Ecological restoration area intelligent identification method based on multi-source data and model integration

Abstract

The invention relates to an ecological restoration area intelligent identification method based on multi-source data and model integration. The method comprises the following steps: acquiring a multi-source remote sensing image, terrain and soil attribute data, and generating a multi-source data base and an ecological source land set through feature extraction; constructing an ecological resistance surface based on the multi-source data base; performing circuit theoretical simulation based on the ecological resistance surface and the ecological source ground set, and generating a reference current density distribution diagram and a reference equivalent connectivity index; generating a connected gain distribution diagram based on the ecological resistance surface and the multi-source data base virtual restoration; calculating a gain-cost ratio based on the connected gain distribution map, the terrain gradient and the road distance data, and generating a preferential restoration potential distribution map; and performing spatial clustering analysis based on the preferential restoration potential distribution map to generate an ecological restoration key area. By adopting the method, benefits and cost can be considered, and the ecological restoration key area can be accurately identified.

Description

Technical Field

[0001] This invention belongs to the field of intelligent identification technology, and in particular relates to an intelligent identification method for ecological restoration areas based on multi-source data and model integration. Background Technology

[0002] With the development of land space ecological restoration and landscape ecology technologies, the use of Geographic Information Systems (GIS) and multi-source remote sensing data for ecological network construction and key area identification has become a mainstream technical approach. This approach can intuitively reflect the spatial distribution characteristics and dynamic changes of regional landscape patterns, providing scientific and quantitative data support for ecological restoration planning, and making it possible to accurately identify and assess fragmented habitats.

[0003] In traditional techniques, identifying key areas for ecological restoration primarily relies on spatial analysis using the minimum cumulative resistance (MCR) model and circuit theory. Specifically, researchers typically construct ecological resistance surfaces based on fundamental geographical elements such as land use type, topography, and vegetation cover. They then extract ecological source areas through morphological spatial analysis and use circuit theory to simulate the movement of ecological flows within the landscape. Based on this, areas with high current density are identified as "ecological pinch points," or areas where connectivity is disrupted due to high resistance are identified as "ecological barrier points." These areas with poor current connectivity or flow bottlenecks are directly designated as priority sites for restoration.

[0004] However, current identification methods typically focus only on the distribution of resistance and flow bottlenecks within the existing landscape pattern, representing a static diagnosis of existing ecological problems. This approach cannot quantify the extent to which the connectivity of the entire ecosystem would improve after restoration of a specific area; in other words, it lacks a mechanism for dynamically simulating and evaluating the expected benefits of restoration actions. Furthermore, existing technologies often neglect the balance between engineering implementation difficulty and ecological restoration benefits during the planning process, potentially leading to selected restoration areas with excessively high costs but limited connectivity improvements. Summary of the Invention

[0005] Therefore, it is necessary to provide an intelligent identification method for ecological restoration areas based on multi-source data and model integration, which can intelligently and accurately identify the areas with the greatest restoration value from the dual perspectives of maximizing connectivity gain and engineering economic feasibility, in order to address the above-mentioned technical problems.

[0006] Firstly, this application provides a method for intelligent identification of ecological restoration areas based on multi-source data and model integration, including:

[0007] S1. Acquire multi-source remote sensing images, topographic data, and soil attribute data, and extract features from the multi-source remote sensing images, topographic data, and soil attribute data to generate a multi-source data base and an ecological source set;

[0008] S2. Construct an ecological resistance surface based on multi-source data;

[0009] S3. Perform circuit theory simulation calculations based on the ecological resistance surface and the set of ecological source areas to generate a reference current density distribution map and a reference equivalent connectivity index.

[0010] S4. Based on the ecological resistance surface and multi-source data base, perform virtual restoration simulation to generate a connectivity gain distribution map;

[0011] S5. Calculate the gain-cost ratio based on the connectivity gain distribution map, terrain slope, and road distance data, and generate a priority repair potential distribution map.

[0012] S6. Based on the priority restoration potential distribution map, perform spatial cluster analysis to generate key areas for ecological restoration.

[0013] In one embodiment, S2 includes:

[0014] S21. Extract resistance parameters based on land use type, topographic factors, intensity of human disturbance and vegetation cover in the multi-source data base, and generate a set of basic resistance coefficients and correction factors.

[0015] S22. Perform multi-dimensional factor coupling calculations based on the basic resistance coefficient and the set of correction factors to generate the final ecological resistance value; the formula for calculating the final ecological resistance value is as follows:

[0016]

[0017] in, For position The final ecological resistance value at the location, This is the basic drag coefficient. The normalized slope factor. The normalized artificial interference intensity factor. The normalized vegetation cover factor. , , These are the topography sensitivity coefficient, disturbance sensitivity coefficient, and vegetation cover correction coefficient, respectively.

[0018] S23. Based on the final ecological resistance value, perform rasterization mapping to generate an ecological resistance surface.

[0019] In one embodiment, S3 includes:

[0020] S31. Based on the set of ecological source areas, perform node mapping and conductivity matrix transformation of ecological resistance surfaces to generate a conductivity matrix;

[0021] S32. Solve the linear equations based on the conductivity matrix to generate voltage and current distributions;

[0022] S33. Based on the voltage and current distribution, extract connectivity indices to generate a reference current density distribution map and a reference equivalent connectivity index.

[0023] In one embodiment, S4 includes:

[0024] S41. Extract obstacle points based on the ecological resistance surface to generate a set of candidate units to be restored;

[0025] S42. Update the resistance value of the candidate unit set to be repaired according to the preset ideal recovery state attributes, and generate a temporary ecological resistance surface;

[0026] S43. Calculate the gain based on the temporary ecological resistance surface and the benchmark equivalent connectivity index, and generate a connectivity gain distribution map.

[0027] In one embodiment, S43 includes:

[0028] S431. Update the local voltage based on the temporary ecological resistance surface to generate the updated voltage distribution;

[0029] S432. Calculate the repaired equivalent connectivity index based on the updated voltage distribution, and generate the repaired equivalent connectivity index.

[0030] S433. Calculate the difference between the repaired equivalent connectivity index and the baseline equivalent connectivity index to generate a connectivity gain distribution map.

[0031] In one embodiment, S5 includes:

[0032] S51. Calculate the implementation cost index based on terrain slope data, road distance data, and land ownership data to generate the implementation cost distribution; the formula for calculating the implementation cost index is:

[0033]

[0034] in, For position The implementation cost index at the location For slope, For the maximum slope, Distance from the road For the maximum distance, For land ownership factors, , , These are the slope weighting coefficient, distance weighting coefficient, and ownership weighting coefficient, respectively.

[0035] S52. Calculate the gain-cost ratio based on the implementation cost distribution and connectivity gain distribution diagram, and generate a gain-cost ratio matrix; the formula for calculating the gain-cost ratio is:

[0036]

[0037] in, For position Gain-cost ratio at the location, This is the connectivity gain value. To implement the cost index, , To adjust the parameters;

[0038] S53. Perform spatial mapping based on the gain-cost ratio matrix to generate a priority repair potential distribution map.

[0039] In one embodiment, S6 includes:

[0040] S61. Based on the priority repair potential distribution map, perform threshold screening to generate a target repair pixel set;

[0041] S62. Perform density clustering based on the target restored pixel set to generate a clustered patch set;

[0042] S63. Prioritize the clustered patch sets according to their average gain-cost ratio to generate key areas for ecological restoration.

[0043] Secondly, this application also provides an intelligent identification device for ecological restoration areas based on multi-source data and model integration, comprising:

[0044] The feature extraction module is used to acquire multi-source remote sensing images, terrain data, and soil attribute data, and to extract features from the multi-source remote sensing images, terrain data, and soil attribute data to generate a multi-source data base and an ecological source set.

[0045] The ecological resistance surface construction module is used to construct and generate ecological resistance surfaces based on multi-source data.

[0046] The circuit theory simulation module is used to perform circuit theory simulations based on the ecological resistance surface and the ecological source area set, and generate a reference current density distribution map and a reference equivalent connectivity index.

[0047] The virtual remediation simulation module is used to perform virtual remediation simulations based on ecological resistance surfaces and multi-source data bases, and generate connectivity gain distribution maps.

[0048] The gain-cost ratio calculation module is used to calculate the gain-cost ratio based on the connectivity gain distribution map, terrain slope and road distance data, and generate a priority repair potential distribution map.

[0049] The spatial clustering analysis module is used to perform spatial clustering analysis based on the priority restoration potential distribution map to generate key areas for ecological restoration.

[0050] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the intelligent identification method for ecological restoration areas based on multi-source data and model integration as described in the first aspect.

[0051] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent identification method for ecological restoration areas based on multi-source data and model integration as described in the first aspect.

[0052] The aforementioned intelligent identification method for ecological restoration areas based on multi-source data and model integration extracts features from multi-source remote sensing images, topographic data, and soil attribute data to generate a multi-source data base and an ecological source area set. Based on this, an ecological resistance surface is constructed, and circuit theory simulations are performed using the ecological source area set to generate a benchmark current density distribution map and a benchmark equivalent connectivity index. Then, based on the ecological resistance surface and the multi-source data base, virtual restoration simulations are conducted to generate a connectivity gain distribution map. Next, the gain-cost ratio is calculated using topographic slope and road distance data to generate a priority restoration potential distribution map. Finally, spatial clustering analysis is used to identify key ecological restoration areas. This method can quantitatively evaluate the connectivity improvement potential of different areas after restoration and comprehensively optimize the selection based on engineering costs, thus effectively solving the problem that existing technologies cannot balance the expected benefits and implementation costs of ecological restoration, achieving intelligent, accurate, and economical identification of ecological restoration areas. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 A flowchart illustrating an intelligent identification method for ecological restoration areas based on multi-source data and model integration provided by this invention;

[0055] Figure 2A flowchart illustrating a method for generating a connectivity gain distribution map in an optional embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of the structure of an intelligent identification device for ecological restoration areas based on multi-source data and model integration, provided by the present invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] In one embodiment, such as Figure 1 As shown, an intelligent identification method for ecological restoration areas based on multi-source data and model integration is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps S1 to S6:

[0059] S1. Acquire multi-source remote sensing images, topographic data, and soil attribute data, and extract features from the multi-source remote sensing images, topographic data, and soil attribute data to generate a multi-source data base and an ecological source set.

[0060] Optionally, multi-source remote sensing images are acquired through a satellite remote sensing platform, topographic data is acquired through a Digital Elevation Model (DEM) database, and soil attribute data is acquired through a soil survey database. Radiometric and geometric correction preprocessing is sequentially performed on the multi-source remote sensing images. A pixel-level fusion algorithm is used to fuse remote sensing images of different resolutions. Then, features such as vegetation cover, Normalized Difference Vegetation Index (NDVI), and surface albedo are extracted using an object-oriented classification method. For the topographic data, a window analysis algorithm is used to extract features such as slope, aspect, and topographic relief. For the soil attribute data, an interpolation algorithm is used to supplement missing values ​​and extract features such as soil texture and organic matter content. All the extracted features are then integrated to generate a multi-source data base. Morphological Spatial Pattern Analysis (MSPA) was used to analyze the land use type characteristics in the multi-source data base. First, the land use types were reclassified to retain ecological land types such as forest land and grassland. Then, core patches were identified through MSPA. Patches with an area greater than a set threshold and a connectivity index that met the standard were selected from the core patches to form an ecological source area set.

[0061] S2. Construct an ecological resistance surface based on multi-source data to generate an ecological resistance surface.

[0062] Optionally, land use type, vegetation cover, slope, and soil texture are selected as ecological resistance factors from a multi-source data base. The Analytic Hierarchy Process (AHP) is used to determine the weight of each resistance factor. Specifically, a resistance factor judgment matrix is ​​first constructed, and the rationality of the judgment matrix is ​​verified by consistency test. Then, the weight value of each factor is calculated. The eigenvalues ​​of each resistance factor are transformed to the resistance value range of [1, 100] using the min-max standardization method, where the resistance value is lower in areas with higher ecological suitability. An ecological resistance surface is generated using a weighted superposition algorithm. Specifically, the resistance value of each raster cell is obtained by multiplying the standardized value of each resistance factor by its corresponding weight value and then summing the results. The formula is: resistance value = Σ (standardized value of resistance factor × weight of resistance factor). Finally, the ecological resistance surface of the entire raster is output.

[0063] S3. Based on the ecological resistance surface and the set of ecological source areas, perform circuit theory simulation calculations to generate a reference current density distribution map and a reference equivalent connectivity index.

[0064] Optionally, each ecological source site in the ecological source site set is considered as a node in the circuit, and the resistance value of the grid cell in the ecological resistance surface is considered as the resistance value between nodes, thus constructing an ecological circuit network model. Circuit theory simulation is performed using Circuitscape software, with the simulation parameters set to global connectivity analysis mode. The ecological source site set and ecological resistance surface data are input, and the software calculates the distribution of ecological flow in the network by solving the Laplace equation, outputting the current density value of each grid cell and generating a baseline current density distribution map. Simultaneously, the software calculates the baseline equivalent connectivity index, which is derived based on the equivalent resistance of the circuit network, using the following formula: ,in The equivalent resistance of the ecological circuit network was obtained by matrix operations using Circuitscape software. The smaller the value, the larger the equivalent resistance EC value of the circuit network, which represents better ecological connectivity.

[0065] S4. Based on the ecological resistance surface and multi-source data base, perform virtual restoration simulation to generate a connectivity gain distribution map.

[0066] Optionally, the ecological resistance surface is divided into regular grid restoration units of equal size. Based on the land use type characteristics in the multi-source data base, a virtual restoration scenario is set for each restoration unit. For example, high-resistance construction land, bare land, and other types within the restoration unit are converted into low-resistance ecological land types. The resistance value of the corresponding restoration unit is adjusted according to the virtual restoration scenario to generate the restored ecological resistance surface. Using the same Circuitscape software and simulation parameters, the ecological source area set and the restored ecological resistance surface are input to calculate the restored current density distribution map and the restored equivalent connectivity index. The connectivity gain of each restoration unit is calculated through grid operations, i.e., connectivity gain = restored current density - reference current density. The connectivity gain value is assigned to the corresponding restoration unit to generate a connectivity gain distribution map. For multiple restoration units, the above virtual restoration and gain calculation process can be executed one by one to finally obtain the global connectivity gain distribution.

[0067] S5. Calculate the gain-cost ratio based on the connectivity gain distribution map, terrain slope, and road distance data, and generate a priority repair potential distribution map.

[0068] Optionally, using the connectivity gain value in the connectivity gain distribution map as the numerator and the repair cost of each repair unit as the denominator, a Benefit-Cost Ratio (BCR) calculation model is constructed, with the formula: BCR = Connectivity Gain / Repair Cost. The repair cost calculation combines terrain slope and road distance data. First, the terrain slope and road distance data are transformed to a cost range of [1, 10] using the min-max normalization method, where a higher slope and a longer road distance result in a higher cost value. The cost weights of terrain slope and road distance are determined using expert scoring, and the repair cost of each repair unit is obtained through weighted summation, with the formula: Repair Cost = α × Slope Normalized Value + β × Road Distance Normalized Value, where α and β are the weights of terrain slope and road distance, respectively, and α + β = 1. The BCR value of each repair unit is assigned to a corresponding grid through raster operations, generating a priority repair potential distribution map. A higher BCR value indicates a higher cost-effectiveness for the repair of that unit.

[0069] S6. Based on the priority restoration potential distribution map, perform spatial cluster analysis to generate key areas for ecological restoration.

[0070] Optionally, the K-means clustering algorithm is used to perform spatial clustering analysis on the priority restoration potential distribution map, using the BCR value and spatial coordinates of each grid cell as clustering feature variables. First, the optimal number of clusters is determined using the elbow rule. After initializing the cluster centers, the Euclidean distance from each grid cell to each cluster center is calculated, and the grid cells are assigned to the cluster containing the nearest cluster center. The cluster centers are iteratively updated until the cluster center positions no longer change or the set number of iterations is reached, completing the clustering process. The cluster with the highest BCR value is extracted from the clustering results, and spatial autocorrelation analysis is used to verify the spatial clustering of this cluster. Regions with significant clustering and spatial continuity are retained, ultimately generating key areas for ecological restoration.

[0071] The aforementioned intelligent identification method for ecological restoration areas based on multi-source data and model integration constructs a comprehensive data foundation through multi-source data fusion and feature extraction. It then uses the analytic hierarchy process (AHP) and weighted superposition to build a scientific ecological resistance surface, leverages circuit theory to achieve accurate simulation of ecological connectivity, quantifies restoration gains through virtual restoration simulation, and constructs a gain-cost ratio model to balance benefits and costs through cost analysis. Finally, it accurately identifies key ecological restoration areas through cluster analysis. This method transforms ecological restoration area identification from static diagnosis to dynamic benefit assessment, effectively improving the scientific rigor and practicality of the identification results and providing precise spatial guidance for ecological restoration planning.

[0072] In one embodiment, S2 includes:

[0073] S21. Extract resistance parameters based on land use type, topographic factors, human disturbance intensity and vegetation cover in the multi-source data base, and generate a set of basic resistance coefficients and correction factors.

[0074] Optionally, four core resistance-related parameters—land use type, topographic factors, anthropogenic disturbance intensity, and vegetation cover—are precisely extracted from the multi-source data base. The land use type parameter directly uses the results obtained from object-oriented classification in the multi-source data base. Basic resistance coefficients are determined based on the ecological suitability of different land use types. Specifically, referring to the resistance assignment standards commonly used in landscape ecology, types with high ecological suitability, such as forest land and grassland, are assigned low resistance values, while types with low ecological suitability, such as construction land and bare land, are assigned high resistance values, forming a gridded basic resistance coefficient across the entire region. For topographic factors, slope is selected as the core indicator. Based on the topographic data in the multi-source data base, a 3×3 window slope calculation algorithm is used to extract slope values. Anthropogenic disturbance intensity is calculated using data such as road distribution and construction land distribution in the multi-source data base, employing a kernel density estimation method. Kernel density estimation traverses the entire grid using a sliding window, calculating the density value of anthropogenic disturbance elements within the window to characterize the disturbance intensity. Vegetation cover is calculated using the Normalized Difference Vegetation Index (NDVI), with the formula: Where NIR is the near-infrared reflectance and R is the red reflectance, both extracted from multi-source remote sensing image features. Slope, human disturbance intensity, and vegetation cover are transformed to the [0,1] interval using the min-max normalization method to generate slope factor, human disturbance intensity factor, and vegetation cover factor, which together constitute the set of correction factors.

[0075] S22. Perform multi-dimensional factor coupling calculations based on the basic resistance coefficient and the set of correction factors to generate the final ecological resistance value; the formula for calculating the final ecological resistance value is as follows:

[0076]

[0077] in, For position The final ecological resistance value at the location, This is the basic drag coefficient. The normalized slope factor. The normalized artificial interference intensity factor. The normalized vegetation cover factor. , , These are the topography sensitivity coefficient, disturbance sensitivity coefficient, and vegetation cover correction coefficient, respectively.

[0078] Optionally, each parameter is preprocessed and confirmed to ensure consistent raster resolution between the basic resistance coefficient and the set of correction factors, guaranteeing spatial compatibility of the coupled computation. The terrain sensitivity coefficient, disturbance sensitivity coefficient, and vegetation cover correction coefficient are determined using the analytic hierarchy process (AHP). Specifically, a judgment matrix is ​​constructed containing terrain slope, anthropogenic disturbance intensity, and vegetation cover. After verifying the matrix's rationality through consistency checks, the sensitivity or correction coefficient corresponding to each factor is calculated. , , All values ​​are positive numbers greater than 0, and their magnitude reflects the degree of influence of the corresponding factor on ecological resistance. The multidimensional factor coupling calculation is performed step-by-step according to the formula. First, each correction term is calculated, and then the normalized slope factor is... With terrain sensitivity coefficient Multiplying yields the terrain correction term. The normalized artificial interference intensity factor With interference sensitivity coefficient Multiplication yields the interference correction term. , and the (1-normalized vegetation cover factor) ) and vegetation cover correction factor Multiplying yields the vegetation correction term. Then, multi-dimensional factor coupling is achieved through multiplication, i.e., the basic drag coefficient. The position of each grid cell is obtained by multiplying it sequentially with the terrain correction, disturbance correction, and vegetation correction. The final ecological resistance value at the location .

[0079] S23. Based on the final ecological resistance value, perform rasterization mapping to generate an ecological resistance surface.

[0080] Optionally, a global raster mapping framework is established based on the raster coordinate system and resolution of the multi-source data base to ensure that the mapped ecological resistance surface is consistent with the spatial reference of the previous multi-source data base, ecological source area set, and other data. The final ecological resistance value is then... Substitute one by one Within the raster cells of the coordinate system, the final ecological resistance value is rasterized and assigned. After assigning values ​​to all raster cells across the entire area, an initial raster dataset is formed. The initial raster dataset undergoes integrity verification, removing any potentially null raster cells. Null areas are then filled using neighboring raster interpolation to ensure the continuity of the raster data. Finally, a raster file conforming to Geographic Information System (GIS) standards, i.e., the ecological resistance surface, is output.

[0081] In the above embodiments, by accurately extracting multi-dimensional resistance parameters, a coupled computational model including basic resistance and multi-factor correction is constructed. Combined with standardized processing and spatial matching rasterized mapping, the ecological resistance surface is constructed. The sensitivity and correction coefficients determined by the analytic hierarchy process improve the scientific nature of resistance value calculation. The coupled computation of multi-dimensional factors more comprehensively reflects the influencing factors of ecological connectivity, and the generated ecological resistance surface can accurately characterize the differences in the distribution of ecological resistance across the entire region.

[0082] In one embodiment, S3 includes:

[0083] S31. Based on the set of ecological source areas, perform node mapping and conductivity matrix transformation of ecological resistance surfaces to generate a conductivity matrix.

[0084] Optionally, node mapping is performed based on the grid coordinate system of the ecological resistance surface. All grid cells corresponding to each ecological source area in the ecological source area set are mapped to nodes in the circuit network, while ecological resistance surface grid cells outside the ecological source area set are retained as connection channels between nodes. The conductivity matrix transformation is based on the inverse relationship between conductivity and resistance in circuit theory. Conductivity is a physical quantity characterizing the ability of a material to conduct current, with the formula σ=1 / R, where σ is conductivity and R is the final ecological resistance value of the ecological resistance surface grid cell. A conductivity matrix is ​​constructed with the number of grid cells in the ecological resistance surface as the dimension. Each element in the matrix corresponds to the conductivity value of the connection channel between two nodes. Specifically, the conductivity values ​​of adjacent grid cells in the ecological resistance surface are assigned to the corresponding positions in the matrix, while the conductivity values ​​between non-adjacent nodes are set to 0. This ultimately generates a conductivity matrix that perfectly matches the spatial structure of the ecological resistance surface.

[0085] S32. Solve the linear equations based on the conductivity matrix to generate the voltage and current distributions.

[0086] Optionally, a system of linear equations describing the voltage distribution in the circuit network is constructed based on the conductivity matrix. This system of equations is derived from Kirchhoff's Current Law (KCL), the core principle of which is that the sum of the inflow currents at any node equals the sum of the outflow currents. The nodes in the ecological source set are divided into two categories: for example, some nodes are designated as voltage source nodes (assigned fixed voltage values), and the remaining nodes are designated as unknown voltage nodes. Substituting these into the conductivity matrix, a system of linear equations Ax=b is constructed, where A is the coefficient matrix derived from the conductivity matrix, x is the voltage value vector of the unknown voltage node, and b is a constant term vector calculated from the voltage values ​​of the voltage source nodes and the conductivity matrix. The Conjugate Gradient Method (CGM) is used to solve the linear equation system. This method iteratively generates a sequence of conjugate directions to gradually approximate the exact solution of the equation system. During the iteration process, the convergence is judged by the residual. When the residual is less than a set threshold, the iteration stops and the voltage values ​​of all nodes are output and a voltage distribution is generated. The current distribution is calculated based on Ohm's law, with the formula I = σ × ∇V, where I is the current and ∇V is the voltage gradient, which is calculated from the voltage difference between adjacent nodes. The current value is calculated channel by channel through grid operation to generate the current distribution.

[0087] S33. Based on the voltage and current distribution, extract connectivity indices to generate a reference current density distribution map and a reference equivalent connectivity index.

[0088] Optionally, the current density can be calculated based on the current distribution and the grid area of ​​the ecological resistance surface. The current density is the current intensity per unit area, and the formula is: ,in For current density, This represents the current value of the grid channel. Given the area of ​​a grid cell, grid operations are used to assign the current density value of each grid cell to its corresponding location, generating a baseline current density distribution map. The baseline equivalent connectivity index is extracted based on the equivalent resistance calculation of the circuit network. First, the total current of the ecological circuit network is calculated using voltage and current distributions; the total current is the sum of the outflow currents from all voltage source nodes. Then, the equivalent resistance is calculated according to Ohm's law. ,in The voltage difference between the voltage source nodes. The total current is given; the formula for calculating the baseline equivalent connectivity (EC) index is: The EC value is calculated using the above formula. This index directly quantifies the overall ecological connectivity level and ultimately outputs a benchmark current density distribution map and a benchmark equivalent connectivity index.

[0089] In the above embodiments, a circuit network model that fits the ecological pattern is constructed through node mapping and conductivity matrix transformation. Kirchhoff's laws and the conjugate gradient method are used to accurately solve for voltage and current distributions, ultimately extracting core indicators characterizing ecological connectivity. This embodiment transforms ecological connectivity simulation into quantifiable circuit network computation, improving the accuracy of ecological flow distribution simulation. The generated benchmark current density distribution map and benchmark equivalent connectivity index objectively reflect the current ecological connectivity characteristics.

[0090] In an optional embodiment, such as Figure 2 As shown, S4 includes:

[0091] S41. Extract obstacle points based on the ecological resistance surface to generate a set of candidate units to be restored.

[0092] Optionally, obstacle points refer to areas that significantly impede ecological flow transport and lead to a decline in ecological connectivity. Their extraction is based on the ecological resistance surface and the baseline current density distribution map. Specifically, obstacle points are first initially identified through grid threshold screening. Grid cells with resistance values ​​higher than a set threshold in the ecological resistance surface and grid cells with current density values ​​lower than a set threshold in the baseline current density distribution map are marked as potential obstacle points. Then, spatial connectivity analysis is used to eliminate isolated and scattered grids, retaining spatially continuous clusters of potential obstacle points with an area reaching a set size. These clusters are then divided into regular grid cells of equal size, with each grid cell serving as an independent candidate cell to be repaired. All candidate cells to be repaired are integrated to form a set of candidate cells to be repaired, while the spatial coordinates of each candidate cell are recorded to ensure that these spatial coordinates are consistent with the spatial reference of the ecological resistance surface.

[0093] S42. Update the resistance value of the candidate unit set to be repaired according to the preset ideal recovery state attributes, and generate a temporary ecological resistance surface.

[0094] Optionally, the preset ideal restoration state attribute is determined based on the ecological suitability principle of landscape ecology. The core is to convert the land use type of the candidate units to be restored into a state with optimal ecological connectivity. For example, high-resistance construction land and bare land are converted into low-resistance forest or grassland. Based on this ideal restoration state attribute, the commonly used ecological resistance assignment standard in landscape ecology is queried to determine the target resistance value corresponding to the ideal state. Based on the spatial coordinates of the candidate units to be restored, the grid position corresponding to each candidate unit is accurately located in the ecological resistance surface, and the original final ecological resistance value at the grid position is replaced with the target resistance value. After the resistance values ​​of all candidate units to be restored are updated, the resistance values ​​of other non-candidate units in the ecological resistance surface are retained unchanged to generate a temporary ecological resistance surface, ensuring that the grid resolution and spatial range of the temporary ecological resistance surface are completely consistent with the original ecological resistance surface.

[0095] S43. Calculate the gain based on the temporary ecological resistance surface and the benchmark equivalent connectivity index, and generate a connectivity gain distribution map.

[0096] Optionally, the temporary ecological resistance surface and the set of ecological source areas are used as input data. Operations such as node mapping, conductivity matrix transformation, and solving linear equations are performed sequentially to calculate the temporary voltage distribution and temporary current distribution, thereby extracting the temporary equivalent connectivity index and temporary current density distribution map. Connectivity gain calculation is divided into two dimensions: equivalent connectivity gain (equivalent connectivity gain = temporary equivalent connectivity index - baseline equivalent connectivity index) and current density gain (current density gain = temporary current density - baseline current density). Through raster operations, the gain values ​​of these two dimensions are assigned to corresponding raster cells. Current density gain is prioritized as the core characterization indicator. After rasterization, a connectivity gain distribution map is generated, which visually reflects the connectivity improvement effect of each candidate cell to be repaired and its surrounding area after virtual repair.

[0097] In the above embodiments, a set of candidate units to be repaired is formed by accurately extracting obstacle points, and a temporary ecological resistance surface is constructed by updating the resistance value based on the ideal recovery state. A mature circuit theory simulation method is used to calculate the connectivity gain and generate a distribution map. This embodiment achieves dynamic quantitative simulation of the repair effect, accurately identifies key areas with repair potential, and quantifies the connectivity improvement benefits of each area after repair. It effectively compensates for the shortcomings of traditional static diagnosis in assessing the expected repair benefits, and enhances the foresight and scientific rigor of ecological restoration area identification.

[0098] In one embodiment, S43 includes:

[0099] S431. Update the local voltage based on the temporary ecological resistance surface to generate the updated voltage distribution.

[0100] Optionally, the local voltage update is based on the spatial variation characteristics of the temporary ecological resistance surface. The update range is precisely defined as a local area consisting of the candidate unit to be restored and its adjacent grid cells. Since the resistance value of the temporary ecological resistance surface only changes in the candidate unit area, updating the voltage of this local area is sufficient to ensure accuracy, without requiring a full recalculation. Using the voltage value of the corresponding local area in the reference voltage distribution as the initial iteration value, a local linear equation system is constructed based on the conductivity matrix of the local area in the temporary ecological resistance surface and Kirchhoff's current law. The coefficient matrix of this equation system is derived from the conductivity matrix of the local area, and the constant term vector is calculated using the initial iteration voltage value. The conjugate gradient method is used to solve the local linear equation system. During the iteration process, convergence is judged by the residual magnitude. Iteration stops when the residual is less than a set threshold, yielding the updated voltage value of the local area. This updated voltage value is then spatially integrated with the voltage values ​​of the unchanged areas in the reference voltage distribution to generate the updated complete voltage distribution.

[0101] S432. Calculate the repaired equivalent connectivity index based on the updated voltage distribution, and generate the repaired equivalent connectivity index.

[0102] Optionally, the calculation of the equivalent connectivity index after repair follows the same core logic as the benchmark equivalent connectivity index to ensure the uniformity of the evaluation criteria and the comparability of the results. Using the updated voltage distribution as the core input data, the fixed voltage difference between voltage source nodes in the circuit network is first determined. This difference is consistent with the voltage difference set during the benchmark equivalent connectivity index calculation. The outflow current of each voltage source node is calculated based on Ohm's law, and the total current of the entire ecological circuit network (after repair) is obtained through grid traversal and summation. Based on the formula relating equivalent resistance to voltage and current: Calculate the equivalent resistance after repair, where U is the voltage difference between the voltage source nodes. This represents the total current after repair. Substituting the repaired equivalent resistance into the equivalent connectivity index calculation formula: The equivalent connectivity index after repair is obtained through numerical calculation.

[0103] S433. Calculate the difference between the repaired equivalent connectivity index and the baseline equivalent connectivity index to generate a connectivity gain distribution map.

[0104] Optionally, a connectivity gain difference calculation model is constructed, clearly defining the core calculation as the difference between the repaired equivalent connectivity index and the baseline equivalent connectivity index. The formula is: Connectivity Gain = Repaired Equivalent Connectivity Index - Baseline Equivalent Connectivity Index. Using the grid coordinate system corresponding to the updated voltage distribution as the baseline, it is ensured that the spatial positions of the grid cells corresponding to the two indices are perfectly matched, avoiding calculation errors caused by spatial misalignment. Grid operations are used to perform difference calculations cell by cell, assigning the connectivity gain value of each grid cell to the corresponding position. For grid cells not included in the local update range, the difference between their repaired equivalent connectivity index and the baseline equivalent connectivity index is minimal, and the connectivity gain value approaches 0, so a value of 0 can be directly assigned. After completing the difference calculation and assignment for all grid cells across the entire domain, a connectivity gain distribution map is generated. This connectivity gain distribution map can intuitively present the distribution characteristics of the improvement in ecological connectivity in different regions after virtual repair.

[0105] In the above embodiments, local voltage updates are used to accurately adapt to local changes in the temporary ecological resistance surface, improving computational efficiency while ensuring calculation accuracy. The equivalent connectivity index after restoration is calculated based on a unified standard, ensuring the objectivity and comparability of connectivity gain assessment. A connectivity gain distribution map is generated through unit-by-unit difference calculations. This embodiment achieves precise quantification of the benefits of improved ecological connectivity after virtual restoration, providing core benefit quantification data for gain-cost ratio calculation, further improving the dynamic evaluation mechanism for expected restoration benefits, effectively supporting the selection of high-cost-effective ecological restoration areas, and enhancing the scientific rigor and accuracy of ecological restoration area identification.

[0106] In one embodiment, S5 includes:

[0107] S51. Calculate the implementation cost index based on terrain slope data, road distance data, and land ownership data to generate the implementation cost distribution; the formula for calculating the implementation cost index is:

[0108]

[0109] in, For position The implementation cost index at the location For slope, For the maximum slope, Distance from the road For the maximum distance, For land ownership factors, , , These are the slope weight coefficient, distance weight coefficient, and ownership weight coefficient, respectively.

[0110] Optionally, the terrain slope data, road distance data, and land ownership data can be preprocessed to ensure that the raster resolution, spatial coordinate system, and connectivity gain distribution map of these three types of data are completely consistent. (Land ownership factor) To quantify the difficulty of restoration for different land ownership types, specific values ​​are assigned based on the nature of land ownership. For example, state-owned land is assigned a lower value, collectively owned land a medium value, and privately owned land a higher value. The assignment standards refer to general specifications in the field of land consolidation engineering. The slope weight coefficient δ, distance weight coefficient ε, and ownership weight coefficient ζ are determined using the analytic hierarchy process (AHP). A judgment matrix including terrain slope, road distance, and land ownership is constructed. After consistency verification to confirm its rationality, the values ​​of each coefficient are calculated. The cost index calculation is performed step-by-step according to the formula, starting with calculating the normalized slope value. ( The current grid slope, (Maximum slope across the entire area, extracted using raster statistics) and distance normalized value. ( This represents the current grid distance from the road. The maximum distance to the road across the entire region is extracted using the grid statistics function. Each cost component is then multiplied by its corresponding weight coefficient, and finally summed to obtain the location. Implementation cost index After performing calculations cell by cell through raster operations, a global implementation cost distribution is generated.

[0111] S52. Calculate the gain-cost ratio based on the implementation cost distribution and connectivity gain distribution diagram, and generate a gain-cost ratio matrix; the formula for calculating the gain-cost ratio is:

[0112]

[0113] in, For position Gain-cost ratio at the location, This is the connectivity gain value. To implement the cost index, , To adjust the parameters.

[0114] Optionally, spatial matching verification is first performed on the implementation cost distribution and connectivity gain distribution maps. This is achieved by checking geographic coordinate alignment and raster resolution consistency to ensure a one-to-one spatial mapping between the raster cells of both maps, thus mitigating computational errors caused by data misalignment at the source. The gain-cost ratio calculation operates on a raster-by-raster-cell basis, calling the raster calculator to load the core parameters corresponding to the two types of data, namely the implementation cost index in the implementation cost distribution. (Right now The connection gain value in the connection gain distribution diagram Substitute the values ​​into the preset gain-cost ratio calculation formula. The adjustment parameters are as follows: , The preset minimum positive number is used to prevent the occurrence of connectivity gain values. =0 or implementation cost index The calculation error is categorized as 0, ensuring the stability of the calculation process. The values ​​of both are determined based on conventional experience standards in the field of ecological restoration benefit assessment and do not affect the final cost-effectiveness ranking result. The calculation process is executed step-by-step: "numerator calculation - denominator calculation - ratio solution." First, the values ​​of each unit are obtained through raster calculations. and Then perform a division operation on the two to obtain the position. Gain-cost ratio at the location ; All grid cells The values ​​are arranged in order according to their spatial coordinates to generate a gain-cost ratio matrix that perfectly matches the entire grid range. Each element in the matrix corresponds precisely to the cost-effectiveness evaluation result of a grid cell.

[0115] S53. Perform spatial mapping based on the gain-cost ratio matrix to generate a priority repair potential distribution map.

[0116] Optionally, a spatial mapping framework is established based on the raster coordinate system and resolution of the implementation cost distribution. This ensures that the mapped data maintains spatial consistency with the spatial references of core data such as the previous multi-source data base and ecological resistance surface, laying the foundation for spatial coherence in further threshold screening and density clustering. Each element in the gain-cost ratio matrix corresponds to... Value, substitute it into its corresponding value one by one. Within the raster cells of the coordinates, the gain-cost ratio data is accurately assigned from the matrix to the spatial raster. Integrity checks are performed on the initial raster dataset after assignment, identifying potentially null cells through raster traversal and using neighboring raster cells. Mean interpolation is used to fill in missing values, ensuring the continuity and integrity of the raster data across the entire area. The final output is a raster file conforming to Geographic Information System standards, namely a priority repair potential distribution map. The value directly reflects the cost-effectiveness of the corresponding area's restoration; a high value indicates a high cost-effectiveness ratio. The value range corresponds to a high repair priority.

[0117] In the above embodiments, the implementation cost is accurately calculated by integrating multi-dimensional cost factors, and a scientific gain-cost ratio model is constructed by combining connectivity gain. A priority restoration potential distribution map is then generated through spatial mapping. This embodiment achieves a quantitative balance between ecological restoration benefits and implementation costs, effectively avoiding the problem of low cost-effectiveness in restoration areas in traditional technologies. The generated priority restoration potential distribution map can accurately pinpoint high-cost-effectiveness restoration areas.

[0118] In one embodiment, S6 includes:

[0119] S61. Based on the priority repair potential distribution map, threshold screening is performed to generate a target repair pixel set.

[0120] Optionally, the core of threshold screening is to accurately define the range of pixels with high restoration potential. First, perform raster statistical analysis on the distribution map of priority restoration potential and extract the global gain-cost ratio. Based on the distribution characteristics of the data, the natural breakpoint method was used to determine the screening threshold. This method, based on the natural grouping patterns of the data itself, maximizes intra-group similarity and inter-group differences, conforming to the spatial distribution characteristics of ecological restoration potential. Using the determined screening threshold as a standard, raster operations were used to select priority restoration potential distribution maps. All raster cells with values ​​higher than the threshold are marked as target restoration cells. Spatial integrity is checked on the marked cells, and scattered isolated cells caused by data errors are removed. Cells with a certain degree of spatial continuity are retained. All target restoration cells that meet the conditions are integrated according to their spatial coordinates to generate a target restoration cell set, ensuring that the cells in the set have a high restoration cost-effectiveness.

[0121] S62. Perform density clustering based on the target repair pixel set to generate a clustered patch set.

[0122] Optionally, a density-based spatial clustering (DBSCAN) algorithm is used to perform cluster analysis on the target restoration pixel set. This algorithm does not require a preset number of clusters and can effectively identify spatially connected pixel clusters. First, the core parameters of the algorithm are set. The neighborhood radius is determined through neighborhood analysis to ensure coverage of adjacent target restoration pixels. The minimum number of points is set to ensure that the clustered patches have a certain size. Both values ​​are based on conventional empirical standards for geospatial clustering. Using the spatial coordinates of the target restoration pixels as clustering feature variables, all target restoration pixels are traversed to identify core pixels. The number of target restoration pixels in the neighborhood of a core pixel is greater than or equal to the minimum number of points. Adjacent target restoration pixels are integrated through neighborhood expansion of the core pixel to form initial clusters. Isolated pixels that cannot be covered by the expansion of the core pixel (i.e., noise points) are removed. The initial clusters are then merged and deduplicated to ensure that the internal pixel space of each cluster is continuous and the external boundary is clear, ultimately generating a set of clustered patches.

[0123] S63. Prioritize the clustered patch sets according to their average gain-cost ratio to generate key areas for ecological restoration.

[0124] Optionally, first extract all target restoration pixels corresponding to each cluster in the clustered patch set, and then use the raster statistics function to calculate the average gain-cost ratio of all pixels within each clustered patch, i.e., average gain-cost ratio = total pixels within the patch. The sum of values ​​divided by the number of pixels within a patch, the average gain-cost ratio, comprehensively reflects the cost-effectiveness of restoration for the entire clustered patch. Based on the average gain-cost ratio, a descending sorting algorithm is used to prioritize all clustered patches in the set. Clustered patches with higher average gain-cost ratios in the sorted results have higher restoration priority. Several top-ranked clustered patches are selected as candidate restoration areas, and their spatial continuity and integrity are verified to ensure their feasibility for practical engineering restoration. The verified candidate restoration areas are integrated to generate key ecological restoration areas, and the average gain-cost ratio and spatial extent information of each key area are output, providing precise spatial guidance for ecological restoration planning.

[0125] In the above embodiments, high-potential restoration pixels are identified through threshold screening using the natural breakpoint method, continuous restoration patches are formed using DBSCAN density clustering, and the optimal restoration area is selected based on the average gain-cost ratio. This embodiment achieves precise focusing from scattered high-potential pixels to concentrated key areas, effectively ensuring the cost-effectiveness and engineering feasibility of key ecological restoration areas, and avoiding the problems of scattered restoration areas and low cost-effectiveness in traditional identification.

[0126] The aforementioned intelligent identification method for ecological restoration areas based on multi-source data and model integration constructs a multi-source data base and an ecological source area set through feature extraction, and generates an ecological resistance surface by coupling calculation with multi-dimensional resistance factors. It obtains benchmark connectivity indicators through circuit theory simulation, quantifies connectivity gain through virtual restoration simulation, and calculates the gain-cost ratio by integrating cost factors to obtain the distribution of priority restoration potential. Finally, it accurately identifies key ecological restoration areas through threshold screening, density clustering, and priority ranking. This technical solution achieves dynamic simulation and evaluation of restoration benefits, balancing the difficulty of engineering implementation with the benefits of ecological restoration, and effectively solves the technical problems of traditional static diagnosis's inability to quantify expected restoration benefits and the low cost-effectiveness of restoration areas.

[0127] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0128] Based on the same inventive concept, this application also provides an apparatus for implementing the above-mentioned intelligent identification method for ecological restoration areas based on multi-source data and model integration. The solution provided by this apparatus is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more embodiments of the intelligent identification apparatus for ecological restoration areas based on multi-source data and model integration provided below can be found in the limitations of the intelligent identification method for ecological restoration areas based on multi-source data and model integration described above, and will not be repeated here.

[0129] In one exemplary embodiment, such as Figure 3 As shown, an intelligent identification device 10 for ecological restoration areas based on multi-source data and model integration is provided, comprising:

[0130] The feature extraction module 11 is used to acquire multi-source remote sensing images, terrain data and soil attribute data, and to extract features from the multi-source remote sensing images, terrain data and soil attribute data to generate a multi-source data base and an ecological source set.

[0131] Ecological resistance surface construction module 12 is used to construct ecological resistance surfaces based on multi-source data base and generate ecological resistance surfaces.

[0132] The circuit theory simulation module 13 is used to perform circuit theory simulation calculations based on the ecological resistance surface and the ecological source area set, and generate a reference current density distribution map and a reference equivalent connectivity index.

[0133] Virtual restoration simulation module 14 is used to perform virtual restoration simulation based on ecological resistance surface and multi-source data base, and generate connectivity gain distribution map;

[0134] The gain-cost ratio calculation module 15 is used to calculate the gain-cost ratio based on the connectivity gain distribution map, terrain slope and road distance data, and generate a priority repair potential distribution map.

[0135] The spatial clustering analysis module 16 is used to perform spatial clustering analysis based on the priority restoration potential distribution map to generate key areas for ecological restoration.

[0136] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the intelligent identification method for ecological restoration areas based on multi-source data and model integration as described above.

[0137] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent identification method for ecological restoration areas based on multi-source data and model integration as described above.

[0138] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and 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 modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0139] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for intelligent identification of ecological restoration areas based on multi-source data and model integration, characterized in that, The method includes: S1. Acquire multi-source remote sensing images, terrain data, and soil attribute data, and extract features from the multi-source remote sensing images, terrain data, and soil attribute data to generate a multi-source data base and an ecological source set; S2. Construct an ecological resistance surface based on the multi-source data base to generate an ecological resistance surface; S3. Perform circuit theory simulation calculations based on the ecological resistance surface and the ecological source area set to generate a reference current density distribution map and a reference equivalent connectivity index. S4. Perform virtual restoration simulation based on the ecological resistance surface and the multi-source data base to generate a connectivity gain distribution map; S5. Calculate the gain-cost ratio based on the connectivity gain distribution map, terrain slope, and road distance data, and generate a priority repair potential distribution map. S6. Perform spatial cluster analysis based on the priority restoration potential distribution map to generate key areas for ecological restoration.

2. The method according to claim 1, characterized in that, S2 includes: S21. Extract resistance parameters based on the land use type, topographic factors, human disturbance intensity and vegetation cover in the multi-source data base, and generate a set of basic resistance coefficients and correction factors. S22. Perform multi-dimensional factor coupling calculations based on the basic resistance coefficient and the set of correction factors to generate the final ecological resistance value; the calculation formula for the final ecological resistance value is as follows: in, For position The final ecological resistance value at the location, This is the basic drag coefficient. The normalized slope factor. The normalized artificial interference intensity factor. The normalized vegetation cover factor. , , These are the topography sensitivity coefficient, disturbance sensitivity coefficient, and vegetation cover correction coefficient, respectively. S23. Perform rasterization mapping based on the final ecological resistance value to generate the ecological resistance surface.

3. The method according to claim 2, characterized in that, S3 includes: S31. Based on the set of ecological source areas, perform node mapping and conductivity matrix transformation of the ecological resistance surface to generate a conductivity matrix; S32. Solve the linear equations based on the conductivity matrix to generate voltage and current distributions; S33. Based on the voltage distribution and the current distribution, the connectivity index is extracted to generate the reference current density distribution map and the reference equivalent connectivity index.

4. The method according to claim 1, characterized in that, S4 includes: S41. Based on the ecological resistance surface, extract obstacle points to generate a set of candidate units to be repaired; S42. Update the resistance value of the set of candidate units to be repaired according to the preset ideal recovery state attributes to generate a temporary ecological resistance surface; S43. Calculate the gain based on the temporary ecological resistance surface and the benchmark equivalent connectivity index to generate the connectivity gain distribution map.

5. The method according to claim 4, characterized in that, S43 includes: S431. Update the local voltage based on the temporary ecological resistance surface to generate the updated voltage distribution; S432. Calculate the repaired equivalent connectivity index based on the updated voltage distribution to generate the repaired equivalent connectivity index. S433. Calculate the difference between the repaired equivalent connectivity index and the benchmark equivalent connectivity index to generate the connectivity gain distribution map.

6. The method according to claim 1, characterized in that, S5 includes: S51. Calculate the implementation cost index based on terrain slope data, road distance data, and land ownership data to generate an implementation cost distribution; the calculation formula for the implementation cost index is: in, For position The implementation cost index at the location, For slope, For the maximum slope, Distance from the road For the maximum distance, For land ownership factors, , , These are the slope weighting coefficient, distance weighting coefficient, and ownership weighting coefficient, respectively. S52. Calculate the gain-cost ratio based on the implementation cost distribution and the connectivity gain distribution diagram to generate a gain-cost ratio matrix; the formula for calculating the gain-cost ratio is: in, For position Gain-cost ratio at the location, This is the connectivity gain value. To implement the cost index, , To adjust the parameters; S53. Perform spatial mapping based on the gain-cost ratio matrix to generate the priority repair potential distribution map.

7. The method according to claim 1, characterized in that, S6 includes: S61. Perform threshold screening based on the priority repair potential distribution map to generate a target repair cell set; S62. Perform density clustering based on the target repair pixel set to generate a clustered patch set; S63. Prioritize the clustered patch sets according to their average gain-cost ratio to generate the key ecological restoration areas.

8. An intelligent identification device for ecological restoration areas based on multi-source data and model integration, characterized in that, The device includes: The feature extraction module is used to acquire multi-source remote sensing images, terrain data and soil attribute data, and to extract features from the multi-source remote sensing images, terrain data and soil attribute data to generate a multi-source data base and an ecological source set. An ecological resistance surface construction module is used to construct an ecological resistance surface based on the multi-source data base and generate an ecological resistance surface. The circuit theory simulation module is used to perform circuit theory simulation calculations based on the ecological resistance surface and the ecological source area set, and generate a reference current density distribution map and a reference equivalent connectivity index. The virtual restoration simulation module is used to perform virtual restoration simulation based on the ecological resistance surface and the multi-source data base, and generate a connectivity gain distribution map. The gain-cost ratio calculation module is used to calculate the gain-cost ratio based on the connectivity gain distribution map, terrain slope and road distance data, and generate a priority repair potential distribution map. The spatial clustering analysis module is used to perform spatial clustering analysis based on the priority restoration potential distribution map to generate key areas for ecological restoration.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images