Method for improving hydrological structure connectivity of ecological water network of silty muddy tidal flat wetland

Abstract

This invention discloses a method for improving the connectivity of the hydrological structure of an ecological water network in silty mudflat wetlands. The method includes: acquiring remote sensing images of the wetland ecological water network and identifying nodes where water channels and bodies intersect to establish a water network topology map; calculating network loop parameters and node connection ratios based on the water network topology map to obtain current connectivity evaluation indicators; constructing a geometrically partitioned network based on node plane coordinate sets and pre-configured engineering boundary constraints, with the improvement indicators as the target, and extracting candidate edge sets from it; prioritizing the candidate edge sets based on distance metrics reflecting engineering costs, and gradually adding corresponding candidate edges to the water network topology map according to priority until a preset indicator threshold is reached to obtain an optimized water network topology map; and outputting a wetland ecological water network structure layout scheme based on the optimized water network topology map. This invention helps reduce the blindness of local water system adjustments and achieves a balance between wetland construction cost control and global spatial connectivity benefits.

Description

Technical Field

[0001] This invention relates to the field of wetland ecological restoration and water resources engineering technology, and more specifically, to a method for improving the connectivity of the hydrological structure of the ecological water network in silty mudflat wetlands. Background Technology

[0002] The stable operation of wetland ecosystems is highly dependent on continuous and dynamic hydrological processes. Hydrological connectivity constitutes the static framework of the water system in physical space, directly determining the potential water exchange paths and material transport efficiency between water bodies. Quantitatively assessing the spatial topology of water networks and implementing appropriate interventions can enhance the spatiotemporal distribution and regulation capacity of regional water resources. This has significant technical support value for maintaining ecological baseflow within the water system, improving overall hydrodynamic conditions, and enhancing the resilience and self-evolution of complex water systems.

[0003] Currently, the construction and connectivity restoration of wetland water systems mainly employ engineering methods such as localized river channel excavation, dredging, or straightening of waterways. Existing conventional solutions are largely guided by meeting single-point water demands or predetermined flood control and drainage targets, typically relying on manual surveys and engineering experience for point-to-point or nearby water flow connections in spatial layout. However, when facing large-scale, complex geographical systems containing various types of areal and linear water bodies, this experience-driven, localized construction model struggles to coordinate overall connectivity benefits. Furthermore, in key areas where multiple rivers converge, existing solutions often lack systematic calculations of the physical carrying capacity of water flow, leading to localized hydrological bottlenecks and even secondary hydrodynamic disasters such as bank erosion and instability after the projects are put into use, due to a lack of comprehensive calculations.

[0004] Existing water system construction models struggle to balance the optimal evolution of the overall connectivity structure with the physical feasibility of local confluence nodes when dealing with large-scale complex water network systems. This often leads to an engineering dilemma where construction costs are high, but the overall water cycle connectivity benefits are limited. Therefore, there is an urgent need to explore a method that can scientifically assess macro-hydrological networks and guide the optimal layout of macro-structures to address the systemic layout problems in large-scale water network planning. Summary of the Invention

[0005] Therefore, this application provides a method for improving the connectivity of the hydrological structure of the ecological water network in silty mudflat wetlands, in order to solve the above-mentioned technical problems existing in the prior art.

[0006] According to one aspect of this application, a method for improving the connectivity of the hydrological structure of an ecological water network in a silty mudflat wetland includes:

[0007] Acquire remote sensing images of wetland ecological water networks, and based on these images, identify water flow channels and water body confluence nodes to establish a water network topology map.

[0008] Based on the water network topology diagram, the network looping parameters and node connection ratios are calculated to obtain the current connectivity evaluation index.

[0009] With the goal of improving the current connectivity evaluation index, a geometric partitioning network is constructed based on the planar coordinate set of nodes in the water network topology diagram and the pre-configured engineering boundary constraints, and a candidate edge set that is different from the existing water flow channels is extracted from it.

[0010] Based on the distance metric that reflects the project cost, the candidate edge set is prioritized and evaluated. The corresponding candidate edges are added to the water network topology graph step by step according to the determined priority order until the connectivity reaches the preset index threshold, thus obtaining the optimized water network topology graph.

[0011] Output a layout scheme for the wetland ecological water network structure based on the topological connections of the optimized water network topology map.

[0012] According to another aspect of this application, a computer-readable storage medium is provided that stores computer instructions thereon, which, when executed by a processor, implement the method for improving the connectivity of the hydrological structure of the ecological water network in silty mudflat wetlands proposed in this invention.

[0013] Based on the above technical solutions, this invention helps to reduce the blindness of local water system adjustments and achieve a balance between wetland construction cost control and global spatial connectivity benefits. Attached Figure Description

[0014] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0015] Figure 1 This is a schematic diagram of a method for improving the connectivity of the hydrological structure of an ecological water network in a silty mudflat wetland, as provided in this application.

[0016] Figure 2 This is a schematic diagram of the process of establishing a water network topology map based on the identification of water flow channels and water body intersection nodes in remote sensing images of wetland ecological water networks, as provided in this application.

[0017] Figure 3 This is a schematic diagram of the process for calculating network loop parameters and node connection ratios based on the water network topology diagram to obtain the current connectivity evaluation index, as provided in this application.

[0018] Figure 4 This application provides a schematic diagram of the process of prioritizing candidate edge sets based on distance metrics that reflect engineering costs, and gradually adding the corresponding candidate edges to the water network topology graph according to the determined priority order.

[0019] Figure 5 This is a schematic diagram of the physical space verification process for intersection nodes provided in this application. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] Example 1

[0022] This embodiment provides a method for improving the connectivity of the hydrological structure of the ecological water network in silty mudflat wetlands, such as... Figure 1 As shown, it may include the following steps:

[0023] Step 101: Obtain remote sensing images of the wetland ecological water network, and identify water flow channels and water body intersection nodes based on the remote sensing images of the wetland ecological water network to establish a water network topology map.

[0024] Specifically, the remote sensing imagery of wetland ecological water networks can be high-resolution satellite or aerial imagery with a spatial resolution of ≤10m. Since wetland environments often involve intermingling aquatic plants, to improve the accuracy of water body extraction, the spectral selection of this remote sensing imagery needs to include bands sensitive to water absorption characteristics, such as blue-green bands, red bands, near-infrared bands, or shortwave infrared bands. Furthermore, the imagery data must be acquired without significant cloud cover to ensure the integrity of surface hydrological features.

[0025] Furthermore, the process of identifying water flow channels and water body intersections can be described as transforming the hydrological morphology within the natural geographical environment into an abstract graph theory structure. Specifically, this step uses remote sensing image processing technology to identify natural or artificial water flow channels within the wetland, such as rivers, streams, tidal channels, and ditches, and abstracts them as edges of a water network. Simultaneously, it identifies the confluences of multiple water flow channels, water body connection points, or river ends, abstracting them as nodes of the water network. By recording the connections between these nodes, an undirected water network topology graph composed of nodes and edges is constructed. This transformation process removes the complex tortuous morphology of the water system while retaining its core spatial connectivity attributes, providing a foundational data structure for subsequent quantitative analysis of structural connectivity.

[0026] Step 102: Calculate the network looping parameters and node connection ratio based on the water network topology diagram to obtain the current connectivity evaluation index.

[0027] Specifically, current connectivity evaluation indicators can be used to quantitatively reflect the spatial continuity and water exchange potential of the current wetland water system. After obtaining the water network topology map, a statistical algorithm is used to traverse the graph structure and extract the network's characteristic parameters. Calculating the network loop formation parameter is to assess the degree to which closed loops are formed within the water system; the more loops, the richer the alternative paths for water flow and the stronger the resistance to siltation risk. The node connection ratio reflects the average number of waterways connected to each node; the larger this value, the more complex the internal interweaving of the water network. By comprehensively calculating these parameters, an evaluation system reflecting the current connectivity characteristics of the water network can be established.

[0028] Step 103: With the goal of improving the current connectivity evaluation index, a geometric partitioning network is constructed based on the planar coordinate set of nodes in the water network topology diagram and the pre-configured engineering boundary constraints, and candidate edge sets that are different from the existing water flow channels are extracted from it.

[0029] In engineering practice, blindly excavating new waterways often fails to achieve optimal connectivity. This step introduces a geometric constraint optimization mechanism. Pre-configured engineering boundary constraints can refer to the outer boundary of the study area and internal non-constructable areas, such as roads, buildings, and dams. The planar two-dimensional coordinates of all current water network nodes are obtained, and under the constraints of the aforementioned boundaries, a geometric partitioning network covering the entire area is constructed. This network tends to establish connections between nearby nodes. Through set difference operations, the connection relationships between this geometric partitioning network and the existing water network topology are compared, extracting connections that exist in the partitioning network but have not yet been constructed in the original water network. These connections constitute the candidate edge set, representing all potential new waterway schemes that conform to the geometric proximity rule.

[0030] In some alternative implementations, before constructing the geometrically partitioned network, new water network nodes are added at selected inlet gate locations, storage node locations, or water diversion node locations within the study area, based on wetland water replenishment needs and engineering plans. The planar coordinates of the new nodes are then added to the node set of the water network topology diagram. If the planned location of the new node is located on a waterway containing an existing water network edge in the water network topology diagram, that water network edge is split into two water network edges at the new node, and the edge set of the water network topology diagram is updated accordingly.

[0031] Step 104: Prioritize the candidate edge set based on the distance metric that reflects the project cost, and add the corresponding candidate edges to the water network topology graph step by step according to the determined priority order until the connectivity reaches the preset index threshold, thus obtaining the optimized water network topology graph.

[0032] This step is further implemented as follows: the candidate edge set is prioritized based on the spatial length between the two ends of each candidate edge in the candidate edge set as a distance metric reflecting the project cost. The corresponding candidate edges are added to the water network topology map step by step according to the determined priority order. After each addition, the current connectivity evaluation index is obtained in the same way as the current connectivity evaluation index, until the current connectivity evaluation index reaches the preset index threshold, and the optimized water network topology map is obtained.

[0033] Distance can be specifically represented by the actual Euclidean length of the candidate edges in the spatial coordinate system. In hydraulic engineering, the excavation length of waterways is directly correlated with earthwork volume and construction cost. Since the generation mechanism of the candidate edge set already ensures that each edge brings a positive improvement to the connectivity index in topological theory, the core of the evaluation shifts to economy. Candidate edges are arranged in ascending order of Euclidean length, that is, shorter edges are given higher addition priority (priority order), so that more node connections can be established under the premise of consuming the same engineering budget. According to this priority order, candidate edges are added one by one in the water network topology diagram and the connectivity state is dynamically updated. Once the updated connectivity state reaches the index threshold preset in the initial planning and design, the optimization process can be stopped, and the optimized water network topology diagram is output. This optimized diagram reflects the best topology scheme that achieves the target connectivity level with the minimum engineering cost under the existing node layout.

[0034] Step 105: Output the wetland ecological water network structure layout scheme based on the topological connection relationship of the optimized water network topology map.

[0035] The final output wetland ecological water network structure layout scheme restores the newly added abstract edges in the optimized water network topology map into engineering layout instructions in geospatial space. This scheme not only includes the existing water flow channels that need to be preserved, but also clearly marks the specific latitude and longitude coordinates between which new communication channels need to be excavated, or where sluice gates and water intake gates need to be added at existing node locations, providing a construction blueprint with rigorous graph theory logic support for subsequent wetland ecological restoration construction.

[0036] Example 2

[0037] Building upon the first embodiment described above, this embodiment further details the process of identifying water flow channels and water body confluence nodes based on remote sensing images of wetland ecological water networks to establish a water network topology map. Addressing the common geomorphic feature of linear ditches and area lakes in natural wetlands, this embodiment provides a method for abstracting the topology of a composite water network, thereby avoiding the extraction of redundant pseudo-skeleton lines from large water bodies.

[0038] In one possible implementation, a water network topology map is established by identifying water flow channels and water body intersection nodes based on remote sensing images of wetland ecological water networks, such as... Figure 2As shown, it includes the following steps:

[0039] Step 201: Perform water body mask generation and morphological refinement extraction on the remote sensing image of the wetland ecological water network to obtain the water body centerline network.

[0040] In some embodiments, water body mask generation and morphological refinement extraction are performed on remote sensing images of wetland ecological water networks to obtain the water body centerline network and water body mask regions.

[0041] Specifically, after acquiring remote sensing images, image segmentation algorithms are used to extract water body pixels, generating a binarized water body mask layer. In this mask layer, pixels identified as water bodies are assigned a value of 1, and non-water body pixels are assigned a value of 0. A morphological thinning algorithm is then applied to the generated binarized water body mask to peel away the boundary pixels of the connected regions of the water body layer by layer. During the pixel peeling process, the topological connectivity of the water body structure is strictly preserved until the entire connected region shrinks to a skeleton line segment with a width of one pixel, resulting in a water body centerline network that reflects the direction of the water flow channel.

[0042] Step 202: Identify the confluence nodes (i.e., water confluence nodes) and endpoint nodes in the water body centerline network based on the connection status, and identify the connection between adjacent nodes as the water network edge.

[0043] Furthermore, in the single-pixel-width water centerline network, each skeleton cell is traversed and the number of connected cells within its eight-neighborhood is calculated. Based on the number of connected cells, the attribute of the skeleton cell in the topology is determined. When the number of connected cells in the eight-neighborhood of a skeleton cell is ≥3, it is identified as an intersection node, indicating that the location is the intersection of multiple water flow paths; when the number of connected cells in the eight-neighborhood of a skeleton cell is equal to 1, it is identified as an endpoint node, indicating that the location is the end or beginning of a water flow path. After identifying all intersection and endpoint nodes, the continuous set of skeleton cells connecting two adjacent nodes is identified as a water network edge, and the actual arc length of this continuous skeleton segment is calculated as the spatial length of the water network edge.

[0044] Step 203: For a planar water body with an area greater than a preset area threshold, extract the geometric centroid of the planar water body as a super node, and use the contact point between the boundary of the planar water body and the water body centerline network as a supplementary endpoint, and establish an edge connection between the supplementary endpoint and the super node.

[0045] Select planar water bodies with an area greater than a preset area threshold from the water mask region, extract the geometric centroid of the planar water body as a super node, and use the contact point between the boundary of the planar water body and the water centerline network as supplementary endpoints to establish water network edge connections between the supplementary endpoints and the super node.

[0046] In wetland environments, planar water bodies can specifically be entities such as lakes, ponds, or reservoirs. Directly performing the aforementioned morphological refinement on large areas of planar water bodies would create a dense, network-like pseudo-skeleton within them, disrupting the overall macroscopic hydrological connectivity. This embodiment introduces a preset area threshold to classify and intercept water body features. When the area of ​​a water body region is detected to be larger than the preset area threshold, the refinement operation within that region is stopped, and the geometric centroid of the polygonal region is directly calculated and extracted, defining this centroid as a supernode. Simultaneously, the outer boundary of the planar water body is scanned, and contact pixels where the boundary line intersects with the network of the outer water body's centerline are extracted; these contact pixels are defined as supplementary endpoints. Topologically, connecting edges are directly added from each supplementary endpoint to the supernode.

[0047] Optionally, when identifying isometric water bodies, the preset area threshold is determined by multiplying the square of the average width of the waterway in the study area obtained in advance with the area threshold coefficient, and the preset area threshold is calculated using the following formula:

[0048] A th =k th *w avg 2 ;

[0049] Among them, A th k is the preset area threshold. th w is the area threshold coefficient. avg The average width of the waterway in the study area.

[0050] For example, the area threshold coefficient in the above formula ranges from 5 to 20. By introducing this area threshold coefficient, the determination scale of the isometric water body can be dynamically adjusted according to the hydrological characteristics of different regions, making the model more adaptable to the environment.

[0051] Optionally, the geometric centroid of the planar water body is extracted as a super node, and the edge connection between the supplementary endpoints and the super node is established, including: abstracting the planar water body into a star-shaped topological radial structure with the geometric centroid as the center and the supplementary endpoints as the edge connection points.

[0052] Specifically, after the above processing, lakes or ponds that originally had complex two-dimensional geometries are reduced in dimensionality and reconstructed into a star-shaped radial topology in the graph theory data structure. This radial structure contains no closed loops; water flowing into the planar water body from any supplementary endpoint is considered to logically flow directly towards the geometric centroid at the center, and can flow out through other supplementary endpoints. This star-shaped radial topology retains the planar water body's multi-directional regulation and connectivity with surrounding waterways while minimizing the number of internal edges, thus reducing the dimensionality of the graph theory network matrix.

[0053] Step 204: Establish an undirected water network topology graph based on the intersection nodes, endpoint nodes, supplementary endpoints, super nodes, and the water network edges connecting each node.

[0054] In this embodiment, all the identified node and edge elements are integrated to construct a node-edge adjacency matrix reflecting the global connectivity state. In the adjacency matrix, if there is an edge connection between nodes, the corresponding matrix element is assigned a value of 1; otherwise, it is assigned a value of 0. Based on this adjacency matrix, an undirected water network topology graph containing a set of nodes and a set of edges is finally formed.

[0055] In some alternative implementations, for short spur branches in the water body centerline network, a pre-configured length threshold can be used for spur trimming and filtering before identifying the intersection nodes and endpoint nodes, in order to further improve the noise resistance of the water network topology and eliminate the error edges caused by the extraction algorithm.

[0056] To address the modeling distortion caused by the interweaving of linear waterways and planar lakes in complex wetland environments, a high-fidelity underlying water flow network matrix can be constructed by employing supernode abstraction based on area thresholds and star-shaped radial topology dimensionality reduction techniques to filter out redundant pseudo-skeletons extracted from large water bodies.

[0057] Example 3

[0058] Based on the above embodiment one, this embodiment further details the process of calculating network looping parameters and node connection ratios based on the water network topology diagram to obtain the current connectivity evaluation index.

[0059] In one possible implementation, network loop parameters and node connection ratios are calculated based on the water network topology diagram to obtain current connectivity evaluation indicators, such as... Figure 3 As shown, it includes the following steps:

[0060] Step 301: Extract the total number of edges and nodes of the water network from the water network topology graph.

[0061] Specifically, based on the water network topology graph constructed in the above embodiments, the total number of water flow channels and the total number of intersections or endpoints in the network are counted by traversing the node set and edge set of the water network topology graph. The total number of water network edges is recorded as variable m, and the total number of nodes is recorded as variable n. These two basic dimension data will serve as the basis for calculating all subsequent topological connectivity physical parameters.

[0062] Step 302: Calculate the ratio of the actual number of loops to the theoretical maximum number of loops based on the total number of edges and the total number of nodes in the water network, and obtain the network loop formation parameters.

[0063] Furthermore, the network loop formation parameter is used to characterize the proportional relationship between the actual number of closed loops in the water network topology graph and the maximum number of loops that can be formed in the physical plane under the same node scale. This parameter objectively reflects the actual loop formation level of the river network system. In the algorithm implementation, the result of subtracting the total number of nodes from the total number of edges in the water network and adding a constant of 1 is used as the numerator, and the result of subtracting a constant of 5 from twice the total number of nodes is used as the denominator. Finally, the ratio of the numerator to the denominator is calculated to obtain the network loop formation parameter.

[0064] α = (m - n + 1) / (2 * n - 5);

[0065] Where α is the network loop formation parameter, m is the total number of edges in the water network, and n is the total number of nodes. The value of the network loop formation parameter is limited to the range of 0 to 1. The closer the value is to 1, the more alternative loop structures are formed between wetland water bodies, and the stronger the connectivity redundancy of the entire water system in the face of local siltation.

[0066] Step 303: Calculate the ratio of the total number of edges in the water network to the total number of nodes to obtain the node connection ratio.

[0067] Specifically, the node connection ratio is used to quantify the complexity of the internal structure of a wetland river network, i.e., the average number of waterway edges distributed across the confluence nodes of various water bodies. It is calculated by directly performing a division operation: the total number of edges in the water network divided by the total number of nodes, and outputting the node connection ratio.

[0068] β=m / n;

[0069] Where β is the node connection ratio, m is the total number of edges in the water network, and n is the total number of nodes. In conventional wetland ecological water network models, the value of this ratio index is usually distributed between 0 and 3, and the upward trend of this value directly reflects the increase in the density of water flow crossing within the region.

[0070] Step 304: Calculate the ratio of the actual number of existing edges to the theoretical maximum number of edges based on the total number of edges and nodes in the water network to obtain the actual connectivity rate. Then, weight and sum the actual connectivity rate, network loop parameters, and node connection ratio based on pre-configured weight coefficients to obtain a normalized comprehensive index as the current connectivity evaluation index.

[0071] Specifically, this can be further implemented by normalizing the node connection ratio by dividing it by its theoretical upper limit under planar conditions; calculating the ratio of the actual number of existing edges to the theoretical maximum number of edges based on the total number of edges in the water network and the total number of nodes to obtain the actual connection rate; and weighting and summing the actual connection rate, network loop parameters, and normalized node connection ratio based on pre-configured weight coefficients to obtain a normalized comprehensive index as an evaluation index of current connectivity.

[0072] In this embodiment, the actual connectivity rate is used to evaluate the true connectivity between heterogeneous nodes in the river network. The calculation process is as follows: the actual connectivity rate is obtained by dividing the total number of edges in the river network by the difference between three times the total number of nodes and a constant of 6.

[0073] γ=m / (3*n-6);

[0074] Where γ is the actual connectivity rate, m is the total number of edges in the water network, and n is the total number of nodes. For a water network in a connected state, the theoretical minimum value of γ approaches 1 / 3, and the maximum value is 1.

[0075] Based on this, the network loop formation parameters, node connection ratio, and actual connectivity rate are processed separately, multiplied by their corresponding weight coefficients, and the products are summed to generate a normalized composite index. Specifically, the node connection ratio is linearly compressed to a standard range of 0 to 1 by dividing it by its theoretical upper limit under planar graph conditions; this theoretical upper limit is equal to 3-6 / n. The compressed node connection ratio, network loop formation parameters, and actual connectivity rate are then multiplied by their corresponding weight coefficients, and the products are summed to generate the normalized composite index.

[0076] S=w α *α+w β *(β / (3-6 / n))+w γ *γ;

[0077] Where S is the normalized composite index, which is the current connectivity evaluation index used as the final output; α is the network loop formation parameter; β is the node connection ratio; γ is the actual connection rate; n is the total number of nodes; and w α w β and w γ These are the weighting coefficients corresponding to the three parameters mentioned above.

[0078] It should be noted that the normalized node connection ratio β / (3-6 / n) is mathematically equal to the actual connectivity rate γ. Therefore, in the weighted calculation of the normalized composite index S, the two components are essentially merged into the same dimension. Under equal weighting, S=(1 / 3)α+(2 / 3)γ, meaning that the normalized composite index is mainly dominated by the actual connectivity rate, while the network looping parameters provide a supplementary dimension of differentiation. Those skilled in the art will understand that in horizontal comparison scenarios where the independence of all three dimensions is required, an alternative normalization method can be used, which involves dividing the node connection ratio by its theoretical upper limit of 3 (rather than 3-6 / n).

[0079] Optionally, when calculating the normalized composite index, if there is no prior weight information in the current environment, the actual connectivity rate, network loop parameters, and node connection ratio are weighted and summed with equal weight coefficients to obtain the normalized composite index.

[0080] In engineering applications, wetland ecosystems of different regional types may exhibit varying sensitivities to various connectivity indicators. When a lack of hydrogeological data at the input end leads to a lack of prior expert experience or regression weight information, the system's underlying logic automatically triggers an equal-weight allocation principle. In this case, the weight coefficients of each feature parameter are configured to be strictly equal proportions for fusion calculation to ensure the generalization robustness of the computational model.

[0081] w α =1 / 3; w β =1 / 3; w γ =1 / 3;

[0082] Among them, w α w represents the weighting coefficients of the network loop formation parameters. β w is the weighting coefficient for the node connection ratio. γ The weighting coefficients are the actual connection rate, and the sum of the above three weighting coefficients is always equal to 1.

[0083] As an alternative to the aforementioned normalized calculation scheme, when the received instruction is only to perform a longitudinal time-series evaluation of a single wetland network before and after topology optimization, rather than a lateral cross-domain comparison between wetlands of different sizes, the division balancing and weighting process can be skipped. In this simplified mode, the controller directly performs a simple algebraic summation on the extracted actual connectivity rate, network loop parameters, and node connection ratios, and outputs the resulting non-normalized sum as a comprehensive evaluation index (i.e., the current connectivity evaluation index). This alternative scheme can effectively reduce the computational overhead of the server in the evaluation module, while being sufficient to characterize the connectivity evolution trend within a single network structure.

[0084] Example 4

[0085] Based on the above embodiments, this embodiment further details the preprocessing steps before constructing the geometrically partitioned network. Due to long-term evolution and human development, natural wetland ecosystems are often fragmented by physical barriers such as road construction, flood control dams, or agricultural reclamation, resulting in spatially isolated waterways. To ensure the successful convergence of the subsequent global topology optimization algorithm, this embodiment provides a topology repair method involving cross-component obstacle avoidance bridging.

[0086] In one possible implementation, before constructing the geometrically partitioned network based on the planar coordinate set of nodes in the water network topology diagram and pre-configured engineering boundary constraints, the following steps are included:

[0087] Step 401: Perform connectivity component traversal detection on the water network topology graph to identify multiple disconnected water network subgraphs.

[0088] Specifically, a connected component can refer to a maximal subgraph structure in a graph theory model where internal nodes are interconnected but have no edge connections to external nodes. In computer program implementation, a breadth-first search or depth-first search algorithm is used to start from any unvisited node in the water network topology graph and probe along existing water network edges until no new associated nodes can be reached. All nodes and associated edges traversed in a single probe are marked as an independent connected component. This traversal and marking process is repeated until all nodes in the water network topology graph have been marked. If the final total number of connected components is greater than 1, the current water network topology graph is determined to be fragmented and isolated, and each of the probed independent connected components is defined as a separate, disconnected water network subgraph. This data dimension detection mechanism can accurately identify fractured areas in the hydrological cycle within wetlands.

[0089] Step 402: For disconnected water network subgraphs, calculate the spatial Euclidean distance between nodes contained in different subgraphs, and screen and filter out cross-component node pairs whose connections do not cross the pre-configured engineering boundary constraints (non-construction areas).

[0090] Furthermore, after obtaining multiple isolated water network sub-maps, it is necessary to quantitatively assess the feasibility of re-establishing hydrophysical connections between them. The two-dimensional planar coordinates of the boundary nodes belonging to different water network sub-maps are extracted sequentially; for example, the longitudinal and transverse coordinate data of nodes in the first and second water network sub-maps are extracted, and the spatial Euclidean distance between them in a two-dimensional planar projection system is calculated. This distance index, as a core parameter, directly reflects the shortest excavation length of the inter-regional water diversion project under theoretical conditions.

[0091] However, the shortest absolute straight-line distance may not be feasible for construction in real-world engineering scenarios. Therefore, pre-stored terrain feature layers are acquired simultaneously, and non-constructable areas are mapped as polygonal boundary layers on a two-dimensional coordinate system. Specifically, non-constructable obstacles can include highways, basic farmland red line areas, existing permanent paved structures, and high-voltage power tower foundations, etc.

[0092] The spatial geometry engine is used to perform line segment and polygon intersection detection logic to determine whether the geometric line segment representing the connection between cross-component node pairs intersects the boundary line of the obstacle polygon, or whether the line segment completely falls within the interior region of the obstacle polygon. Node pairs with geometric intersection or containment relationships are automatically eliminated, retaining only legal node pairs whose connection paths completely avoid all obstacle polygons. This accurately filters out the set of cross-component node pairs that are feasible in terms of both engineering geological conditions and land use planning.

[0093] Step 403: Select the cross-component node pair with the shortest spatial Euclidean distance and add interconnecting edges. Iterate and merge the water network subgraphs until a simply connected water network topology graph is formed.

[0094] Specifically, in the pool of feasible candidate nodes, all retained cross-component node pairs are sorted in ascending order based on the spatial Euclidean distance values ​​extracted in the aforementioned steps. The cross-component node pair with the smallest distance value is selected as the optimal bridging unit for the current stage. In the underlying storage matrix, an interconnecting edge representing a new water flow channel is added to the two endpoints corresponding to this optimal bridging unit. After performing this operation, the two originally separate independent water network subgraphs are integrated into a larger new connected component through this interconnecting edge.

[0095] After completing a single topology merging operation, the entire network state is refreshed and the total number of connected components is updated. If the underlying matrix detects that the number of remaining independent water network subgraphs is still greater than 1, the above bridging process is repeated, and the spatial Euclidean distance calculation and geometric obstacle avoidance logic are triggered again based on the newly generated connected component structures, repeatedly executing the operation of selecting the optimal bridging point and merging the network.

[0096] This iterative process continues until the system detector confirms that there is only one connected component within the graph structure that contains all nodes. At this point, the loop terminates, and the matrix data is output as a simply connected water network topology graph. This repair mechanism uses a spatial greedy strategy to eliminate the island effect in the hydrological network, establishes an initial framework for global regulation and storage linkage, and provides the necessary simply connected input preconditions for subsequent constraint partitioning algorithms.

[0097] In some alternative implementations, if multiple candidate cross-component node pairs with the same spatial Euclidean distance exist in the computation sequence, the node elevation data can be introduced as an evaluation parameter for the second dimension. Specifically, the topographic elevation difference between two nodes can be extracted and calculated, and node pairs with larger elevation differences that can promote the natural flow of water in the gravity field can be selected to establish interconnection edges, thereby reducing the energy load of the pumping station for water extraction in the later stage.

[0098] In addition, if the distribution of non-constructable obstacles in the study area is relatively dense, resulting in the absence of a direct connecting line segment that does not cross the polygon, the grid pathfinding algorithm module is automatically attached to calculate the shortest physical length of the polyline that bypasses the obstacle, and the obstacle bypass length is used to replace the straight-line Euclidean distance for subsequent sorting, optimization and filtering operations.

[0099] Faced with the fragmentation of wetlands into isolated islands caused by human intervention, the shortest obstacle avoidance spatial distance across component node pairs was used for optimization and merging, eliminating water system breaks and establishing a single-connected skeleton to support subsequent global simulations.

[0100] Example 5

[0101] Based on the above embodiments, this embodiment further details the core optimization mechanism for constructing the geometrically partitioned network, extracting candidate edges, and dynamic optimization. The optimization algorithm in this stage integrates planar graph geometric constraints and greedy cost metric rules.

[0102] In one possible implementation, a geometric partitioning network is constructed based on the planar coordinate set of nodes in the water network topology diagram and pre-configured engineering boundary constraints, and a candidate edge set that is different from the existing water flow channels is extracted from it, including the following steps:

[0103] Step 501: Using the set of planar coordinates of all nodes in the water network topology diagram as the input point set, and the boundary of the study area and the pre-defined non-constructable area as the pre-configured engineering boundary constraints, perform constrained triangulation calculations to obtain a geometrically partitioned network that meets the engineering geological conditions.

[0104] Specifically, all nodes in the water network topology map constructed in the previous steps are extracted, and their coordinate data are combined into a two-dimensional input point set. The computational geometry module is then invoked to perform constrained Delaunay triangulation. During this process, pre-configured engineering boundary constraints act as mandatory blocking conditions, ensuring that no triangle edges generated by the triangulation cross these boundaries.

[0105] The technical mechanism of using triangulation as the optimal base graph lies in Euler's formula for planar graphs in graph theory. Euler's formula states that the number of vertices minus the number of edges plus the number of faces equals a constant of 2. Adding a water flow channel to a wetland water network results in three possible geometric connection topologies.

[0106] The first mode is where both ends of the channel are connected to existing nodes, in which case the total number of nodes increases by 0 and the total number of edges in the water network increases by 1. The second mode is where one end of the channel is connected to a node and the other end is connected to the middle section of a waterway, where a new converging node is generated, in which case the total number of nodes increases by 1 and the total number of edges in the water network increases by 2. The third mode is where both ends of the channel are connected to the middle section of the waterway, in which case the total number of nodes increases by 2 and the total number of edges in the water network increases by 3.

[0107] Based on the calculation logic of the aforementioned connectivity evaluation index, the first mode can maximize the increase in connectivity constant while minimizing the change in the denominator. In this embodiment, constrained Delaunay triangulation ensures that all connections are generated between existing nodes, locking the first mode in pure mathematical logic and guaranteeing that the water network possesses the theoretically maximum potential for topology improvement each time a connection is added.

[0108] In other alternative implementations, the geometrically partitioned network can also be constructed using other geometric proximity networks such as Gabriel graphs, relative neighborhood graphs, or spatial connectivity graphs based on K-nearest neighbors. Those skilled in the art can select an appropriate partitioning method based on the specific distribution density of water network nodes.

[0109] Step 502: Compare the edge set contained in the geometric partitioning network with the current edge set of the water network topology graph, and extract the connecting edges that exist in the geometric partitioning network but not in the water network topology graph to form a candidate edge set.

[0110] Furthermore, a first set consisting of all edge data generated by the geometrically partitioned network is obtained, and a second set consisting of existing waterway connections in the water network topology map is obtained simultaneously. The computation unit performs a subtraction operation between the first and second sets. Edge elements existing within the first set and simultaneously outside the second set are retained and stored. These remaining edge elements represent new channel planning lines that have not yet been excavated and constructed in the current water network topology but have optimal potential in the geometrically partitioned model. These planning lines constitute the output candidate edge set.

[0111] Optionally, before adding the corresponding candidate edges to the water network topology graph step by step according to a determined priority order, the following may also be included:

[0112] Step S01: Based on the difference between the current connectivity evaluation index and the index threshold, and combined with the total number of nodes in the water network topology map, calculate the expected minimum number of edges required to achieve the connectivity goal.

[0113] In this step, reverse reasoning logic is introduced to pre-calculate the minimum construction scale required to achieve the engineering planning goals under ideal conditions. Initial values ​​and pre-set expected values ​​of various current connectivity evaluation indicators are extracted, and the increment in the number of edges is calculated through the product of the difference and a constant factor. The minimum expected number of edges is calculated using the following formula:

[0114] k * =max(ceil((α t -α0)*(2*n-5)),ceil((β t -β0)*n),ceil((γ t-γ0)*(3*n-6)));

[0115] Where, k * The minimum expected number of edges, max is the maximum value function, ceil is the floor function, and α is the minimum expected number of edges. t Let α0 be the desired target value of the network loop formation parameters, n be the total number of nodes, and β be the initial value of the network loop formation parameters. t Let β0 be the desired target for the node connection ratio, and γ be the initial value of the node connection ratio. t γ0 represents the expected target for the actual connection rate, and γ0 is the initial value of the actual connection rate.

[0116] Step S02: Use the expected value of the minimum number of sides to estimate the scale of material input in the early stage of the project, obtain the material input budget parameters, and output them as the engineering budget reference information for the wetland ecological water network structure layout scheme.

[0117] Specifically, after obtaining the expected value of the minimum number of sides, the expected value of the minimum number of sides is converted into total earthwork volume parameters and budget parameters by combining the average geological excavation unit price of the project area and the earthwork volume of the standard waterway cross section. This conversion result serves as a constraint during the construction drawing stage, guiding the preliminary budget preparation and machinery shift allocation scale setting in the bidding documents.

[0118] In one possible implementation, candidate edge sets are prioritized based on a distance metric reflecting project cost, and corresponding candidate edges are added to the water network topology graph step by step according to the determined priority order, such as... Figure 4 As shown, the process includes the following:

[0119] Step S1a: Since any candidate edge in the candidate edge set connects to existing water network nodes and thus has the same theoretical connectivity improvement benefits, the actual spatial length of each candidate edge in the candidate edge set is extracted as a distance metric reflecting the project cost.

[0120] In the algorithm architecture of this embodiment, since each edge in the candidate edge set follows the first geometric connection pattern of the aforementioned triangulation, when added to the network, the increment of the numerator of the fraction and the change in the denominator are consistent. From a topological perspective, the improvement benefit of arbitrarily selecting a candidate edge is equivalent. Based on this condition, the optimization objective function is reduced from multivariate to single-variable, and the coordinates of the nodes at both ends of each edge are directly extracted and its absolute spatial straight-line length is calculated. In civil construction, physical length and excavation cost are positively correlated; therefore, this length value is directly used as a distance metric reflecting the project cost and construction difficulty.

[0121] In other alternative implementations, distance metrics may also be weighted path lengths that take into account differences in terrain elevation or shortest path lengths around obstacles, to more accurately reflect the actual construction costs under different geological conditions.

[0122] Step S2a: Sort the edges in the candidate edge set according to the ascending order of actual space length from shortest to longest, and determine the priority of adding each candidate edge.

[0123] The system calls the sorting algorithm module, which performs a strict ascending order sorting based on the length value attached to each candidate edge. In the reorganized candidate edge set, the smaller the spatial length value, the earlier the sorting position, and the higher the corresponding addition priority. This sorting strategy conforms to the local optimum selection principle of greedy algorithms, that is, in each iteration, it prioritizes the combination of locally optimal solutions with the lowest cost, in order to achieve a construction scheme that minimizes the global cost.

[0124] Step S3a: Add the corresponding candidate edges to the connection relationship of the water network topology graph one by one in order of addition priority from high to low, and dynamically update the water network status.

[0125] This step can also be implemented as follows: add the corresponding candidate edges to the connection relationship of the water network topology graph one by one in order of addition priority from high to low, so as to obtain the water network topology graph after the dynamic update of the water network status.

[0126] Based on the priority queue generated by the sorting, the system pointer points to the first candidate edge at the head of the queue, changing the cross-value of its two endpoints in the underlying adjacency matrix from 0 to 1, completing a virtual addition operation. This operation changes the basic input data, driving the system to reconstruct the water network topology in memory in real time. The pointer then moves to the next high-priority edge, repeating the connectivity operation.

[0127] In one possible implementation, the process of gradually adding the corresponding candidate edges to the water network topology graph according to a determined priority order until the connectivity reaches a preset threshold to obtain the optimized water network topology graph includes the following steps:

[0128] Step a: While adding candidate edges step by step according to the addition priority, simultaneously calculate and monitor the current structural connectivity parameters (i.e., the current connectivity evaluation index) of the water network topology and the total length of the newly added waterways formed by the accumulation of all added candidate edges.

[0129] After each virtual addition operation and update of the underlying matrix, the monitoring module is automatically activated. The monitoring module obtains the latest distribution of node edges and substitutes it into the metric calculation logic described in Example 3 above to calculate the actual connection rate, network loop parameters, and node connection ratio in real time. Simultaneously, an accumulator variable is set, and the spatial length values ​​of the candidate edges added this time are added to this accumulator to output the total length of the newly generated waterways in real time.

[0130] Step b: When the current structural connectivity parameter is greater than or equal to the preset index threshold, or when the total length of the newly added waterway reaches the pre-configured upper limit of the project budget length, or when all candidate edges in the candidate edge set have been added, the termination condition is triggered, the addition operation is stopped, and the final optimized water network topology is output.

[0131] The preset threshold values ​​can be determined by the water conservancy project planner and designer based on the ecological water replenishment needs of the target area, the hydrodynamic improvement goals, and the project budget. In this embodiment, the threshold value is set to trigger termination when the current normalized composite index reaches 0.55. Those skilled in the art will understand that the specific threshold value can be adaptively adjusted according to the ecological function goals of different wetlands.

[0132] Specifically, the following three termination conditions are monitored simultaneously: whether the dynamically calculated value of the connectivity parameter has reached or exceeded the qualification line set in the initial planning stage; whether the total length in the accumulator variable exceeds the limit threshold that the construction funds can bear; and whether the priority queue has been emptied. At any point in the iteration loop, as long as any one of the above three termination conditions is met, the addition operation stops, the current water network topology map state is saved, and the node and connection data in this state are encapsulated and output as an optimized water network topology map. This multi-interception mechanism can effectively suppress over-planning and computational dead loops, ensuring that the output scheme has both ecological compliance and engineering economic feasibility.

[0133] To break the dilemma of balancing construction costs and connectivity benefits in traditional water network construction, a constrained triangulation network is used to identify candidate grids that can bring the maximum theoretical connectivity increment. The spatial excavation length is used as an economic lever to perform greedy, dynamic expansion of each grid. In this way, under the premise of strictly controlling the project cost, the number of closed loops and the node connection ratio of the overall hydrological network are systematically increased to the preset target level or the achievable upper limit under the current project constraints.

[0134] Example 6

[0135] Building upon Embodiment 5 above, this embodiment further details the steps involved in physical spatial verification of confluence nodes during the process of outputting a wetland ecological water network structure layout scheme based on the topological connectivity of the optimized water network topology graph. The aforementioned topology optimization pertains to mathematical optimization at the graph theory level; however, in actual water conservancy projects, the physical water surface space constraints required when multiple rivers converge must be considered.

[0136] In one possible implementation, such as Figure 5 As shown, the process of outputting the wetland ecological water network structure layout scheme based on the topological connection relationship of the optimized water network topology map also includes physical spatial verification of the intersection nodes:

[0137] Step 601: Count the number of waterways connected to each intersection node in the optimized water network topology diagram to obtain the node degree.

[0138] Specifically, the optimized water network topology graph output after topology optimization is traversed to extract node elements with water flow convergence characteristics. For each extracted convergence node, the total number of edge elements directly connected to it is calculated, and this value is defined as the node degree of the convergence node. This node degree reflects how many water flow channels will flow into or out of the convergence area in future physical engineering.

[0139] Step 602: Based on the node degree and the pre-obtained average width of each channel corresponding to the connecting node, calculate the minimum inscribed circle radius of the connecting node; wherein, the minimum inscribed circle radius is determined according to the proportional relationship between the product of the node degree and the average width of each channel and pi. Specifically, the minimum inscribed circle radius is equal to the quotient obtained by dividing the product of the node degree and the average width of each channel by twice pi.

[0140] Furthermore, after obtaining the network features, they are transformed into two-dimensional geometric constraint boundaries. The future node convergence region is approximated as a circular water surface area, and it is assumed that all connected water flow channels are uniformly arranged and connected along the radial direction of this circular area. To prevent water flow congestion, the theoretical minimum length of the circumference of this node region must be greater than or equal to the sum of the widths of all connected water channels. The design parameters or current measurement data of each connected water flow channel are extracted to calculate the average width of each water channel. The product of the node degree and the average width of each water channel is calculated, and the result is divided by twice the constant pi to obtain the minimum inscribed circle radius under the extreme compact state.

[0141] r min =(d*w) / (2*π);

[0142] Where, r minLet d be the minimum inscribed circle radius of the confluence node, d be the node degree of the confluence node in the optimized water network topology diagram, w be the average width of each waterway connecting the confluence node, and π be the constant pi.

[0143] Step 603: Determine the minimum water surface area of ​​the intersection node based on the pre-configured safety factor and the minimum inscribed circle radius.

[0144] In actual wetland construction, sufficient spacing must be reserved between different water flow channels to ensure bank slope stability. Therefore, a pre-configured safety factor is needed to provide a safety margin for the project. This safety factor is extracted and, combined with the previously obtained minimum inscribed circle radius, the required final safe water surface area is calculated. The specific calculation logic is as follows: calculate the product of the square of the safety factor, the square of the node degree, and the square of the average width of each waterway; divide this product by 4 times pi; and output the calculated result as the minimum water surface area of ​​the node.

[0145] A min =(k s 2 *d 2 *w 2 ) / (4*π);

[0146] Among them, A min k represents the minimum water surface area at the intersection node. s Here, d is the pre-configured safety factor, w is the node degree, w is the average width of each waterway, and π is the constant of pi.

[0147] Optionally, when calculating the minimum water surface area of ​​a node, the safety factor is limited to a range of 2.0 to 4.0.

[0148] For example, let's assume a normalized calculation case is used for illustration. After topology optimization, the newly added degree d of a certain intersection node is determined to be 4. The normalized value of the average width w connecting the waterways in this region is set to 1.0. First, the inscribed circle radius formula is used to calculate r. min =(4*1.0) / (2*3.1416)=0.637. The security factor k is obtained from the configuration library. s And set its value to 2.0. Substitute the above parameters into the area formula to calculate A. min =(2.0 2 *4 2 *1.0 2 ) / (4*3.1416)=64 / 12.566=5.093. Therefore, the theoretical lower limit of the safe physical water surface area of ​​this confluence node with 4 normalized width waterways is 5.093 area units.

[0149] Step 604: Compare the current actual water surface area of ​​the intersection node obtained from the remote sensing image of the wetland ecological water network with the minimum water surface area of ​​the node. If the actual water surface area is insufficient, output the widening or dredging suggestions for the corresponding intersection area in the wetland ecological water network structure layout scheme.

[0150] Specifically, the system obtains the current actual water surface area of ​​the intersection node from the remote sensing image analysis module or the mapping database. Through numerical comparison logic, it determines whether the actual water surface area is greater than or equal to the calculated minimum water surface area of ​​the node. When the actual water surface area is less than the minimum water surface area of ​​the node, it is determined that there is a potential hydrological bottleneck in the area. Continuing with the aforementioned normalization example, if the current actual water surface area of ​​the node is calculated to be 3.0 area units, since 3.0 < 5.093, it is determined that it does not meet the physical connectivity requirements. In this case, the system automatically highlights the intersection coordinate point in the layer attributes of the final generated wetland ecological water network structure layout scheme, adds engineering construction parameters, and explicitly outputs a dredging suggestion that the area needs to be widened and that an additional 2.093 area units of earthwork need to be excavated. This mechanism effectively bridges the purely abstract graph theory optimization algorithm with the real physical constraints of wetland water network transformation projects.

[0151] In addition, to mitigate the risk of hydrodynamic congestion that may be caused by the confluence of multiple new waterways, the abstract topological connectivity degree is mapped to the minimum water surface area requirement of the inscribed circle in the physical dimension. This provides accurate dredging and widening parameter verification for the actual geographical space and opens up a transformation path from graph theory optimization to the safe implementation of water conservancy projects.

[0152] Based on the same inventive concept as the above-described method embodiments, this embodiment provides a computer-readable storage medium. This computer-readable storage medium can be used to execute the methods provided in Embodiments 1 to 6 above. Its implementation mechanism and technical output form are consistent, and the same processing procedures will not be repeated here. The following focuses on describing the comprehensive implementation effect of the hardware system loading computer instructions running the algorithm model in a specific large-scale ecological restoration project.

[0153] When computer instructions are executed by the processor, they implement the method for improving the connectivity of the hydrological structure of the ecological water network in silty mudflat wetlands proposed in this invention.

[0154] Specifically, the computer-readable storage medium can be implemented using non-volatile storage elements, such as internal solid-state drives or flash memory chips. When computer instructions are fetched and loaded into the processor of the local workstation for execution, the system automatically completes the entire link of graph theory operations, from image recognition and topology extraction to mathematical optimization. To verify the actual engineering physical performance of the aforementioned connectivity enhancement method executed by computer instructions, a wetland area at the mouth of a river was selected as the target test area. According to the input geological survey data, before system intervention, this test area exhibited objective defects such as insufficient channel distribution density, local water shortage, missing water storage nodes, and interrupted water replenishment in the shallow waters.

[0155] Further, the data source acquisition action is performed to acquire multispectral remote sensing images covering the test area with sufficient spatial resolution (this embodiment uses Sentinel-2 satellite imagery as an example). Based on the established water body extraction mask operation and morphological skeleton stripping mechanism, the processor constructs the current existing water network topology matrix and calls the background evaluation function to calculate the current status characteristic parameters. The operation log shows that the total number of edges m in the current existing water network is 73, and the total number of nodes n is 47. According to the index system in Embodiment 3, the division operation is performed, and the output network loop parameter α is 0.30, the node connection ratio β is 1.55, and the actual connection rate γ is 0.54. According to the normalized comprehensive index calculation logic in Embodiment 3, the fusion calculation is performed using equal weight configuration, and the output current status normalized comprehensive index S is 0.46.

[0156] After completing the current status assessment, the processor automatically triggers the core topology extension instruction set based on constrained Delaunay triangulation. Before executing the above topology optimization method, engineers can add new nodes such as water intake gates within the study area according to water replenishment needs; this node expansion is a conventional engineering planning method in the field. After adding new nodes, the updated node set is substituted into the topology optimization process of this invention. Those skilled in the art will understand that even without adding new nodes, the constrained triangulation and greedy edge addition method of this invention can still improve the water network connectivity index under the existing node layout.

[0157] Based on the engineering design boundary input instructions, multiple water inlets are added as new water body confluence nodes within the specified spatial coordinate region, and non-constructable polygonal obstacle constraint layers are checked. Under the geometric constraints of Eulerian plane graph theory, the processor continuously executes a greedy edge-adding algorithm based on ascending spatial length cost, gradually generating a triangular partitioning structure layout that maximizes connectivity benefits. After triggering the preset connectivity index threshold termination condition, the memory state is locked and the optimized water network topology map is output.

[0158] Based on this, the system monitoring module synchronously output the final optimized network attributes. After the optimization operation, the total number of edges m in the water network increased to 139. According to the project plan, the system added 27 water intake gate nodes to the water network topology map, expanding the total number of nodes from 47 to 74. The evaluation function engine was reactivated for post-verification calculations, yielding the optimized network loop formation parameter α of 0.46, the node connection ratio β of 1.88, and the actual connection rate γ of 0.64. The optimized normalized comprehensive index S increased to 0.58. When the termination condition was triggered, the total length parameter of the newly added waterway was output, providing a quantitative basis for subsequent project budget preparation.

[0159] By comparing the output data before and after optimization, it can be concluded that the water network topology scheme generated by executing computer instructions improved the normalized comprehensive connectivity evaluation index of the target test area by approximately 26%. This comparative data objectively verifies that the algorithm model built into this invention can effectively increase the number of spatial connection channels and the scale of closed loops within the wetland water network, demonstrating its technical feasibility in guiding actual ecological water replenishment channel excavation projects.

[0160] According to one aspect of this application, a method for improving the connectivity of the hydrological structure of an ecological water network in a silty mudflat wetland may further include the following steps:

[0161] S1. Acquire remote sensing images of wetland ecological water networks and construct a spatial dataset of wetland water networks. The wetland ecological water network is a composite system of wetlands and water networks with connectivity and functional synergy, constructed based on natural wetlands through artificial water systems, water storage projects, and ecological restoration. Preferably, the remote sensing images are high-resolution images with a spatial resolution ≤10m. Preferably, the remote sensing images are free from significant cloud cover. Preferably, the image bands include blue-green bands, red bands, near-infrared bands, or shortwave infrared bands—bands sensitive to water bodies.

[0162] S2 identifies water network nodes and edge structures, and establishes water network topology relationships; water network edges are water flow channels in wetland water networks, including natural or artificial water flow channels such as rivers, streams, tidal channels, and ditches; water network nodes are the confluence of multiple water flow channels, water body connection points, or river endpoints in wetland water networks.

[0163] The establishment of water network topology includes the following sub-steps:

[0164] S21, extract water bodies from the remote sensing image obtained in step S1 and generate a binary water body mask layer, in which water body pixels are assigned a value of 1 and non-water body pixels are assigned a value of 0.

[0165] S22. Morphological skeleton extraction is performed on the binarized water mask to obtain the water centerline network. The skeleton extraction uses a morphological thinning algorithm to peel off the water boundary pixels layer by layer while preserving topological connectivity until a skeleton line with a single pixel width is obtained.

[0166] S23, Identify nodes and edges in the skeleton network: Calculate the number of eight neighbors N8 for each skeleton cell, identify cells with N8≥3 as intersection nodes, and identify cells with N8=1 as endpoint nodes; identify continuous skeleton line segments between two adjacent nodes as an edge, and the length of the edge is the actual arc length of the skeleton line segment.

[0167] S24, for areas larger than a preset area threshold A th For planar water bodies, such as lakes, ponds, and reservoirs, a supernode approach is used for topological abstraction: the geometric centroid of the planar water body is taken as the supernode v. s And take all the contact points on the boundary of the planar water body that are connected to the skeleton line network as the endpoints of the edges, and connect them to the super node v respectively. s Establish edge connections. Area threshold A. th The optimal value is determined based on the characteristics of the water body in the study area, and is 5 to 20 times the square of the average width of the waterway in the study area.

[0168] S25, construct the node-edge adjacency matrix A, where A ij =1 indicates that there is an edge connecting node i and node j. ij =0 indicates that there are no edge connections, thus forming an undirected topological graph G(V,E) of the water network, where V is the set of nodes, E is the set of edges, the number of nodes n=|V|, and the number of edges m=|E|.

[0169] S3, calculate the current connectivity evaluation index of the water network to evaluate the connectivity of the current water network structure. Preferably, the current connectivity evaluation index is the classic connectivity index, including the actual loop formation rate (network loop formation parameter) α, the edge-to-node ratio (node ​​connection ratio) β, and the actual connection rate γ, whose expression is:

[0170] , , ,

[0171] In the formula, m is the number of water network edges, and n is the number of water network nodes. Preferably, the structural connectivity evaluation is based on the value ranges of α, β, and γ, where α represents the actual loop level of the river network system, with a value range of 0~1; β represents the internal complexity of the river network, with a value range of 0~3; and γ represents the actual connectivity between different nodes in the river network system. For connected water networks, the value range of γ is (n-1) / (3n-6)~1, and when the number of nodes is sufficiently large, it is approximately 1 / 3~1. The larger the values ​​of the three indicators, the better the structural connectivity of the water network. For the comparison before and after optimization of a single network, the above three indicators α, β, and γ can be added together as a comprehensive evaluation index (i.e., the current connectivity evaluation index). When comparing water networks of different scales, it can be seen that the normalized comprehensive index S is used, and each index is divided by its theoretical maximum value and then weighted and summed.

[0172] ;

[0173] Where 3-6 / n is the theoretical upper limit of β under planar graph conditions, corresponding to the maximum number of edges in the planar graph m=3n-6, w α +w β +w γ =1 represents the weighting coefficient. When there is no prior information, equal weights w are used. α =w β =w γ =1 / 3. After normalization, all components are within the interval [0,1], and the comprehensive index S∈[0,1], eliminating the scaling bias caused by the different value ranges of each index.

[0174] S4. Based on the structural connectivity results, optimize the water network topology. The optimization of the water network topology, based on the principle of planar graphs in graph theory, ensures that the number of vertices, edges, and faces satisfies Euler's formula.

[0175] V-E+F=C+1;

[0176] In the formula, F is the number of faces, E is the number of edges (i.e., the total number of edges of the water network m), V is the number of vertices (i.e., the total number of nodes n), and C is the number of connected parts that make up the graph. When the graph is a simply connected graph, the formula simplifies to: V-E+F=2.

[0177] Assuming the wetland network is a relatively ideal planar diagram, adding any edge (not a multiple edge or cycle) to the network will result in a decrease in its connectivity index value if only one end of the edge is connected to the network. Table 1 shows the changes in the overall number of nodes, edges, and connectivity indices of the network, considering the case where both ends are connected to the network.

[0178] Table 1. Changes in the number of nodes, edges, and connectivity indices when an edge is added to the water network.

[0179]

[0180] For complex water networks in general engineering areas, with a sufficiently large number of nodes (n>3), the denominators of the columns for the network loop parameter change Δα, node connection ratio change Δβ, and actual connection rate change Δγ in Table 1, such as 2n-5, n-1, 3n-6, etc., are all greater than 0. Substituting factors of the form m-n+p (p=1, 2, 3, 4) into Euler's formula, we get m-n+p=F-2+p, where F is the number of faces. Since a typical wetland water network has at least one face, it is clear that F≥1 and F-2+p≥0. Finally, we obtain the following table... When the term is less than 1, meaning a new edge is added to connect to a node, and Δm=1 and Δn=0, the classic connectivity index of the wetland network increases the most. Therefore, the following principles apply when planning or optimizing a water network:

[0181] (1) When planning or constructing water channels such as ditches, it is necessary to avoid connecting the channel to the water network at only one end;

[0182] (2) The two ends of the side should be connected to the existing water network nodes as much as possible, so that each face of the water network is enclosed by three sides, i.e., a triangular partitioned loop structure;

[0183] (3) Increase the water surface area at the nodes of the water network so that it can accommodate more connecting edges;

[0184] (4) When the water network already has many loops (m is large), the addition of waterways in the second and third connection modes may cause a decrease in some connectivity indicators. Such connection methods should be avoided in the project.

[0185] The above principles are based on the assumption of a simply connected graph. When a wetland water network has multiple disconnected subnetworks formed by barriers such as roads and dams, the disconnected components should be eliminated first before performing the above optimization analysis.

[0186] To increase the water surface area at nodes in the water network, the following quantification method is used: For nodes in the optimized water network with an increased degree (number of connecting edges), estimate the minimum required water surface area. When d waterways converge at a node, the node area is approximated as a circular region, with the waterways evenly arranged radially. The circumference of the node should not be less than the sum of the widths of all waterways. Therefore, the minimum inscribed circle radius of the node is:

[0187] ;

[0188] Considering the spacing between waterways and the safety margin of the project, a safety factor k is introduced. s (Recommended value: 2.0~4.0), then the minimum water surface area of ​​the node is:

[0189] ;

[0190] Where d is the degree of the node in the optimized water network, and w is the average width of the waterways connecting the node. When the degree d of a node increases after optimization, while the existing water surface area of ​​the node is less than A... min At that time, the area of ​​the node should be widened or dredged to ensure that its water surface area meets the requirements of A. min Require.

[0191] Optimizing the water network topology includes the following sub-steps:

[0192] S41, Connectivity Component Detection: Perform a breadth-first traversal of the water network topology graph G(V,E) to identify all connected components. Record the number of connected components k (i.e., C in the aforementioned Euler formula).

[0193] S42, Node Set Expansion: Based on the water network replenishment needs and engineering planning, new water intake gates, storage nodes, or water distribution nodes are added to the study area, and their coordinates are added to the node set V. If the newly added node is located on the waterway where an existing edge is located, the edge is split into two edges (a new node is formed at the split point), and G(V,E) is updated accordingly.

[0194] S43, Cross-component priority bridging (executed when k>1): For each pair of disconnected components (C i C j ), calculate the Euclidean distance between all node pairs between components, and select the node pair with the shortest distance and whose connection does not cross any inaccessible obstacles (v i ,v j Add edge (v) i ,v j Merge the two components. Repeat the above process until all components are merged into a single connected component (k=1). Update the topology graph G and the connected component information after each merge.

[0195] S44, Constrained Delaunay Triangulation Candidate Edge Generation: Using the planar coordinates of all nodes V in the current water network as the input point set, and the boundary of the study area and non-constructable areas (roads, embankments, buildings, etc.) as constraint boundaries, calculate the constrained Delaunay triangulation CDT(V). Extract the candidate edge set E. c :

[0196] ;

[0197] This refers to edges that exist in the triangulation but not in the current water network. Here, 'e' represents a candidate edge (a connection between two nodes); and 'E' represents the set of existing edges in the current water network topology.

[0198] S45, Candidate edge priority sorting: Due to the candidate edge set E c In this case, each edge has two existing nodes at both ends, satisfying the optimal connection conditions of Δm=1 and Δn=0. In this case, the improvement of each candidate edge on the three connectivity indices is a fixed value:

[0199] ,

[0200] The aforementioned improvement depends only on the total number of nodes n (which remains constant in the pure node-to-node connection mode), and is independent of the current number of edges m and which specific pair of nodes is connected. Therefore, the connectivity improvement benefits of each candidate edge are the same, and their priority is determined by engineering cost. Using the Euclidean length L(e) of the edge as the distance metric, candidate edges are arranged in ascending order of L(e), with shorter edges having higher priority.

[0201] S46, Greedy Stepwise Addition: Add candidate edges to the water network topology graph in descending order of priority. Let z be the optimized connectivity indices after adding z candidate edges:

[0202] ,

[0203] Where α z To optimize the loop formation parameters of the network, β z To optimize the node connection ratio, γ z The optimized actual connection rate is represented by α0, β0, and γ0, which are the initial values ​​of the metrics before optimization. Addition stops when any of the following termination conditions are met:

[0204] (a) The connectivity index reaches the preset target threshold, i.e., α z ≥α t And β z ≥β t And γ z ≥γ t ;

[0205] (b) The total length of the newly added waterway reaches the upper limit of the project budget, i.e. , where L(e i Let ) be the i-th edge e i The length, that is, the actual length of this section of the waterway, L max q is the upper limit of the project budget, which is the maximum total length of waterways that can be constructed, and q is the upper limit of the summation, which is the total number of edges (waterways) that have been added so far.

[0206] (c) Candidate edge set E c All added (the water network has achieved a triangular segmentation structure).

[0207] The minimum number of edges required to reach the goal can be estimated in advance as follows:

[0208] ;

[0209] Where ┌ ┐ is the floor function.

[0210] S47, Output optimization results: Output the optimized water network topology graph G'(V,E'), the set of newly added edges E'\E, the length of each edge, and the optimized connectivity index α. z ,β z γz .

[0211] S5 forms an optimized layout scheme for the wetland ecological water network structure. This layout scheme is a reference scheme under ideal conditions. In actual projects, the final water network layout scheme should be determined by comprehensively considering various factors such as the current status of the water network, water replenishment needs, project costs, and boundary constraints.

[0212] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A method for improving the connectivity of the hydrological structure of an ecological water network in a silty mudflat wetland, characterized in that, include: Acquire and identify water flow channels and water body confluence nodes based on remote sensing images of wetland ecological water networks, and establish a water network topology map; Based on the water network topology diagram, the network looping parameters and node connection ratios are calculated to obtain the current connectivity evaluation index. With the goal of improving the current connectivity evaluation index, a geometric partitioning network is constructed based on the planar coordinate set of nodes in the water network topology diagram and the pre-configured engineering boundary constraints, and a candidate edge set that is different from the existing water flow channels is extracted from it. Based on the distance metric that reflects the project cost, the candidate edge set is prioritized and evaluated. The corresponding candidate edges are added to the water network topology graph step by step according to the determined priority order until the connectivity reaches the preset index threshold, thus obtaining the optimized water network topology graph. Output a wetland ecological water network structure layout scheme based on the topological connection relationship of the optimized water network topology map.

2. The method according to claim 1, characterized in that, Based on remote sensing images of wetland ecological water networks, water flow channels and water body confluence nodes are identified, and a water network topology map is established, including: Water body mask generation and morphological refinement extraction were performed on remote sensing images of wetland ecological water networks to obtain a water body centerline network. Based on the connection status, intersecting nodes and endpoint nodes are identified in the centerline network of the water body, and the connection between adjacent nodes is identified as the edge of the water network. For a planar water body with an area greater than a preset area threshold, the geometric centroid of the planar water body is extracted as a super node, and the contact point between the boundary of the planar water body and the water body centerline network is used as a supplementary endpoint. An edge connection is established between the supplementary endpoint and the super node. An undirected water network topology graph is constructed based on the intersection nodes, endpoint nodes, supplementary endpoints, super nodes, and the water network edges connecting the nodes.

3. The method according to claim 1, characterized in that, Based on the water network topology diagram, network loop parameters and node connection ratios are calculated to obtain current connectivity evaluation indicators, including: The total number of edges and nodes of the water network are extracted from the water network topology diagram. Based on this, the ratio of the actual number of loops to the theoretical maximum number of loops is calculated to obtain the network loop parameters. Calculate the ratio of the total number of edges to the total number of nodes in the water network to obtain the node connection ratio; The actual connectivity rate is obtained by calculating the ratio of the actual number of existing edges to the theoretical maximum number of edges based on the total number of edges and nodes in the water network. The actual connectivity rate, network loop parameters, and node connection ratio are then weighted and summed to obtain a normalized comprehensive index, which serves as an evaluation index for the current connectivity.

4. The method according to claim 1, characterized in that, Before constructing the geometrically partitioned network based on the planar coordinate set of nodes in the water network topology diagram and pre-configured engineering boundary constraints, the following steps are also included: Perform connectivity component traversal detection on the water network topology graph to identify multiple disconnected water network subgraphs; For disconnected water network subgraphs, calculate the spatial Euclidean distance between nodes contained in different subgraphs, and identify and filter cross-component node pairs whose connections do not cross non-construction obstacles. Select the cross-component node pairs with the shortest spatial Euclidean distance and add interconnecting edges. Iterate and merge the water network subgraphs until a simply connected water network topology graph is formed.

5. The method according to claim 1, characterized in that, A geometrically partitioned network is constructed based on the planar coordinate set of nodes in the water network topology diagram and pre-configured engineering boundary constraints. A candidate edge set, distinct from existing water flow channels, is extracted from this network, including: Using the planar coordinate set of all nodes in the water network topology diagram as the input point set, and the boundary of the study area and the pre-defined non-constructable area as the pre-configured engineering boundary constraints, constrained triangulation calculation is performed to obtain a geometrically partitioned network that conforms to the engineering geological conditions. The edge set contained in the geometrically partitioned network is compared with the current edge set of the water network topology graph. Connecting edges that exist in the geometrically partitioned network but not in the water network topology graph are extracted to form a candidate edge set.

6. The method according to claim 1, characterized in that, Based on a distance metric reflecting project cost, the candidate edge set is prioritized, and the corresponding candidate edges are added to the water network topology graph step by step according to the determined priority order, including: Extract the actual spatial length of each candidate edge in the candidate edge set as a distance metric reflecting the project cost; The edges in the candidate edge set are sorted in ascending order of actual space length from shortest to longest to determine the priority of adding each candidate edge. Following the order of priority from high to low, the corresponding candidate edges are added to the connection relationship of the water network topology graph one by one to dynamically update the water network status.

7. The method according to claim 1, characterized in that, Based on the topological connections of the optimized water network topology map, a wetland ecological water network structure layout scheme is output, which also includes physical spatial verification of the confluence nodes: The number of waterways connected to each intersection node in the optimized water network topology is counted to obtain the node degree. The minimum inscribed circle radius of the intersection node is calculated based on the node degree and the average width of each waterway connecting the corresponding intersection node; wherein, the minimum inscribed circle radius is determined according to the proportional relationship between the product of the node degree and the average width of each waterway and pi. The minimum water surface area of ​​the intersection node is determined based on the pre-configured safety factor and the minimum inscribed circle radius; The actual water surface area of ​​the confluence node is compared with the minimum water surface area of ​​the node. If the actual water surface area is insufficient, the wetland ecological water network structure layout plan will output suggestions for widening or dredging the corresponding confluence area.

8. The method according to claim 3, characterized in that, When calculating the normalized composite index, if there is no prior weight information in the current environment, the actual connectivity rate, network loop parameters, and node connection ratio are weighted and summed with equal weight coefficients.

9. The method according to claim 2, characterized in that, Extract the geometric centroid of the planar water body as a super node, and establish edge connections between the supplementary endpoints and the super node, including: The planar water body is abstracted as a star-shaped topological radial structure with the geometric centroid as the center and the supplementary endpoints as the edge connection points.

10. A computer-readable storage medium storing computer instructions thereon, characterized in that, When the computer instructions are executed by the processor, they implement the method for improving the connectivity of the ecological water network hydrological structure of silty mudflat wetlands as described in any one of claims 1 to 9.

Images