System and method for simulating basin inundation with fused multi-dimensional features

Abstract

The application relates to the field of water conservancy information technology, and particularly discloses a basin submerging simulation system and method fusing multi-dimensional features, which realizes dynamic adaptation of flood submerging process simulation and optimal evacuation route planning through a simulation, planning, monitoring and adjustment closed loop, achieves the synergistic effect of accurate early warning and dynamic evacuation, can accurately simulate the flood submerging process to perform early warning and dynamically adjust the optimal evacuation route, and improves the pertinence and effectiveness of flood disaster prevention and reduction.

Description

Technical Field

[0001] This invention relates to the field of water conservancy information technology, and in particular to a watershed inundation simulation system and method that integrates multi-dimensional features. Background Technology

[0002] As one of the most destructive natural disasters globally, torrential rain and floods have a profound impact on the safety of people's lives and property, socio-economic development, and the ecological environment within river basins. Simulating flood evolution, enabling early warning, and supporting efficient emergency evacuation are core requirements for disaster prevention and mitigation. Currently, technologies related to river basin flood prevention and mitigation mainly focus on two core areas: flood inundation simulation monitoring and emergency evacuation route planning.

[0003] In the field of flood inundation simulation, the method largely relies on fixed monitoring data from hydrological stations (such as water level and flow data) and measured maps of the watershed topography to simulate flood processes using hydrological models. However, this approach suffers from several drawbacks. Firstly, relying on manual measurements to obtain topographic data is inefficient. For watersheds with complex and extensive topography, the measurement cycle is long and costly, and real-time updates of topographic data are difficult. Secondly, simulation accuracy is limited by the accuracy of the topographic data and the density of hydrological station distribution; simulation errors are significant in areas without station coverage. Thirdly, it cannot correlate the flood inundation process with the floor information of buildings within the watershed, resulting in insufficient targeted early warning information and difficulty in effectively guiding personnel evacuation and property relocation. In the field of emergency evacuation route planning, while it is possible to predict flooded and affected areas, evacuation route planning often relies on pre-set fixed routes. These routes are typically planned on an administrative regional basis, without dynamic adjustments based on real-time flood evolution. Fixed routes may be submerged by sudden rises in floodwater during evacuation, leading to evacuation interruptions and secondary hazards. Furthermore, the method fails to consider the differences in evacuation capacity among different routes within the flooded area, easily causing traffic congestion and reducing evacuation efficiency.

[0004] Therefore, there is an urgent need for a watershed inundation simulation system and method that integrates multi-dimensional features, can improve the accuracy of flood inundation simulation, and can dynamically adjust evacuation routes. Summary of the Invention

[0005] This invention provides a watershed inundation simulation system and method that integrates multi-dimensional features, which can accurately simulate flood inundation for early warning and dynamically adjust evacuation routes.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution: A watershed inundation simulation method incorporating multidimensional features includes the following steps: Step 1, Collect and preprocess multi-source data: Obtain image data and corresponding positioning data of the target watershed through drone aerial photography; collect hydrological data, road network data, and shelter data of the target watershed in real time; establish a building database; Step 2, Constructing a watershed flood inundation model: Based on image data and location data, a watershed flood inundation model including topography, landforms and three-dimensional building morphology is constructed through feature point extraction, image matching and multi-view reconstruction algorithms. The model restores the floor structure of the buildings and the elevation information of each floor. Step 3: Simulate the flood inundation process and plan the optimal evacuation route: Based on hydrological data, evaluate and predict parameters, input these parameters into the watershed flood inundation model to initiate simulation. Dynamically simulate the real-time changes in the flooded area within the target watershed during the water level rise, simultaneously displaying the real-time flooded area. Match flooded buildings and the number of people awaiting rescue from the building database based on the flooded area. Use Dijkstra's algorithm to plan the optimal evacuation route from the flooded building to the shelter for the people awaiting rescue. Set the objective function: Min(T) g ×α)+Min(P×β); Among them, T g To normalize evacuation time, T g =T / T b T is the evacuation time. b The maximum possible evacuation time is preset, where α is the safety factor weight, P is the congestion probability, and β is the traffic efficiency weight. Set constraints: The route is not flooded, and the route's traffic capacity is greater than or equal to the evacuation flow rate. Generate the optimal set of evacuation routes based on the objective function and constraints; Step 4, dynamically adjust the optimal evacuation route: monitor the congestion and flooding status of the optimal evacuation route in real time. If congestion or flooding occurs, automatically trigger the replanning of the optimal evacuation route, update the constraints, and call the Dijkstra algorithm to generate the latest optimal evacuation route. Step 5: Issue warnings and optimal evacuation routes: Visualize the flooded areas, flooded buildings, and optimal evacuation routes.

[0007] The basic principles and beneficial effects of this solution are as follows: Compared to relying solely on fixed hydrological station data or low-resolution remote sensing imagery, this solution can simultaneously collect high-resolution imagery data from UAVs, along with positioning data, hydrological data, road network data, and shelter data, among other multi-source information. This avoids simulation biases caused by missing data and provides data support for accurately identifying flooded areas, associating flooded buildings with the number of people awaiting rescue. The watershed flood inundation model constructed through feature point extraction, image matching, and multi-view reconstruction algorithms can accurately reconstruct the terrain, landforms, and three-dimensional morphology, building structure, and elevation information of buildings compared to two-dimensional drawings or low-precision models. This restores the real-world scene and intuitively presents flooded buildings at different flood stages, making early warning information more targeted. Simultaneously, it can dynamically simulate changes in flooded areas during water level rise, simultaneously presenting key data such as flooded streets and backflow points. Compared to static simulations, it can more accurately capture the real-time evolution of the flood, providing dynamic and accurate decision-making basis for early warning issuance.

[0008] In this scheme, Dijkstra's algorithm is used to plan the optimal evacuation route for people awaiting rescue from flooded buildings to shelters. By using an objective function that minimizes evacuation time multiplied by a safety factor weight and minimizes congestion probability multiplied by traffic efficiency weight, along with constraints such as the route not being flooded and its traffic capacity being greater than or equal to the evacuation flow, the scheme ensures that the optimal evacuation route is both safe and efficient, avoiding the shortcomings of fixed routes that are detached from the actual scenario. Simultaneously, the scheme continuously monitors the traffic status (congestion) and flooding status (submersion) of the optimal evacuation route, enabling the immediate capture of key variables affecting evacuation safety and efficiency, providing real-time data support for adjusting the optimal evacuation route. When congestion or flooding is detected on the optimal evacuation route, the scheme automatically triggers a replanning of the optimal evacuation route. By updating the constraints and re-invoking Dijkstra's algorithm, the scheme generates a new optimal evacuation route, ensuring that the optimal evacuation route always adapts to the flood evolution and road traffic conditions, preventing problems such as evacuation interruptions and secondary dangers caused by fixed routes. In addition, by visually displaying the flooded areas, flooded buildings, and optimal evacuation routes, the system can simultaneously push out early warning information and evacuation guidance, preventing the disconnect between early warnings and routes and avoiding unexpected problems such as getting lost or taking detours due to unclear warnings or routes.

[0009] In summary, this solution achieves dynamic adaptation between flood inundation process simulation and optimal evacuation route planning through a closed loop of simulation, planning, monitoring, and adjustment. It realizes the synergistic effect of accurate early warning and dynamic evacuation, accurately simulates the flood inundation process for early warning, and dynamically adjusts the optimal evacuation route, thereby improving the pertinence and effectiveness of flood prevention and mitigation.

[0010] Furthermore, in step 2, when constructing the watershed flood inundation model, the image data is first subjected to distortion correction, noise reduction, and uniform illumination preprocessing. Then, the SIFT algorithm is used to extract a preset number of feature points from each image. A dense point cloud is generated through multi-view image matching. A TIN triangulation network is constructed based on the dense point cloud, and the watershed flood inundation model is generated through texture mapping. The watershed flood inundation model includes a terrain submodule and a building submodule. The terrain submodule is used to accurately restore the topographic undulation of the watershed, and the building submodule is used to present the floor structure of the buildings and mark the elevation of each floor. It is also associated with the emergency information in the building database. Clicking on a building in the model allows you to view the emergency information.

[0011] The beneficial effects are as follows: By performing distortion correction, noise reduction, and uniform illumination preprocessing, image distortion caused by equipment errors and ambient light during UAV aerial photography is effectively eliminated, ensuring the clarity of image data. Simultaneously, limiting the number of feature points extracted from each image, combined with the strong robustness of the SIFT algorithm, improves the accuracy of multi-view image matching. The generated dense point cloud can accurately restore the detailed features of the watershed terrain and buildings, avoiding modeling deviations caused by poor image quality and insufficient feature points. Through TIN triangulation and texture mapping, the elevation accuracy of the terrain submodule is improved, enabling more accurate restoration of the watershed terrain undulations and ensuring the matching accuracy of flood inundation range, water level, and terrain, avoiding misjudgments of inundated areas due to terrain errors. It can also quickly locate flooded floors, retrieve key information such as the number of people awaiting rescue in the corresponding building, and emergency response personnel, improving the accuracy of early warning information.

[0012] Furthermore, in step 3, the prediction parameters are input through a visual touch screen all-in-one machine. The prediction parameters include the minimum water level, the maximum water level, and the rate of water level rise. After clicking the analysis and simulation button on the visual touch screen all-in-one machine, the flood inundation simulation process is automatically started.

[0013] The benefits are as follows: the simplified operation of parameter input and one-click operation lowers the operating threshold compared to complex parameter configuration and process initiation; in flood emergency scenarios, it can save the time cost of parameter setting and process initiation, realize rapid response to flood inundation simulation, and buy valuable time for subsequent optimal evacuation route planning and adjustment. This convenient triggering mechanism improves emergency response efficiency.

[0014] Furthermore, in step 3, during the dynamic simulation of the water level rise, the water level at the inlet of the main stream and the water level at the confluence of tributaries in the target basin for the future preset time period are calculated using the flood rise prediction algorithm. The propagation time of the flood peak from the inlet of the main stream to the confluence of each tributary is determined using the flood peak extrapolation algorithm. The water level rise process is dynamically presented using the flood inundation algorithm. When the water level reaches the first preset warning level, the flooded buildings are marked. When the water level reaches the second preset warning level, the flood status of the flooded buildings is marked, and the emergency information of the flooded buildings is displayed in conjunction with the information.

[0015] The beneficial effects are as follows: by combining the precise propagation time of the flood peak at the confluence of the main stream and various tributaries, it is possible to accurately predict the pace of flood evolution and grasp the flooding sequence of different areas in advance; based on the propagation time of the flood peak, the evacuation window period of different tributary areas can be planned, and the route can be dynamically adjusted according to the flooded range of different water levels to avoid risks, so that the optimal evacuation route planning can better meet the actual rescue needs and ensure that the route avoids the flooded areas.

[0016] Furthermore, in step 3, during the dynamic simulation of the water level rise, if the water level rises to the highest level, the flood inundation simulation process will automatically terminate, and a flood inundation analysis report for the target basin will be generated simultaneously. The flood inundation analysis report for the target basin includes the flooded area, the flooded buildings, and the number of people awaiting rescue.

[0017] The benefits include: automatic generation of analysis reports, eliminating the need for manual data compilation and organization, saving time in emergency report preparation, and providing standardized data retention.

[0018] Furthermore, in step 5, when issuing an early warning, staff members view the visualized flooded buildings, click the one-click response button, and notify the emergency response personnel of the flooded buildings via VoIP. The call time and content are automatically recorded. At the same time, an early warning notification is pushed to the residents of the flooded buildings. The early warning notification includes information about the flooded buildings, the optimal evacuation route, and the emergency response personnel.

[0019] The beneficial effects are: it replaces the notification model of relaying information through multiple layers, avoiding delays or distortions in information transmission; at the same time, it accurately informs the flooded buildings and the optimal evacuation routes, enabling emergency personnel to quickly initiate personnel transfer work and improving evacuation efficiency.

[0020] Furthermore, in step 5, when the optimal evacuation route is released, the precise navigation information of the optimal evacuation route, which can be viewed offline, is pushed to the mobile APP of the residents of the flooded building. Real-time voice navigation prompts are pushed to the positioning wristbands of special groups in the flooded building, and the map and guidance signs of the optimal evacuation route are displayed on the visual terminal.

[0021] The beneficial effects are as follows: Accurate navigation information that can be viewed offline is pushed to residents' mobile apps, preventing navigation failures caused by network interruptions or poor signals during flood emergencies. Residents can check routes at any time without relying on a real-time network, and the navigation information is updated synchronously with dynamically adjusted optimal routes, ensuring that the safest and most efficient path is always followed during evacuation, avoiding getting lost or taking detours due to unclear routes or network problems. Positioning wristbands equipped for special groups provide real-time voice navigation prompts, providing continuous route guidance without manual operation. This is tailored to the characteristics of people with limited mobility and operational abilities, and the immediacy of voice navigation helps them evacuate quickly in emergency situations.

[0022] Furthermore, in step 3, during the dynamic simulation of the water level rise, inundation data and geographical environment data of the target watershed are recorded; before planning the optimal evacuation route, an inundation risk assessment is conducted. Using the inundation data and geographical environment data, the weights of the assessment indicators are determined using the analytic hierarchy process (AHP). The weights of the assessment indicators include inundation depth weight, rise rate weight, terrain slope weight, and distance from the danger point weight; based on the weights of the assessment indicators, the risk value of each sub-region of the target watershed is calculated, and the risk level is divided. Based on the risk level, the evacuation priority is determined; in step 5, the optimal evacuation route is allocated according to the evacuation priority.

[0023] The beneficial effects are as follows: evacuation priorities are determined based on risk levels, and an evacuation order is established that prioritizes high-risk areas and postpones low-risk areas, preventing resource misallocation caused by indiscriminate coordination; when allocating the optimal evacuation routes, routes with high safety coefficients and fast passage efficiency can be allocated to high-risk areas first, ensuring that evacuation can be carried out in the shortest possible time and improving overall evacuation efficiency.

[0024] Furthermore, in step 2, a fluid dynamics model between buildings, a silt and debris impact load model, and a multi-wave dynamic immersion corrosion damage model are also constructed: The impact intensity of water flow on structures under each wave of flood is calculated, and the bypass effect is determined using the following formula:

[0025] in, Let k be the water flow impact intensity of the k-th flood wave, where k is the flood wave number. Let the water density of the k-th flood wave be . Let S be the real-time water velocity of the k-th flood wave, and S be the water-facing area of ​​the building. This is the drag coefficient around the flow; The impact load of sediment and debris on buildings during each flood wave is calculated using the following formula:

[0026] in, The impact load of sediment and debris from the k-th wave flood. Let be the impact correction factor for the k-th wave flood. Let be the volume fraction of sediment and debris in the k-th flood wave. Let be the average density of sediment and debris in the k-th wave flood. Let be the average particle size of sediment and debris in the k-th wave flood; The cumulative dynamic corrosion damage coefficient of a building after multiple waves of flooding is calculated to quantify the dynamic impact of multi-wave immersion corrosion. The formula is as follows:

[0027] in, To calculate the cumulative dynamic corrosion damage coefficient, Let be the corrosion rate coefficient of the k-th wave flood. The duration of the inundation of the k-th wave of floodwater. , , These represent the normalized pH, normalized dissolved oxygen content, and normalized salinity of the water body during the k-th flood wave. , , Corrosion affects weight. The corrosion resistance coefficient of building materials. Let be the corrosion superposition coefficient of the k-th wave flood; The formula for calculating the overall damage coefficient of a building is as follows:

[0028] in, For wave-order corrosion damage amplification terms, The ultimate impact load of a building. The duration of the k-th flood wave For time step; In step 3, the real-time changes of the flooded buildings under the action of each wave of flood are dynamically simulated, and the simulation results are presented simultaneously.

[0029] The beneficial effects are as follows: By establishing historical-current-multi-wave data correlation through a temporal attention fusion algorithm, it can not only capture the differences in water quality and debris concentration among each wave of flood, but also correlate the state of buildings after the previous wave, providing continuous data support for dynamic cumulative effect simulation and facilitating the refined breakdown of impact mechanics effects. This scheme constructs independent inter-building fluid dynamics models and sediment / debris impact load models to quantify the dual effects of water flow impact and solid particle impact, respectively. The formula for water flow impact intensity is detailed below. The formula for impact load of sediment and debris considers the dynamic changes in water density and flow velocity at different wave levels. By using parameters such as impact correction coefficient and volume fraction, the influence of debris shape and collision angle on impact force is accurately characterized. Compared to the logic of judging flooding solely based on water level and depth, this approach more closely reflects the physical destructive mechanism of real floods. Multi-wave dynamic immersion corrosion damage model Introducing wave-specific corrosion rate coefficient To adapt to the differences in water quality across different waves, through α, The corrosion contribution of each chemical factor is quantified by weighting, and a corrosion superposition coefficient is added. This demonstrates the cumulative effect of initial corrosion leading to decreased material strength and subsequent accelerated corrosion, recreating the actual corrosion process of long-term flood immersion. It constructs a coupling logic between impact and corrosion, using the comprehensive damage coefficient formula for buildings. This achieves the synergistic quantification of dual damage. The coupling relationship between impact-induced surface cracks expanding corrosion area and corrosion-induced material degradation reducing impact resistance can be addressed through (1+) The coefficient incorporates the amplification effect of corrosion damage on overall damage, avoiding errors caused by simulating impact and corrosion separately, thus accurately and dynamically quantifying the cumulative effect of corrosion.

[0030] Meanwhile, reducing dynamic process errors is crucial because the multi-wave nature and dynamic parameters of floods are key factors hindering impact and corrosion simulation. This solution achieves accurate dynamic capture through time-series processing. The flood process is broken down into waves, with each wave independently assigned parameters such as water quality, debris, and flow velocity, rather than treating multiple flood waves as a single entity and ignoring dynamic parameter changes. This ensures that the calculated impact intensity and corrosion rate at different stages are realistic. Based on historical, current, and multi-wave fused data, a water level prediction model is constructed using an improved LSTM network. This model accurately captures the water level change trends and inundation area expansion patterns of each flood wave, providing a precise time-series reference for the dynamic calculation of impact force and corrosion duration. This avoids the errors of static models failing to adapt to flood evolution and improves prediction accuracy. Within the duration of each flood wave, the calculation is based on... To refine the time-step iterative calculation of impact force and perform time-series dynamic simulation, the mechanical parameters of key scenarios such as instantaneous strong impacts are not averaged out, thus improving the time accuracy of impact simulation. Attached Figure Description

[0031] Figure 1 The flowchart is for Example 1 of a watershed inundation simulation method that integrates multi-dimensional features. Detailed Implementation

[0032] The following detailed description illustrates the specific implementation method: Example 1 This scheme integrates a multi-dimensional watershed inundation simulation method, as shown in the appendix.Figure 1 As shown, the specific implementation process is as follows: Step 1: Collect and preprocess multi-source data: Acquire image data and corresponding location data of the target watershed through drone aerial photography; collect hydrological data, road network data, and shelter data of the target watershed in real time; establish a building database. Hydrological data such as water level, flow rate, and rainfall are collected in real time through sensors such as water level gauges, flow meters, and rain gauges deployed within the watershed and transmitted wirelessly to the system; real-time road conditions are obtained by connecting to the traffic management system API, and road network data is collected by combining drone aerial photography with updated road traffic status; shelter data is collected by synchronizing shelter location, capacity, and supplies information through the emergency management platform, and is regularly verified and updated on-site. Integrate building archives from the Natural Resources Bureau and 3D building information from drone aerial photography to collect building location, floor, use, and emergency information, and form a dynamically updated database after standardized input, thus establishing a building database.

[0033] Step 2, Constructing a watershed flood inundation model: Based on image data and location data, a watershed flood inundation model including topography, landforms and three-dimensional building morphology is constructed through feature point extraction, image matching and multi-view reconstruction algorithms. The model restores the floor structure of the buildings and the elevation information of each floor.

[0034] In step 2, when constructing the watershed flood inundation model, the image data is first preprocessed with distortion correction, noise reduction, and uniform illumination. Then, the SIFT algorithm is used to extract at least a preset number of feature points from each image. A dense point cloud is generated through multi-view image matching. A TIN triangulation network is constructed based on the dense point cloud, and the watershed flood inundation model is generated through texture mapping. The watershed flood inundation model includes a terrain submodule and a building submodule. The terrain submodule is used to accurately restore the topographic undulations of the watershed, and the building submodule is used to present the floor structure of buildings and label the elevation of each floor. It is also linked to emergency information from the building database; clicking on a building in the model allows users to view emergency information. Through distortion correction, noise reduction, and uniform illumination preprocessing, image distortion caused by equipment errors, ambient light, and other factors during UAV aerial photography is effectively eliminated, ensuring the clarity of the image data. Simultaneously, limiting the number of feature points extracted from each image, combined with the strong robustness of the SIFT algorithm, improves the accuracy of multi-view image matching. The generated dense point cloud can accurately restore the detailed features of the watershed topography and buildings, avoiding modeling deviations caused by poor image quality and insufficient feature points. By constructing TIN triangulation and mapping textures, the elevation accuracy of the terrain submodule is improved, which can more accurately restore the topographic undulation of the watershed, ensure the matching accuracy of flood inundation range, water level and terrain, and avoid misjudgment of inundation area due to terrain error; it can quickly locate flooded floors, retrieve key information such as the number of people waiting to be rescued in the corresponding building and emergency personnel in charge, and improve the accuracy of early warning information.

[0035] Step 3: Simulate the flood inundation process and plan the optimal evacuation route: Based on hydrological data, evaluate and predict parameters, input these parameters into the watershed flood inundation model to initiate simulation. Dynamically simulate the real-time changes in the flooded area within the target watershed during the water level rise, simultaneously displaying the real-time flooded area. Match flooded buildings and the number of people awaiting rescue from the building database based on the flooded area. Use Dijkstra's algorithm to plan the optimal evacuation route from the flooded building to the shelter for the people awaiting rescue. Set the objective function: Min(T) g ×α)+Min(P×β); Among them, T g To normalize evacuation time, T g =T / T b T is the evacuation time. b The maximum possible evacuation time is preset, where α is the safety factor weight, P is the congestion probability, and β is the traffic efficiency weight. Set constraints: The route is not flooded, and the route's traffic capacity is greater than or equal to the evacuation flow rate. Generate the optimal evacuation route set based on the objective function and constraints.

[0036] In this embodiment, the evacuation time T and congestion probability P are calculated based on multi-source data. The calculation of evacuation time T is as follows: The actual length L (in meters) of the road segment is extracted from the preprocessed road network data. Combined with road safety standards under flood scenarios (such as avoiding slips and trampling caused by rapid passage), a safe passage speed V is set (taken as 1.2-1.5 m / s, which can be adjusted according to road type: 1.5 m / s for flat main roads and 1.2 m / s for narrow side roads). The calculation formula is as follows: T is in seconds; II. The congestion probability P is calculated as follows: Extract the design capacity C of the road segment from the road network data (unit: people / minute, i.e., the maximum number of evacuees the road segment can carry per unit time), and extract the corresponding evacuation flow Q of the road segment from the personnel distribution data (unit: people / minute, i.e., the number of evacuees who need to pass through the road segment). The calculation formula is as follows:

[0037] The value of P can be adjusted according to the actual needs of the scenario. The core logic is that the closer the pedestrian flow is to the traffic capacity, the higher the probability of congestion.

[0038] Dijkstra's algorithm is a classic single-source shortest path algorithm. Its core application is point-to-point evacuation route planning between flooded buildings and emergency shelters. Before the algorithm is launched, key preparations have been completed: road network topology data, flooded area data, road capacity data, personnel distribution data, and shelter capacity data have been acquired, and multi-source data fusion and standardization have been completed.

[0039] Step a: Construct a weighted directed road network graph for evacuation route planning

[0040] The core computational vehicle of the algorithm is the road network topology graph. The watershed's road network, flooding status, and traffic attributes must first be transformed into a graph structure that the algorithm can recognize. This is the foundation of route planning, defined as follows: Vertex set Abstract all core nodes into vertices of a graph, containing three types of key nodes: the starting point set, which includes the entrances and exits of all flooded buildings, such as the main entrance of Building 1 in XX Community; the passing nodes, which include road intersections, bridges, tunnels, and key turning points in the road network; and the ending point set, which includes all available emergency shelters within the basin, including community shelters and public venue shelters. edge set Abstract the passable roads between any two adjacent vertices into edges of the graph, retain only the roads that have not been flooded, satisfy the constraint that the route is not flooded, and remove the edges corresponding to flooded or collapsed road sections. Weight set Assign a composite weight value to each valid edge. The weight values ​​are used to calculate two types of core factors, as shown in the following formula: ; in, Let be the edge weight from vertex i to vertex j; g represents the normalized evacuation time for this road section. g= / T b , T is the evacuation time for this section of road. b The maximum possible evacuation time is preset; Weighted by the safety factor; This represents the probability of congestion on this road segment (values ​​range from 0 to 1). Weighted by traffic efficiency; (Weight allocation can be adjusted as needed. Evacuation time) Actual distance of road segment ÷ safe road speed; probability of congestion Based on calculations combining road capacity and regional pedestrian flow, the closer the pedestrian flow is to the capacity, the higher the probability of congestion.

[0041] Step b: Define the core constraints of the algorithm input. Before Dijkstra's algorithm is started, two rigid constraints are fixed and must be met by all planned routes to filter out invalid routes: Core constraint 1, the planned evacuation route has no flooded sections, that is, all edges corresponding to the route are valid edges retained in step 1; Core constraint 2, for each road segment along the route, the actual evacuation flow is less than or equal to the road's design capacity to avoid congestion and stampedes due to overloading; Additional constraint, if a flood risk assessment has been completed, an evacuation priority constraint must be added simultaneously, and buildings in high-risk areas should be given priority for routes with higher weights.

[0042] Step c: Initialize the core parameters of the algorithm We perform basic assignments to Dijkstra's algorithm to prepare for shortest path search. The unified initialization rule is as follows: set a distance value for each vertex (node). The distance value of the starting point (the flooded building) is initialized to... (No distance between itself), the distance values ​​of all other vertices are initialized to... (Infinity represents an initially unknown path); set a predecessor node identifier for each vertex. All vertex predecessor nodes are initialized to (Empty), used for subsequent backtracking of the complete route; construct a set S of unvisited nodes, include all vertices in set S, and gradually remove nodes with the determined shortest path from the set during the algorithm iteration process.

[0043] Steps: Iteratively search for the shortest weighted path for each node. Starting with a single flooded building (the starting point), the following operations are performed iteratively until all accessible shelters (the ending point) are traversed. Essentially, this is a greedy strategy to select the optimal edge, ensuring that each step chooses the path with the smallest current weight: The vertex u with the smallest distance value is selected from the unvisited set S, marked as visited, and removed from S; all adjacent vertices v of vertex u are traversed. If v is still in the unvisited set S, and the edge from vertex u to v is a valid edge (satisfying the constraints), then the candidate distance value is calculated: [Candidate...] g; Let 'a' be the normalized composite weight value of the edge between vertex u and vertex v. For normalized distance values, = / disb, where disb is the preset maximum distance value; = / , Let be the composite weight of the edge between vertex u and vertex v. The preset maximum composite weight value; If alternatives Then update the alternatives. Simultaneously update (Record the predecessor node of v as u, and lock in a better path); if alternatives are available... If the original distance value is not updated, the original distance value is retained. Repeat the above steps until the unvisited set S is empty, thus completing the path search for a single starting point (a single flooded building).

[0044] Step e: Backtrack to generate the complete evacuation route Based on the results obtained after algorithm iteration (Predecessor node array), tracing back from the target endpoint (shelter) to the starting point (flooded building), reconstructing the complete route trajectory. The operation is as follows: Select the optimal endpoint, filter all reachable shelters by distance value. The node with the smallest weight is selected as the optimal shelter (optimal weight = optimal route); path backtracking begins from the optimal shelter, and predecessor nodes are retrieved sequentially. The process continues until the starting point (the flooded building) is reached. The path is then sorted in the forward direction, and the node sequence obtained from the backtracking is reversed to form a complete and forward optimal evacuation route: flooded building → passing node 1 → passing node 2 → … → optimal shelter.

[0045] Step f: Batch generate the optimal evacuation route set Repeat step ce to complete the two batch processing steps and finally output the optimal evacuation route set: For each flooded building in the watershed, execute the algorithm as the starting point to generate the optimal route for each building; for each building, if there are multiple accessible shelters, select the top N routes with the best weights and include them in the route set; the route set includes the starting point, ending point, nodes passed through, total weight value, estimated evacuation time, safety factor, and congestion probability of each route.

[0046] Step g: Route verification and secondary optimization The initial route set generated by the algorithm undergoes compliance verification to filter out routes that do not meet the constraints, ensuring that the routes are feasible for implementation. The verification and optimization rules are as follows: Constraint verification: Each route is checked to see if it is not flooded and if the road capacity is greater than or equal to the pedestrian flow, and routes that violate the rules are removed; Shelter adaptation verification: The remaining capacity of the shelter at the end of the route is checked to see if it is greater than or equal to the number of people to be evacuated from the building. If the shelter is saturated, the route is removed and a second-best route is selected from the route set; Special population adaptation optimization: For special populations such as the elderly, children, and patients in the building, the routes are fine-tuned, such as prioritizing routes with a slope of ≤5°, no narrow sections, and shorter distances.

[0047] Step h: Output the final optimal evacuation route The verified optimal evacuation route set is simultaneously output to two main modules to complete the planning loop: output to the dynamic adjustment module to provide an initial route benchmark for subsequent route replanning after congestion / flooding; output to the visualization publishing module to convert the routes into intuitive map trajectories and push them to residents' apps, visualization terminals, and positioning wristbands to realize route publicity.

[0048] When route congestion or flooding is detected, the Dijkstra algorithm can be quickly reused to update the route without re-initializing all parameters, significantly improving adjustment efficiency. When an adjustment is triggered, only the weights of the affected road segments are modified; congested road segments will have their congestion probability adjusted. The system adjusts the parameters accordingly, marking flooded road sections as invalid edges. Using the original route's starting point (flooded building) as a reference, step dg is re-executed, only iterating and updating the distance values ​​of affected nodes without traversing the entire road network. The system automatically replaces the original route by comparing the weight values ​​of the old and new routes, generating the latest optimal evacuation route and updating the visualization content simultaneously.

[0049] In step 3, the predicted parameters are input via a visual touchscreen all-in-one machine. These parameters include the minimum water level, the maximum water level, and the rate of water level rise. Clicking the analysis and simulation button on the all-in-one machine automatically initiates the flood inundation simulation process. In flood emergency scenarios, this saves time on parameter setting and process initiation, enabling rapid response to flood inundation simulations and providing valuable time for subsequent optimal evacuation route planning and adjustments. This convenient triggering mechanism improves emergency response efficiency. The predicted parameters are obtained through evaluation by a flood peak projection intelligent agent. The evaluation includes: acquiring real-time data on current flood peak flow, interval rainfall, and station rainfall, as well as historical flood peak corresponding factors and historical data on the maximum water level; performing preprocessing on the data, including interpolation, format unification, and normalization; and using formulas based on the preprocessed data.

[0050] Calculate the data connectivity, where, Data connectivity is used to measure the similarity between current data A and historical data B. These are the homoegenic components (terms with consistent features) of the current data and historical data after normalization. These are the heterogeneous components (items with different features) between the current data and historical data after normalization. These are the inverse components (terms with completely opposite characteristics) of the current data and historical data after normalization. is the total number of data items; i is used to identify anomalous components (representing the parts where the current data differs from historical data); Used to identify reversal components (representing the part that is completely opposite to the historical data); The coefficients represent the proportions of the isotropic, heterotropic, and antitropic components, respectively. , , ; Historical water level data corresponding to the maximum connectivity is selected, and the prediction parameters are obtained by combining them with the current data. Based on the prediction parameters, a three-dimensional dynamic simulation of flood inundation is realized.

[0051] In step 3, during the dynamic simulation of the water level rise, the water level at the inlet of the main stream and the water level at the confluence of tributaries in the target basin for the future time period are calculated using a rise estimation algorithm. This rise estimation algorithm is a specific implementation of the "similarity analogy method" in hydrological forecasting. Its core logic is to quantify data similarity using "set pair analysis" and match historical flood scenarios to estimate future water levels. The correlation between "current flood data (A)" and "historical flood data (B)" is calculated through set pair analysis (using the formula...). The system identifies the most relevant historical flood scenarios and uses the water level of these similar scenarios as a benchmark. By combining real-time data such as current peak flow and rainfall, it can correct the water levels at the main stream inlet and tributary confluence in future periods, enabling rapid flood prediction and meeting the timeliness requirements of dynamic water level simulation in the basin.

[0052] The propagation time of the flood peak from the main stream inlet to the confluence of tributaries is determined by a flood peak projection algorithm, and the water level rise process is dynamically presented by a flood inundation algorithm. The flood peak projection algorithm selects historical data with the highest correlation, uses its water level as a benchmark, and combines it with current data to obtain the water levels at the main stream inlet and tributary confluences for future periods. This data is ultimately used for three-dimensional inundation simulation. The correlation calculation formula is as follows:

[0053] By leveraging the similarity of historical data, complex hydrological processes can be simplified, and future water levels within a watershed can be quickly predicted, such as at the inlet of the main stream and the confluence of tributaries. When the water level reaches the first preset warning level, the flooded buildings are marked; when the water level reaches the second preset warning level, the flooding status of the flooded buildings is marked, and emergency information for the flooded buildings is displayed in conjunction with the warning. This enables accurate prediction of the flood's evolution and advance understanding of the flooding sequence in different areas; it allows for planning evacuation windows for different tributary areas based on the propagation time of the flood peak, and dynamic adjustment of routes to avoid risks based on the flooded area at different water levels. This ensures that the optimal evacuation route planning aligns with actual rescue needs and that routes avoid flooded areas.

[0054] In step 3, during the dynamic simulation of the rising water level, if the water level reaches its highest point, the flood inundation simulation process automatically terminates, and a flood inundation analysis report for the target basin is generated simultaneously. This report includes the inundated areas, inundated buildings, and the number of people awaiting rescue. The automatic generation of the analysis report eliminates the need for manual data compilation, saving time in emergency scenario report preparation and providing standardized data retention.

[0055] Step 4, dynamically adjust the optimal evacuation route: monitor the congestion and flooding status of the optimal evacuation route in real time. If congestion or flooding occurs, automatically trigger the replanning of the optimal evacuation route, update the constraints, and call the Dijkstra algorithm to generate the latest optimal evacuation route.

[0056] Step 5: Issue warnings and optimal evacuation routes: Visualize the flooded areas, flooded buildings, and optimal evacuation routes.

[0057] In step 5, when issuing an early warning, staff view the visualized flooded buildings and click the one-click response button to notify the emergency response personnel of the flooded buildings via VoIP. The call time and content are automatically recorded. Simultaneously, an early warning notification is sent to residents of the flooded buildings, containing information about the flooded buildings, the optimal evacuation route, and the emergency response personnel. This replaces the traditional multi-level notification system, avoiding information delays or distortions. Furthermore, by accurately informing residents of the flooded buildings and the optimal evacuation route, the emergency response personnel can quickly initiate evacuation operations, improving evacuation efficiency.

[0058] When the optimal evacuation route is released, precise navigation information for the optimal evacuation route, which can be viewed offline, is pushed to the mobile apps of residents in flooded buildings. Real-time voice navigation prompts are also pushed to the location wristbands of vulnerable groups in flooded buildings, and a map and directional markers for the optimal evacuation route are displayed on a visual terminal. Pushing precise navigation information to residents' mobile apps, which can be viewed offline, prevents navigation failure due to network interruptions or poor signal in flood emergencies. Residents can view the route at any time without relying on a real-time network, and the navigation information is updated synchronously with the dynamically adjusted optimal route, ensuring that the safest and most efficient path is always followed during evacuation, avoiding getting lost or going around due to unclear routes or network problems. The location wristbands equipped for vulnerable groups push real-time voice navigation prompts, providing continuous route guidance without manual operation. This is suitable for the characteristics of vulnerable groups with limited mobility and operational abilities, and the immediacy of voice navigation helps them evacuate quickly in emergency situations.

[0059] Example 2 The only difference from Example 1 is that in step 3, during the dynamic simulation of the water level rise, inundation data and geographical environment data of the target watershed are recorded; before planning the optimal evacuation route, an inundation risk assessment is conducted. Using the inundation data and geographical environment data, the weights of the assessment indicators are determined using the analytic hierarchy process (AHP). These weights include inundation depth weight, rise rate weight, terrain slope weight, and distance from the danger point weight. Based on these weights, the risk value of each sub-region of the target watershed is calculated, and risk levels are assigned. Evacuation priorities are then determined based on these risk levels. In step 5, the optimal evacuation route is allocated according to the evacuation priority. Determining the evacuation priority based on risk levels establishes an evacuation order that prioritizes high-risk areas and postpones low-risk areas, preventing resource misallocation caused by indiscriminate planning. When allocating the optimal evacuation route, routes with high safety coefficients and fast passage efficiency can be prioritized for high-risk areas to ensure evacuation in the shortest possible time and improve overall evacuation efficiency.

[0060] In step 2, a hydrodynamic model between buildings, a silt and debris impact load model, and a multi-wave dynamic immersion corrosion damage model are also constructed: The impact intensity of water flow on structures under each wave of flood is calculated, and the bypass effect is determined using the following formula:

[0061] in, The water flow impact intensity of the k-th flood wave is expressed in nanometers (N), where k is the flood wave number. The density of the water in the k-th flood wave is expressed in kg / m³. 3 S represents the real-time water velocity of the k-th flood wave, in m / s, and S represents the water-facing area of ​​the structure, in m². 2 This is the drag coefficient around the flow; The water density of the k-th wave flood is collected in real time using hydrological monitoring sensors (such as densitometers); The water velocity of the k-th wave flood is monitored in real time using equipment such as hydrological station current meters and Doppler current sensors. The projected area of ​​the water-facing side is calculated by retrieving the facade dimension data of the flooded building from the building's comprehensive database; The value is determined based on industry standards (such as the "Code for Design of Building Structures") and the shape of the building's water-facing side (such as rectangular or circular). (For example, a value of 1.2-1.5 is typically used for rectangular buildings.) The impact load of sediment and debris on buildings during each flood wave is calculated using the following formula:

[0062] in, The sediment and debris impact load of the k-th flood wave is expressed in nanometers (N). Let be the impact correction factor for the k-th wave flood. Let be the volume fraction of sediment and debris in the k-th flood wave. The average density of sediment and debris in the kth wave flood is given in kg / m³. 3 The average particle size of sediment and debris in the k-th flood wave is given in meters. Calibration can be performed through indoor simulation experiments (simulating debris collision scenarios of different flood waves), or by referring to the measured data of debris impact from historical floods in the basin. Water samples were collected from the kth wave of flood using a sampler, and the volume percentage of sediment and debris in the water was calculated through laboratory analysis. Density measurements were performed on the collected mud and sand debris samples, and the average value was taken. The particle size distribution of sediment samples was measured and the average value was calculated using equipment such as sieving and laser particle size analyzers. The cumulative dynamic corrosion damage coefficient of buildings after multiple waves of flooding was calculated to quantify the dynamic impact of multi-wave immersion corrosion. No. The corrosion damage caused by wave flooding alone is as follows:

[0063] Calculate the sum of all waves of corrosion:

[0064] in, For the first Corrosion damage caused solely by wave floods To calculate the cumulative dynamic corrosion damage coefficient, Let be the corrosion rate coefficient of the k-th wave flood. The duration of immersion in the k-th wave of floodwater, in hours. , , These represent the normalized pH, normalized dissolved oxygen content, and normalized salinity of the water body during the k-th flood wave. , , Corrosion affects weight. The corrosion resistance coefficient of building materials. Let be the corrosion superposition coefficient of the k-th wave flood; The determination is based on the type of building material (such as reinforced concrete and wood) and combined with laboratory corrosion experiments (simulating immersion scenarios in corresponding water quality); The duration of the flooding of buildings by the kth wave of floodwaters was determined using water level monitoring data. The pH value, dissolved oxygen, and salinity of the k-th wave flood were collected in real time using water quality monitoring sensors. The weighting percentages were determined by analyzing the influence of different chemical parameters on corrosion using orthogonal experiments. It is determined based on material standards (such as the "Code for Design of Concrete Durability"), combined with the type and strength grade of the building materials; The acceleration coefficient of the preceding corrosion on the subsequent corrosion is determined by multiple immersion experiments. It is usually taken as 1.1-1.3. The more severe the corrosion, the higher the superposition coefficient.

[0065] The formula for calculating the overall damage coefficient of a building is as follows:

[0066] in, The overall damage coefficient of the building. The corrosion damage amplification term is divided into waves, meaning that the impact damage of the kth flood is only affected by the accumulated corrosion damage of the previous k floods (including the current one). The impact damage of each wave of flood is amplified based on the corrosion damage accumulated up to the current wave, rather than using the total corrosion damage of all waves to amplify all impact damage. This is consistent with the temporal causal relationship that corrosion occurs first and then impact damage is amplified. The ultimate impact load of a building, in N. The duration of the k-th flood wave, in seconds. The time step is expressed in seconds (s). The maximum impact load that a building can withstand is determined based on its structural design documents (such as construction drawings and structural calculation reports). The duration of the k-th flood wave is equal to the soaking time. ; The time interval for simulation is typically 1-10 seconds, depending on the required accuracy.

[0067] The theoretical basis of this formula is the principle of cumulative damage, which combines the cumulative effect of impact load in structural mechanics with the superposition effect of corrosion damage in materials science. The core includes two points: (1) the accumulation of impact damage, that is, the cumulative term in the inner layer of the formula. Based on the theory of cumulative damage of the structure under alternating load, the building is subjected to multiple waves of flood impact, and the water flow impact of each wave of flood ( ), sediment impact ( The load is a dynamic load that varies with time, and it is necessary to integrate the impact load of each wave over the duration of the load (using a time step). Discrete-cumulative approximation), then combined with the building's ultimate impact load ( (1) By comparing the impact damage of a single wave, the proportion of impact damage in a single wave is obtained, and the proportion of total impact damage is obtained by summing the impact damage of multiple waves; (2) The superposition of corrosion damage, i.e. ( ) item, based on the dynamic superposition theory of material corrosion, corrosion damage caused by multiple waves of flood immersion ( This will reduce the structural strength of the building, amplifying the actual impact of impact damage. The percentage of total impact damage is multiplied by the corrosion damage amplification factor. The final comprehensive damage coefficient is obtained. In short, this formula is a combination of "accumulated damage from impact loads" and "superimposed amplification of corrosion damage", quantifying the comprehensive damage of multiple waves of floods to buildings from two dimensions: mechanical impact and material corrosion. This is consistent with the analytical logic of multi-factor coupled damage in engineering structures.

[0068] In step 3, the real-time changes of the flooded buildings under the action of each wave of flood are dynamically simulated, and the simulation results are presented simultaneously.

[0069] In this embodiment, a wave-specific corrosion rate coefficient is introduced. To adapt to the differences in water quality across different waves, through α, The corrosion contribution of each chemical factor is quantified by weighting, and a corrosion superposition coefficient is added. This study demonstrates the cumulative effect of prior corrosion leading to a decrease in material strength and subsequent accelerated corrosion, recreating the actual corrosion process after long-term immersion in floodwater. It constructs a coupling logic between impact and corrosion to achieve the synergistic quantification of dual damage. The coupling relationship between impact-induced surface cracks expanding corrosion area and corrosion-induced material degradation reducing impact resistance is established through (1+ The coefficient incorporates the amplification effect of corrosion damage on overall damage, avoiding errors caused by simulating impact and corrosion separately, thus accurately and dynamically quantifying the cumulative effect of corrosion.

[0070] The above are merely embodiments of the present invention. The invention is not limited to the fields covered by these embodiments. Commonly known structures and characteristics in the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are able to access all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A watershed inundation simulation method integrating multi-dimensional features, characterized in that, Including the following steps: Step 1, Collect and preprocess multi-source data: Obtain image data and corresponding positioning data of the target watershed through drone aerial photography; collect hydrological data, road network data, and shelter data of the target watershed in real time; establish a building database; Step 2, Constructing a watershed flood inundation model: Based on image data and location data, a watershed flood inundation model including topography, landforms and three-dimensional building morphology is constructed through feature point extraction, image matching and multi-view reconstruction algorithms. The model restores the floor structure of the buildings and the elevation information of each floor. Step 3: Simulate the flood inundation process and plan the optimal evacuation route: Based on hydrological data, evaluate and predict parameters, input these parameters into the watershed flood inundation model to initiate simulation. Dynamically simulate the real-time changes in the flooded area within the target watershed during the water level rise, simultaneously displaying the real-time flooded area. Match flooded buildings and the number of people awaiting rescue from the building database based on the flooded area. Use Dijkstra's algorithm to plan the optimal evacuation route from the flooded building to the shelter for the people awaiting rescue. Set the objective function: Min(T) g ×α)+Min(P×β); Among them, T g To normalize evacuation time, T g =T / T b T is the evacuation time. b The maximum possible evacuation time is preset, where α is the safety factor weight, P is the congestion probability, and β is the traffic efficiency weight. Set constraints: The route is not flooded, and the route's traffic capacity is greater than or equal to the evacuation flow rate. Generate the optimal set of evacuation routes based on the objective function and constraints; Step 4, dynamically adjust the optimal evacuation route: monitor the congestion and flooding status of the optimal evacuation route in real time. If congestion or flooding occurs, automatically trigger the replanning of the optimal evacuation route, update the constraints, and call the Dijkstra algorithm to generate the latest optimal evacuation route. Step 5: Issue warnings and optimal evacuation routes: Visualize the flooded areas, flooded buildings, and optimal evacuation routes.

2. The watershed inundation simulation method integrating multi-dimensional features according to claim 1, characterized in that, In step 2, when constructing the watershed flood inundation model, the image data is first preprocessed with distortion correction, noise reduction, and uniform illumination. Then, the SIFT algorithm is used to extract a preset number of feature points from each image. Dense point clouds are generated through multi-view image matching. A TIN triangulation network is constructed based on the dense point cloud, and the watershed flood inundation model is generated through texture mapping. The watershed flood inundation model includes a terrain submodule and a building submodule. The terrain submodule is used to accurately restore the topographic undulations of the watershed, and the building submodule is used to present the floor structure of buildings and mark the elevation of each floor. It is also associated with emergency information in the building database. Clicking on a building in the model allows users to view emergency information. In step 3, the prediction parameters are input through a visual touch screen all-in-one machine. The prediction parameters include the minimum water level, the maximum water level, and the rate of water level rise. Clicking the analysis and simulation button on the visual touch screen all-in-one machine automatically starts the flood inundation simulation process.

3. The watershed inundation simulation method integrating multi-dimensional features according to claim 2, characterized in that, The prediction parameters mentioned in step 3 are obtained through evaluation by a flood peak projection agent. The evaluation of the flood peak projection agent includes: acquiring real-time data of the current flood peak flow, interval rainfall, and station rainfall, as well as historical data on corresponding factors and highest water levels from historical flood peaks; performing preprocessing on the data, including interpolation, format standardization, and normalization; and using formulas based on the preprocessed data. ； Calculate the data connectivity, where, Data connectivity is used to measure the similarity between current data A and historical data B. These are the homoegenic components (terms with consistent features) of the current data and historical data after normalization. These are the heterogeneous components (items with different features) between the current data and historical data after normalization. These are the inverse components (terms with completely opposite characteristics) of the current data and historical data after normalization. is the total number of data items; i is used to identify anomalous components (representing the parts where the current data differs from historical data); Used to identify reversal components (representing the part that is completely opposite to the historical data); The coefficients represent the proportions of the isotropic, heterotropic, and antitropic components, respectively. , , ; Historical water level data corresponding to the maximum connectivity is selected, and the prediction parameters are obtained by combining them with the current data. Based on the prediction parameters, a three-dimensional dynamic simulation of flood inundation is realized.

4. The watershed inundation simulation method integrating multi-dimensional features according to claim 3, characterized in that, In step 3, during the dynamic simulation of the water level rise, the water level at the inlet of the main stream and the water level at the confluence of tributaries in the target basin for the future preset time period are calculated using the flood rise prediction algorithm. The propagation time of the flood peak from the inlet of the main stream to the confluence of each tributary is determined using the flood peak extrapolation algorithm. The water level rise process is dynamically presented using the flood inundation algorithm. When the water level reaches the first preset warning level, the flooded buildings are marked. When the water level reaches the second preset warning level, the flood status of the flooded buildings is marked, and the emergency information of the flooded buildings is displayed in conjunction with the data.

5. The watershed inundation simulation method integrating multi-dimensional features according to claim 4, characterized in that, In step 3, during the dynamic simulation of the rising water level, if the water level rises to the highest level, the flood inundation simulation process will automatically terminate and a flood inundation analysis report for the target basin will be generated simultaneously. The flood inundation analysis report for the target basin includes the flooded area, the flooded buildings, and the number of people awaiting rescue.

6. The watershed inundation simulation method integrating multi-dimensional features according to claim 5, characterized in that, In step 5, when issuing an early warning, staff members view the visualized flooded buildings, click the one-click response button, and notify the emergency response personnel of the flooded buildings via VoIP. The call time and content are automatically recorded. At the same time, an early warning notification is pushed to the residents of the flooded buildings. The early warning notification includes information about the flooded buildings, the optimal evacuation route, and the emergency response personnel.

7. The watershed inundation simulation method integrating multi-dimensional features according to claim 6, characterized in that, In step 5, when the optimal evacuation route is released, the precise navigation information of the optimal evacuation route, which can be viewed offline, is pushed to the mobile APP of the residents of the flooded building. Real-time voice navigation prompts are pushed to the positioning wristbands of special groups in the flooded building, and the map and directional signs of the optimal evacuation route are displayed on the visual terminal.

8. The watershed inundation simulation method integrating multi-dimensional features according to claim 7, characterized in that, In step 3, during the dynamic simulation of the water level rise, the inundation data and geographical environment data of the target watershed are recorded. Before planning the optimal evacuation route, an inundation risk assessment is conducted. The weights of the assessment indicators are determined by using the inundation data and geographical environment data, including the weights of inundation depth, rise rate, terrain slope, and distance from the danger point. The risk value of each sub-region of the target watershed is calculated based on the weight of the evaluation index, and the risk level is divided. The evacuation priority is determined based on the risk level. In step 5, the optimal evacuation route is allocated according to the evacuation priority.

9. The watershed inundation simulation method integrating multi-dimensional features according to claim 8, characterized in that, In step 2, a hydrodynamic model between buildings, a silt and debris impact load model, and a multi-wave dynamic immersion corrosion damage model are also constructed: The impact intensity of water flow on structures under each wave of flood is calculated, and the bypass effect is determined using the following formula: ； in, Let k be the water flow impact intensity of the k-th flood wave, where k is the flood wave number. Let the water density of the k-th flood wave be . Let S be the real-time water velocity of the k-th flood wave, and S be the water-facing area of ​​the building. This is the drag coefficient around the flow; The impact load of sediment and debris on buildings during each flood wave is calculated using the following formula: ； in, The impact load of sediment and debris from the k-th wave flood. This is the impact correction factor for the k-th wave flood. Let be the volume fraction of sediment and debris in the k-th flood wave. Let be the average density of sediment and debris in the k-th flood wave. The average particle size of sediment and debris in the k-th wave flood; Calculate the cumulative dynamic corrosion damage coefficient of the structure after multiple waves of flooding, quantify the dynamic impact of multiple waves of immersion corrosion, and the first... The corrosion damage caused by wave flooding alone is as follows: Calculate the sum of all waves of corrosion: in, For the first Corrosion damage caused solely by wave floods To calculate the cumulative dynamic corrosion damage coefficient, Let be the corrosion rate coefficient of the k-th flood wave. The duration of the inundation of the k-th wave of floodwater. , , These represent the normalized pH, normalized dissolved oxygen content, and normalized salinity of the water body during the k-th flood wave, respectively. , , Corrosion affects weight. The corrosion resistance coefficient of building materials. Let be the corrosion superposition coefficient of the k-th wave flood; The formula for calculating the overall damage coefficient of a building is as follows: ； in, The overall damage coefficient of the building. The ultimate impact load of a building. The duration of the k-th flood wave For time step; In step 3, the real-time changes of the flooded buildings under the action of each wave of flood are dynamically simulated, and the simulation results are presented simultaneously.

10. A watershed inundation simulation system integrating multi-dimensional features, characterized in that, Used to perform the method according to any one of claims 1-9.

Images