresponding kernel bandwidth is reduced; for areas with low local density of sample points, the corresponding kernel bandwidth is increased. This allows the geographic weighted regression model to adapt to the complex mountainous terrain changes within the study area, accurately capture the spatial heterogeneity characteristics of different regions, and avoid the problem that fixed bandwidth cannot adapt to differences in terrain density.

[0049] S2012. Spatial heterogeneity at multiple scales, including local topographic scale and regional ecological function scale, is analyzed by a geographically weighted regression model. In some embodiments, two core analysis scales are set: the local topographic scale and the regional ecological function scale. The analysis scope of the local topographic scale is a single spatial unit and its surrounding adjacent spatial units, which is used to capture the impact of micro-features such as topographic undulations and slope changes on the needs of mountain protection within a small area.

[0050] The analysis scope at the regional ecological function scale covers complete ecological sub-basins or mountain units, used to capture the impact of macro-characteristics such as ecosystem integrity, water conservation function, and geological disaster risk on mountain protection needs over a large area. Through a multi-scale geographically weighted regression model, independent bandwidth parameters are set for variables corresponding to different analysis scales, and regression calculations are performed at different scales to obtain spatial relationship calculation results for each scale.

[0051] S2013. Output the regression coefficients on each spatial unit as spatial heterogeneity parameters.

[0052] For each independent spatial unit obtained from the S100 step, the regression coefficient of each indicator in the corresponding multi-dimensional indicator system is calculated. The regression coefficient is used to characterize the degree and direction of the influence of the change of the indicator on the mountain protection needs within the corresponding spatial unit.

[0053] The regression coefficients of all indicators corresponding to each spatial unit are combined to form the spatial heterogeneity parameter corresponding to that spatial unit. The spatial heterogeneity parameter is a multi-dimensional vector that corresponds one-to-one with the spatial unit. The spatial heterogeneity parameters of all spatial units are output to the feature fusion module of the multi-scale hybrid model for subsequent feature fusion processing.

[0054] S202. Combining spatial heterogeneity parameters, variable importance weights, and characteristics of mountain boundaries and land cover categories, a fused feature vector is obtained, and a preliminary protection line is output.

[0055] The feature fusion module of the multi-scale hybrid model receives spatial heterogeneity parameters output by the geographically weighted regression model branch, variable importance weights output by the random forest model branch, and mountain boundary and land cover category features output by the deep learning network branch, thus completing the fusion processing of multi-source features.

[0056] In some embodiments, the spatial heterogeneity parameters are first weighted and corrected by the importance weight of variables to obtain the weighted spatial heterogeneity features. The weighting correction method is to multiply the regression coefficient of each index in the spatial heterogeneity parameters of each spatial unit with the importance weight of the variable for that index.

[0057] The weighted spatial heterogeneity features are matched with the mountain boundary and land cover category features in terms of spatial location and dimension to obtain the fused feature vector corresponding to each spatial unit. The fused feature vector fully represents the mountain protection priority of the corresponding spatial unit.

[0058] Based on the values ​​of the fused feature vectors, all spatial units are classified into mountain protection priority levels. Spatial units with protection priority exceeding the preset priority threshold are classified as mountain protection units, while spatial units with protection priority not exceeding the preset priority threshold are classified as non-protection units.

[0059] The outer boundaries of all mountain protection units are extracted. After boundary smoothing and topology correction, closed vector boundary lines are generated as preliminary protection lines for the mountains in the study area. The generated preliminary protection lines are then output to the subsequent optimization processing steps.

[0060] In some embodiments, the preset priority threshold can be 0.6, and spatial units with fused feature vector values ​​exceeding 0.6 will be classified as mountain protection units.

[0061] S300: Input the preliminary protection line into the dynamic optimization algorithm model for spatial layout optimization, and output the final mountain protection line.

[0062] In some embodiments, a dynamic optimization algorithm model is constructed to refine the spatial layout of the preliminary protection line, thereby addressing issues such as boundary fragmentation, conflict with existing planning and control requirements, and insufficient spatial accessibility of the preliminary protection line.

[0063] The preliminary protection line of the mountain output from step S200 is input into the constructed dynamic optimization algorithm model. The dynamic optimization algorithm model uses the preliminary protection line as the optimization basis and combines the preset optimization objectives and constraints to complete the iterative optimization calculation of the spatial layout of the protection line. Finally, the final mountain protection line of the study area is output. The final mountain protection line is an optimized closed vector boundary line with the characteristics of spatial continuity, control feasibility, and ecological protection rationality.

[0064] Based on the above embodiments, as another optional embodiment, the preliminary protection line is input into a dynamic optimization algorithm model for spatial layout optimization, and the final mountain protection line is output, specifically including: S301. Use the initial protection line as the initial particle swarm and input it into the particle swarm optimization algorithm model. In some embodiments, the dynamic optimization algorithm model includes a particle swarm optimization (PSO) algorithm model, which first discretizes the preliminary protection line output in step S202, and discretizes the boundary of the preliminary protection line into multiple continuous boundary control points, each boundary control point corresponding to a set of two-dimensional spatial coordinates.

[0065] Each boundary control point is treated as an independent particle. All particles corresponding to the boundary control points are combined to form an initial particle swarm. The number of particles in the initial particle swarm is the same as the number of boundary control points. The position parameter of each particle is the two-dimensional spatial coordinates of the corresponding boundary control point, and the velocity parameter of each particle is the spatial position adjustment step size of the corresponding boundary control point. The constructed initial particle swarm is input into the particle swarm optimization algorithm model as the initial solution for iterative optimization operations.

[0066] S302. Define the fitness function and search for the guard line space location that maximizes the fitness function; In some embodiments, in the particle swarm optimization algorithm model, a fitness function is defined to evaluate the merits of the spatial layout of the protection line. The optimization objective of the fitness function may include maximizing the ecological benefits of mountain protection, minimizing the impact of protection line delineation on legal development activities in the area, and maximizing the spatial continuity and feasibility of the protection line boundary.

[0067] The input parameter of the fitness function is the boundary coordinates of the protection line corresponding to the particle swarm, and the output value of the fitness function is the comprehensive fitness score of the corresponding protection line layout. The comprehensive fitness score is positively correlated with ecological benefits, negatively correlated with development impact, and positively correlated with boundary continuity.

[0068] The particle swarm optimization algorithm model is used to perform multiple rounds of iterative calculations. In each round of iteration, the model updates the velocity and position parameters of each particle, calculates the fitness function value of each particle, and records the position information of the global best particle and the individual best particle. The global best particle is the particle with the largest fitness function value in all iterations, and the individual best particle is the position of the single particle with the largest fitness function value in the iteration.

[0069] Continue iterative calculations until the preset maximum number of iterations is reached, or the change in the fitness function value is less than the preset convergence threshold, then stop the iterative calculations and output the guard line space position that maximizes the fitness function obtained during the iteration process.

[0070] In some embodiments, the preset maximum number of iterations can be 200, and the preset convergence threshold can be 0.001.

[0071] S303. In the optimization process, constraints are introduced, and the protection line that satisfies the constraints and has the largest fitness function value is output as the final mountain protection line.

[0072] In some embodiments, during the iterative optimization process of the particle swarm optimization algorithm model, hard constraints are introduced. These hard constraints may include constraints related to land spatial planning and control, ecological protection red line connection, land ownership boundaries, geological disaster prevention and control, and mountain ecological integrity.

[0073] In each round of iterative calculation, the constraint conditions of the protection line boundary corresponding to the updated particle position are checked. Particle positions that do not meet any hard constraint conditions are eliminated, and only particle positions that meet all hard constraint conditions are retained to participate in subsequent iterative calculations and fitness score calculations.

[0074] After completing the iterative calculation, the combination of particle positions with the largest fitness function value is selected from all particle positions that meet the constraints, and restored to a continuous protection line vector boundary. After topology checking, boundary smoothing and attribute assignment, the final mountain protection line of the study area is generated, and the final mountain protection line is output to the storage module and display module, thus completing the delineation of the mountain protection line.

[0075] Based on the above embodiments, as another optional embodiment, after outputting the final mountain protection line, the method further includes: S304. Continuously access real-time mountain data, which includes at least remote sensing image data, meteorological data, and human activity data. In some embodiments, after outputting the final mountain protection line, a dynamic monitoring process for mountain protection is initiated, continuously accessing real-time mountain data for the study area through a preset data interface. The accessed real-time mountain data may include time-series high-resolution remote sensing imagery, real-time meteorological monitoring data, and human activity monitoring data.

[0076] Among them, the update cycle of time-series high-resolution remote sensing image data does not exceed the preset image update cycle, real-time meteorological monitoring data includes real-time monitoring values ​​and cumulative values ​​of rainfall, wind force, and temperature, and human activity monitoring data includes data on building construction activities, road network traffic, vegetation destruction activities, and land use change within the protection line area.

[0077] S305. Input real-time mountain data into a spatiotemporal geographic weighted regression model for dynamic monitoring, and output the risk index of the area inside and outside the protection line. In some embodiments, a geographically and temporally weighted regression (GTWR) model is constructed. The GTWR model introduces a time dimension on the basis of the geographically weighted regression model, which can simultaneously capture spatial heterogeneity and temporal dynamic change characteristics.

[0078] The real-time mountain data preprocessed in step S304 is input into the spatiotemporal weighted regression model. The spatiotemporal weighted regression model combines historical monitoring data with real-time data to comprehensively calculate the ecological risk, geological disaster risk, and human activity disturbance risk of each spatial unit inside and outside the protection line, and outputs the comprehensive risk index corresponding to each spatial unit. The value range of the comprehensive risk index is 0 to 1, and the higher the value, the higher the risk level of the corresponding spatial unit.

[0079] The distribution of average risk index, highest risk index and high-risk spatial units in the areas inside and outside the protection line were statistically analyzed to complete the dynamic monitoring of the mountain protection status in the study area.

[0080] S306. If the risk index of the area within the protection line is detected to exceed the preset risk threshold, the protection line optimization process is triggered, and the process returns to the step of building a multi-scale hybrid model.

[0081] In some embodiments, a risk threshold for mountain protection is preset, which corresponds to the highest permissible risk index within the protection zone.

[0082] The highest risk index and average risk index of the area within the protection line output by step S305 are compared with the preset risk thresholds. If the highest risk index of the area within the protection line exceeds the preset risk threshold, or the average risk index of the area within the protection line exceeds the preset risk threshold for a consecutive preset monitoring period, the mountain protection line optimization process is automatically triggered.

[0083] After triggering the protection line optimization process, the latest real-time mountain data and updated planning and control data are obtained. The process then returns to the step of building a multi-scale hybrid model in step S200. The multi-scale hybrid model is retrained and calculated using the latest data. The existing mountain protection line is iteratively optimized, and the updated mountain protection line is output to ensure that the protection line can adapt to the dynamic changes in the mountain environment and regional development.

[0084] In some embodiments, the preset risk threshold can be 0.7, and the preset monitoring period can be 3 consecutive monthly periods.

[0085] Based on the above embodiments, as another optional embodiment, the method further includes: S307. Select typical areas for field reconnaissance, obtain field survey data, and verify the accessibility and rationality of the protection line boundary; In some embodiments, based on the final delineation of the mountain protection line and in conjunction with the topographic features, risk distribution, and land feature distribution of the study area, typical areas are selected as field reconnaissance areas. These typical areas include protection line boundary transition zones, high-risk areas, areas with dense human activity, ecologically sensitive areas, and areas with complex land feature types.

[0086] Field surveyors conduct field surveys based on the areas selected by the system, acquiring field survey data through on-site measurements, photographic recording, ownership verification, and boundary accessibility testing. This field survey data includes the actual coordinates of the survey points, current topography, current land features, land ownership information, boundary accessibility information, and current ecological status information.

[0087] Obtain field survey data from field reconnaissance, compare and verify the field survey data with the protection line boundary delineated by the model, verify the spatial location accuracy, accessibility, and rationality of the protection line boundary, and make local optimizations and adjustments for boundary sections that do not conform to the actual situation, thereby improving the feasibility of implementing the protection line.

[0088] S308. By comparing changes in land features inside and outside the protection line using real-time mountain data, the protection effect can be assessed.

[0089] In some embodiments, the time-series real-time mountain data continuously accessed through step S304 is used to extract information on changes in land features in the area within and outside the protection line. This information includes changes in vegetation cover, built-up land use, water area, bare land area, and soil erosion intensity.

[0090] Using the time when the protection line was delineated as the baseline time, the change data of ground features were compared with the baseline time and the current monitoring time to calculate the magnitude and trend of changes in ground features in the area inside and outside the protection line, respectively.

[0091] Based on the comparison results of land feature changes, combined with changes in ecosystem service functions and the occurrence of geological disasters, the protection effect of the mountain protection line is quantitatively assessed, and a protection effect assessment report is generated. The protection effect assessment report includes the effectiveness of protection line management, existing risks and problems, and subsequent optimization suggestions, providing data support for the long-term management and policy formulation of mountain protection.

[0092] Based on the above embodiments, as another optional embodiment, this application also provides a mountain protection line delineation system based on a multi-scale hybrid model, the system comprising: The data acquisition module is used to collect multi-source data from the study area; The training module is used to construct a multi-scale hybrid model, inputting multi-source datasets into the multi-scale hybrid model for training, and outputting an initial protection line; The dynamic optimization module is used to input the preliminary protection line into the dynamic optimization algorithm model for spatial layout optimization and output the final mountain protection line.

[0093] In some embodiments, the system further includes: a data preprocessing module for preprocessing the multi-source dataset to obtain a standardized multi-source dataset; and a verification and evaluation module for verifying and evaluating the final mountain protection line. If the verification is successful, the final mountain protection line is output to the land and space planning platform.

[0094] It should be noted that the system provided in the above embodiments is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0095] Based on the above embodiments, as another optional embodiment, the present application embodiment may further include a computer storage medium, which may store multiple instructions adapted for loading by a processor and executing a method of the above embodiments. For the specific execution process, please refer to the detailed description of the above embodiments, which will not be repeated here.

[0096] Based on the above embodiments, as another optional embodiment, this application embodiment may further include an electronic device. The electronic device may include: at least one processor, at least one communication bus, a user interface, at least one network interface, and a memory.

[0097] The communication bus is used to enable communication between these components.

[0098] The user interface may include a display screen and a camera. Optional user interfaces may also include standard wired interfaces and wireless interfaces.

[0099] The network interface may include standard wired interfaces and wireless interfaces (such as Wi-Fi interfaces).

[0100] The processor may include one or more processing cores. It connects to various parts of the server via various interfaces and lines, executing instructions, programs, code sets, or instruction sets stored in memory, and accessing data stored in memory to perform various server functions and process data. Optionally, the processor may be implemented using at least one of the following hardware forms: Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor may integrate one or more of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip without being integrated into the processor.

[0101] The memory may include random access memory (RAM) or read-only memory. Optionally, the memory may include a non-transitory computer-readable storage medium. The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor. As a computer storage medium, the memory may include an operating system, a network communication module, a user interface module, and an application program of one method.

[0102] In electronic devices, the user interface is primarily used to provide an input interface for users and to acquire user input data; while the processor can be used to call an application program stored in memory that represents a method. When executed by one or more processors, this causes the electronic device to perform one or more methods as described in the above embodiments. It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps can be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0103] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0104] In the various embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between apparatuses or units may be electrical or other forms.

[0105] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0106] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0107] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0108] The above are merely exemplary embodiments of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will readily conceive of those skilled in the art upon consideration of the specification and the disclosure of practical truths.

[0109] This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for delineating a mountain protection line based on a multi-scale hybrid model, characterized in that, The method includes: Based on a preset multi-dimensional indicator system, multi-source data of the study area are collected to obtain a multi-source dataset, and the study area is divided into multiple regular grids according to a preset resolution, with each grid serving as a spatial unit; Construct a multi-scale hybrid model, input the multi-source dataset into the multi-scale hybrid model for training, and output a preliminary protection line; The initial protection line is input into the dynamic optimization algorithm model for spatial layout optimization, and the final mountain protection line is output.

2. The method for delineating mountain protection lines based on a multi-scale hybrid model according to claim 1, characterized in that, The construction of the multi-scale hybrid model, which involves inputting the multi-source dataset into the multi-scale hybrid model for training and outputting an initial protection line, specifically includes: The multi-source datasets are respectively input into the geographic weighted regression model, the random forest model, and the deep learning network; The spatial relationship between mountain protection-related variables and protection needs at different spatial scales is analyzed using the geographically weighted regression model, and spatial heterogeneity parameters are output. The random forest model is used to evaluate the importance of variables in the multi-source dataset, select key indicators, and output the importance weights of the variables. The deep learning network is used to perform semantic segmentation on the remote sensing data of the multi-source dataset, and to extract features of mountain boundaries and land cover categories. By combining the spatial heterogeneity parameters, the importance weights of the variables, and the characteristics of the mountain boundary and land cover categories, a fused feature vector is obtained, and a preliminary protection line is output.

3. The method for delineating mountain protection lines based on a multi-scale hybrid model according to claim 2, characterized in that, The method analyzes the spatial relationship between mountain protection-related variables and protection needs at different spatial scales using the geographic weighted regression model, and outputs spatial heterogeneity parameters, specifically including: An adaptive kernel function is used to determine the bandwidth of the geographic weighted regression model. The adaptive kernel function dynamically adjusts the bandwidth according to the local density of the sample points. The spatial heterogeneity at multiple scales, including local topographic scale and regional ecological function scale, is analyzed using the geographically weighted regression model. Output the regression coefficients on each spatial unit as spatial heterogeneity parameters.

4. The method for delineating mountain protection lines based on a multi-scale hybrid model according to claim 1, characterized in that, The step of inputting the preliminary protection line into a dynamic optimization algorithm model for spatial layout optimization and outputting the final mountain protection line specifically includes: The initial protection line is used as the initial particle swarm and input into the particle swarm optimization algorithm model. Define a fitness function and search for the guard line space location that maximizes the fitness function; During the optimization process, constraints are introduced, and the protection line that satisfies the constraints and has the largest fitness function value is output as the final mountain protection line.

5. The method for delineating mountain protection lines based on a multi-scale hybrid model according to claim 1, characterized in that, Before collecting multi-source data of the study area according to a preset multi-dimensional indicator system, the method further includes: Based on the preset ecological prevention and control goals, a multi-dimensional indicator system is constructed; Each indicator in the multi-dimensional indicator system is assigned a graded value to determine the weight of each indicator.

6. The method for delineating mountain protection lines based on a multi-scale hybrid model according to claim 1, characterized in that, After outputting the final mountain protection line, the method further includes: Continuously access real-time mountain data, which includes at least remote sensing image data, meteorological data, and human activity data; The real-time data of the mountain is input into a spatiotemporal geographic weighted regression model for dynamic monitoring, and the risk index of the area inside and outside the protection line is output. If the risk index of the area within the protection line exceeds the preset risk threshold, the protection line optimization process is triggered, and the process returns to the step of constructing the multi-scale hybrid model.

7. The method for delineating mountain protection lines based on a multi-scale hybrid model according to claim 6, characterized in that, The method further includes: Select typical locations for field reconnaissance to obtain field survey data and verify the accessibility and rationality of the protection line boundary; By comparing the changes in land features inside and outside the protection line using real-time data of the mountain, the effectiveness of the protection can be assessed.

8. A mountain protection line delineation system based on a multi-scale hybrid model, characterized in that, The system includes: The data acquisition module is used to collect multi-source data from the study area; The training module is used to construct a multi-scale hybrid model, inputting multi-source datasets into the multi-scale hybrid model for training, and outputting an initial protection line; The dynamic optimization module is used to input the preliminary protection line into the dynamic optimization algorithm model for spatial layout optimization and output the final mountain protection line.

9. An electronic device, characterized in that, It includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted to be loaded by a processor and executed as described in any one of claims 1-7.

Images