Water conservancy disaster dynamic monitoring and accurate early warning method and system

Abstract

The invention discloses a water conservancy disaster dynamic monitoring and accurate early warning method, and relates to the technical field of intelligent water conservancy and emergency management, and the method comprises the following steps: carrying out the multi-source data fusion modeling in advance: obtaining high-precision DEM and point cloud data through a high-resolution satellite and an unmanned aerial vehicle LiDAR, and carrying out the multi-source data fusion modeling according to hydrological monitoring and communication signaling data; constructing a multi-dimensional database including terrain, hydrology and personnel distribution; and carrying out small watershed disaster evolution simulation. Multi-source data fusion is achieved, monitoring comprehensiveness and real-time performance are improved, the problems that a traditional single data source is limited in monitoring range and insufficient in precision are solved, meanwhile, high-precision positioning and trajectory prediction are achieved, dynamic tracking of personnel coordinates with the precision ranging from the centimeter level to 200 m is achieved, and the system is suitable for popularization and application. Smooth prediction is carried out on a track in a signal shielding scene in combination with a state space model, positioning interruption is avoided, and real-time and continuous personnel distribution data support is provided for disaster influence assessment.

Description

Technical Field

[0001] The present invention relates to the field of smart water conservancy and emergency management technology, and specifically to a method for dynamic monitoring and precise early warning of water conservancy disasters. Background Art

[0002] In the field of smart water conservancy and emergency management technology, monitoring and early warning of water disasters such as floods and waterlogging have always been crucial for protecting people's lives and property and ensuring social stability and development. Currently, traditional water disaster monitoring and early warning methods rely primarily on a single type of monitoring data and simple model analysis, which presents numerous limitations.

[0003] From a data collection perspective, traditional methods often rely solely on a limited number of ground-based monitoring stations to obtain hydrological data. This results in limited coverage, low accuracy, and delayed updates, making it difficult to fully and real-timely reflect the complex evolution of disasters. For example, in small mountainous watersheds with complex terrain, ground-based stations are easily obscured by the terrain, resulting in missing or distorted data and an inability to accurately capture the risk of sudden disasters in localized areas. Furthermore, traditional methods lack effective integration and dynamic monitoring of human distribution information, making it difficult to quickly assess the threat level to personnel and develop targeted early warning and evacuation plans when disasters occur.

[0004] In terms of model simulation, existing hydrological models are mostly based on simplified assumptions and single physical processes. Consequently, they lack the accuracy to simulate flood runoff, confluence, and inundation. This is particularly true when dealing with hydrodynamic coupling in the complex terrain of small watersheds, making it difficult to accurately predict flood paths and inundation ranges. Furthermore, traditional risk assessment methods typically only consider the single dimension of disaster intensity, ignoring social factors such as population density. This results in inaccurate warning level classifications and fails to meet the refined and personalized needs of emergency management.

[0005] Traditional methods for delivering warning information lack efficient spatial indexing and base station matching algorithms, making it difficult to accurately deliver warning information. Broadcasting is often used, which not only results in inefficient information delivery but can also waste resources and cause confusion due to excessive amounts of warning information being sent to unrelated areas, impacting the effectiveness of warnings in key areas.

[0006] In summary, the existing technologies in the process of water disaster monitoring and early warning have problems such as incomplete data collection, insufficient model simulation accuracy, single risk assessment dimension, and inaccurate warning information push. There is an urgent need for a new method that can integrate multi-source data, achieve high-precision disaster simulation and accurate early warning, so as to improve the scientificity and effectiveness of water disaster emergency management.

[0007] Currently, no effective solutions have been proposed for the problems in related technologies. Summary of the Invention

[0008] In response to the problems in the related technologies, the present invention proposes a method for dynamic monitoring and accurate early warning of water disasters to overcome the above-mentioned technical problems existing in the existing related technologies.

[0009] The technical solution of the present invention is achieved as follows:

[0010] A method for dynamic monitoring and accurate early warning of water disasters, comprising the following steps:

[0011] Step S1: Preliminary multi-source data fusion modeling: high-precision DEM and point cloud data are obtained through high-resolution satellites and UAV LiDAR, and a multi-dimensional database including topography, hydrology, and personnel distribution is constructed based on hydrological monitoring and communication signaling data;

[0012] Step S2, simulating the evolution of small watershed disasters, including: dividing the target small watershed into microgrids of 500m×500m, calculating the terrain index and runoff rate based on TOPMODEL; simultaneously constructing a one- and two-dimensional hydrodynamic coupling model using a one-dimensional river channel model and a two-dimensional inundation model, achieving energy conservation through flow interpolation, simulating the flood evolution process, and outputting an inundation grid map and flow velocity distribution data;

[0013] Step S3, using the DPC clustering algorithm and KDE kernel density estimation method to process the personnel coordinate data, generate the personnel thermal field distribution, and calculate the personnel thermal value;

[0014] Step S4, constructing a three-dimensional risk assessment matrix based on water depth, flow velocity, and personnel thermal value, and determining the warning level based on the three-dimensional risk assessment matrix;

[0015] Step S5: Rapidly match communication base stations associated with the disaster area based on the R-tree spatial index algorithm, sort the associated base stations using the TOPSIS algorithm, and prioritize pushing warning information to base stations covering high-risk areas;

[0016] In step S6, a weighted directed graph model is constructed, and an improved A* algorithm is used to generate a personnel transfer path based on terrain cost, flood parameters, and personnel density. The personnel positions are monitored in real time, and a secondary warning is triggered for personnel who stay in unsafe areas for more than a preset time.

[0017] The communication signaling data includes: parsing the communication signaling and locating personnel, including the following steps:

[0018] The coordinates of the three base stations are calibrated as (x1, y1), (x2, y2), and (x3, y3), and the signal arrival time difference is expressed as Δt 21 , Δt 31 , the speed of light c = 3 × 10 8 m / s, calculate the distance difference, expressed as:

[0019] d 31 =c·Δt 31 ;

[0020] d 21 =c·Δt 21 ;

[0021] Construct the hyperbola equation and obtain the user coordinates (x, y);

[0022] Among them, base station 1 and base station 2 form a hyperbola, which is expressed as:

[0023]

[0024] Among them, base station 1 and base station 3 form a hyperbola, which is expressed as:

[0025]

[0026] The method also includes: performing trajectory smoothing and movement trend prediction on personnel positioning data, including the following steps:

[0027] Build a state space model, including: calibrating the four-dimensional state vector Among them, (x t ,y t ) is the real-time coordinate obtained through communication signaling analysis, expressed as the initial value, is the horizontal moving speed, and the state transfer matrix F is based on uniform motion and is expressed as:

[0028]

[0029] Where Δt = 60s, corresponding to a high-frequency sampling interval of 1 time / minute, and the observation matrix H extracts the coordinate information in the state vector to match the actual observation value, which is expressed as:

[0030]

[0031] Perform recursive prediction, including the following steps:

[0032] Calibrate the state X at the previous moment t-1 , predict the current state, expressed as:

[0033] Update the prediction error covariance matrix, expressed as: P t - =F·P t-1 ·F T +Q;

[0034] Among them, Q is the process noise matrix, which represents the uncertainty of the motion model, i.e., the speed change of personnel;

[0035] Among them, if a valid observation value Z is received t =[x t′ ,y t′ ] T , calculate the observation residual, expressed as:

[0036] By Kalman gain K t Fusion predictions and observations are expressed as:

[0037]

[0038] Where R = diag([200 2 ,200 2 ]) is the observation noise covariance matrix;

[0039] The modified state vector and covariance matrix are expressed as:

[0040]

[0041] Among them, when the signal is blocked and the observation value is missing, the predicted state is used As a current estimate, avoid trajectory interruptions.

[0042] The small watershed disaster evolution simulation includes: TOPMODEL simulation, including the following steps:

[0043] The small watershed is divided into 500m×500m microgrids and the terrain index is calculated, which is expressed as:

[0044]

[0045] Determine the critical terrain index TI c , through the water balance equation Iterative solution;

[0046] Calculate the flow rate, expressed as:

[0047]

[0048] Where β is the slope, A is the catchment area, θ is the soil moisture, i is the rainfall intensity, and K s is the saturated hydraulic conductivity.

[0049] It also includes: a one- and two-dimensional hydrodynamic coupling model simulation, including the following steps:

[0050] One-dimensional river channel model simulation: including: based on the Saint-Venant equations, using the Preissmann four-point implicit format discretization equation, dividing the river channel into N sections, with a time step of Δt and a space step of Δx;

[0051] The equation is discretized and expressed as:

[0052]

[0053] Among them, Q is the cross-sectional flow, A is the cross-sectional area of ​​water flow, S0 is the riverbed slope, S f is the friction slope, expressed as: Where n is the Manning roughness coefficient, R is the hydraulic radius;

[0054] The two-dimensional flooding model simulation includes the following steps:

[0055] The finite volume method is calibrated to discretize the area into 500m×500m grids, and the flux is solved using the Roe scheme, considering the ground roughness n=0.03;

[0056] The coupling interface is used to achieve energy conservation by interpolating and matching the one-dimensional river section flow and the two-dimensional grid boundary flow, which can be expressed as:

[0057]

[0058] Where M is the number of boundary cells in the two-dimensional grid, u i ,h i is the grid flow velocity and water depth.

[0059] The step of calculating the thermal value of a person includes the following steps:

[0060] Use density peak clustering algorithm to identify core points where people gather, including:

[0061] The calibration base station signaling coordinate set is Cutoff distance d c =500m;

[0062] Calculate the local density, expressed as:

[0063]

[0064] Among them, ρ i is the local density, reflecting the number of people within 500m around point i, ρ i The larger the value, the higher the degree of aggregation;

[0065] Calculate the distance metric:

[0066] Among them, the core point is expressed as: δ i =min{d ij |ρ j >ρ i};

[0067] Among them, the edge point is represented by: δi =max{d ij};

[0068] Among them, δ i As the distance index, the core point takes the minimum distance of the adjacent high-density points, and the edge point takes the global maximum distance, which is used to identify the cluster center;

[0069] Perform cluster center identification and select ρ i and δ i The points with higher values ​​are taken as cluster centers to divide the gathering areas of people;

[0070] Performing kernel density estimation includes the following steps:

[0071] Calibrate the coordinates of people after clustering Bandwidth h = 500m;

[0072] The thermal value is calculated as:

[0073]

[0074] The method of preferentially pushing warning information to base stations covering high-risk areas includes the following steps:

[0075] The base station set B obtained by R-tree query is input as the candidate base station set;

[0076] The calibration evaluation indicators are expressed as: number of covered users C1, mean signal strength C2, historical warning success rate C3, and their normalized matrices;

[0077] Among them, the number of covered users C1, the number of users covered by the base station, is standardized as:

[0078]

[0079] The mean signal strength C2, base station signal strength, is normalized to:

[0080]

[0081] The historical warning success rate C3, a percentage value, is standardized as follows:

[0082]

[0083] Perform weighted matrix and ideal solution calculations;

[0084] Among them, the weighted matrix is ​​expressed as: W = [0.5, 0.3, 0.2], where the number of users has the highest weight;

[0085] Positive ideal solution, expressed as: Z + =(max(v i1 ),max(vi2 ),max(v i3 );

[0086] Negative ideal solution, expressed as: Z - =(min(v i1 ),min(v i2 ),min(v i3 );

[0087] Perform distance calculation and sorting, where the distances between base station i and the positive and negative ideal solutions are:

[0088]

[0089] Get comprehensive score S i , and press S i Arrange in descending order to obtain the base station priority for accurate push of warning information, expressed as:

[0090]

[0091] The triggering of a secondary warning for users who have stayed for longer than a preset time includes the following steps:

[0092] Real-time monitoring of personnel location changes. If a person stays in a non-safe area for longer than the preset time and does not move according to the planned route, it is determined to be abnormal detention, and a secondary warning is triggered.

[0093] Beneficial effects of the present invention:

[0094] 1. The present invention achieves multi-source data fusion to enhance the comprehensiveness and real-time performance of monitoring through aerospace-ground-signaling data collaboration: integrating aerospace data such as high-resolution satellite DEM, drone LiDAR point cloud, and InSAR surface deformation monitoring, combined with high-frequency ground sampling data from NB-IoT rain gauges and Beidou water level gauges, and personnel positioning data analyzed through communication signaling, to construct a multidimensional database covering topography, hydrology, and personnel distribution. This achieves full coverage of factors from macro-topography to micro-personnel dynamics, solving the problems of limited monitoring range and insufficient accuracy of traditional single data sources. Simultaneously, high-precision positioning and trajectory prediction are achieved: through personnel positioning and Kalman filter models, dynamic tracking of personnel coordinates with an accuracy of centimeters to 200 meters is achieved. Combined with the state-space model, the trajectory in signal-blocked scenarios is smoothly predicted to avoid positioning interruptions, providing real-time and continuous personnel distribution data support for disaster impact assessment.

[0095] 2. The present invention realizes refined model simulation to improve the accuracy of disaster prediction, through TOPMODEL and hydrodynamic coupling model: based on 500m×500m micro-grid division, TOPMODEL calculates terrain index and runoff rate, combines one- and two-dimensional hydrodynamic coupling model to simulate flood evolution process, outputs high-precision flooding grid map and flow velocity distribution data, realizes dynamic deduction of the whole process of flood runoff generation-confluence-flooding in small watershed, and improves the prediction accuracy of flood diffusion path and flooding depth under complex terrain compared with traditional models (such as water depth error ≤0.5m, flow velocity error ≤0.2m / s). At the same time, by adopting multi-physical process coupling: the one-dimensional river channel model uses the Saint-Venant equations to describe water flow movement, the two-dimensional flooding model considers ground roughness and flux conservation, and realizes energy conservation through flow interpolation, which effectively solves the problem of insufficient simulation of the coupling effect between river channel and flooding area in traditional models, and more realistically reflects the propagation characteristics of floods between different geomorphological units.

[0096] 3. The present invention implements dynamic risk assessment to achieve accurate classification of warning levels. It does this by constructing a three-dimensional risk assessment matrix: using water depth, flow velocity, and personnel thermal value as three-dimensional assessment dimensions, a discretized matrix containing 45 risk combinations is established. Through quantitative scoring (1-5 points) and level mapping (low risk, yellow warning, red warning), it achieves a transition from "one-way assessment of disaster intensity" to "comprehensive assessment of human-disaster-environment interaction." For example, when the water depth is 2.5m (medium level), the flow velocity is 1.8m / s (medium level), and the population is dense (level 5 thermal value), a red warning can be accurately triggered, which is more in line with the actual risk than the traditional warning mechanism based only on water depth or flow velocity. At the same time, combined with dynamic thermal field generation: using DPC clustering and KDE kernel density estimation, the distribution of personnel thermal values ​​is updated every 5 minutes, reflecting changes in the gathering of personnel in real time, ensuring that the risk assessment results are synchronized with the dynamics of the disaster site, and avoiding warning deviations caused by lagging personnel flow.

[0097] 4. The present invention realizes precise communication and path planning to improve the efficiency of emergency response, and adopts R-tree spatial index and TOPSIS base station sorting: through R-tree, it quickly matches the base stations associated with the disaster area, and combines the number of covered users, signal strength, historical warning success rate and other indicators to perform TOPSIS sorting, so as to realize the priority push of warning information to high-risk areas (such as sending red warnings to base stations covering 1,500 people first), which shortens the invalid information transmission time by more than 50% compared with broadcast push, and improves the warning response speed in key areas. At the same time, it realizes personalized transfer path generation: based on the weighted directed graph model and the improved A* algorithm, the optimal transfer route is generated by comprehensively considering the terrain cost, flood parameters, and population density, avoiding dangerous areas with water depth ≥1.5m or flow rate ≥2m / s, and providing estimated time and turning point coordinates to provide customized escape guidance for people in different locations. For example, the path planning for the transfer of people from a village to a highland shelter took only 25 minutes, which was 30% shorter than the traditional empirical route. A secondary warning mechanism was also introduced: by real-time monitoring of the status of stranded personnel, voice calls or emergency contact notifications were automatically triggered for abnormal personnel who stayed in non-safe areas for more than 30 minutes, and drones were linked to provide guidance, effectively solving the blind spot problem of traditional warnings for people who "did not receive information" or "had their movements blocked". BRIEF DESCRIPTION OF THE DRAWINGS

[0098] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0099] Figure 1 The figure is a flow chart of a method for dynamic monitoring and precise early warning of water disasters according to an embodiment of the present invention. DETAILED DESCRIPTION

[0100] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention are within the scope of protection of the present invention.

[0101] According to an embodiment of the present invention, a method for dynamic monitoring and precise early warning of water disasters is provided.

[0102] like Figure 1 As shown, the method for dynamic monitoring and accurate early warning of water disasters according to an embodiment of the present invention includes the following steps:

[0103] Step S1: Pre-construct multi-source data fusion modeling: obtain high-precision DEM (Digital Elevation Model) and point cloud data through high-resolution satellites and UAV LiDAR (Light Detection and Ranging), and build a multidimensional database including topography, hydrology, and personnel distribution based on hydrological monitoring and communication signaling data;

[0104] The acquisition of aerospace data includes the following steps:

[0105] The Gaofen-7 satellite (GF-7) was used to obtain a 5m resolution DEM, and terrain modeling with a plane accuracy of 0.5m and an elevation accuracy of 1.5m was achieved using the rational function model (RFM).

[0106] Use UAV LiDAR (such as DJI Mavic 3 Multispectral) to generate 1m precision point cloud data, and use the moving surface fitting algorithm to construct a 3D terrain model, including parameters such as river curvature and vegetation coverage;

[0107] Based on the InSAR technology of the Sentinel-1SAR satellite, the Small Baseline Assemblies (SBAS) can achieve millimeter-level monitoring of surface deformation (accuracy ±3mm / year)

[0108] Among them, ground data collection includes the following steps:

[0109] Deploy NB-IoT smart rain gauges (±2% accuracy) and Beidou water level gauges (centimeter-level accuracy), using time division multiplexing (TDM) technology to achieve high-frequency sampling of once per minute, with data transmitted back over the LoRaWAN network;

[0110] This also includes communication signaling analysis and personnel positioning, including the following steps:

[0111] The coordinates of the three base stations are calibrated as (x1, y1), (x2, y2), and (x3, y3), and the signal arrival time difference is expressed as Δt 21 , Δt 31 , the speed of light c = 3 × 10 8 m / s, calculate the distance difference, expressed as:

[0112] d 31 =c·Δt 31 ;

[0113] d 21 =c·Δt 21 ;

[0114] Construct the hyperbola equation and obtain the user coordinates (x, y);

[0115] Among them, base station 1 and base station 2 form a hyperbola, which is expressed as:

[0116]

[0117] Among them, base station 1 and base station 3 form a hyperbola, which is expressed as:

[0118]

[0119] Specifically, by combining the two sets of hyperbolic equations, the quadratic equation system is solved to obtain the user coordinates (x, y).

[0120] In addition, to solve the signal obstruction problem caused by mountainous terrain, trajectory smoothing and movement trend prediction are performed on the personnel positioning data. The specific steps are as follows:

[0121] Build a state space model, including: calibrating the four-dimensional state vector Among them, (x t ,y t ) is the real-time coordinate obtained through communication signaling analysis, expressed as the initial value, is the horizontal moving speed, and the state transfer matrix F is based on uniform motion and is expressed as:

[0122]

[0123] Where Δt = 60s, corresponding to a high-frequency sampling interval of 1 time / minute, and the observation matrix H extracts the coordinate information in the state vector to match the actual observation value, which is expressed as:

[0124]

[0125] Perform recursive prediction, including the following steps:

[0126] Calibrate the state X at the previous moment t-1 , predict the current state, expressed as:

[0127] Update the prediction error covariance matrix, expressed as: P t - =F·P t-1 ·F T +Q;

[0128] Among them, Q is the process noise matrix, which represents the uncertainty of the motion model, i.e., the speed change of personnel;

[0129] Among them, if a valid observation value Z is received t =[x t′ ,y t′ ] T , positioning error ≤ 200m, calculate the observation residual, expressed as:

[0130] By Kalman gain K t Fusion predictions and observations are expressed as:

[0131] K t =P t - ·H T ·(H·P t - ·H T +R) -1 ;

[0132] Where R = diag([200 2 ,200 2 ]) is the observation noise covariance matrix.

[0133] The modified state vector and covariance matrix are expressed as:

[0134]

[0135] Among them, when the signal is blocked and the observation value is missing, the predicted state is used As a current estimate, avoid trajectory interruptions.

[0136] Step S2: Simulate the evolution of small watershed disasters, including: simulating the flood runoff generation, confluence, and inundation process based on TOPMODEL (Topography-Based Hydrological Model) and a one- and two-dimensional hydrodynamic coupling model, and outputting a 500m resolution inundation grid map;

[0137] Among them, TOPMODEL simulation includes the following steps:

[0138] The small watershed is divided into 500m×500m microgrids and the terrain index is calculated, which is expressed as:

[0139]

[0140] Determine the critical terrain index TI c , through the water balance equation Iterative solution;

[0141] Calculate the flow rate, expressed as:

[0142]

[0143] Where β is the slope, A is the catchment area, θ is the soil moisture, i is the rainfall intensity, and K s is the saturated hydraulic conductivity;

[0144] The one- and two-dimensional hydrodynamic coupling model simulation includes the following steps:

[0145] One-dimensional river channel model simulation: including: based on the Saint-Venant equations, using the Preissmann four-point implicit format discretization equation, dividing the river channel into N sections, with a time step of Δt and a space step of Δx;

[0146] The equation is discretized and expressed as:

[0147]

[0148] Among them, Q is the cross-sectional flow, A is the cross-sectional area of ​​water flow, S0 is the riverbed slope, S f is the friction slope, expressed as: Where n is the Manning roughness coefficient, which is 0.02-0.04 for natural rivers and 0.01-0.02 for artificial channels; R is the hydraulic radius.

[0149] The two-dimensional flooding model simulation includes the following steps:

[0150] The finite volume method is calibrated to discretize the area into 500m×500m grids, and the flux is solved using the Roe scheme, considering the ground roughness n=0.03;

[0151] The coupling interface is used to achieve energy conservation by interpolating and matching the one-dimensional river section flow and the two-dimensional grid boundary flow, which can be expressed as:

[0152]

[0153] Where M is the number of boundary cells in the two-dimensional grid, u i ,h i is the grid flow velocity and water depth;

[0154] Step S3, performing dynamic identification of personnel thermal fields, including: generating personnel thermal values ​​through DPC clustering and KDE kernel density estimation;

[0155] The generation of dynamic personnel thermal value includes the following steps:

[0156] Perform density peak clustering (DPC) to identify core points where people gather, including:

[0157] The calibration base station signaling coordinate set is Cutoff distance d c =500m;

[0158] Calculate the local density, expressed as:

[0159]

[0160] Among them, ρi is the local density, reflecting the number of people within 500m around point i, ρ i The larger the value, the higher the degree of aggregation.

[0161] Calculate the distance metric:

[0162] Among them, the core point is expressed as: δ i =min{d ij |ρ j >ρ i};

[0163] Among them, the edge point is represented by: δ i =max{d ij};

[0164] Among them, δ i As the distance index, the core point takes the minimum distance to the adjacent high-density points, and the edge point takes the global maximum distance to identify the cluster center.

[0165] Perform cluster center identification and select ρ i and δ i The points with higher values ​​are taken as cluster centers to divide the areas where people gather.

[0166] Perform kernel density estimation (KDE), which includes the following steps:

[0167] Calibrate the coordinates of people after clustering Bandwidth h = 500m;

[0168] The thermal value is calculated as:

[0169]

[0170] The thermal values ​​of a 500m×500m grid were obtained and divided into 5 levels after normalization (level 1 is sparse and level 5 is dense).

[0171] When applied, this technical solution realizes the real-time inflow of signaling data based on the message queue Kafka and recalculates the thermal field every 5 minutes.

[0172] Step S4 is to predict the disaster level, including: establishing a three-dimensional risk assessment matrix and obtaining the warning level based on the disaster intensity (water depth, flow rate) and population density (thermal value). The details are as follows:

[0173] Step S401: Extract disaster intensity features and population density features in advance, where disaster intensity includes water depth and flow velocity, and population density includes thermal value;

[0174] The water depth data is derived from a 500m-resolution inundation grid map generated by a one- and two-dimensional hydrodynamic coupling model, with each grid recording the flood depth. The flow velocity data is calculated using a one-dimensional river channel model and a two-dimensional inundation model, reflecting the flow velocity in the river channel and in the inundated area.

[0175] Among them, the personnel thermal value: generated based on DPC clustering and KDE kernel density estimation, the study area is divided into 500m×500m grids, and each grid corresponds to 1-5 levels of thermal value (level 1 is sparse and level 5 is dense).

[0176] Step S402: constructing a three-dimensional risk assessment matrix, including: using water depth, flow velocity, and personnel thermal value as three-dimensional coordinate axes to establish a discretized risk assessment matrix. The specific steps are as follows:

[0177] Pre-discretize the data, including:

[0178] The water depth is divided into three levels: low (<1m), medium (1-3m), and high (≥3m).

[0179] The flow velocity is divided into three levels: low (<0.5m / s), medium (0.5-2m / s), and high (≥2m / s).

[0180] The thermal values ​​of personnel are divided into 5 levels: Level 1 (sparse), Level 2, Level 3, Level 4, and Level 5 (dense).

[0181] Perform matrix dimension mapping to generate a three-dimensional risk assessment matrix, and obtain the current risk level based on the three-dimensional risk assessment matrix.

[0182] Among them, the three-dimensional risk assessment matrix is ​​shown in Table 1:

[0183] Table 1 Three-dimensional risk assessment matrix

[0184]

[0185]

[0186] Step S5: Matching communication base stations, including using an R-tree spatial index algorithm to match base stations associated with the disaster area, and sorting them using the TOPSIS method for accurate push of warning information;

[0187] Specifically, using the R-tree spatial index algorithm to match disaster area-related base stations includes the following steps:

[0188] Build R-tree spatial index, including spatial object modeling and hierarchical index construction;

[0189] Among them, spatial object modeling: each communication base station is abstracted as a two-dimensional space point object with coordinates (x i,y i ), and construct the minimum bounding rectangle MBR for each base station, that is, the minimum rectangle containing the point.

[0190] The hierarchical index is constructed by storing the MBRs of base stations and their corresponding base station IDs. Each leaf node can hold up to M base station entries (M is the node capacity threshold, usually an adaptive value based on the disk page size, such as 50-100). Non-leaf nodes store the MBRs of their child nodes. The MBR of the parent node is the minimum bounding rectangle of the MBRs of all its child nodes. The details are as follows:

[0191] Initialize and group all base stations according to their spatial locations. Generate a leaf node for each group and calculate its MBR.

[0192] Aggregate upwards, taking leaf nodes as child nodes, and recursively construct upper-level non-leaf nodes until the root node is generated.

[0193] Dynamic maintenance: When a new base station is added or an old one is removed, the tree structure is balanced through node splitting or merging. Node splitting strategy: If the insertion causes a leaf node to exceed its capacity M, the node is split using the "minimum overlap first" principle: the base stations within the node are divided into two groups to minimize the overlap between the MBRs of the two groups. If the overlap is the same, the partitioning method with the smaller total area is selected.

[0194] Node splitting strategy: If the insertion causes the leaf node to exceed the capacity M, the "minimum overlap first" principle is selected to split the node: the base stations in the node are divided into two groups so that the overlapping area of ​​the two groups' MBRs is minimized; if the overlapping areas are the same, the division method with the smaller total area is selected.

[0195] Querying base stations associated with disaster areas includes the following steps:

[0196] The spatial extent of the disaster area, i.e., a polygonal or rectangular area, is recorded as the query window Q as input;

[0197] Root node traversal, starting from the root node of the R-tree, checks whether the MBR of the current node intersects with the query window Q. If not, skip the node and its subtree; if so, proceed to the following steps;

[0198] Recursively search for child nodes. For intersecting child nodes, if they are leaf nodes, directly extract the base stations that intersect with Q, that is, the point object intersects with Q and is inside Q; if they are non-leaf nodes, recursively execute the above steps.

[0199] Generate the result set, including collecting all base station IDs that intersect with Q to form a candidate base station set, expressed as: B = {b1, b2, ..., b n}.

[0200] Among them, checking whether the MBR of the current node intersects with the query window Q includes: for two rectangles A(x A1 ,y A1 ,x A2 ,y A2 ) and B(x B1 ,y B1 ,x B2 ,y B2 ), the intersection condition is expressed as:

[0201] That is, the "complete separation" condition is not met in any dimension.

[0202] Among them, to determine whether a point object intersects with Q, that is, the base station coordinates (x, y) are located within the query window Q (polygon), the ray method is used: a horizontal ray is emitted from the point (x, y) to the right, and the number of intersections with the polygon boundary is counted. If it is an odd number, the point is within the area.

[0203] Among them, the ranking by TOPSIS method includes the following steps:

[0204] The base station set B obtained by R-tree query is used as the evaluation object of TOPSIS method, that is, as the input of candidate base station set;

[0205] The calibration evaluation indicators are expressed as: number of covered users C1, mean signal strength C2, historical warning success rate C3, and their normalized matrices;

[0206] Among them, the number of covered users C1, the number of users covered by the base station, is standardized as:

[0207]

[0208] The mean signal strength C2, base station signal strength (unit: dBm), is normalized to:

[0209]

[0210] The historical warning success rate C3, a percentage value, is standardized as follows:

[0211]

[0212] Perform weighted matrix and ideal solution calculations;

[0213] Among them, the weighted matrix is ​​expressed as: W = [0.5, 0.3, 0.2], where the number of users has the highest weight;

[0214] Positive ideal solution, expressed as: Z + =(max(v i1 ),max(v i2 ),max(vi3 );

[0215] Negative ideal solution, expressed as: Z - =(min(v i1 ),min(v i2 ),min(v i3 );

[0216] Perform distance calculation and sorting, where the distances between base station i and the positive and negative ideal solutions are:

[0217]

[0218] Get comprehensive score S i , and press S i Arrange in descending order to obtain the base station priority for accurate push of warning information, expressed as:

[0219]

[0220] Step S6: Plan a transfer route and trigger a secondary warning for users who stay longer than a preset time.

[0221] This technical solution, after completing the disaster level prediction and base station matching, builds a personnel transfer path planning model based on multi-source data, combines real-time personnel distribution and terrain characteristics to generate the optimal transfer route, and triggers a secondary warning for abnormally stranded personnel. The details are as follows:

[0222] A weighted directed graph model is constructed based on the multidimensional database constructed in step S1, which contains high-precision DEM, point cloud data, and personnel distribution thermal values. The geographic space is abstracted into a weighted directed graph G(V,E), where the nodes V represent safe areas (such as high ground, shelters), key landmarks (such as bridges, road intersections), and the real-time locations of personnel. The coordinates are obtained by communication signaling analysis or GNSS positioning. The edges E represent accessible paths, such as roads and river embankments. Each edge is accompanied by a weight w. ij , calculated based on the following parameters:

[0223] Among them, terrain cost: extract slope, roughness and other parameters based on DEM data, and use formula w terrain =α·slope+β·roughness quantification, α and β are empirical coefficients, the greater the slope and the higher the roughness, the higher the cost;

[0224] Among them, the travel efficiency is calculated by combining real-time hydrological data, such as flow velocity and water depth, with a hydrodynamic model to calculate the travel time of the flooded road section, which is expressed as:

[0225] w flood =γ·depth+δ·velocity;

[0226] Specifically, when applied, passage is prohibited when the water depth exceeds 1.5m or the flow rate exceeds 2m / s;

[0227] Among them, population density: thermal value level generated by the association step, path average thermal value HeatMap avg The higher the density, such as level 4-5 dense area, the weight is w density =η·HeatMap avg Increases gradually, η is the density weight coefficient, and its value ranges from 0.5 to 1.0.

[0228] Generate a personalized path, including the following steps:

[0229] The improved A* algorithm is used to search for the optimal path in a weighted directed graph. The heuristic function h(n) and the actual cost g(n) are combined to achieve efficient path finding, which can be expressed as:

[0230] f(n)=g(n)+h(n)·(1+λ·HeatMap avg );

[0231] The actual cost g(n) represents the cumulative weight from the starting point to the current node n, including the sum of terrain, flood, and density costs. The heuristic function h(n) uses Euclidean distance to estimate the straight-line distance from the current node to the end point, and is corrected based on the priority of safe areas, such as high ground.

[0232] Add the starting point, i.e. the person's current location, to the open list, set g(start) = 0, and h(start) to be the straight-line distance to the nearest safe area;

[0233] Expand the node, traverse the node with the lowest cost in the open list, generate adjacent nodes and calculate f(n). If the adjacent node is a prohibited area, such as water depth ≥ 3m, skip it;

[0234] Update the path. If the cost of reaching the adjacent node through the current node is lower, update the parent node of the adjacent node to the current node and add it to the open list.

[0235] When the end node, i.e. the safe area, is added to the open list, the optimal path is generated by tracing back to the parent node and outputting guidance instructions including the coordinates of the turning point and the estimated time.

[0236] The second warning is triggered for users who stay beyond the preset time, including the following steps:

[0237] Through the communication signaling analysis and trajectory prediction in step S1, the personnel position changes are monitored in real time:

[0238] If a person remains in an unsafe area, such as a flooded area or near a high-risk river, for more than 30 minutes and does not move according to the planned route, it will be considered "abnormal detention." A secondary warning will be triggered, such as making a voice call or automatically extracting the emergency contact information corresponding to the user's anonymized ID, and sending a helpline notification containing the location coordinates.

[0239] In addition, in a specific embodiment, taking a small watershed flood disaster warning scenario as an example, the area of ​​a small watershed in a mountainous area is about 20km 2 The river in the basin is 15 km long, with three villages (approximately 2,000 people) scattered along the riverbank. The terrain is mainly mountainous and hilly, with a vegetation coverage rate of approximately 60%. During the rainy season, the area experiences sudden heavy rainfall, as follows:

[0240] Step S1, pre-collect and process multi-source data;

[0241] Among them, aerospace data: DEM data acquired by the Gaofen-7 satellite revealed an average slope of 15° within the basin and a minimum river curvature radius of 50 meters. UAV LiDAR generated point cloud data, identifying three landslide-prone locations (vegetation cover <30%). InSAR monitoring showed a surface deformation rate of +2 mm / year on the right bank of the river (slight uplift, not currently posing a risk).

[0242] Among them, ground data: the NB-IoT rain gauge monitored the rainfall intensity in real time at 80mm / h (rainstorm level), and the Beidou water level meter showed that the river water level rose by 1.2m within 1 hour.

[0243] Among them, personnel positioning: through signaling analysis of three base stations (coordinates are A(0,0), B(1000,0), and C(500,800), the coordinates of 2000 people were obtained, of which 1500 people were concentrated in the village center (thermal value level 5) and 500 people were distributed in farmland (thermal value level 3).

[0244] Step S2: Simulate the evolution of small watershed disasters:

[0245] Among them, TOPMODEL calculations: the average terrain index ln(A / tanβ) is 4.2, the critical terrain index is 3.8, and the runoff rate f = 0.05m / s (significant over-infiltration runoff under heavy rainfall).

[0246] Among them, the one-dimensional and two-dimensional hydrodynamic coupling model: the maximum flow velocity calculated by the one-dimensional river model is 3m / s (downstream of the river), and the cross-sectional area A = 20m 2 The 2D inundation model outputs an inundation grid map, showing that the water depth in the area downstream of the village is 2.5 m (medium level) and the flow velocity is 1.8 m / s (medium level).

[0247] Step S3, dynamic identification of personnel thermal field, wherein DPC clustering, cutoff distance d c =500m, 3 cluster centers (corresponding to 3 villages) are identified, and the local density ρ i The values ​​are 200, 150, and 100 (people / 500m radius), respectively. KDE kernel density estimation generates a 500m×500m grid with a thermal value level of 5 for the village center and level 3 for the farmland area.

[0248] Step S4: Disaster severity prediction. Based on the three-dimensional risk assessment matrix, the downstream area of ​​the village has a water depth of 2.5m (medium), a flow velocity of 1.8m / s (medium), and a human thermal value of level 5 (dense). Table 1 indicates a risk score of 4.5, corresponding to a red alert (immediate evacuation required). The farmland area has a water depth of 1.2m (low), a flow velocity of 0.8m / s (medium), and a human thermal value of level 3 (medium density). The risk score is 2.5, corresponding to a yellow alert (requires enhanced monitoring).

[0249] Step S5, performing communication base station matching and early warning push, includes the following steps:

[0250] The disaster area (a 2 km × 2 km rectangle centered on the village) is matched to two base stations through the R-tree (base station 1 covers the village center and base station 2 covers the farmland area).

[0251] Using the TOPSIS method, we find that base station 1 covers 1,500 users (C1 = 1,500), has a signal strength of -80 dBm (C2 = -80), and a historical warning success rate of 95% (C3 = 95%). Base station 2 covers 500 users (C1 = 500), has a signal strength of -75 dBm (C2 = -75), and a historical warning success rate of 90% (C3 = 90%). The resulting comprehensive scores, S1 = 0.56 and S2 = 0.44, indicate a priority of base station 1 over base station 2. Base station 1 prioritizes sending red warnings to the village center, while base station 2 prioritizes sending yellow warnings to farmland areas.

[0252] Step S6: Perform transfer path planning and secondary warning, as follows:

[0253] Among them, the optimal path is generated, and people in the village center are transferred to high-altitude shelters (straight-line distance 1.5km). The path avoids river areas with water depths greater than 1.5m, and the ridgeline with lower terrain cost is the preferred route. It is expected to take 25 minutes.

[0254] At the same time, monitoring found that three users were stranded in the flooded area for more than 40 minutes. The system automatically sent a help notice containing coordinates (x=800, y=600) to their emergency contacts and started the drone to shout and guide them.

[0255] It's important to note that in this technical solution, collected personnel location data is encrypted for transmission and storage using AES-256, and user IDs are anonymized using a hashing algorithm to ensure compliance with relevant regulations. Communication signaling data is used only for disaster warnings and not for other commercial purposes. The storage period does not exceed 72 hours.

[0256] In summary, with the help of the above technical solution of the present invention, the following effects can be achieved:

[0257] 1. The present invention achieves multi-source data fusion to enhance the comprehensiveness and real-time performance of monitoring through aerospace-ground-signaling data collaboration: integrating aerospace data such as high-resolution satellite DEM, drone LiDAR point cloud, and InSAR surface deformation monitoring, combined with high-frequency ground sampling data from NB-IoT rain gauges and Beidou water level gauges, and personnel positioning data analyzed through communication signaling, to construct a multidimensional database covering topography, hydrology, and personnel distribution. This achieves full coverage of factors from macro-topography to micro-personnel dynamics, solving the problems of limited monitoring range and insufficient accuracy of traditional single data sources. Simultaneously, high-precision positioning and trajectory prediction are achieved: through personnel positioning and Kalman filter models, dynamic tracking of personnel coordinates with an accuracy of centimeters to 200 meters is achieved. Combined with the state-space model, the trajectory in signal-blocked scenarios is smoothly predicted to avoid positioning interruptions, providing real-time and continuous personnel distribution data support for disaster impact assessment.

[0258] 2. The present invention realizes refined model simulation to improve the accuracy of disaster prediction, through TOPMODEL and hydrodynamic coupling model: based on 500m×500m micro-grid division, TOPMODEL calculates terrain index and runoff rate, combines one- and two-dimensional hydrodynamic coupling model to simulate flood evolution process, outputs high-precision flooding grid map and flow velocity distribution data, realizes dynamic deduction of the whole process of flood runoff generation-confluence-flooding in small watershed, and improves the prediction accuracy of flood diffusion path and flooding depth under complex terrain compared with traditional models (such as water depth error ≤0.5m, flow velocity error ≤0.2m / s). At the same time, by adopting multi-physical process coupling: the one-dimensional river channel model uses the Saint-Venant equations to describe water flow movement, the two-dimensional flooding model considers ground roughness and flux conservation, and realizes energy conservation through flow interpolation, which effectively solves the problem of insufficient simulation of the coupling effect between river channel and flooding area in traditional models, and more realistically reflects the propagation characteristics of floods between different geomorphological units.

[0259] 3. The present invention implements dynamic risk assessment to achieve accurate classification of warning levels. It does this by constructing a three-dimensional risk assessment matrix: using water depth, flow velocity, and personnel thermal value as three-dimensional assessment dimensions, a discretized matrix containing 45 risk combinations is established. Through quantitative scoring (1-5 points) and level mapping (low risk, yellow warning, red warning), it achieves a transition from "one-way assessment of disaster intensity" to "comprehensive assessment of human-disaster-environment interaction." For example, when the water depth is 2.5m (medium level), the flow velocity is 1.8m / s (medium level), and the population is dense (level 5 thermal value), a red warning can be accurately triggered, which is more in line with the actual risk than the traditional warning mechanism based only on water depth or flow velocity. At the same time, combined with dynamic thermal field generation: using DPC clustering and KDE kernel density estimation, the distribution of personnel thermal values ​​is updated every 5 minutes, reflecting changes in the gathering of personnel in real time, ensuring that the risk assessment results are synchronized with the dynamics of the disaster site, and avoiding warning deviations caused by lagging personnel flow.

[0260] 4. The present invention realizes precise communication and path planning to improve the efficiency of emergency response, and adopts R-tree spatial index and TOPSIS base station sorting: through R-tree, it quickly matches the base stations associated with the disaster area, and combines the number of covered users, signal strength, historical warning success rate and other indicators to perform TOPSIS sorting, so as to realize the priority push of warning information to high-risk areas (such as sending red warnings to base stations covering 1,500 people first), which shortens the invalid information transmission time by more than 50% compared with broadcast push, and improves the warning response speed in key areas. At the same time, it realizes personalized transfer path generation: based on the weighted directed graph model and the improved A* algorithm, the optimal transfer route is generated by comprehensively considering the terrain cost, flood parameters, and population density, avoiding dangerous areas with water depth ≥1.5m or flow rate ≥2m / s, and providing estimated time and turning point coordinates to provide customized escape guidance for people in different locations. For example, the path planning for the transfer of people from a village to a highland shelter took only 25 minutes, which was 30% shorter than the traditional empirical route. A secondary warning mechanism was also introduced: by real-time monitoring of the status of stranded personnel, voice calls or emergency contact notifications were automatically triggered for abnormal personnel who stayed in non-safe areas for more than 30 minutes, and drones were linked to provide guidance, effectively solving the blind spot problem of traditional warnings for people who "did not receive information" or "had their movements blocked".

[0261] In summary, the present invention breaks through the bottlenecks of traditional water disaster warning in data fusion, simulation accuracy, evaluation dimension and response efficiency, and provides a set of high-precision, intelligent and feasible solutions for smart water conservancy and emergency management, which has significant social value and engineering application prospects.

[0262] The foregoing is merely a preferred embodiment of the present invention and is not intended to limit the present invention. A person skilled in the art will readily appreciate other embodiments of the present invention after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses, or adaptations of the present invention, which follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered merely exemplary, and the true scope and spirit of the present invention are indicated by the claims.

[0263] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for dynamic monitoring and accurate early warning of water disasters, characterized in that: The following steps are involved: Conduct multi-source data fusion modeling in advance: Obtain high-precision DEM and point cloud data through high-resolution satellites and UAV LiDAR, and build a multi-dimensional database covering topography, hydrology, and personnel distribution based on hydrological monitoring and communication signaling data; Conduct small watershed disaster evolution simulations, including: dividing the target small watershed into microgrids, calculating terrain indices and runoff rates based on TOPMODEL; constructing a one- and two-dimensional hydrodynamic coupling model using a one-dimensional river channel model and a two-dimensional inundation model, achieving energy conservation through flow interpolation, simulating flood evolution, and outputting inundation grid maps and flow velocity distribution data; The DPC clustering algorithm and KDE kernel density estimation method are used to process the personnel coordinate data, generate the personnel thermal field distribution, and calculate the personnel thermal value; Construct a three-dimensional risk assessment matrix based on water depth, flow velocity and personnel thermal value, and determine the warning level based on the three-dimensional risk assessment matrix; The R-tree spatial index algorithm is used to quickly match communication base stations associated with disaster areas, and the TOPSIS algorithm is used to sort the associated base stations, giving priority to pushing warning information to base stations covering high-risk areas. A weighted directed graph model is constructed, and an improved A* algorithm is used to comprehensively consider terrain cost, flood parameters, and population density to generate personnel transfer paths. The personnel locations are monitored in real time, and a secondary warning is triggered for personnel who stay in unsafe areas for more than a preset time.

2. The method for dynamic monitoring and accurate early warning of water disasters according to claim 1 is characterized in that: The communication signaling data includes: parsing the communication signaling and locating personnel, including the following steps: The coordinates of the three base stations are calibrated as (x1, y1), (x2, y2), and (x3, y3), and the signal arrival time difference is expressed as Δt 21 , Δt 31 , the speed of light c = 3 × 10 8 m / s, calculate the distance difference, expressed as: d 31 =c·Δt 31 ; d 21 =c·Δt 21 ; Construct the hyperbola equation and obtain the user coordinates (x, y); Among them, base station 1 and base station 2 form a hyperbola, which is expressed as: Among them, base station 1 and base station 3 form a hyperbola, which is expressed as:

3. The method for dynamic monitoring and accurate early warning of water disasters according to claim 2 is characterized in that: Also includes: Trajectory smoothing and movement trend prediction of personnel positioning data include the following steps: Build a state space model, including: calibrating the four-dimensional state vector Among them, (x t ,y t ) is the real-time coordinate obtained through communication signaling analysis, expressed as the initial value, is the horizontal moving speed, and the state transfer matrix F is based on uniform motion and is expressed as: Where Δt = 60s, corresponding to a high-frequency sampling interval of 1 time / minute, and the observation matrix H extracts the coordinate information in the state vector to match the actual observation value, which is expressed as: Perform recursive prediction, including the following steps: Calibrate the state X at the previous moment t-1 , predict the current state, expressed as: Update the prediction error covariance matrix, expressed as: P t - =F·P t-1 ·F T +Q; Among them, Q is the process noise matrix, which represents the uncertainty of the motion model, i.e., the speed change of personnel; Among them, if a valid observation value Z is received t =[x t′ ,y t′ ] T , calculate the observation residual, expressed as: By Kalman gain K t Fusion predictions and observations are expressed as: K t =P t - ·H T ·(H·P t - ·H T +R) -1 ; Where R = diag([200 2 ,200 2 ]) is the observation noise covariance matrix; The modified state vector and covariance matrix are expressed as: Among them, when the signal is blocked and the observation value is missing, the predicted state is used As a current estimate, avoid trajectory interruptions.

4. The method for dynamic monitoring and accurate early warning of water disasters according to claim 1 is characterized in that: The small watershed disaster evolution simulation includes: TOPMODEL simulation, including the following steps: The small watershed is divided into 500m×500m microgrids and the terrain index is calculated, which is expressed as: Determine the critical terrain index TI c , through the water balance equation Iterative solution; Calculate the flow rate, expressed as: Where β is the slope, A is the catchment area, θ is the soil moisture, i is the rainfall intensity, and K s is the saturated hydraulic conductivity.

5. The method for dynamic monitoring and accurate early warning of water disasters according to claim 1 is characterized in that: Also includes: The simulation of the one- and two-dimensional hydrodynamic coupling model includes the following steps: One-dimensional river channel model simulation: including: based on the Saint-Venant equations, using the Preissmann four-point implicit format discretization equation, dividing the river channel into N sections, with a time step of Δt and a space step of Δx; The equation is discretized and expressed as: Among them, Q is the cross-sectional flow, A is the cross-sectional area of ​​water flow, S0 is the riverbed slope, S f is the friction slope, expressed as: Where n is the Manning roughness coefficient, R is the hydraulic radius; The two-dimensional flooding model simulation includes the following steps: The finite volume method is calibrated to discretize the area into 500m×500m grids, and the flux is solved using the Roe scheme, considering the ground roughness n=0.03; The coupling interface is used to achieve energy conservation by interpolating and matching the one-dimensional river section flow and the two-dimensional grid boundary flow, which can be expressed as: Where M is the number of boundary cells in the two-dimensional grid, u i ,h i is the grid flow velocity and water depth.

6. The method for dynamic monitoring and accurate early warning of water disasters according to claim 1 is characterized in that: The calculation of the personnel thermal value comprises the following steps: Perform density peak clustering algorithm to identify core points where people gather, including: The calibration base station signaling coordinate set is Cutoff distance d c =500m; Calculate the local density, expressed as: Among them, ρ i is the local density, reflecting the number of people within 500m around point i, ρ i The larger the value, the higher the degree of aggregation; Calculate the distance metric: Among them, the core point is expressed as: δ i =min{d ij |ρ j >ρ i }; Among them, the edge point is represented by: δ i =max{d ij }; Among them, δ i As the distance index, the core point takes the minimum distance of the adjacent high-density points, and the edge point takes the global maximum distance, which is used to identify the cluster center; Perform cluster center identification and select ρ i and δ i The points with higher values ​​are taken as cluster centers to divide the gathering areas of people; Performing kernel density estimation includes the following steps: Calibrate the coordinates of people after clustering Bandwidth h = 500m; The thermal value is calculated as:

7. The method for dynamic monitoring and accurate early warning of water disasters according to claim 1 is characterized in that: The method of preferentially pushing warning information to base stations covering high-risk areas includes the following steps: The base station set B obtained by R-tree query is input as the candidate base station set; The calibration evaluation indicators are expressed as: number of covered users C1, mean signal strength C2, historical warning success rate C3, and their normalized matrices; Among them, the number of covered users C1, the number of users covered by the base station, is standardized as: The mean signal strength C2, base station signal strength, is normalized to: The historical warning success rate C3, a percentage value, is standardized as follows: Perform weighted matrix and ideal solution calculations; Among them, the weighted matrix is ​​expressed as: W = [0.5, 0.3, 0.2], where the number of users has the highest weight; Positive ideal solution, expressed as: Z + =(max(v i1 ),max(v i2 ),max(v i3 ); Negative ideal solution, expressed as: Z - =(min(v i1 ),min(v i2 ),min(v i3 ); Perform distance calculation and sorting, where the distances between base station i and the positive and negative ideal solutions are: Get comprehensive score S i , and press S i Arrange in descending order to obtain the base station priority for accurate push of warning information, expressed as:

8. The method for dynamic monitoring and accurate early warning of water disasters according to claim 1 is characterized in that: The triggering of a secondary warning for users who have stayed for longer than a preset time includes the following steps: Real-time monitoring of personnel location changes. If a person stays in a non-safe area for longer than the preset time and does not move according to the planned route, it is determined to be abnormal detention, and a secondary warning is triggered.

Images