A method and system for automatically generalizing virtual river channels of a plain river network hydrodynamic model

Abstract

The present application relates to a kind of plain river network hydrodynamic model virtual river automatic generalization method and system, it is related to hydrological model construction technical field, it utilizes modeling river center line to divide water production subarea, establishes the topological mapping relationship of point in subarea and river intersection point, generates virtual river center line after collision detection;With water production subarea as statistical unit, the river width is obtained by the virtual river length mean inversion of water area in subarea according to river width;By shoreline extraction, fixed-distance section connection, in-pull buffer intersection positioning and elevation interpolation correction, section line and section point are automatically generated;Finally, virtual river and modeling river data are diagnosed, self-adapting cutting and topological network encapsulation, and complete hydrodynamic topological network is automatically formed, fundamentally solve the problem of low efficiency, unstable quality of artificial generalization.

Description

Technical Field

[0001] This invention relates to the field of hydrological model construction technology, specifically to a method and system for automatic generalization of virtual river channels in a plain river network hydrodynamic model. Background Technology

[0002] Plain river network areas are characterized by flat terrain, dense and crisscrossing waterways, and close hydraulic connections between rivers. Water flow direction is influenced by multiple factors such as tides and sluice gate / pump scheduling, exhibiting a reciprocating flow characteristic. Accurate simulation of these hydrodynamic processes plays a crucial supporting role in flood control and drainage scheduling, optimal water resource allocation, and water environment management. With the deepening of smart water conservancy and digital twin watershed construction, efficiently constructing numerical models that accurately reflect the hydrodynamic characteristics of plain river networks has become an urgent need for the development of water conservancy informatization. The construction of hydrodynamic models relies on a complete river geometric topology network, including basic data such as river centerlines, cross-sectional lines, and cross-sectional points. Only by forming a closed, connected, and computable hydraulic topology can the model solve and operate normally. However, plain river network areas have numerous rivers, and some small and medium-sized rivers often lack detailed measured cross-sectional data and accurate river mapping data, resulting in topological gaps and data blanks during modeling, making it impossible to directly construct a complete hydrodynamic calculation network. To address this issue, a virtual river generalization method is typically used to supplement the missing river segments. This involves generating virtual river geometric elements to fill the topological gaps, while ensuring that the model can reasonably reflect the overall hydraulic connectivity of the region, thereby forming a complete and computable river network structure.

[0003] Currently, the generalization of virtual waterways mainly relies on manual operations by technical personnel based on experience. This includes manually drawing the centerline of the virtual waterway, setting the width and cross-sectional shape of each waterway individually, manually judging topological connections, and manually editing the elevation of cross-sectional points. This manual generalization method has significant shortcomings: First, when the modeling scope involves a large number of missing river sections, the workload of manual processing is enormous, the modeling preprocessing cycle is long and inefficient, and it is difficult to meet the timeliness requirements of rapid modeling in the construction of digital twin watersheds. Second, key aspects such as determining the width of the virtual waterway, assigning cross-sectional shapes, and interpolating elevations lack unified quantitative criteria, resulting in significant differences in the processing results among different operators, making it difficult to guarantee the consistency and objectivity of the generalization quality. Third, problems such as topological connection errors and cross-sectional line intersection conflicts are prone to occur during manual operation, leading to high costs for subsequent troubleshooting and correction.

[0004] Therefore, there is an urgent need for a systematic generalization scheme that can automatically complete the division of water production zones, generation of virtual channels, construction of cross sections, and assignment of elevation values, so as to achieve efficient, standardized, and accurate generalization of virtual channels in the hydrodynamic model of plain river networks. Summary of the Invention

[0005] This application provides a method and system for automatic generalization of virtual river channels in a plain river network hydrodynamic model, in order to at least solve the above-mentioned technical problems existing in the prior art.

[0006] According to a first aspect of this application, a method for automatically generalizing virtual channels in a plain river network hydrodynamic model is provided, comprising: S1, acquiring a basic dataset for river network modeling, the basic dataset including modeling channel water surface data, modeling channel centerline data, water surface data within zones, digital elevation model data, modeling channel cross-sectional line data, and modeling channel cross-sectional point data; S2, performing water yield zone division and topological connection on the modeling channel centerline data based on the modeling channel water surface data to obtain water yield zone data and virtual channel centerline data; S3, based on the spatial statistical range of the water yield zone data, constructing and allocating virtual channel water surfaces and areas on the water surface data within zones and the virtual channel centerline data to obtain virtual channel width values ​​and virtual... S4. Based on the spatial orientation constraints of the water production zone data, shoreline extraction and fixed-distance cross-section connection are performed on the virtual river water surface data and virtual river centerline data to obtain shoreline cross-section point data and virtual river cross-section line data; S5. Using the virtual river centerline data, virtual river width value and digital elevation model data, the shoreline cross-section point data and virtual river cross-section line data are adjusted by inward buffering intersection location and elevation interpolation correction to obtain virtual river cross-section point data; S6. The virtual river centerline data, virtual river cross-section line data and virtual river cross-section point data are fully fused with the modeled river centerline data, modeled river cross-section line data and modeled river cross-section point data to obtain hydrodynamic topology network fused data.

[0007] According to a second aspect of this application, an automatic virtual river channel generalization system for a plain river network hydrodynamic model is provided, comprising: a river network modeling data acquisition module for acquiring a basic dataset for river network modeling, the basic dataset including modeling river channel water surface data, modeling river channel centerline data, water surface data within zones, digital elevation model data, modeling river channel cross-section line data, and modeling river channel cross-section point data; a water yield zone division and connection module for performing water yield zone division and topological connection on the modeling river channel centerline data based on the modeling river channel water surface data to obtain water yield zone data and virtual river channel centerline data; and a virtual river channel construction module for constructing and allocating virtual river channel water surfaces and areas based on the spatial statistical range of the water yield zone data and the virtual river channel centerline data to obtain virtual river channel width values ​​and virtual river channel centerline data. The system includes: a channel water surface data module; a virtual channel cross-section line construction module, used to extract shoreline data and construct fixed-distance cross-section lines from virtual channel water surface data and virtual channel centerline data based on spatial orientation constraints of water production zone data to obtain shoreline cross-section point data and virtual channel cross-section line data; a virtual channel cross-section point construction module, used to locate intersection points and perform elevation interpolation correction on shoreline cross-section point data and virtual channel cross-section line data using virtual channel centerline data, virtual channel width values, and digital elevation model data to obtain virtual channel cross-section point data; and a topology network generation module, used to perform global fusion of virtual channel centerline data, virtual channel cross-section line data, and virtual channel cross-section point data with modeled channel centerline data, modeled channel cross-section line data, and modeled channel cross-section point data to obtain hydrodynamic topology network fused data.

[0008] Compared with existing technologies, this application aims to construct an automatic virtual channel generalization scheme for plain river network hydrodynamic models, addressing the technical problems of low efficiency, poor consistency, and susceptibility to errors caused by the reliance on manual operation in the construction of virtual channel generalization in existing plain river network hydrodynamic models. Its core lies in: generating Thiessen polygons using the modeled channel centerline and clipping them with the outer boundary of the water surface to automatically divide the water-producing zones; establishing a topological mapping relationship between points within the zones and the river intersections; generating virtual channel centerlines through collision detection to eliminate the subjectivity of manual judgment of topological connections; using the water-producing zones as statistical units, inverting the average water area within the zones according to the virtual channel length to obtain the channel width, achieving quantitative and standardized width determination; automatically generating cross-section lines and cross-section points through shoreline extraction, fixed-distance cross-section connection, inward buffer intersection location, and elevation interpolation correction to eliminate the arbitrariness of manually setting cross-section shapes; and finally, performing conflict diagnosis, adaptive clipping, and topological network encapsulation on the virtual channel and modeled channel data to automatically form a complete hydrodynamic topological network. This concept transforms the entire process of virtual river generalization into an automated spatial computing process, fundamentally solving the problems of low efficiency and unstable quality in manual generalization. Attached Figure Description

[0009] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, wherein: in the drawings, the same or corresponding reference numerals denote the same or corresponding parts.

[0010] Figure 1 This is a flowchart of the method for automatic generalization of virtual channels in a plain river network hydrodynamic model according to an embodiment of this application; Figure 2 This is a flowchart illustrating the generation of virtual channel centerline data in the virtual channel automatic generalization method for a plain river network hydrodynamic model according to an embodiment of this application. Figure 3 This is a flowchart illustrating the construction of virtual river surface data in the automatic generalization method for virtual river channels in the hydrodynamic model of a plain river network according to an embodiment of this application. Figure 4 This is a flowchart illustrating the construction of virtual channel cross-section data in the virtual channel automatic generalization method for a plain river network hydrodynamic model according to an embodiment of this application. Figure 5 This is a flowchart of step S5 in the method for automatic generalization of virtual river channels in a plain river network hydrodynamic model according to an embodiment of this application. Detailed Implementation

[0011] To further illustrate the technical means and effects adopted by this application in order to achieve the intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of this application is provided in conjunction with the accompanying drawings and preferred embodiments.

[0012] To address the shortcomings in the aforementioned technical fields, this application provides an automatic generalization method for virtual river channels in a plain river network hydrodynamic model, such as... Figure 1 As shown, Figure 1 This is a flowchart of a method for automatically generalizing virtual river channels in a plain river network hydrodynamic model according to an embodiment of this application.

[0013] Specifically, step S1 involves acquiring the basic dataset for river network modeling. This dataset includes modeling river channel water surface data, modeling river channel centerline data, water surface data within the zones, digital elevation model data, modeling river channel cross-section data, and modeling river channel cross-section point data. It should be understood that the construction of a plain river network hydrodynamic model requires a complete river channel geometric topology, and the formation of this topology relies on the collaborative support of multi-source spatial data. Plain river network areas are characterized by dense river systems and crisscrossing channels. The modeling area includes both major channels with existing measured data and numerous small and medium-sized channels lacking detailed survey data. The subsequent automatic generalization of virtual channels requires relying on the spatial geometric information of existing modeled channels to deduce and generate topological elements for the missing river segments. Simultaneously, the elevation assignment of virtual river channel cross-section points requires topographic data as an interpolation benchmark, and the determination of virtual river channel width requires the water area within the zones as a statistical basis. Therefore, step S1 is executed.

[0014] In one specific implementation, step S1 is performed as follows: The modeled river water surface data is a planar vector feature, representing the planar water surface coverage of each river channel included in the hydrodynamic model calculation within the modeling area. Its geometric shape is a closed polygon extending along the river channel direction, and the attribute fields include river channel code, river channel name, and water surface area. This data is obtained by extracting water bodies from remote sensing images and correcting them with measured river channel boundary lines, or it can be directly retrieved from the existing river management scope vector database of the water resources department. Taking a plain river network modeling area as an example, the modeled river water surface data includes water surface polygons of 56 rivers, with the water surface area of ​​a single river channel ranging from 8,000 square meters to 350,000 square meters.

[0015] The modeled river centerline data are linear vector elements, representing the planar orientation skeleton of each modeled river. Their geometry is a polyline extending along the main channel direction, and attribute fields include river code, river name, and centerline length. This data is obtained by extracting the centerline from the water surface of the modeled river, or by incorporating river planar alignment data provided by the water conservancy surveying department after topological checking. During the data preparation phase, a topological check of the centerlines is required to ensure the absence of short segments, overhanging ends, and overlaps, and that each centerline is located within the corresponding water surface polygon. In the aforementioned modeling area, the total length of the 56 modeled river centerlines is approximately 187 kilometers, with individual centerline lengths ranging from 320 meters to 12,600 meters.

[0016] The water surface data within the partition is represented by isometric vector features, indicating the planar water area of ​​small and medium-sized rivers, ponds, and ditches within the modeled area that are not included in the modeled river channels but actually have water surface coverage. Attribute fields include water area code and water area. This data is derived from the water body classification extraction results of high-resolution remote sensing imagery, and is obtained by performing a spatial difference operation with the water surface data of the modeled river channels, i.e., removing the water areas that belong to the modeled river channels from the overall water body extraction results.

[0017] The digital elevation model (DEM) data is raster-based terrain data, recording the surface elevation information of the modeling area in a regular grid format, with each raster cell storing one elevation value. The spatial resolution of this data is set according to the required modeling accuracy; in plain and river network areas, a DEM with a resolution of 5 to 10 meters is used, achieving an elevation accuracy better than 0.5 meters. This data originates from airborne lidar measurements or aerial photogrammetry.

[0018] The modeled river channel cross-section line data are linear vector elements, representing the positions of the cross-sectional measurement lines laid out on each modeled river channel. Their geometric shape is a line segment perpendicular to the river channel centerline, and the attribute fields include the river channel code and cross-section code. At least two cross-section lines are laid out on each modeled river channel, and each cross-section line intersects the river channel water surface boundary at two points. The modeled river channel cross-section point data are point vector elements, representing the measured elevation control points on each cross-section line. The attribute fields include the cross-section code, point number, and elevation value. Each cross-section line contains 4 to 8 cross-section points, covering both on-shore and underwater points. Both types of data are derived from actual river channel cross-section measurements.

[0019] After data acquisition, the above six types of data need to undergo rationality analysis and quality checks. Specific checks include: verifying the spatial matching relationship between the modeled river centerline and the water surface, ensuring the centerline is located within the corresponding water surface; verifying the orthogonality between the cross-section line and the river centerline, ensuring each cross-section line intersects with exactly one river centerline; verifying the spatial subordination relationship between cross-section points and cross-section lines, ensuring cross-section points are within 5 meters of the corresponding cross-section line and contain valid elevation values; and checking for surface overlap or gaps in the water surface data. Data that does not meet the above conditions should be repaired or removed to form the final basic dataset for river network modeling.

[0020] Specifically, in step S2, the water production zone is divided and topologically connected based on the modeled river water surface data to obtain water production zone data and virtual river centerline data. Figure 2 This is a flowchart illustrating the generation of virtual channel centerline data in the automatic generalization method for virtual channels in a plain river network hydrodynamic model according to an embodiment of this application. Figure 2As shown, the specific data flow and judgment rules are clearly illustrated. The modeling of river centerline data involves two aspects: firstly, dividing the river into regions and encoding the water production zones, then extracting internal points and encoding them as the starting points (internal points of the zones) for subsequent topological connections; secondly, extracting and encoding intersection points through global intersection operations as the ending points (river intersections) for subsequent topological connections. Subsequently, a one-to-many relationship is established between the starting and ending points using spatial inclusion mapping, ultimately generating coherent virtual river centerline data based on this topological relationship. Correspondingly, the construction of a plain river network hydrodynamic model requires the formation of a complete and closed hydraulic topological network. However, the river network modeling basic dataset obtained in step S1 only contains geometric elements of the modeling rivers from existing measured data. A large number of small and medium-sized rivers within the modeling area lack centerline and cross-sectional data, resulting in blank sections in the topological network. The automatic generation of virtual rivers requires first clarifying their spatial scope and topological connection objects, i.e., determining which catchment area each virtual river should serve and which node of the existing river network it should connect to. Therefore, step S2 is introduced.

[0021] In one specific implementation, step S2 includes: generating a Thiessen polygon network using the modeled river centerline data as the generator; using the outer boundary of the modeled river water surface data as a geometric clipping constraint, spatially clipping and encoding the Thiessen polygon network to obtain water production zone data; calculating the geometric centroid of the vertex coordinate sequence of the closed polygon boundary of the water production zone data to extract the generated zone's internal point data. More specifically, step S2 further includes: performing global spatial intersection on the modeled river centerline data to obtain generated river intersection point data; using the water production zone data as a spatial inclusion constraint, performing a one-to-many topological mapping association between the zone's internal point data and the river intersection point data to obtain internal and external node mapping data; based on the internal and external node mapping data, generating directed straight line connections with the zone's internal point data as the topological starting point and the river intersection point data as the topological ending point; and then using the modeled river centerline data as a collision detection benchmark to perform spatial intersection filtering and elimination on the generated preliminary connections to obtain virtual river centerline data.

[0022] The process is as follows: First, a Thiessen polygon network is generated. The Thiessen polygon network is a region partitioning method based on spatial proximity. Its principle is to use discrete nodes on the input linear features as seed points for generators, performing a Voronoi partition on the planar space so that the distance from any position within each generated polygon region to its corresponding generator is less than the distance to any other generator. In this implementation, all polyline nodes on the 56 modeled river centerlines are used as the set of seed points for generators and input into the Thiessen polygon generation algorithm. This algorithm spatially partitions the planar area covered by the modeling region, generating a Thiessen polygon network with each river centerline node as its core. Since the modeled river centerlines are continuous polyline features with densely distributed nodes, the Thiessen polygons generated on both sides of the same river will naturally aggregate into strip-shaped regions extending along the river's direction. The areas between adjacent rivers are divided into independent polygonal units, which form the initial water production zoning prototype.

[0023] Next, spatial clipping and encoding are performed. The outer boundary of the modeled river water surface data refers to the boundary of the region enclosed by the outer contour lines of the 56 modeled river water surface polygons. This boundary divides the modeling area into the internal and external regions of the river water surface. Using this outer boundary as a geometric clipping constraint, a spatial clipping operation is performed on the Thiessen polygon network, retaining the portion of the Thiessen polygon network located outside the outer boundary of the modeled river water surface and removing the portion located inside the river water surface. The discrete polygons obtained after clipping are the water production zone data, with each water production zone representing an independent catchment unit. A unique identifier is assigned to each water production zone polygon generated after clipping, using sequentially increasing integer numbering. In this embodiment, after the 56 modeled river centerlines are segmented and clipped using Thiessen polygons, a total of 42 water production zone polygons are generated, coded as zone 01 to zone 42, with each zone's area ranging from 12,000 square meters to 580,000 square meters. During the encoding and assignment process, the area value of each water production zone polygon needs to be calculated. This area value is calculated using the shoelace formula based on the coordinate sequence of the polygon boundary vertices, and its expression is as follows:

[0024] in The area of ​​the polygon. The total number of vertices on the polygon boundary. and Let x and y be the x and y coordinates of the i-th vertex, respectively. and Let x and y be the x and y coordinates of the (i+1)th vertex, respectively. When i=k, take... , This achieves a closed loop. The formula obtains the area by summing the cross products of the side vectors of the polygon and taking half of the absolute value. The sign of the result depends on the rotation direction of the vertex sequence: positive when arranged counterclockwise and negative when arranged clockwise. The absolute value is the actual area.

[0025] The geometric centroid is calculated again. The point data within each zone are point vector features, representing the geometric center position of each water production zone. For each closed polygon feature in the water production zone data, the centroid coordinates are calculated using a polygon geometric centroid algorithm based on its boundary vertex coordinate sequence. The formula for calculating the centroid's abscissa can be specifically expressed as: The formula for calculating the centroid's ordinate can be specifically expressed as follows: ,in The x-coordinate of the centroid The ordinate of the centroid is y. The directed area of ​​the polygon calculated using the aforementioned shoelace formula. The total number of vertices on the polygon boundary. , Let i be the coordinates of the i-th vertex. , Let be the coordinates of the (i+1)th vertex. This decomposes the polygon into several triangles sharing the origin as a common vertex. The centroids of these triangles are then weighted by area to obtain the geometric centroid of the entire polygon. Taking the water production zone coded as zone 15 as an example, the polygon boundary of this zone contains 23 vertices. The centroid coordinates calculated using the above formula are 456832.6 meters x-coordinate and 3412567.3 meters y-coordinate. This coordinate point is the internal point data corresponding to zone 15. The above calculation is performed on each of the 42 water production zones, generating 42 internal point data points. The code of each internal point corresponds to the code of its respective water production zone.

[0026] Then, a global spatial intersection is performed. River intersection data are point-like vector elements, representing the intersection points between the modeled river centerlines. All segments of the 56 modeled river centerlines are traversed, and each pair is checked to determine if there is a geometric intersection relationship between any two centerlines, including cases where the endpoints of the segments coincide and the segments intersect in the middle. For pairs of segments with an intersection relationship, their intersection coordinates are extracted and the associated river coding information is recorded. After extraction, all intersection coordinates are deduplicated; that is, if the coordinate distance between multiple intersection points is less than 0.1 meters, only one is retained to avoid redundant nodes due to data precision issues. In this embodiment, after global spatial intersection of the 56 modeled river centerlines, 73 river intersection data points are generated. Each intersection point is assigned a unique code, and the coding rule is a combination of the intersection point number and the associated river code.

[0027] Next, a one-to-many topology mapping association is performed. The mapping data for internal and external nodes is an attribute association table, recording the correspondence between points within each partition and their reachable river crossings. For the 42 water production partitions, a spatial inclusion relationship test is performed one by one to determine which of the 73 river crossings are located on or adjacent to the polygon boundary of the current water production partition. The judgment criterion is: if a river crossing is located on the polygon boundary line of the current water production partition, or if the shortest distance from the crossing to the boundary of the water production partition is less than the set adjacency tolerance threshold, then the crossing is considered to have an adjacency relationship with the current partition. The adjacency tolerance threshold is set based on the average width of the modeled river water surface, and in this embodiment, it is set to 15 meters. After the spatial inclusion relationship test, points within the same water production partition are paired with river crossings adjacent to that partition to establish a star-shaped one-to-many association relationship. Taking partition 15 as an example, the internal point code of this partition is internal point 15. After spatial inclusion relationship testing, it is determined that the river intersections adjacent to it are intersection point 08, intersection point 12, and intersection point 35, a total of 3 nodes. Therefore, the internal and external node mapping data records the one-to-three mapping relationship between internal point 15 and intersection points 08, 12, and 35. The above operation is performed on each of the 42 water production partitions, and finally, internal and external node mapping data containing 42 records is generated. Each record contains an internal point code of the partition and a list of several corresponding river intersection code.

[0028] Finally, directed line connection generation and spatial intersection filtering and removal are performed. In the directed line connection generation stage, all records in the internal and external node mapping data are traversed. For each record, the coordinates of the internal points of the partition are used as the starting point of the line segment, and the coordinates of the river intersections associated with that record are used as the ending point. The geometric principle that two points determine a straight line is used to generate directed line elements. Taking partition 15 as an example, the coordinates of internal point 15 are used as the starting point, and the coordinates of intersection points 08, 12, and 35 are connected as the ending points to generate three directed lines as preliminary connections. This operation is performed on each of the 42 water production partitions, generating a total of several preliminary connections. In the collision detection filtering stage, the centerline data of 56 modeled river channels are used as the collision detection benchmark, and spatial intersection judgment is performed on all generated preliminary connections one by one. The judgment criteria are as follows: if a preliminary connection line has a geometric intersection with any modeled river centerline, i.e., the two line segments intersect at non-endpoint positions, the preliminary connection line is considered to have a topological conflict and is discarded; if a preliminary connection line only contacts the modeled river centerline at its endpoints and does not intersect in the middle, the preliminary connection line is considered compliant and is retained. In this embodiment, a total of 126 preliminary connections were generated for 42 water production zones. After collision detection filtering, 47 line segments that had unreasonable intersections with the modeled river centerline were removed, and 79 compliant line segments were retained. The 79 compliant line segments were coded and assigned values. The coding rule is a combination of virtual river identifier prefix, zone internal point code, and river intersection point code, which ultimately generates virtual river centerline data.

[0029] Specifically, in step S3, based on the spatial statistical range of the water production zone data, the virtual river water surface is constructed and the area is allocated for the water surface data and virtual river centerline data within the zone to obtain the virtual river width value and virtual river water surface data. Figure 3 This is a flowchart illustrating the construction of virtual river channel water surface data in the automatic generalization method for virtual river channels in a plain river network hydrodynamic model according to embodiments of this application. (Combined with...) Figure 3It can be seen that this process adopts a two-pronged statistical strategy: firstly, calculating the total area of ​​the internal water bodies within the corresponding water production zone spatial retrieval range; and secondly, calculating the total length of the virtual river channel based on the virtual river centerline. After obtaining the aforementioned two statistical parameters, they are combined to inversely calculate and deduce the virtual river channel width. Then, based on this width value, the centerline is extended to generate a buffer zone, ultimately transforming it into the corresponding virtual river channel water surface data. It is understandable that after step S2 completes the water production zone division and topological connection, 42 water production zone data and 79 virtual river channel centerline data have been generated. However, the virtual river channel centerline is only a one-dimensional linear element and does not possess geometric attributes such as river channel width and water area, so it cannot be directly used for subsequent cross-sectional line construction and hydrodynamic model calculation. In the plain river network area, there are actually a large number of small and medium-sized rivers, ditches, and ponds in each water production zone that are not included in the modeled river channels. The water surface data within these zones reflects the actual water surface coverage of the area, and their total area can be used as the physical basis for determining the virtual river channel width. Step S3 is then introduced.

[0030] In one specific implementation, step S3 includes: using the spatial polygon boundaries of the water production zone data as the basis for grouping statistics, performing polygon surface accumulation and summation statistics on the water surface data within each zone to obtain the total water area data of the zone; then performing line segment length accumulation and summation statistics on the virtual river centerline data belonging to each zone to obtain the total length data of the zone centerline; performing virtual width mean inversion on the total water area data and the total length data of the zone centerline according to the same zone index to obtain the virtual river width value; and performing bidirectional geometric buffer expansion on the virtual river centerline data based on the virtual river width value to obtain the virtual river water surface data.

[0031] The process is as follows: First, perform cumulative summation statistics on polygon faces and cumulative summation statistics on line segment lengths. The total area data of the water area of ​​each zone is attribute-value data, recording the sum of the areas of all water surface polygons within each zone; the total length data of the zone centerline is attribute-value data, recording the sum of the lengths of all virtual river centerline segments within each zone. Using the spatial boundaries of the 42 water production zone polygons as the grouping search range, each water production zone is traversed one by one, and a spatial inclusion relationship query is performed on each zone to retrieve all water surface features within the current zone polygon boundary. The geometric area of ​​each retrieved independent water surface polygon is calculated one by one, using the same shoelace formula as in step S2. The area values ​​of all water surfaces within the same zone are cumulatively summed to obtain the total area of ​​the zone's water area. Taking partition 15 as an example, if this partition contains 12 independent water surface polygons within its polygonal area, with areas of 1260 square meters, 890 square meters, 2150 square meters, 560 square meters, 1780 square meters, 3200 square meters, 670 square meters, 1450 square meters, 980 square meters, 2360 square meters, 1120 square meters, and 1580 square meters respectively, the total water area of ​​partition 15 is 18000 square meters after summing. Using the same spatial grouping criteria, the virtual river centerline data belonging to each water production partition is retrieved, and the geometric length of each line segment is calculated. The line segment length is calculated using the Euclidean distance formula. The distances between adjacent nodes in the broken line element are accumulated segment by segment to obtain the total length of a single centerline. Then, the lengths of all virtual river centerlines within the same partition are summed to obtain the total length of the partition centerline. Taking partition 15 as an example, if there are three virtual river centerlines belonging to this partition, with lengths of 620 meters, 480 meters, and 700 meters respectively, the total length of the partition centerlines of partition 15 is 1800 meters after summing. The above statistical operation is performed on each of the 42 water production partitions, ultimately generating data on the total area of ​​the partition's water body and the total length of the partition centerlines, containing 42 records. Each record uses the water production partition code as its index key.

[0032] Next, the virtual width of the river is averaged and inverted. The total area of ​​the actual water body within the water production zone is considered as the equivalent area that the virtual river surface should cover. This area is then evenly distributed along the total length of the virtual river centerline, thus inverting to obtain the average width of the virtual river. The calculation formula is as follows: ,in This represents the virtual channel width value corresponding to the i-th water production zone, in meters; This represents the total area of ​​the watershed in the i-th water-producing zone, in square meters. This represents the total length of the centerline of the i-th water-producing zone, in meters. The formula converts the two-dimensional area dimension into a one-dimensional width dimension through the ratio of area to length. The result represents the equivalent average width of the virtual channel while maintaining the conserved water area. Taking zone 15 as an example... It is 18,000 square meters. The value is 1800 meters. Substituting this into the formula, we get... =18000 / 1800=10 meters, meaning the theoretical equivalent width of the virtual channel within zone 15 is 10 meters. The above calculation is performed on each of the 42 water production zones. After the calculation, the virtual channel width value needs to be verified for reasonableness. If the calculated width value for a zone is less than 2 meters, it is corrected to the lower limit of 2 meters to avoid generating an excessively narrow virtual channel surface that could lead to subsequent cross-section construction failure. If the width value is greater than 30 meters, it is necessary to check whether there are any anomalies in the water surface data within that zone, and if necessary, correct it to the upper limit of 30 meters.

[0033] Finally, a bidirectional geometric buffer expansion is performed. Using the water production zone data as the mapping index key, the virtual channel width values ​​corresponding to each zone are mapped one-to-one to the virtual channel centerline features belonging to that zone, ensuring that each virtual channel centerline acquires the width attribute of its respective zone. The buffer expansion radius is calculated by uniformly dividing the virtual channel width values ​​by 2, using the following formula: ,in The buffer extension radius of the virtual channel centerline within the i-th water production zone is expressed in meters. This represents the virtual channel width value corresponding to this partition, in meters. The physical meaning of halving the width value in this formula is: the buffer expansion operation expands the channel by the same distance on both sides with the centerline as the axis of symmetry; the sum of the expansion distances on both sides is the full width of the channel. Therefore, the radius of expansion on one side is half the width value. Taking partition 15 as an example... If it is 10 meters, then =10 / 2=5 meters, meaning that the centerlines of the three virtual river channels within zone 15 are each extended 5 meters to both sides. Based on the calculated buffer extension radius, a geometric bidirectional buffer extension algorithm is performed on the virtual river channel centerline data. This algorithm generates arcs with each node on the centerline as the center and the buffer extension radius as the radius, and then connects the arcs of adjacent nodes through common tangents to form a continuous closed polygon outline, thereby converting one-dimensional linear features into two-dimensional planar features. During the buffer extension process, the ends of the line segments are truncated to avoid the end arcs protruding and overlapping with the modeled river water surface. The above buffer extension operation is performed on each of the 79 virtual river channel centerlines, generating a total of 79 virtual river water surface polygons. The attribute fields of each polygon include the water production zone code, virtual river code, and virtual river width value to obtain the virtual river water surface data.

[0034] Specifically, in the basic implementation of step S3, a uniform-width bidirectional buffering strategy is used to generate the virtual river channel water surface by applying a uniform buffer expansion radius to the entire line segment, thus expanding the one-dimensional line segment into a uniform-width rectangular polygon. However, this strategy ignores the spatial confluence and adjacency relationship between the virtual river and the modeled river at their confluence. The endpoint of the virtual river centerline is the river intersection, which is located on the boundary or inside the modeled river water surface. When bidirectional expansion is performed with a constant uniform-width buffer radius, the rectangular polygon generated by the buffer will inevitably intrude into the modeled river water surface in the vicinity of the confluence, forming a topologically illegal geometric overlay region. This directly leads to the hydrodynamic model repeatedly calculating the water area, causing numerical errors in the water conservation equation. At the same time, since the same river intersection may be associated with multiple virtual river centerlines, the uniform-width buffering causes multiple virtual river water surfaces to overlap at the confluence, further exacerbating the degree of topological violation. Furthermore, natural river channels generally exhibit a narrowing effect as they converge into the main channel, while the rectangular water surface generated by the equal-width buffer maintains a constant width at the confluence, significantly deviating from the actual hydraulic morphology. To address these issues, this preferred embodiment employs an adaptive variable-width buffer based on confluence adjacency distance attenuation and a spatial difference pruning method. This method constructs a buffer radius sequence with nonlinear attenuation along the confluence, allowing the virtual river water surface to naturally narrow as it approaches the confluence, eliminating spatial intrusion area from the geometric source. Furthermore, spatial difference operations are used to thoroughly remove residual overlap, ensuring that the virtual river water surface and the modeled river water surface are topologically strictly mutually exclusive and seamlessly adjacent.

[0035] In a specific preferred embodiment, based on the virtual river channel width value, the virtual river channel centerline data is subjected to bidirectional geometric buffering expansion to obtain virtual river channel water surface data, including: Based on the normalized arc length distance ratio from the virtual river centerline data to the endpoint and the adjacency distance of the outer boundary of the modeled river water surface data, a node attenuation weight sequence is generated. First, a dimensionless spatial scale needs to be established for each node on the virtual river centerline to quantify its proximity to the confluence, serving as the basis for subsequent differentiated buffer control. This ensures that nodes closer to the confluence receive smaller weight values, thereby driving the buffer radius to adaptively attenuate in the terminal region. The specific implementation process involves traversing the 28 polygonal nodes on the virtual river centerline and calculating the actual arc length distance accumulated along the arc length direction from the starting point for each node. Then calculate the normalized sink-adjacency distance ratio, the formula of which is: ,in is the normalized convergent adjacency distance ratio at the t-th node, with a value range of [0,1]. The total arc length of the entire segment of the virtual river centerline is 620 meters in this example; This represents the actual arc length accumulated from the starting point of the centerline along the arc length direction to the t-th node, expressed in meters. This formula normalizes spatial location information into a dimensionless ratio by calculating the remaining arc length from the current node to the endpoint and dividing it by the total arc length. A value approaching 1 indicates that the node is located near the starting point of the centerline and far from the intersection. When the value approaches 0, it indicates that the node is extremely close to the endpoint of the intersection.

[0036] Taking the 25th node as an example, its cumulative arc length =558 meters, then =(620-558) / 620=0.1, indicating that the node is only 62 meters away from the junction and is in a high-risk intrusion zone. After obtaining the normalized distance ratio, it is substituted into the nonlinear decay function to generate the node decay weight sequence. The formula for calculating the decay function is: ,in is the width decay weight factor at the t-th node, with a value range of (0,1), representing the proportion of the local width at that node relative to the full width. The steepness control parameter controls the rate at which the attenuation curve transitions from a gentle slope to a steep one; the larger the value, the shorter the transition interval of the river channel narrowing. The normalized distance threshold that triggers width decay represents the inflection point of the decay curve and determines how far from the confluence the river channel begins to narrow significantly. The setting is based on the ratio of the virtual river length to the desired transition zone length, and in this embodiment, it is set to 8. The setting is based on the relative position at which the virtual river channel needs to begin narrowing. In this embodiment, it is set to 0.3, meaning that significant attenuation begins to be triggered when the normalized distance from the node to the endpoint is less than 0.3 (approximately 186 meters). Taking the 25th node as an example... =0.1, substituting into the formula, we get This indicates that the buffer width at that node is only 16.8% of its full width. For the 10th node, its cumulative arc length... =220 meters = (620-220) / 620 = 0.645, substituting into the formula gives... This indicates that the buffer width at this node remains at 94.1% of its full width, almost unaffected by attenuation. The reason for using this S-shaped nonlinear function instead of a simple linear decreasing function is that the narrowing process of a natural river channel does not occur uniformly; it often maintains a stable width in the middle and upper reaches, only showing significant convergence near the confluence at the end. This function can accurately fit this hydraulic law. The above calculations were performed on each of the 28 nodes, generating a node attenuation weight sequence containing 28 weighting factors.

[0037] Using half the virtual channel width as the theoretical maximum buffer radius, the weight factors of each node in the node attenuation weight sequence are scaled point-by-point to obtain an adaptive buffer radius sequence that gradually varies along the channel. Correspondingly, the abstract dimensionless weight factors obtained in the previous step need to be transformed into buffer radius values ​​with actual geometric scale significance, so that each node obtains an independent buffer extension distance, thereby achieving differentiated width control along the channel. Using the virtual channel width value... Half of the theoretical maximum one-sided buffer radius is used as the basis for scaling the weight factors of each node in the node decay weight sequence through point-by-point product scaling. The calculation formula is as follows: ,in Let be the adaptive one-sided buffer expansion radius at the t-th node, in meters; This is the theoretical maximum single-sided buffer radius of the virtual river channel, in this example. =10 meters, then the theoretical maximum radius is 5 meters; This is the width attenuation weight factor for the t-th node calculated in the previous step. This formula maps the dimensionless proportional value of the weight factor to the actual geometric dimension control value through point-by-point multiplication. Taking the 10th node as an example... =5 × 0.941 = 4.705 meters. The full width of the virtual river at this node is 2 × 4.705 = 9.41 meters, close to the full width of 10 meters. Taking the 25th node as an example... =5 × 0.168 = 0.84 meters. The full width of the virtual river channel at this node is 2 × 0.84 = 1.68 meters, which is significantly narrowed. Perform the above calculation on each of the 28 nodes to generate an adaptive buffer radius sequence containing 28 buffer radius values.

[0038] The virtual river centerline data is constructed using a node-by-node widening buffer based on an adaptive buffer radius sequence to obtain initial widened water surface data. Next, the node-by-node differentiated buffer radii obtained in the previous step need to be transformed into actual two-dimensional polygon geometry to generate a spindle-shaped water surface with gradually varying width along the river's course, replacing the rectangular water surface generated by the constant-width buffer. The 28 ordered node coordinate sequences of the virtual river centerline are read, and each node is processed according to its corresponding adaptive buffer radius. Offset along the normal direction to both sides of the centerline. For adjacent nodes... and For the line segment between them, calculate its normal direction unit vector, and offset it along the left normal direction. Obtain the left offset point Offset along the right normal direction Obtain the right offset point Because adjacent nodes have different buffer radii, the offset boundary lines of adjacent line segments are not parallel and equidistant. At the turning point, the intersection of two adjacent offset boundary lines is taken as the actual offset position to ensure the continuity of the contour line. The left offset point sequence is connected in forward order, and the right offset point sequence is connected in reverse order, forming a closed polygon contour. Taking this example, the buffer radius of the area from node 1 to node 15 is all above 4.5 meters, and the width of the generated polygon is close to 10 meters; the buffer radius of the area from node 15 to node 28 gradually decreases from 4.5 meters to 0.84 meters, and the width of the generated polygon gradually narrows from 9 meters to 1.68 meters, presenting an overall spindle-shaped geometric shape that is wide at the starting end and gradually narrows towards the ending end. The area of ​​this initial widened water surface data is approximately 4800 square meters, which is about 22% smaller than the 6200 square meter rectangular water surface generated by the equal-width buffer, effectively reducing the spatial intrusion at the confluence.

[0039] The virtual river surface data is obtained by spatial difference pruning of the initial widening water surface data using modeled river surface data. Although the adaptive widening buffer has significantly reduced the intrusion area at the confluence, due to limitations in discretization sampling accuracy and floating-point operation errors, a small amount of residual overlapping region may still intrude into the modeled river surface at the end boundary of the initial widening water surface. Therefore, a rigorous geometric difference operation is needed to completely eliminate any residual topological violations. The specific implementation process is as follows: using the modeled river surface data as an inviolable spatial mask, a spatial difference pruning operation is performed on the initial widening water surface data. The calculation formula is as follows: ,in The two-dimensional geometric space domain of the final output virtual river surface data after trimming; This refers to the two-dimensional geometric space domain of the initial widened water surface data generated in the previous step. To model the two-dimensional geometric space domain of river water surface data; This is a spatial geometric intersection operator used to extract the common area where the initial widened water surface overlaps with the modeled river channel water surface; It is a spatial geometric difference operator used to peel off the aforementioned overlapping portion from the initial widened water surface.

[0040] First, the geometric intersection of the initial widened water surface and the modeled river surface is calculated to identify all polygonal regions that intrude into the modeled river surface. Then, these intruding regions are precisely trimmed from the initial widened water surface, retaining the compliant portion outside the modeled river surface. In this example, approximately 12 square meters of the initial widened water surface still intrudes into the modeled river surface within about 15 meters of its end. After spatial difference trimming, this intruding region is completely removed, resulting in a virtual river surface area of ​​approximately 4788 square meters. This step ensures that the final output virtual river surface data and the modeled river surface data are topologically absolutely mutually exclusive, with no geometric overlap, meeting the basic requirement of non-overlapping computational domain space in hydrodynamic models. The above four steps are performed on each of the 79 virtual river centerlines, generating 79 virtual river surface polygons. Each polygon exhibits a spindle-shaped geometry that gradually narrows from the starting point to the ending point, and is strictly mutually exclusive and seamlessly adjacent to the modeled river surface.

[0041] Specifically, in step S4, based on the spatial orientation constraints of the water production zone data, shoreline extraction and fixed-distance cross-section connection are performed on the virtual river water surface data and the virtual river centerline data to obtain the shore cross-section point data and the virtual river cross-section line data. Figure 4 This is a flowchart illustrating the construction of virtual channel cross-section data in the automatic generalization method for virtual channels in a plain river network hydrodynamic model according to an embodiment of this application. Figure 4As shown, the virtual river water surface is first transformed into a closed-loop virtual river water line through dimensionality reduction. Simultaneously, a 10m buffer zone is generated at the endpoint of the virtual river centerline. This buffer zone is used to erase closed arc segments on the water line, thus breaking off an undirected virtual river shoreline. Then, directional constraints are used to determine whether the shoreline starting point falls within the corresponding water production zone. If so, the original shoreline direction is maintained to align with the flow direction and assigned the attribute "Shoreline 1". Otherwise, the original shoreline direction is reversed and reconstructed, and the attribute is assigned "Shoreline 2". Based on this, discrete shoreline points are generated parametrically along the line, and according to their odd / even number in the generation order, they are divided into independent starting points and boundary endpoints for sections 1 and 2. Finally, the aligned shoreline points are connected to generate precisely constrained virtual river section line data. It is understandable that the numerical solution of the hydrodynamic model requires the establishment of several cross-sections on each river channel. The cross-section lines are perpendicular to the river channel direction and run through both banks. The cross-section points carry elevation information to describe the cross-sectional morphology of the river channel, and the model constructs discrete computation nodes for the one-dimensional Saint-Venant equations based on this. Step S3 has generated 79 virtual river channel water surface polygons and corresponding virtual river channel width values, but the virtual river channels do not yet have any cross-sectional geometric elements and cannot participate in the numerical discretization and solution of the hydrodynamic equations. The establishment of cross-section lines requires the riverbank as a spatial positioning reference, and the direction definition of the bank requires the spatial orientation of the water production zone as a criterion to ensure that the banks on both sides of the same river channel can be correctly paired to form cross-section lines. Therefore, step S4 is introduced.

[0042] In one specific implementation, step S4 includes: performing dimensionality reduction and stripping of the outer contour boundary of the virtual river water surface data to obtain water area line data; extracting the start and end points of the virtual river centerline data to generate an endpoint buffer with a fixed radius; and using the endpoint buffer to geometrically erase and truncate the water area line data to obtain undirected virtual shoreline data; using the spatial area range of the water production zone data as a Boolean decision constraint, performing spatial inclusion analysis and vector direction orientation assignment on the starting point of the undirected virtual shoreline data to obtain oriented attribute shoreline data; performing arc length parameterization and fixed distance dispersion extraction on the oriented attribute shoreline data to obtain shoreline cross-section point data; and then using the corresponding shoreline cross-section point data on both sides of the same virtual river as the start and end points to perform topological straight line connection and encoding assignment to obtain virtual river cross-section line data.

[0043] The process is as follows: First, dimensionality reduction and geometric erasure are performed on the outer contour boundary. For each of the 79 virtual river surface polygons, a polygon-to-line conversion operation is performed, transforming each closed two-dimensional polygon feature into a one-dimensional closed loop feature that completely coincides with its outer boundary. This loop retains the coordinate sequence and geometric shape of all boundary nodes of the original polygon. Taking the virtual river coded as intersection point 08 of virtual river interior point 15 within partition 15 as an example, its water surface polygon has a width of 10 meters and a length of 620 meters. After the polygon-to-line conversion operation, a closed loop with a perimeter of approximately 1260 meters is obtained, which is the water area boundary line data. Subsequently, the topological start and end node coordinates of the 79 virtual river centerlines are extracted, generating two endpoints for each centerline, for a total of 158 endpoints. A circular buffer operation is performed on each of these 158 endpoints with a fixed radius, generating 158 independent circular endpoint buffers. =10 meters. This fixed radius is set based on the average width of the virtual river channel. A radius of 10 meters ensures that the endpoint buffer completely covers the closed arc segment of the water area boundary line at the center line end. Using the generated endpoint buffer as an erasure mask, a geometric erasure operation is performed on the closed water area boundary line data. This involves deleting the segments of the water area boundary line that fall within the endpoint buffer, while retaining the segments outside the buffer. After the erasure operation, the originally closed circular water area boundary line is cut off at both ends, breaking into two independent, parallel, undirected line segments extending along both sides of the river channel, generating undirected virtual shoreline data. Taking the aforementioned virtual river channel as an example, after approximately 10 meters of arc segments are erased at both the starting and ending points of the closed loop, it breaks into two independent line segments, each approximately 600 meters long, on the left and right banks. Each line segment has a starting and ending point but its direction attribute has not yet been defined.

[0044] Next, spatial inclusion analysis and vector direction orientation assignment are performed. The starting coordinates of each independent line segment in the undirected virtual shoreline data are extracted. Using the spatial region of the water production zone polygon to which the virtual river channel belongs as a Boolean constraint, spatial inclusion analysis is performed on the starting points. The mathematical expression of the decision rule is as follows: ,in For line segments Direction determination value, For line segments The starting coordinates, This refers to the spatial area of ​​the polygon belonging to the water production zone of the virtual river channel. When =+1, meaning the starting point is within the water production zone area, it indicates that the digitized direction of the shoreline is consistent with the flow direction of the virtual river from the internal point of the zone to the river intersection point. The original vector direction of the line segment remains unchanged, and the attribute label "Shoreline 1" is assigned. When the value is -1, it means the starting point is outside the water production zone area, indicating that the digitization direction of the shoreline is opposite to the river flow direction. A geometric reversal reconstruction operation needs to be performed on this line segment, that is, the node coordinate sequence is reversed to flip its direction, and the attribute label "Shoreline 2" is assigned. Taking the two undirected shorelines of the aforementioned virtual river as examples, the starting coordinates of one shoreline are determined to be within the polygon of zone 15 after spatial inclusion analysis, with a determination value of +1, maintaining its original direction and being labeled as shoreline 1; the starting coordinates of the other shoreline are outside the zone 15 area, with a determination value of -1, and after direction flipping, it is labeled as shoreline 2, resulting in directional attribute shoreline data. After the above processing, the shorelines on both sides of the same virtual river have a unified vector direction pointing from inside the zone towards the river intersection point.

[0045] Then, perform arc length parameterized fixed-distance sampling extraction and topological straight-line connection and encoding assignment again. The specific process of arc length parameterized fixed-distance sampling extraction is as follows: traverse each shoreline in the directional attribute shoreline data and obtain its total arc length. Using a fixed percentage =33.3% is used as the discrete spacing parameter to calculate the arc length position of each interpolation node along the shoreline. The calculation formula is as follows: ,in Let be the arc length distance from the k-th interpolation node to the starting point of the shoreline. This is the interpolation node number. For fixed percentage parameters, Let be the total arc length of the shoreline. The planar coordinates of each interpolation node are obtained through curve parametric interpolation, and their expression is: ,in Let be the planar coordinates of the k-th interpolated node. This is a function that parametrically modulates the coastline by its arc length. It returns the coordinates of the corresponding position on the coastline, taking the arc length as the independent variable. Percentage parameter. The basis for setting it to 33.3% is as follows: Without including the start and end points, a spacing of 33.3% generates two interpolation nodes on each shoreline, corresponding to arc length positions of... and This creates two cross-sections on each virtual river channel, satisfying the basic requirement of the hydrodynamic model that each river channel must have at least two cross-sections.

[0046] Taking the aforementioned virtual river channel as an example, the total arc length of shoreline 1 is 600 meters. The first interpolation node is located 200 meters from the starting point, and the second interpolation node is located 400 meters from the starting point. Similarly, two interpolation nodes are generated at corresponding positions for shoreline 2. The generated interpolation nodes are classified as odd or even according to their generation order: the first interpolation node (odd-numbered position) is marked as section 1, and the second interpolation node (even-numbered position) is marked as section 2. These are then merged to generate shoreline section point data. Under the same virtual river channel identifier index, the interpolation node marked as section 1 on shoreline 1 is used as the starting point of the topological connection, and the corresponding interpolation node marked as section 1 on shoreline 2 is used as the ending point of the topological connection. The first cross-sectional line segment is generated using a geometric method that determines a straight line from two points. Similarly, the second cross-sectional line segment is generated using the section 2 node on shoreline 1 as the starting point and the section 2 node on shoreline 2 as the ending point. The generated cross-sectional segments are standardized and coded according to the combination rule of virtual river channel code plus cross-sectional sequence number. Taking the above virtual river channel as an example, the generated two cross-sectional line codes are cross-sectional 1 at intersection point 08 of virtual river channel internal point 15 and cross-sectional 2 at intersection point 08 of virtual river channel internal point 15. The above operation is performed on each of the 79 virtual rivers, generating two cross-sectional lines for each virtual river channel, for a total of 158 virtual river channel cross-sectional lines and 316 on-bank cross-sectional points.

[0047] Specifically, in step S5, using the virtual river channel centerline data, virtual river channel width values, and digital elevation model data, the cross-sectional point data on the bank and the virtual river channel cross-sectional line data are used to perform inward buffering intersection point positioning and elevation interpolation correction to obtain the virtual river channel cross-sectional point data. Figure 5 As shown, Figure 5 This is a flowchart of step S5 in the automatic generalization method for virtual river channels in a plain river network hydrodynamic model according to an embodiment of this application. Accordingly, solving the one-dimensional Saint-Venant equations of the hydrodynamic model requires the placement of several cross-sectional points carrying three-dimensional elevation information along each cross-section line. These cross-sectional points collectively describe the geometry of the river channel cross-section, including the elevation of the land surface on the bank and the elevation of the underwater riverbed. Based on this, the model constructs key hydraulic parameters such as the cross-sectional area and hydraulic radius. Step S4 has generated 316 on-bank cross-sectional points and 158 virtual river channel cross-section lines, but the on-bank cross-sectional points only have planar coordinates and do not carry elevation values. Furthermore, cross-sectional points representing the location of the underwater riverbed have not yet been placed along the cross-section lines, thus failing to constitute a complete cross-sectional geometric description. The underwater topography of rivers with missing measurements in the plain river network area cannot be directly obtained from the digital elevation model to obtain the true riverbed elevation; it needs to be reasonably estimated through inward positioning and depth compensation. Therefore, step S5 is introduced.

[0048] In one specific implementation, step S5 includes: subtracting a fixed inward offset from the virtual channel width value to obtain the underwater buffer control radius; using the underwater buffer control radius to geometrically buffer the virtual channel centerline data to obtain an underwater simulated polygon; then spatially intersecting the outer boundary of the underwater simulated polygon with the virtual channel cross-section line data to generate elevationless underwater cross-section point data; using digital elevation model data as a three-dimensional reference surface, performing bilinear elevation interpolation extraction and thresholding deviation correction on the shore cross-section point data and the elevationless underwater cross-section point data to obtain elevation-bearing shore cross-section point data and initial elevation underwater cross-section point data; uniformly subtracting the deep-water compensation constant from the initial elevation underwater cross-section point data to obtain corrected underwater cross-section point data; then performing point set calculation on the elevation-bearing shore cross-section point data and the corrected underwater cross-section point data to obtain virtual channel cross-section point data.

[0049] The process is as follows: First, geometric buffering and spatial intersection positioning are performed. A fixed inward offset is set to a preset constant value. Its physical meaning is the distance the underwater cross-section point shrinks inward from the shoreline towards the river channel. This ensures the underwater cross-section point is located inside the river's water surface rather than at the shoreline edge, resulting in a reasonable trapezoidal or trough-shaped cross-section. The fixed inward offset is set based on empirical values ​​for the bank slope width of small channels in plain river networks. In this embodiment, it is set to 1 meter, meaning the underwater cross-section point shrinks inward from the shoreline by 0.5 meters (on one side). For each virtual river channel, the virtual river channel width value corresponding to its water production zone is read. Subtract the fixed inward offset from it. The underwater buffer control radius is obtained by the following formula: ,in The underwater buffer control radius of the virtual channel within the i-th water production zone is expressed in meters. This represents the virtual channel width for this zone, in meters. To fix the inward offset, a value of 1 meter is used. Taking partition 15 as an example... =10 meters, then =10-1=9 meters. Using half of this underwater buffer control radius as the actual buffer distance, that is... / 2=9 / 2=4.5 meters. A geometric buffer operation is performed on the centerlines of the three virtual river channels within partition 15, expanding them by 4.5 meters on each side of the centerline axis to generate an underwater simulated polygon with a width of 9 meters. This underwater simulated polygon is 10 meters narrower than the width of the virtual river channel surface polygon, with each side indented by 0.5 meters. Subsequently, the outer boundary lines of the underwater simulated polygons are extracted, and spatial intersection operations are performed between them and the data of 158 virtual river channel cross-section lines. This calculates the geometric intersection point between each cross-section line and the corresponding outer boundary line of the underwater simulated polygon. Since each cross-section line spans the river channel, it intersects with the outer boundary line of the underwater simulated polygon on both the left and right sides, thus generating two underwater cross-section points without elevation for each cross-section line. Taking section 1 at intersection point 08 of virtual river channel 15 within partition 15 as an example, this section line is 10 meters long. After intersecting with the outer boundary line of the underwater simulated polygon with a width of 9 meters, two intersection points are generated at 0.5 meters from the start point and 0.5 meters from the end point of the section line. These two intersection points are the underwater section points without elevation on this section line, possessing only planar coordinates and not carrying elevation values. The above operation is performed on each of the 158 virtual river channel section lines, generating a total of 316 underwater section points without elevation.

[0050] Next, bilinear elevation interpolation extraction and thresholding deviation correction are performed. Using the regular grid of the digital elevation model data as a three-dimensional reference surface, the planar coordinates of the points to be interpolated are... Locate the four neighboring raster cells of the point within the raster grid, read the elevation values ​​of the four corner points, and calculate the elevation value of that point using a bilinear interpolation algorithm. The calculation formula is as follows: ,in For point Interpolated elevation at the location, This is the bilinear interpolation function for the digital elevation model (DEM). It first performs linear interpolation along the horizontal direction on two sets of adjacent grid points, and then performs linear interpolation along the vertical direction on the two horizontal interpolation results, thus obtaining continuous elevation values ​​for any planar location. The above interpolation operation was performed on 316 onshore cross-section points and 316 underwater cross-section points without elevation data to obtain the initial interpolated elevations for each point. For the onshore cross-section points, after obtaining the interpolated elevations, a thresholding deviation correction is required to eliminate the influence of local outliers in the DEM on the cross-section point elevations. The mathematical expression of the correction rule is as follows: ,in The corrected elevation of the onshore cross-section point. The initial elevation is obtained through bilinear interpolation. This serves as the ground benchmark elevation, or reference elevation, for the water-producing zone where this cross-section point is located. This is the set tolerance threshold. Reference elevation. The method for determining the threshold is as follows: All raster values ​​of the digital elevation model within the water production zone are statistically analyzed, and the median value is taken as the ground reference elevation for the zone. Tolerance threshold. Based on the topographic relief characteristics of the plains area, this embodiment uses a value of 1.5 meters. That is, when the absolute value of the deviation between the interpolated elevation and the zoning baseline elevation exceeds 1.5 meters, the interpolation result is considered to be affected by local anomalies and is forcibly corrected to the zoning baseline elevation value. Taking zoning 15 as an example, the ground baseline elevation of this zoning... =3.2 meters, tolerance threshold =1.5 meters. If the interpolated elevation of a certain cross-section point on the bank is 3.5 meters, the deviation is 0.3 meters, which is less than 1.5 meters, then the interpolated elevation of 3.5 meters is retained; if the interpolated elevation of another cross-section point on the bank is 5.8 meters, the deviation is 2.6 meters, which is greater than 1.5 meters, then it is corrected to 3.2 meters. After the above processing, all 316 cross-section points on the bank have obtained corrected elevation values, generating cross-section point data with elevation. The same bilinear interpolation operation is performed on 316 underwater cross-section points without elevation to obtain initial elevation values, generating initial elevation underwater cross-section point data. The elevation values ​​of the underwater cross-section points in this data are the results of surface elevation interpolation and do not yet reflect the actual depth of the underwater riverbed.

[0051] Finally, corrections and point set / merging are performed. The deep-water compensation constant is a preset fixed value, which physically represents the estimated depth of the underwater riverbed relative to the surface elevation of the virtual river channel. It is used to shift the interpolated surface elevation downwards to simulate the actual position of the underwater riverbed. The deep-water compensation constant is set based on the empirical value of the average water depth of small and medium-sized rivers in the plain river network area. In this embodiment, it is set to 2.5 meters, which is applicable to small and medium-sized rivers in the plain with a virtual river channel width ranging from 4 meters to 18 meters. The specific calculation process is as follows: extract the elevation values ​​of each point in the initial elevation underwater cross-section point data, and uniformly subtract the deep-water compensation constant. =2.5 meters, the calculation formula is: ,in The corrected underwater cross-section point elevation. The initial elevation is obtained through bilinear interpolation. This represents the deep-water compensation constant. Taking an underwater cross-section point within zone 15 as an example, its initial interpolated elevation is 3.1 meters. Subtracting 2.5 meters yields a corrected elevation of 0.6 meters, which is the simulated riverbed elevation for that underwater cross-section point. This subtraction operation is performed on each of the 316 initial elevation underwater cross-section points to generate corrected underwater cross-section point data. Subsequently, a point set union operation is performed on the elevation-bearing onshore cross-section point data and the corrected underwater cross-section point data, that is, arranging and integrating the onshore and underwater cross-section points on the same cross-section line according to their spatial location from the starting point to the ending point. Each cross-section line contains 2 onshore cross-section points (located at both ends of the cross-section line) and 2 underwater cross-section points (located inside the cross-section line). After merging, each cross-section line has a total of 4 cross-section points, arranged in the order of onshore cross-section point, underwater cross-section point, underwater cross-section point, and onshore cross-section point from the start to the end, forming a trapezoidal cross-section shape. The above merging operation was performed on each of the 158 virtual river cross-section lines, generating a total of 632 virtual river cross-section points. Each cross-section point carries the cross-section code, point number, and elevation value attributes.

[0052] Specifically, in step S6, the virtual river centerline data, virtual river cross-section data, and virtual river cross-section point data are fully fused with the modeled river centerline data, modeled river cross-section data, and modeled river cross-section point data to obtain the hydrodynamic topology network fused data. That is, the numerical solver for the hydrodynamic model requires the input data to be a complete, closed, and topologically connected hydraulic calculation network. All river centerlines, cross-section lines, and cross-section points in this network must have clear spatial hierarchical and topological adjacency relationships. Any topological breaks or element conflicts will prevent the solver from properly discretizing and iteratively calculating. After processing in steps S2 to S5, virtual river centerline data, virtual river cross-section data, and virtual river cross-section point data have been generated respectively. However, these virtual river elements are still independent of the modeled river elements obtained in step S1, and a unified topological network structure has not yet been formed. At the same time, virtual river cross-section lines may intersect and conflict with modeled river cross-section lines in spatial location. Direct merging without diagnosis and elimination will lead to topological errors. Finally, step S6 is introduced to unify and merge all river channel elements and establish graph theory topological relationships, and finally encapsulate them into complete topological network fusion data that can be directly read by the hydrodynamic solver.

[0053] In one specific implementation, step S6 includes: using the modeled river cross-section data as a static reference, performing spatial topology cross-diagnosis and adaptive pruning and elimination of conflict sections on the virtual river cross-section data to obtain conflict-free virtual cross-section data; merging data elements of the virtual river centerline data, modeled river centerline data, conflict-free virtual cross-section data, modeled river cross-section data, virtual river cross-section point data, and modeled river cross-section point data to obtain global fusion centerline data, global fusion cross-section data, and global fusion cross-section point data; and binding the global fusion centerline data, global fusion cross-section data, and global fusion cross-section point data with a graph theory adjacency matrix and encapsulating them in a topological network to obtain hydrodynamic topological network fusion data.

[0054] The process is as follows: First, spatial topology intersection diagnosis and adaptive clipping and elimination of conflict sections are performed. Using all cross-section lines in the modeled river cross-section data as static reference benchmarks, spatial intersection judgment is performed on each of the 158 virtual river cross-section lines, and the Boolean conflict flag between each virtual cross-section line and all modeled river cross-section lines is calculated. The specific formula for determining the conflict flag is expressed as follows. ,in For conflict flag values, The geometric elements of the virtual river cross-section line to be detected. To model the geometric features of any cross-section line in the river channel cross-section data. This represents the result of the spatial intersection operation between two line segments. When the intersection result is a non-empty set, it indicates that the two cross-sectional lines have a geometric intersection relationship, and the conflict flag is assigned a value of 1; when the intersection result is an empty set, it indicates that the two cross-sectional lines do not have a geometric intersection relationship, and the conflict flag is assigned a value of 0. In this embodiment, after checking each of the 158 virtual river cross-sectional lines, a total of 12 virtual cross-sectional lines had a conflict flag of 1, meaning that they had spatial intersection conflicts with the modeled river cross-sectional lines.

[0055] Perform adaptive clipping and elimination of conflict sections on virtual cross-section lines with a conflict flag of 1. Specifically, for each virtual cross-section line with a conflict flag... Determine the intersection area between it and the modeled river channel cross-section line, and define the conflict influence domain. The spatial extent is defined by the intersection point and a radius equal to half the length of the modeled river channel cross-section. A clipping operation is performed on the virtual cross-section line that causes spatial interference; the calculation formula is as follows: ,in The virtual cross-section line after clipping. For the original virtual cross-section line, To determine the spatial extent of the conflict's impact domain. This refers to the portion of the virtual cross-section line that falls within the conflict influence domain. The formula uses spatial difference operations to remove the portion of the virtual cross-section line that overlaps with the conflict influence domain, retaining the portion unaffected by the conflict. When the length of the remaining segment after trimming is less than one-third of the virtual river channel width, the virtual cross-section line is considered unable to carry valid cross-sectional information and is therefore removed entirely. In this embodiment, after trimming 12 conflicting virtual cross-section lines, 8 retained valid portions, and 4 were removed entirely due to their short remaining length. The 146 conflict-free virtual cross-section lines are merged with the 8 trimmed and retained virtual cross-section lines, totaling 154, to generate conflict-free virtual cross-section line data.

[0056] Secondly, a data element merging operation is performed. In one specific implementation, data elements are merged on the virtual river centerline data, modeled river centerline data, conflict-free virtual cross-section line data, modeled river cross-section line data, virtual river cross-section point data, and modeled river cross-section point data to obtain global fused centerline data, global fused cross-section line data, and global fused cross-section point data. This includes: merging the virtual river centerline data and modeled river centerline data by adding and merging feature layers to obtain global fused centerline data; merging the conflict-free virtual cross-section line data and modeled river cross-section line data by adding and merging feature layers to obtain global fused cross-section line data; and merging the virtual river cross-section point data and modeled river cross-section point data by adding and merging feature layers to obtain global fused cross-section point data. The specific implementation process is divided into three merging operations: First, the system performs element layer appending and merging on 79 virtual river centerline data and 56 modeled river centerline data. The vector element set of the virtual river centerlines is directly merged into the element set of the modeled river centerlines, resulting in a total of 135 centerline elements, generating fused centerline data covering the planar orientation skeleton of all existing and virtual rivers within the modeling area. Second, the system performs element layer appending and merging on 154 conflict-free virtual cross-section line data and modeled river cross-section line data. In this embodiment, there are a total of 186 modeled river cross-section lines, resulting in a total of 340 cross-section line elements, generating fused cross-section line data. Third, before performing cross-section point element appending and integration, the system first performs topological dirty data cleaning on virtual river-related cross-section points that have been completely removed due to spatial conflicts. Because four conflicting virtual cross-section lines that had lost their spatial validity were precisely removed in the preceding conflict resolution steps, and each cross-section line was rigidly attached with four cross-section point parameters (two onshore points and two underwater points) during generation, the system synchronously retrieved, unbound, and removed 16 invalid virtual cross-section points that had become spatially isolated nodes based on the topology dependency mapping table. This cleaning process accurately corrected the number of remaining valid virtual river channel cross-section point data across the entire domain from the original 632 to 616. After cleaning, the system then performed feature layer appending and merging on the retained 616 virtual river channel cross-section point data and the modeled river channel cross-section point data. In this embodiment, there are a total of 1024 modeled river channel cross-section points, which are merged into a total of 1640 cross-section point features, generating global fused cross-section point data. During the merging process, attribute fields need to be standardized to ensure that the coding fields, elevation fields, and subordinate relationship fields of virtual river elements and modeled river elements adopt the same naming rules and data format. Virtual river cross-section points that have lost their corresponding cross-section line affiliation due to cross-section line clipping should be removed simultaneously.

[0057] Finally, graph-theoretic adjacency matrix binding and topological network encapsulation are performed. The hydrodynamic topological network fusion data is a structured topological network data packet, whose data architecture adopts a graph-theoretic triplet representation, denoted as . ,in This serves as the overall structural identifier for the fusion of hydrodynamic topology network data. It is a graph theory node set composed of cross-sectional point data fused across the entire domain. For a graph theory-based set of connected edges composed of cross-sectional line data fused across the entire domain, This is a set of directed principal axes composed of centerline data from the entire domain. The physical meaning of this triplet structure is as follows: The cross-sectional point data from the entire domain serves as the basic computational nodes of the topological network. Each cross-sectional point carries three-dimensional coordinates and elevation attributes, used for solving the cross-sectional area and hydraulic radius in the discrete calculation of the hydrodynamic equations; the cross-sectional line data from the entire domain serves as the spatial connecting edges constraining the geometry of the cross-sections. Each cross-sectional line is associated with several cross-sectional points, defining the spatial range of a single computational cross-section; and the centerline data from the entire domain serves as the directed principal axes defining the flow topology and hydraulic connectivity. Each centerline is associated with several cross-sectional lines, defining the computational path of the water flow propagating longitudinally along the river channel.

[0058] The specific process of graph theory adjacency matrix binding is as follows: Traverse the 135 centerlines in the globally integrated centerline data, retrieve the cross-section lines that have a spatial intersection relationship with each centerline, and establish a one-to-many dependency relationship between the centerline and the cross-section lines, that is, determine which cross-section lines are laid out on each centerline and their arrangement order along the course. Then traverse the 340 cross-section lines in the globally integrated cross-section line data, retrieve the cross-section points on each cross-section line, and establish a one-to-many dependency relationship between the cross-section lines and the cross-section points, that is, determine which cross-section points are included on each cross-section line and their lateral arrangement order. After the dependency relationship is established, an adjacency matrix is ​​constructed to describe the topological connectivity between the centerlines: when two centerlines share the same river intersection coordinates at their endpoints, the two centerlines are considered to have a topological adjacency relationship, and the corresponding position in the adjacency matrix is ​​assigned a value of 1; otherwise, it is assigned a value of 0. In this embodiment, the adjacency matrix formed by the 135 centerlines has a dimension of 135 multiplied by 135, where non-zero elements represent the connectivity between river channels. The specific process of topological network encapsulation is as follows: The aforementioned graph theory node set, edge set, directed principal axis set, and adjacency matrix are serialized and assembled according to the data format required by the hydrodynamic solver, including a node coordinate table, cross-sectional geometric parameter table, channel topology connectivity table, and flow direction definition table, encapsulating them into a complete and closed numerical computation data package. In this data package, each centerline is associated with at least two cross-sectional lines, each cross-sectional line contains at least four cross-sectional points, and all centerlines form a connected graph structure through the adjacency matrix, with no isolated nodes or broken paths. The final generated hydrodynamic topological network fusion data can be directly used as input data for numerical discretization and solution calculation of a one-dimensional hydrodynamic model.

[0059] In summary, the automatic virtual channel generalization method for plain river network hydrodynamic models based on the embodiments of this application is clarified, addressing the technical problems of low efficiency, poor consistency, and susceptibility to errors caused by manual operation in the construction of existing plain river network hydrodynamic models. It utilizes the modeled river centerline to divide water-producing zones, establishes a topological mapping relationship between points within the zones and river intersections, and generates a virtual river centerline through collision detection. Using the water-producing zones as statistical units, the river width is obtained by averaging the water area within the zones according to the virtual river length. Cross-sectional lines and points are automatically generated through shoreline extraction, fixed-distance cross-section connection, inward buffer intersection location, and elevation interpolation correction. Finally, conflict diagnosis, adaptive pruning, and topological network encapsulation are performed on the virtual river and modeled river data to automatically form a complete hydrodynamic topological network, fundamentally solving the problems of low efficiency and unstable quality in manual generalization.

[0060] The virtual channel automatic generalization system for the hydrodynamic model of a plain river network according to this application includes: a river network modeling data acquisition module, used to acquire a basic dataset for river network modeling, which includes modeled river channel water surface data, modeled river channel centerline data, water surface data within zones, digital elevation model data, modeled river channel cross-section line data, and modeled river channel cross-section point data; a water yield zone division and connection module, used to perform water yield zone division and topological connection on the modeled river channel centerline data based on the modeled river channel water surface data to obtain water yield zone data and virtual river channel centerline data; and a virtual river channel construction module, used to construct and allocate the virtual river channel water surface based on the spatial statistical range of the water yield zone data, on the water surface data within the zones and the virtual river channel centerline data to obtain the virtual river channel width value and the virtual river channel water surface. The system includes: a virtual river channel cross-section construction module, used to extract shoreline data and construct fixed-distance cross-sections from virtual river water surface data and virtual river centerline data based on spatial orientation constraints of water production zone data; a virtual river channel cross-section point construction module, used to perform inward buffering intersection location and elevation interpolation correction on the shore cross-section point data and virtual river channel cross-section line data using virtual river centerline data, virtual river width values, and digital elevation model data; and a topology network generation module, used to perform full-domain fusion of virtual river centerline data, virtual river cross-section line data, and virtual river cross-section point data with modeled river centerline data, modeled river cross-section line data, and modeled river cross-section point data to obtain hydrodynamic topology network fused data.

[0061] Here, those skilled in the art will understand that the specific operations of each step in the above-mentioned automatic generalization system for virtual river channels in the plain river network hydrodynamic model have been referenced above. Figures 1 to 5The method for automatic generalization of virtual channels in the hydrodynamic model of plain river networks has been described in detail, and therefore, its repeated description will be omitted.

[0062] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for automatic generalization of virtual river channels in a plain river network hydrodynamic model, characterized in that, include: S1. Obtain the basic dataset for river network modeling. The basic dataset for river network modeling includes water surface data of modeled river channels, centerline data of modeled river channels, water surface data within the zones, digital elevation model data, cross-sectional line data of modeled river channels, and cross-sectional point data of modeled river channels. S2, Based on the modeled river water surface data, the modeled river centerline data is divided into water production zones and connected topologically to obtain water production zone data and virtual river centerline data. S3, based on the spatial statistical range of water production zone data, constructs and allocates virtual river water surface and area for water surface data and virtual river centerline data within the zone to obtain virtual river width value and virtual river water surface data. S4. Based on the spatial orientation constraints of the water production zone data, shoreline extraction and fixed-distance cross-section connection are performed on the virtual river water surface data and virtual river centerline data to obtain the shoreline cross-section point data and virtual river cross-section line data. S5. Using virtual river channel centerline data, virtual river channel width value and digital elevation model data, the onshore cross-section point data and virtual river channel cross-section line data are subjected to inward buffering intersection point positioning and elevation interpolation correction to obtain virtual river channel cross-section point data. This includes: subtracting a fixed inward offset from the virtual river channel width value to obtain the underwater buffer control radius, and using the underwater buffer control radius to perform geometric buffering on the virtual river channel centerline data to obtain an underwater simulated polygon, and then spatially intersecting the outer boundary of the underwater simulated polygon with the virtual river channel cross-section line data to generate elevation-free underwater cross-section point data. Using digital elevation model data as a three-dimensional reference surface, bilinear elevation interpolation and thresholding deviation correction are performed on the onshore cross-section data and the underwater cross-section data without elevation to obtain onshore cross-section data with elevation and underwater cross-section data with initial elevation. The deep-water compensation constant is uniformly subtracted from the underwater cross-section data with initial elevation to obtain corrected underwater cross-section data. Then, the onshore cross-section data with elevation and the corrected underwater cross-section data are combined to obtain virtual river channel cross-section data. S6. The virtual river centerline data, virtual river cross-section data, and virtual river cross-section point data are fused with the modeled river centerline data, modeled river cross-section data, and modeled river cross-section point data in the whole domain to obtain the hydrodynamic topology network fused data.

2. The method for automatic virtual channel generalization in the hydrodynamic model of a plain river network according to claim 1, characterized in that, Step S2 includes: The Thiessen polygon network is generated using modeling river centerline data as generators. Using the outer boundary of the modeled river water surface data as a geometric clipping constraint, the Thiessen polygon network is spatially clipped and encoded to obtain water production zone data. Geometric centroid calculation is performed on the vertex coordinate sequence of the closed polygon boundary of the water production zone data to generate point data inside the zone.

3. The method for automatic virtual channel generalization in the hydrodynamic model of a plain river network according to claim 2, characterized in that, Step S2 also includes: Global spatial intersection is performed on the modeled river centerline data to obtain the river intersection point data; Using water production zone data as spatial inclusion constraints, a one-to-many topological mapping association is performed between the point data within the zone and the river intersection data to obtain the mapping data of internal and external nodes. Based on the mapping data of internal and external nodes, directed straight line connections are generated with the internal point data of the partition as the topology starting point and the river intersection data as the topology ending point. Then, the modeled river centerline data is used as the collision detection benchmark to perform spatial intersection filtering and elimination on the generated preliminary connection to obtain virtual river centerline data.

4. The method for automatic virtual channel generalization in the hydrodynamic model of a plain river network according to claim 1, characterized in that, Step S3 includes: Using the spatial polygon boundaries of the water production zone data as the basis for grouping statistics, the polygon surface data of the water area within each zone is accumulated and summed to obtain the total water area data of the zone. Then, the line segment lengths of the virtual river centerline data belonging to each zone are accumulated and summed to obtain the total length data of the zone centerline. The virtual river width value is obtained by inverting the total area data of the water area of ​​the zone and the total length data of the center line of the zone according to the same zone index and performing mean inversion on the virtual river width. Based on the virtual river width value, the virtual river centerline data is expanded by bidirectional geometric buffering to obtain the virtual river water surface data.

5. The method for automatic generalization of virtual river channels in a plain river network hydrodynamic model according to claim 1, characterized in that, Step S4 includes: The outer contour boundary of the virtual river water surface data is reduced and stripped to obtain the water area line data. Then, the start and end points of the virtual river center line data are extracted and an endpoint buffer is generated with a fixed radius. The endpoint buffer is then used to geometrically erase and truncate the water area line data to obtain undirected virtual shoreline data. Using the spatial area range of the water production zone data as a Boolean constraint, spatial inclusion analysis and vector direction orientation assignment are performed on the starting point of the undirected virtual shoreline data to obtain directional attribute shoreline data. The shoreline data with directional attributes is extracted by arc length parameterization at a fixed distance to obtain the shoreline cross-section point data. Then, the corresponding shoreline cross-section point data on both sides of the same virtual river channel are used as the starting and ending points to perform topological straight line connection and encoding assignment to obtain the virtual river channel cross-section line data.

6. The method for automatic generalization of virtual river channels in a plain river network hydrodynamic model according to claim 1, characterized in that, Step S6 includes: Using the modeled river cross-section data as a static reference, spatial topology intersection diagnosis and adaptive clipping and elimination of conflict sections are performed on the virtual river cross-section data to obtain conflict-free virtual cross-section data. Data elements are merged from virtual river centerline data, modeled river centerline data, conflict-free virtual cross-section line data, modeled river cross-section line data, virtual river cross-section point data, and modeled river cross-section point data to obtain global fusion centerline data, global fusion cross-section line data, and global fusion cross-section point data. The centerline data, cross-sectional line data, and cross-sectional point data of the whole domain are bound by graph theory adjacency matrix and encapsulated in topology network to obtain hydrodynamic topology network fused data.

7. The method for automatic virtual channel generalization in a plain river network hydrodynamic model according to claim 6, characterized in that, Data elements are merged from virtual river centerline data, modeled river centerline data, conflict-free virtual cross-section line data, modeled river cross-section line data, virtual river cross-section point data, and modeled river cross-section point data to obtain globally fused centerline data, globally fused cross-section line data, and globally fused cross-section point data, including: The virtual river centerline data and the modeled river centerline data are merged and appended to feature layers to obtain the global fused centerline data. The conflict-free virtual cross-section line data and the modeled river cross-section line data are merged and added to feature layers to obtain the global fused cross-section line data. The virtual river cross-section point data and the modeled river cross-section point data are merged and added to the feature layer to obtain the global fused cross-section point data.

8. The method for automatic virtual channel generalization in the hydrodynamic model of a plain river network according to claim 4, characterized in that, Based on the virtual river width value, the virtual river centerline data is expanded using a bidirectional geometric buffer to obtain the virtual river water surface data, including: Based on the normalized arc length ratio from the virtual river centerline data to the endpoint and the outer boundary adjacency distance of the modeled river water surface data, a node attenuation weight sequence is generated. Using half the virtual river width as the theoretical maximum buffer radius, the weight factors of each node in the node attenuation weight sequence are scaled point by point to obtain an adaptive buffer radius sequence that gradually changes along the river. The virtual river centerline data is constructed by a node-by-node widening buffer based on the adaptive buffer radius sequence to obtain the initial widened water surface data. Virtual river surface data is obtained by spatial difference pruning of the initial widened water surface data using modeled river surface data.

9. A virtual river channel automatic generalization system for a plain river network hydrodynamic model, characterized in that, include: The river network modeling data acquisition module is used to acquire the basic dataset for river network modeling. The basic dataset for river network modeling includes modeling river water surface data, modeling river centerline data, water surface data within the zones, digital elevation model data, modeling river cross-section line data, and modeling river cross-section point data. The water production zone division and connection module is used to divide the water production zone and connect the topology of the modeled river centerline data based on the modeled river water surface data to obtain water production zone data and virtual river centerline data. The virtual river construction module is used to construct and allocate the virtual river water surface based on the spatial statistical range of water production zone data and the virtual river centerline data within the zone to obtain the virtual river width value and virtual river water surface data. The virtual river cross-section line construction module is used to extract shoreline data and construct fixed-distance cross-section lines based on the spatial orientation constraints of water production zone data and virtual river centerline data to obtain shoreline cross-section point data and virtual river cross-section line data. The virtual river channel cross-section point construction module is used to perform inward buffering intersection point positioning and elevation interpolation correction on the onshore cross-section point data and virtual river channel cross-section line data using virtual river channel centerline data, virtual river channel width values, and digital elevation model data to obtain virtual river channel cross-section point data. This includes: subtracting a fixed inward offset from the virtual river channel width value to obtain the underwater buffer control radius; using the underwater buffer control radius to geometrically buffer the virtual river channel centerline data to obtain an underwater simulated polygon; and then interpolating the outer boundary of the underwater simulated polygon with the virtual river channel cross-section line. Spatial intersection positioning of the data is performed to generate underwater cross-section point data without elevation; using digital elevation model data as a three-dimensional reference surface, bilinear elevation interpolation extraction and thresholding deviation correction are performed on the onshore cross-section point data and the underwater cross-section point data without elevation to obtain onshore cross-section point data with elevation and underwater cross-section point data with initial elevation; the deep-water compensation constant is uniformly subtracted from the underwater cross-section point data with initial elevation to obtain corrected underwater cross-section point data; and the onshore cross-section point data with elevation and the corrected underwater cross-section point data are then combined to obtain virtual river channel cross-section point data. The topology network generation module is used to perform global fusion of virtual river centerline data, virtual river cross-section data, and virtual river cross-section point data with modeled river centerline data, modeled river cross-section data, and modeled river cross-section point data to obtain hydrodynamic topology network fused data.

Images