A GIS-based intelligent identification method for ecological benefit areas

Abstract

A GIS-based intelligent identification method for ecological benefit areas relates to the technical field of data feature identification. A three-dimensional regional digital twin model of a target area is constructed by means of GIS to obtain the ecological benefit area type label of each three-dimensional triangulated network model; each three-dimensional triangulated network model is divided into a fixed ecological benefit area and a changing ecological benefit area, and non-intersection type feature identification indicators, intersection type feature identification indicators and association type feature identification indicators in the changing ecological benefit area are extracted; a cross-topological map of the changing ecological benefit area is constructed to obtain the prediction results of the intersection type feature identification indicators and the association type feature identification indicators; the future evolution ecological benefit area type label of the changing ecological benefit area is obtained according to the prediction results of the non-association type feature identification indicators, the intersection type feature identification indicators and the association type feature identification indicators of the changing ecological benefit area, so as to ensure the accuracy and real-time performance of the ecological benefit area identification.

Description

Technical Field

[0001] The invention relates to the technical field of data feature recognition, and in particular to an intelligent recognition method for ecological benefit areas based on GIS. Background Art

[0002] Ecological benefit areas refer to areas with important ecological functions and ecological service functions in natural ecosystems, and play an important role in maintaining ecosystem stability, protecting biodiversity, regulating climate, and ensuring ecological security. Ecological benefit areas usually include areas with specific ecological and environmental conditions, rich biological resources and key ecological functions, and are important spatial units for biodiversity protection, ecological and environmental restoration, and ecosystem management.

[0003] The delineation and protection of ecological benefit zones are crucial to the protection of the ecological environment, the performance of ecosystem service functions, and the sustainable development of human society. How to effectively protect the ecological environment, maintain ecological balance, and promote biodiversity conservation and sustainable use through scientific delineation and management of ecological benefit zones is an issue that we urgently need to address. Summary of the invention

[0004] In order to solve the above technical problems, the object of the present invention is to provide a GIS-based intelligent identification method for ecological benefit areas, comprising the following steps:

[0005] Step s1: construct a three-dimensional regional digital twin model of the target area by means of GIS, and obtain the ecological benefit area type label of each three-dimensional triangulated network model according to the various types of feature identification index layers superimposed on each three-dimensional triangulated network model in the three-dimensional regional digital twin model;

[0006] Step s2: Divide each 3D triangulated network model in the current acquisition period into a fixed ecological benefit area and a changing ecological benefit area according to the historical ecological benefit area type label of each 3D triangulated network model, and extract the non-intersecting type feature identification index, the intersecting type feature identification index and the associated type feature identification index in the changing ecological benefit area;

[0007] Step s3: constructing a cross-topology map of the changing ecological benefit area according to the cross-type feature identification index of the changing ecological benefit area, obtaining the prediction result of the cross-type feature identification index according to the cross-topology map of the changing ecological benefit area, and obtaining the prediction result of the association type feature identification index at the same time;

[0008] Step s4: Obtain the future evolution ecological benefit area type label of the changing ecological benefit area according to the non-correlated type feature identification index, the cross type feature identification index prediction results and the correlated type feature identification index prediction results within the current acquisition period of the changing ecological benefit area.

[0009] Furthermore, the process of constructing a three-dimensional regional digital twin model of the target area by means of GIS includes:

[0010] The terrain features, digital elevation data and ecological environment data of the target area at the current moment are obtained through GIS means and the collection cycle is set. The plane triangulation model of the target area is constructed according to the terrain features. The elevations of each triangle vertex in the plane triangulation model are determined according to the digital elevation data to obtain a three-dimensional triangular model.

[0011] Classify the ecological and environmental data of the target area and extract various types of feature identification indicators. Then, construct various types of feature identification indicator layers based on the various types of feature identification indicators. Superimpose the various types of feature identification indicator layers on the three-dimensional triangulated network model to generate a three-dimensional regional digital twin model.

[0012] Furthermore, according to various types of feature identification index layers superimposed on each three-dimensional triangulated network model in the three-dimensional regional digital twin model, the process of obtaining the ecological benefit area type label of each three-dimensional triangulated network model includes:

[0013] Preset threshold intervals corresponding to various types of feature identification indicators corresponding to several types of ecological benefit areas, obtain various types of feature identification indicator layers superimposed on various three-dimensional triangulated network models in the three-dimensional regional digital twin model within the current acquisition period, extract the numerical values ​​corresponding to various types of feature identification indicators in the various types of feature identification indicator layers, match the numerical values ​​corresponding to various types of feature identification indicators of various three-dimensional triangulated network models with the threshold intervals corresponding to various types of feature identification indicators corresponding to various types of ecological benefit areas, obtain the feature similarity between each three-dimensional triangulated network model and each type of ecological benefit area, preset a feature similarity threshold, and if the feature similarity between the three-dimensional triangulated network model and the type of ecological benefit area is greater than or equal to the feature similarity threshold, add an ecological benefit area type label corresponding to the type of ecological benefit area to the three-dimensional triangulated network model;

[0014] If the feature similarities between the three-dimensional triangulated network model and each type of ecological benefit area are less than the feature similarity threshold, a non-ecological benefit area label is added to the three-dimensional triangulated network model.

[0015] The calculation formula for obtaining the characteristic similarity between the three-dimensional triangulated network model and the ecological benefit area is:

[0016]

[0017] Vds(G DK , G LM )=δ(1-JKC(G DK , G LM ));

[0018]

[0019] Among them, G DK Represents various types of feature recognition indicators of 3D triangulated network models, G LM Represents the identification index of various types of ecological benefit areas, JKC (G DK , G LM ) represents the non-numerical attribute similarity of various types of feature identification indicators between the 3D triangulated network model and the ecological benefit area, SUM(G DK , G LM ) represents the total number of types of characteristic identification indicators of the three-dimensional triangulated network model and ecological benefit area, SAMESUNM(G DK , G LM )) represents the number of the same type of feature identification indicators between the 3D triangulated network model and the ecological benefit area, Vds(G DK , G LM ) represents the Euclidean distance between the 3D triangulated network model and the ecological benefit area, P DKi represents the value corresponding to the i-th type feature recognition index of the 3D triangulated network model, P LMi represents the value corresponding to the characteristic identification index of the i-th type of ecological benefit area, dist(P DKi , P LMi ) represents the Euclidean distance corresponding to the numerical difference between the 3D triangulated network model and the ecological benefit area, Rds(DK, LM) represents the overall similarity between the 3D triangulated network model and the ecological benefit area, N is the number of the same type of feature identification indicators between the 3D triangulated network model and the ecological benefit area, δ, θ, σ represent conversion coefficients, and a1, a2 represent weight factors.

[0020] Furthermore, the process of dividing each 3D triangulated network model in the current acquisition period into a fixed ecological benefit area and a variable ecological benefit area according to the historical ecological benefit area type label of each 3D triangulated network model includes:

[0021] Obtain the historical ecological benefit area type labels of each 3D triangulated network model in several historical acquisition cycles in the 3D regional digital twin model, and determine whether the ecological benefit area type labels of each 3D triangulated network model in the current acquisition cycle are consistent with the historical ecological benefit area type labels of each 3D triangulated network model. If they are consistent, the 3D triangulated network model is marked as a fixed ecological benefit area; if not, the 3D triangulated network model is marked as a changing ecological benefit area.

[0022] Furthermore, the process of extracting non-intersection type feature identification indicators, intersection type feature identification indicators and association type feature identification indicators in the changed ecological benefit area includes:

[0023] Obtain historical ecological benefit area type labels that are inconsistent with ecological benefit area type labels in the current collection period in several historical collection periods of the changed ecological benefit area, extract each type of characteristic identification index contained in the ecological benefit area type label and the corresponding numerical range of each type of characteristic identification index, and at the same time extract each type of characteristic identification index contained in the historical ecological benefit area type label and the corresponding numerical range of each type of characteristic identification index;

[0024] The type feature identification indicators included in both the ecological benefit area type label and the historical ecological benefit area type label are marked as non-crossing type feature identification indicators, and the type feature identification indicators not included in both the ecological benefit area type label and the historical ecological benefit area type label are marked as crossing type feature identification indicators;

[0025] If the numerical range corresponding to the non-intersecting type feature identification indicator contained in the ecological benefit area type label is consistent with the numerical range corresponding to the non-intersecting type feature identification indicator contained in the historical ecological benefit area type label, the non-intersecting type feature identification indicator is marked as a non-associated type feature identification indicator; if the numerical range corresponding to the non-intersecting type feature identification indicator contained in the ecological benefit area type label is inconsistent with the numerical range corresponding to the non-intersecting type feature identification indicator contained in the historical ecological benefit area type label, the non-intersecting type feature identification indicator is marked as an associated type feature identification indicator.

[0026] Furthermore, a cross-type feature identification index of the changed ecological benefit area is used to construct a cross-topology map of the changed ecological benefit area. The process of obtaining the prediction result of the cross-type feature identification index according to the cross-topology map of the changed ecological benefit area includes:

[0027] Obtain cross-type feature identification indicators for each change ecological benefit area, and construct a topological connection relationship between the change ecological benefit areas containing the same cross-type feature identification indicators;

[0028] Constructing a cross-topological map of the changing ecological benefit areas, taking each changing ecological benefit area as a node of the cross-topological map of the changing ecological benefit areas, and taking the topological connection relationship between each changing ecological benefit area as the connection relationship between the nodes;

[0029] Construct a temporal convolutional neural network, obtain the numerical time series sequence corresponding to the cross-type feature identification indicators of each changing ecological benefit area in several historical collection cycles as training data to train the temporal convolutional neural network, input the numerical time series sequence corresponding to the cross-type feature identification indicators of each changing ecological benefit area in the current collection cycle into the trained temporal convolutional neural network, and obtain the change feature trend corresponding to the cross-type feature identification indicators of each changing ecological benefit area;

[0030] The graph attention network is used to learn the cross-topology graph of the changing ecological benefit area, and the change feature trend corresponding to the cross-type feature recognition index of each changing ecological benefit area is imported into the graph attention network. The attention weight between each node is obtained through the attention mechanism, and the spatiotemporal feature sequence of the cross-type feature recognition index of each node is generated using the neighbor aggregation mechanism.

[0031] The changing ecological benefit areas are taken as topological connection relationships to construct the edge structure of the cross-topology graph of the changing ecological benefit areas. The cross-topology graph of the changing ecological benefit areas is learned and represented by the graph attention network. The point structure of the cross-topology graph of the changing ecological benefit areas between nodes j and k in the cross-topology graph of the changing ecological benefit areas is noted. According to the attention value f between each changing ecological benefit area jk for:

[0032] ˊˊ

[0033] H jk =ELU(FULLY((we j ))×FULLY((we k )));

[0034] f jk (1) =Tanh(A 1 H jk +b 1 );

[0035] f jk (2) =Tanh(A 2 f jk (1) +b 2 ); ...

[0036] f jk (x+1) =Tanh(A (x+1) f jk (x) +b (x+1) );

[0037] Among them, e j 、e k represents the initial feature vector of node j and node k, w represents the feature transformation parameter matrix, ELU and Tanh represent activation functions, FULLY represents the fully connected layer of the graph attention network, and A (x+1) and b (x+1) Respectively represent the weight matrix and bias corresponding to the x+1th iteration, f jk (x+1) represents the attention value between node j and node k after the x+1th iteration;

[0038] According to the attention value f between node j and node k jk The attention weight R is determined by the ratio of the sum of the attention values ​​between node j and each other node connected to node j. jk , the formula for updating the characteristics of the changing ecological benefit area through the neighbor aggregation mechanism is:

[0039]

[0040] Among them, e j * represents the final updated representation of node j, ReLu represents the activation function, k represents the number of adjacent nodes of node j, and u j represents the set of nodes connected to node j;

[0041] Based on the RBF neural network, a cross-type feature recognition index prediction model is constructed. The historical ecological environment data of several historical collection periods in the target area are obtained as training data to train the cross-type feature recognition index prediction model. The spatiotemporal feature sequence of the cross-type feature recognition index of each node and the numerical time series sequence corresponding to the cross-type feature recognition index are input into the trained cross-type feature recognition index prediction model. According to the cross-type feature recognition index prediction model, the predicted average values ​​corresponding to the cross-type feature recognition indicators of each changed ecological benefit area are output simultaneously.

[0042] Furthermore, the process of obtaining the prediction result of the association type feature recognition index includes:

[0043] Based on the long short-term memory network, an association type feature recognition index prediction model is constructed. The numerical time series sequence corresponding to the association type feature recognition indicators of each changing ecological benefit area in several historical collection cycles is used as training data to train the association type feature recognition index prediction model. The numerical time series sequence corresponding to the association type feature recognition indicators of each changing ecological benefit area in the current collection cycle is input into the trained association type feature recognition index prediction model to obtain the predicted average value corresponding to the association type feature recognition indicators of each changing ecological benefit area.

[0044] Furthermore, the process of obtaining the type label of the future evolving ecological benefit area of ​​the changing ecological benefit area according to the prediction results of the non-correlated type feature identification index, the cross-type feature identification index and the related type feature identification index in the current collection period of the changing ecological benefit area includes:

[0045] Obtain the average value corresponding to the non-associated type feature identification index of the changing ecological benefit area within the current collection period, match the predicted average value corresponding to the associated type feature identification index of the changing ecological benefit area, the predicted average value corresponding to the cross type feature identification index and the average value corresponding to the non-associated type feature identification index with the threshold interval corresponding to each type of feature identification index of each type of ecological benefit area, obtain the feature similarity between the changing ecological benefit area and each type of ecological benefit area, obtain the type of ecological benefit area whose feature similarity with the changing ecological benefit area is greater than or equal to the feature similarity threshold, and add the future evolution ecological benefit area type label corresponding to the type of ecological benefit area to the changing ecological benefit area.

[0046] Compared with the prior art, the beneficial effect of the present invention is that the functions and values ​​of ecological benefit areas may be changed due to environmental changes, resulting in the original ecological benefit areas no longer having their typical ecological services and functions, or even no longer being recognized as ecological benefit areas. Environmental change factors include climate change, land use change, human activity impact, etc. These factors may have an impact on ecological benefit areas, thereby changing their ecological characteristics and functions. Therefore, on the basis of identifying the ecological benefit areas in the current target area, the present invention takes into account the impact of some environmental changes on the ecological benefit areas, divides the ecological benefit areas into fixed ecological benefit areas and changing ecological benefit areas, extracts cross-type feature identification indicators in the changing ecological benefit areas, and obtains the future evolution ecological benefit area type label of the changing ecological benefit area according to the prediction results of the cross-type feature identification indicators and the prediction results of the associated type feature identification indicators of the changing ecological benefit areas, thereby ensuring the accuracy and real-time nature of the identification of ecological benefit areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a schematic diagram of a GIS-based intelligent identification method for ecological benefit areas in an embodiment of the present application. DETAILED DESCRIPTION

[0048] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0049] like Figure 1 As shown, a GIS-based intelligent identification method for ecological benefit areas includes the following steps:

[0050] Step s1: construct a three-dimensional regional digital twin model of the target area by means of GIS, and obtain the ecological benefit area type label of each three-dimensional triangulated network model according to the various types of feature identification index layers superimposed on each three-dimensional triangulated network model in the three-dimensional regional digital twin model;

[0051] Step s2: Divide each 3D triangulated network model in the current acquisition period into a fixed ecological benefit area and a changing ecological benefit area according to the historical ecological benefit area type label of each 3D triangulated network model, and extract the non-intersecting type feature identification index, the intersecting type feature identification index and the associated type feature identification index in the changing ecological benefit area;

[0052] Step s3: constructing a cross-topology map of the changing ecological benefit area according to the cross-type feature identification index of the changing ecological benefit area, obtaining the prediction result of the cross-type feature identification index according to the cross-topology map of the changing ecological benefit area, and obtaining the prediction result of the association type feature identification index at the same time;

[0053] Step s4: Obtain the future evolution ecological benefit area type label of the changing ecological benefit area according to the non-correlated type feature identification index, the cross type feature identification index prediction results and the correlated type feature identification index prediction results within the current acquisition period of the changing ecological benefit area.

[0054] It should be further explained that, in the specific implementation process, the process of constructing a three-dimensional regional digital twin model of the target area by means of GIS includes:

[0055] The terrain features, digital elevation data and ecological environment data of the target area at the current moment are obtained through GIS means and the collection cycle is set. The plane triangulation model of the target area is constructed according to the terrain features. The elevations of each triangle vertex in the plane triangulation model are determined according to the digital elevation data to obtain a three-dimensional triangular model.

[0056] Classify the ecological and environmental data of the target area and extract various types of feature identification indicators. Then, construct various types of feature identification indicator layers based on the various types of feature identification indicators. Superimpose the various types of feature identification indicator layers on the three-dimensional triangulated network model to generate a three-dimensional regional digital twin model.

[0057] It should be further explained that various types of feature identification indicators include geographical identification indicators, vegetation identification indicators, climate identification indicators, water resource identification indicators and biological identification indicators;

[0058] The geographic identification indicators include:

[0059] Land coverage and land use type: remote sensing image data and geographic data collected by GIS platform are used to obtain land coverage and land use type information, including cultivated land, forest land, grassland, etc., to provide basic data for evaluating ecosystem types and functions;

[0060] Vegetation identification indicators include:

[0061] Vegetation type and vegetation index (NDVI), which is calculated from remote sensing data and reflects the growth and coverage of vegetation, providing important information for assessing the vegetation status and productivity of ecosystems;

[0062] Climate identification indicators include: various climate element data including temperature, precipitation, wind speed, etc., which are used to evaluate the regional climate conditions and climate regulation functions.

[0063] Water resource identification indicators include:

[0064] Water body types and coverage, including data on rivers, lakes, reservoirs, etc., to assess water resource distribution and water resource protection capabilities;

[0065] Biometric indicators include: Biodiversity Index, which is used to assess the biodiversity status and biodiversity maintenance capacity of a region.

[0066] It should be further explained that, in the specific implementation process, according to the various types of feature identification index layers superimposed on each 3D triangulated network model in the 3D regional digital twin model, the process of obtaining the ecological benefit area type label of each 3D triangulated network model includes:

[0067] Preset threshold intervals corresponding to various types of feature identification indicators corresponding to several types of ecological benefit areas, obtain various types of feature identification indicator layers superimposed on various three-dimensional triangulated network models in the three-dimensional regional digital twin model within the current acquisition period, extract the numerical values ​​corresponding to various types of feature identification indicators in the various types of feature identification indicator layers, match the numerical values ​​corresponding to various types of feature identification indicators of various three-dimensional triangulated network models with the threshold intervals corresponding to various types of feature identification indicators corresponding to various types of ecological benefit areas, obtain the feature similarity between each three-dimensional triangulated network model and each type of ecological benefit area, preset a feature similarity threshold, and if the feature similarity between the three-dimensional triangulated network model and the type of ecological benefit area is greater than or equal to the feature similarity threshold, add an ecological benefit area type label corresponding to the type of ecological benefit area to the three-dimensional triangulated network model;

[0068] If the feature similarities between the three-dimensional triangulated network model and each type of ecological benefit area are less than the feature similarity threshold, a non-ecological benefit area label is added to the three-dimensional triangulated network model.

[0069] It should be further explained that, in the specific implementation process, the types of ecological benefit areas include:

[0070] Wetland ecological benefit areas: including swamps, wetlands, rivers, lakes and other water ecosystems. Wetland ecological benefit areas can provide important ecological services such as water resource regulation, flood control, and water purification.

[0071] Forest ecological benefit areas: including different types of forest ecosystems such as natural forests and artificial forests. Forest ecological benefit areas have ecological functions such as maintaining soil, purifying air, and regulating climate.

[0072] Grassland ecological benefit area: an ecosystem covered by herbaceous vegetation such as grassland and meadow. Grassland ecological benefit areas can provide ecological services such as forage production, soil conservation, and ecotourism.

[0073] Coastal ecological benefit areas: including coastal ecosystems such as coastal zones, beaches, and coral reefs. Coastal ecological benefit areas are of great significance to coastal protection and marine biodiversity conservation.

[0074] Urban green space ecological benefit area: green spaces such as parks, green spaces, greenways, etc. in the city. Urban green space ecological benefit areas help improve the quality of the urban environment, regulate the urban climate, and enhance the quality of life of residents.

[0075] Farmland ecological benefit area: including agricultural production systems such as cultivated land, forest land, and farmland. Farmland ecological benefit area plays a role in soil conservation, agricultural production, and ecological planting.

[0076] Mountain ecological benefit areas: including mountain ecosystems such as high mountains and hills. Mountain ecological benefit areas play an important role in water conservation, soil conservation, and ecological protection.

[0077] The calculation formula for obtaining the characteristic similarity between the three-dimensional triangulated network model and the ecological benefit area is:

[0078]

[0079] Vds(G DK , G LM )=δ(1-JKC(G DK , G LM ));

[0080]

[0081] Among them, G DK Represents various types of feature recognition indicators of 3D triangulated network models, G LM Represents the identification index of various types of ecological benefit areas, JKC (G DK , G LM) represents the non-numerical attribute similarity of various types of feature identification indicators between the 3D triangulated network model and the ecological benefit area, SUM(G DK , G LM ) represents the total number of types of characteristic identification indicators of the three-dimensional triangulated network model and ecological benefit area, SAMESUNM(G DK , G LM )) represents the number of the same type of feature identification indicators between the 3D triangulated network model and the ecological benefit area, Vds(G DK , G LM ) represents the Euclidean distance between the 3D triangulated network model and the ecological benefit area, P DKi represents the value corresponding to the i-th type feature recognition index of the 3D triangulated network model, P LMi represents the value corresponding to the characteristic identification index of the i-th type of ecological benefit area, dist(P DKi , P LMi ) represents the Euclidean distance corresponding to the numerical difference between the 3D triangulated network model and the ecological benefit area, Rds(DK, LM) represents the overall similarity between the 3D triangulated network model and the ecological benefit area, N is the number of the same type of feature identification indicators between the 3D triangulated network model and the ecological benefit area, δ, θ, σ represent conversion coefficients, and a1, a2 represent weight factors.

[0082] It should be further explained that, in the specific implementation process, the process of dividing each 3D triangulated network model in the current acquisition period into a fixed ecological benefit area and a changing ecological benefit area according to the historical ecological benefit area type label of each 3D triangulated network model includes:

[0083] Obtain the historical ecological benefit area type labels of each 3D triangulated network model in several historical acquisition cycles in the 3D regional digital twin model, and determine whether the ecological benefit area type labels of each 3D triangulated network model in the current acquisition cycle are consistent with the historical ecological benefit area type labels of each 3D triangulated network model. If they are consistent, the 3D triangulated network model is marked as a fixed ecological benefit area; if not, the 3D triangulated network model is marked as a changing ecological benefit area.

[0084] It should be further explained that, in the specific implementation process, the process of extracting non-intersection type feature identification indicators, intersection type feature identification indicators and association type feature identification indicators in the changed ecological benefit area includes:

[0085] Obtain historical ecological benefit area type labels that are inconsistent with ecological benefit area type labels in the current collection period in several historical collection periods of the changed ecological benefit area, extract each type of characteristic identification index contained in the ecological benefit area type label and the corresponding numerical range of each type of characteristic identification index, and at the same time extract each type of characteristic identification index contained in the historical ecological benefit area type label and the corresponding numerical range of each type of characteristic identification index;

[0086] The type feature identification indicators included in both the ecological benefit area type label and the historical ecological benefit area type label are marked as non-crossing type feature identification indicators, and the type feature identification indicators not included in both the ecological benefit area type label and the historical ecological benefit area type label are marked as crossing type feature identification indicators;

[0087] If the numerical range corresponding to the non-intersecting type feature identification indicator contained in the ecological benefit area type label is consistent with the numerical range corresponding to the non-intersecting type feature identification indicator contained in the historical ecological benefit area type label, the non-intersecting type feature identification indicator is marked as a non-associated type feature identification indicator; if the numerical range corresponding to the non-intersecting type feature identification indicator contained in the ecological benefit area type label is inconsistent with the numerical range corresponding to the non-intersecting type feature identification indicator contained in the historical ecological benefit area type label, the non-intersecting type feature identification indicator is marked as an associated type feature identification indicator.

[0088] It should be further explained that, in the specific implementation process, the cross-type feature identification indicators include:

[0089] Geographic identification indicators, such as farmland reclamation and desertification in a certain ecological benefit area, which in turn affects the ecological processes such as vegetation growth and water resource allocation in other ecological benefit areas, reducing the original ecological services and functions;

[0090] Vegetation identification indicators, such as the invasion of vegetation in a certain ecological benefit area into vegetation in other ecological benefit areas, thereby affecting the ecological processes in other ecological benefit areas and reducing the original ecological services and functions;

[0091] Climate identification indicators: Climate is one of the important factors affecting ecological benefit areas. Climate elements such as temperature, precipitation, and humidity directly affect ecological processes such as vegetation growth and water resource allocation. Climate change will affect the stability and health of the ecosystem, and thus affect the ecological services and functions of the ecological benefit areas.

[0092] Water resource identification indicators, for example, flooding of a river in a certain ecological benefit zone will cause a sharp rise in the water level of water areas, wetlands and other ecosystems in other ecological benefit zones, causing serious threats to the living space of wetland vegetation and animals. Some wetland vegetation may be destroyed, and some wetland animals may die because they cannot avoid floods, affecting the stability and biodiversity of wetland ecosystems in other ecological benefit zones;

[0093] Biological identification indicators; for example, there are rich species populations in the ecological beneficiary areas with rich biodiversity. Some of these species may migrate to other ecological beneficiary areas through migration and diffusion, thereby affecting the species diversity of the latter. This species migration and diffusion can bring new species and may also affect the survival and reproduction of the original species.

[0094] It should be further explained that, in the specific implementation process, a cross-topology map of the changing ecological benefit area is constructed according to the cross-type characteristic identification index of the changing ecological benefit area, and the process of obtaining the prediction result of the cross-type characteristic identification index according to the cross-topology map of the changing ecological benefit area includes:

[0095] Obtain cross-type feature identification indicators for each change ecological benefit area, and construct a topological connection relationship between the change ecological benefit areas containing the same cross-type feature identification indicators;

[0096] Constructing a cross-topological map of the changing ecological benefit areas, taking each changing ecological benefit area as a node of the cross-topological map of the changing ecological benefit areas, and taking the topological connection relationship between each changing ecological benefit area as the connection relationship between the nodes;

[0097] Constructing a temporal convolutional neural network, obtaining a numerical time series sequence corresponding to the cross-type feature identification index of each changing ecological benefit area in several historical collection cycles, using the numerical time series sequence as training data, using the training data to train the temporal convolutional neural network, inputting the numerical time series sequence corresponding to the cross-type feature identification index of each changing ecological benefit area in the current collection cycle into the trained temporal convolutional neural network, and obtaining the change feature trend corresponding to the cross-type feature identification index of each changing ecological benefit area according to the temporal convolutional neural network;

[0098] The graph attention network is used to learn the cross-topology graph of the changing ecological benefit area, and the change feature trend corresponding to the cross-type feature recognition index of each changing ecological benefit area is imported into the graph attention network. The attention weight between each node is obtained through the attention mechanism, and the spatiotemporal feature sequence of the cross-type feature recognition index of each node is generated using the neighbor aggregation mechanism.

[0099] The changing ecological benefit areas are taken as topological connection relationships to construct the edge structure of the cross-topology graph of the changing ecological benefit areas. The cross-topology graph of the changing ecological benefit areas is learned and represented by the graph attention network. The point structure of the cross-topology graph of the changing ecological benefit areas between nodes j and k in the cross-topology graph of the changing ecological benefit areas is noted. According to the attention value f between each changing ecological benefit area jk for:

[0100] ˊˊ

[0101] H jk =ELU(FULLY((we j ))×FULLY((we k )));

[0102] f jk (1) =Tanh(A 1 H jk +b 1 );

[0103] f jk (2) =Tanh(A 2 f jk (1) +b 2 ); ...

[0104] f jk (x+1) =Tanh(A (x+1) f jk (x) +b (x+1) );

[0105] Among them, e j 、e k represents the initial feature vector of node j and node k, w represents the feature transformation parameter matrix, ELU and Tanh represent activation functions, FULLY represents the fully connected layer of the graph attention network, and A (x+1) and b (x+1) Respectively represent the weight matrix and bias corresponding to the x+1th iteration, f jk (x+1) represents the attention value between node j and node k after the x+1th iteration;

[0106] According to the attention value f between node j and node k jk The attention weight R is determined by the ratio of the sum of the attention values ​​between node j and each other node connected to node j. jk , the formula for updating the characteristics of the changing ecological benefit area through the neighbor aggregation mechanism is:

[0107]

[0108] Among them, e j * represents the final updated representation of node j, ReLu represents the activation function, k represents the number of adjacent nodes of node j, and u j Represents the set of nodes connected to node j.

[0109] Based on the RBF neural network, a cross-type feature recognition index prediction model is constructed, and the historical ecological environment data of several historical collection cycles in the target area are obtained as training data. The cross-type feature recognition index prediction model is trained using the training data, and the spatiotemporal feature sequence of the cross-type feature recognition index of each node and the numerical time series sequence corresponding to the cross-type feature recognition index are input into the trained cross-type feature recognition index prediction model. According to the cross-type feature recognition index prediction model, the predicted average values ​​corresponding to the cross-type feature recognition indicators of each changed ecological benefit area are simultaneously output.

[0110] It should be further explained that, in the specific implementation process, the process of obtaining the prediction results of the association type feature recognition index includes:

[0111] Based on the long short-term memory network, a prediction model for the correlation type feature recognition index was constructed, and the numerical time series corresponding to the correlation type feature recognition index of each changing ecological benefit area in several historical collection cycles were used as the test set and training set.

[0112] Input the training set into the association type feature recognition index prediction model for training until the loss function training is stable, and save the model parameters, test the association type feature recognition index prediction model through the test set until it meets the preset requirements, and output the association type feature recognition index prediction model;

[0113] The numerical time series corresponding to the association type characteristic identification index of each changed ecological benefit area in the current collection period is input into the association type characteristic identification index prediction model to obtain the predicted average value corresponding to the association type characteristic identification index of each changed ecological benefit area.

[0114] It should be further explained that, in the specific implementation process, the process of obtaining the type label of the future evolution ecological benefit area of ​​the changing ecological benefit area according to the prediction results of the non-correlated type feature identification indicators, the cross-type feature identification indicators and the related type feature identification indicators in the current collection period of the changing ecological benefit area includes:

[0115] Obtain the average value corresponding to the non-associated type feature identification index of the changing ecological benefit area within the current collection period, match the predicted average value corresponding to the associated type feature identification index of the changing ecological benefit area, the predicted average value corresponding to the cross type feature identification index and the average value corresponding to the non-associated type feature identification index with the threshold interval corresponding to each type of feature identification index of each type of ecological benefit area, obtain the feature similarity between the changing ecological benefit area and each type of ecological benefit area, obtain the type of ecological benefit area whose feature similarity with the changing ecological benefit area is greater than or equal to the feature similarity threshold, and add the future evolution ecological benefit area type label corresponding to the type of ecological benefit area to the changing ecological benefit area.

[0116] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A GIS-based intelligent identification method for ecological benefit areas, characterized in that: The following steps are involved: Step s1: construct a three-dimensional regional digital twin model of the target area by means of GIS, and obtain the ecological benefit area type label of each three-dimensional triangulated network model according to the various types of feature identification index layers superimposed on each three-dimensional triangulated network model in the three-dimensional regional digital twin model; Step s2: Divide each 3D triangulated network model in the current acquisition period into a fixed ecological benefit area and a changing ecological benefit area according to the historical ecological benefit area type label of each 3D triangulated network model, and extract the non-intersecting type feature identification index, the intersecting type feature identification index and the associated type feature identification index in the changing ecological benefit area; Step s3: constructing a cross-topology map of the changing ecological benefit area according to the cross-type feature identification index of the changing ecological benefit area, obtaining the prediction result of the cross-type feature identification index according to the cross-topology map of the changing ecological benefit area, and obtaining the prediction result of the association type feature identification index at the same time; Step s4: Obtain the future evolution ecological benefit area type label of the changing ecological benefit area according to the non-correlated type feature identification index, the cross type feature identification index prediction results and the correlated type feature identification index prediction results within the current acquisition period of the changing ecological benefit area.

2. The method for intelligently identifying ecological benefit areas based on GIS according to claim 1, characterized in that: The process of constructing a three-dimensional regional digital twin model of the target area through GIS means includes: The terrain features, digital elevation data and ecological environment data of the target area at the current moment are obtained through GIS means and the collection cycle is set. The plane triangulation model of the target area is constructed according to the terrain features. The elevations of each triangle vertex in the plane triangulation model are determined according to the digital elevation data to obtain a three-dimensional triangular model. Classify the ecological and environmental data of the target area and extract various types of feature identification indicators. Then, construct various types of feature identification indicator layers based on the various types of feature identification indicators. Superimpose the various types of feature identification indicator layers on the three-dimensional triangulated network model to generate a three-dimensional regional digital twin model.

3. The method for intelligently identifying ecological benefit areas based on GIS according to claim 2 is characterized in that: According to the various types of feature identification index layers superimposed on each 3D triangulated network model in the 3D regional digital twin model, the process of obtaining the ecological benefit area type label of each 3D triangulated network model includes: Preset threshold intervals corresponding to various types of feature identification indicators corresponding to several types of ecological benefit areas, obtain various types of feature identification indicator layers superimposed on various three-dimensional triangulated network models in the three-dimensional regional digital twin model within the current acquisition period, extract the numerical values ​​corresponding to various types of feature identification indicators in the various types of feature identification indicator layers, match the numerical values ​​corresponding to various types of feature identification indicators of various three-dimensional triangulated network models with the threshold intervals corresponding to various types of feature identification indicators corresponding to various types of ecological benefit areas, obtain the feature similarity between each three-dimensional triangulated network model and each type of ecological benefit area, preset a feature similarity threshold, and if the feature similarity between the three-dimensional triangulated network model and the type of ecological benefit area is greater than or equal to the feature similarity threshold, add an ecological benefit area type label corresponding to the type of ecological benefit area to the three-dimensional triangulated network model; If the feature similarities between the three-dimensional triangulated network model and each type of ecological benefit area are less than the feature similarity threshold, a non-ecological benefit area label is added to the three-dimensional triangulated network model.

4. The method for intelligently identifying ecological benefit areas based on GIS according to claim 3 is characterized in that: The process of dividing each 3D triangulated network model in the current collection period into a fixed ecological benefit area and a variable ecological benefit area according to the historical ecological benefit area type label of each 3D triangulated network model includes: Obtain the historical ecological benefit area type labels of each 3D triangulated network model in several historical acquisition cycles in the 3D regional digital twin model, and determine whether the ecological benefit area type labels of each 3D triangulated network model in the current acquisition cycle are consistent with the historical ecological benefit area type labels of each 3D triangulated network model. If they are consistent, the 3D triangulated network model is marked as a fixed ecological benefit area; if not, the 3D triangulated network model is marked as a changing ecological benefit area.

5. The method for intelligently identifying ecological benefit areas based on GIS according to claim 4 is characterized in that: The process of extracting non-intersection type feature identification indicators, intersection type feature identification indicators and association type feature identification indicators in the changed ecological benefit area includes: Obtain historical ecological benefit area type labels that are inconsistent with ecological benefit area type labels in the current collection period in several historical collection periods of the changed ecological benefit area, extract each type of characteristic identification index contained in the ecological benefit area type label and the corresponding numerical range of each type of characteristic identification index, and at the same time extract each type of characteristic identification index contained in the historical ecological benefit area type label and the corresponding numerical range of each type of characteristic identification index; The type feature identification indicators included in both the ecological benefit area type label and the historical ecological benefit area type label are marked as non-crossing type feature identification indicators, and the type feature identification indicators not included in both the ecological benefit area type label and the historical ecological benefit area type label are marked as crossing type feature identification indicators; If the numerical range corresponding to the non-intersecting type feature identification indicator contained in the ecological benefit area type label is consistent with the numerical range corresponding to the non-intersecting type feature identification indicator contained in the historical ecological benefit area type label, the non-intersecting type feature identification indicator is marked as a non-associated type feature identification indicator; if the numerical range corresponding to the non-intersecting type feature identification indicator contained in the ecological benefit area type label is inconsistent with the numerical range corresponding to the non-intersecting type feature identification indicator contained in the historical ecological benefit area type label, the non-intersecting type feature identification indicator is marked as an associated type feature identification indicator.

6. The method for intelligently identifying ecological benefit areas based on GIS according to claim 5 is characterized in that: The process of constructing a cross-topology map of the changing ecological benefit area according to the cross-type characteristic identification index of the changing ecological benefit area and obtaining the prediction result of the cross-type characteristic identification index according to the cross-topology map of the changing ecological benefit area includes: Obtain cross-type feature identification indicators for each change ecological benefit area, and construct a topological connection relationship between the change ecological benefit areas containing the same cross-type feature identification indicators; Constructing a cross-topological map of the changing ecological benefit areas, taking each changing ecological benefit area as a node of the cross-topological map of the changing ecological benefit areas, and taking the topological connection relationship between each changing ecological benefit area as the connection relationship between the nodes; Construct a temporal convolutional neural network, obtain the numerical time series sequence corresponding to the cross-type feature identification indicators of each changing ecological benefit area in several historical collection cycles as training data to train the temporal convolutional neural network, input the numerical time series sequence corresponding to the cross-type feature identification indicators of each changing ecological benefit area in the current collection cycle into the trained temporal convolutional neural network, and obtain the change feature trend corresponding to the cross-type feature identification indicators of each changing ecological benefit area; The graph attention network is used to learn the cross-topology graph of the changing ecological benefit area, and the change feature trend corresponding to the cross-type feature recognition index of each changing ecological benefit area is imported into the graph attention network. The attention weight between each node is obtained through the attention mechanism, and the spatiotemporal feature sequence of the cross-type feature recognition index of each node is generated using the neighbor aggregation mechanism. Based on the RBF neural network, a cross-type feature recognition index prediction model is constructed. The historical ecological environment data of several historical collection periods in the target area are obtained as training data to train the cross-type feature recognition index prediction model. The spatiotemporal feature sequence of the cross-type feature recognition index of each node and the numerical time series sequence corresponding to the cross-type feature recognition index are input into the trained cross-type feature recognition index prediction model. According to the cross-type feature recognition index prediction model, the predicted average values ​​corresponding to the cross-type feature recognition indicators of each changed ecological benefit area are output simultaneously.

7. The method for intelligently identifying ecological benefit areas based on GIS according to claim 6, characterized in that: The process of obtaining the prediction results of the association type feature recognition index includes: Based on the long short-term memory network, an association type feature recognition index prediction model is constructed. The numerical time series sequence corresponding to the association type feature recognition indicators of each changing ecological benefit area in several historical collection cycles is used as training data to train the association type feature recognition index prediction model. The numerical time series sequence corresponding to the association type feature recognition indicators of each changing ecological benefit area in the current collection cycle is input into the trained association type feature recognition index prediction model to obtain the predicted average value corresponding to the association type feature recognition indicators of each changing ecological benefit area.

8. The method for intelligently identifying ecological benefit areas based on GIS according to claim 7, characterized in that: The process of obtaining the type label of the future evolution ecological benefit area of ​​the changing ecological benefit area according to the prediction results of the non-correlated type feature identification index, the cross-type feature identification index and the related type feature identification index in the current collection period of the changing ecological benefit area includes: Obtain the average value corresponding to the non-associated type feature identification index of the changing ecological benefit area within the current collection period, match the predicted average value corresponding to the associated type feature identification index of the changing ecological benefit area, the predicted average value corresponding to the cross type feature identification index and the average value corresponding to the non-associated type feature identification index with the threshold interval corresponding to each type of feature identification index of each type of ecological benefit area, obtain the feature similarity between the changing ecological benefit area and each type of ecological benefit area, obtain the type of ecological benefit area whose feature similarity with the changing ecological benefit area is greater than or equal to the feature similarity threshold, and add the future evolution ecological benefit area type label corresponding to the type of ecological benefit area to the changing ecological benefit area.

Images