Water conservancy project information management method and system based on big data

Abstract

The invention provides a hydraulic engineering information management method and system based on big data. The method comprises the following steps: firstly, uploading flood flow velocity dynamic data, terrain elevation data and spectral feature data to a cloud server, then carrying out multi-dimensional superposition on the flood flow velocity dynamic data, associating with a historical inundation path, and generating a superposition association result; constructing a three-dimensional geographic model based on terrain elevation data and spectral feature data, generating a spatial superposition map according to boundary features of a water body and a terrain abrupt change area and a flood routing path, performing position association with a material reserve position, and performing dynamic priority ranking according to a submerging probability gradient, and finally, according to the resource scheduling instruction sequence, correcting the flight path of the unmanned aerial vehicle cluster and the data acquisition area to form a closed-loop feedback mechanism. According to the technical scheme provided by the invention, the accuracy and timeliness of real-time data updating are ensured.

Description

Technical Field

[0001] This application relates to the technical field of water conservancy project management, and particularly to a water conservancy project information management method and system based on big data. Background Art

[0002] With the intensification of global climate change, the frequency and intensity of natural disasters such as floods are increasing, bringing unprecedented challenges to water conservancy project management; in flood warning and emergency response, obtaining dynamic information of flood areas in real time and accurately is crucial for reducing disaster losses;

[0003] Currently, in the field of water conservancy projects, there are already some technical means for monitoring flood conditions and conducting emergency management; for example, traditional hydrological station networks can provide water flow velocity and water level data at some key locations, while remote sensing technology can cover surface changes over a larger area; in addition, some simulation models based on historical data are also used to predict the development trend of floods; however, these methods usually rely on fixed monitoring points or low-resolution data sources, and it is difficult to comprehensively capture the dynamic changes and complexity of flood development; at the same time, existing systems often lack effective mechanisms to integrate data from multiple sources, limiting their efficiency and accuracy in practical applications. Summary of the Invention

[0004] This application provides a water conservancy project information management method and system based on big data to solve the problems of incomplete data acquisition, insufficient prediction accuracy, and low emergency response efficiency in the prior art.

[0005] In a first aspect, this application provides a water conservancy project information management method based on big data, including:

[0006] Receiving the dynamic data of flood flow velocity and terrain elevation data collected by an unmanned aerial vehicle (UAV) cluster and the spectral feature data of the flood area transmitted by satellite remote sensing equipment, and uploading the dynamic data of flood flow velocity, terrain elevation data, and spectral feature data to a cloud server;

[0007] Performing multi-dimensional superposition processing on the dynamic data of flood flow velocity in the cloud server, associating and matching the superimposed dynamic data of flood flow velocity with the inundation paths in historical flood events to generate a superimposed association result;

[0008] Constructing a three-dimensional geographical model of the flood area based on the terrain elevation data and spectral feature data, identifying the boundary features of the water body and terrain mutation areas in the three-dimensional geographical model, and combining the flood evolution path in the superimposed association result to generate a spatial superposition map containing the inundation probability gradient;

[0009] Associate the spatial superposition map with the material reserve locations in the preset emergency resource distribution map, perform dynamic priority sorting on the associated material reserve locations according to the flood submergence probability gradient, and generate a resource scheduling instruction sequence linked to the changes in the flood area based on the sorting result;

[0010] Feed the resource scheduling instruction sequence back to the control terminal of the UAV cluster, correct the flight trajectories of the UAVs in the UAV cluster and the data acquisition area, and generate a closed-loop feedback mechanism.

[0011] Optionally, construct a three-dimensional geographic model of the flood area based on the terrain elevation data and spectral feature data, identify the boundary features of the water body and terrain mutation areas in the three-dimensional geographic model, and generate a spatial superposition map including the flood submergence probability gradient in combination with the flood evolution path in the superposition association result, including:

[0012] Perform multi-scale wavelet decomposition on the spectral feature data of the flood area, extract the image texture features of different frequency bands, and generate a denoised high-resolution flood area image according to the spectral differences between the water body reflectance and surface coverings in the image texture features;

[0013] Divide the terrain elevation data into continuous terrain units according to the elevation mutation threshold, and construct a three-dimensional geographic grid model with terrain slope attributes based on the elevation gradient change rate of the continuous terrain units and the water body coverage range in the denoised high-resolution flood area image;

[0014] In the three-dimensional geographic grid model, identify the boundary features between the water body coverage range and adjacent terrain units in the continuous terrain units through a convolutional neural network, and extract the curvature parameter, slope mutation parameter, and water body diffusion direction parameter in the boundary features;

[0015] Perform spatio-temporal weight allocation on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the dynamic data of flood flow velocity, and generate a dynamic boundary feature vector representing the flood diffusion trend;

[0016] Align the dynamic boundary feature vector with the historical flood submergence path in the superposition association result, and calculate the submergence probability weight of each terrain unit in the dynamic boundary feature vector according to the proportional relationship between the flood diffusion rate in the historical flood submergence path and the current dynamic data of flood flow velocity;

[0017] Generate a submergence probability gradient parameter covering the flood area based on the submergence probability weight and the terrain slope attributes in the three-dimensional geographic grid model, and map the submergence probability gradient parameter to the high-resolution flood area image to generate a spatial superposition map including the flood submergence probability gradient.

[0018] Optionally, perform spatio-temporal weight assignment on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the dynamic data of flood flow velocity to generate a dynamic boundary feature vector representing the flood diffusion trend, including:

[0019] Perform time series decomposition on the dynamic data of flood flow velocity, extract the extreme value of the change rate and the deflection angle of the diffusion direction of flood flow velocity within a preset time window, and generate a velocity dynamic feature set;

[0020] According to the extreme value of the change rate in the velocity dynamic feature set, assign time weight factors to the curvature parameter, slope mutation parameter, and water body diffusion direction parameter;

[0021] Based on the spatial position relationship of continuous terrain units in the three-dimensional geographic grid model, calculate the cosine value of the angle between each terrain unit and the current flood diffusion direction, and assign spatial weight factors to the curvature parameter and slope mutation parameter according to the cosine value;

[0022] Couple and superimpose the time weight factor and the spatial weight factor according to the terrain unit to generate a comprehensive curvature weight, a comprehensive slope mutation weight, and a comprehensive water body diffusion direction weight;

[0023] Use the comprehensive curvature weight, the comprehensive slope mutation weight, and the comprehensive water body diffusion direction weight to perform weighted fusion on the curvature parameter, the slope mutation parameter, and the water body diffusion direction parameter to obtain the feature component of each terrain unit in the dynamic boundary feature vector;

[0024] According to the comprehensive weight of the water body diffusion direction in the feature component and the change rate of the elevation gradient of adjacent terrain units in the continuous terrain unit, correct the continuity of the diffusion trend of the dynamic boundary feature vector, and output a dynamic boundary feature vector matching the real-time evolution of the flood area.

[0025] Optionally, perform position association on the material storage positions in the spatial superposition map and the preset emergency resource distribution map, perform dynamic priority sorting on the associated material storage positions according to the flood probability gradient, and generate a resource scheduling instruction sequence linked to the changes in the flood area according to the sorting result, including:

[0026] Perform geographical coding on the material storage positions in the emergency resource distribution map, extract the longitude and latitude coordinates and material type identifiers of each material storage position, and generate a material storage feature set;

[0027] Perform spatial grid matching between the inundation probability gradient parameters of each topographic unit in the spatial overlay map and the longitude and latitude coordinates of the material reserve feature set, calculate the real-time inundation threat level of the spatial grid unit where the material reserve location is located, and assign a dynamic material demand weight coefficient to each material reserve location according to the product relationship between the material type identifier and the real-time inundation threat level;

[0028] Based on the topographic slope attribute of the three-dimensional geographic grid model, calculate the cumulative value of the shortest passage path slope from the material reserve location to the high inundation probability gradient area in the spatial overlay map, and generate a topographic passage resistance coefficient according to the cumulative value of the shortest passage path slope;

[0029] Construct a dynamic priority scoring function for each material reserve location according to the real-time inundation threat level, material demand weight coefficient, and topographic passage resistance coefficient;

[0030] Obtain the update result of the dynamic flood flow velocity data at a preset time interval, trigger the recalculation of the inundation probability gradient parameter according to the update result, dynamically adjust the real-time inundation threat level based on the recalculated inundation probability gradient parameter, trigger the iterative calculation of the priority scoring function, and generate a resource scheduling instruction sequence sorted by urgency according to the iterative calculation result.

[0031] Optionally, obtain the update result of the dynamic flood flow velocity data at a preset time interval, trigger the recalculation of the inundation probability gradient parameter according to the update result, dynamically adjust the real-time inundation threat level based on the recalculated inundation probability gradient parameter, trigger the iterative calculation of the priority scoring function, and generate a resource scheduling instruction sequence sorted by urgency according to the iterative calculation result, including:

[0032] Divide a preset time window according to the preset flood disaster emergency response level, obtain the latest dynamic flood flow velocity data collected by the UAV cluster at the end of each time window, and extract the timestamp and the extreme value of the flow velocity change rate in the latest dynamic flood flow velocity data;

[0033] Compare and analyze the extreme value of the flow velocity change rate with the historical data of the same period. When the extreme value of the flow velocity change rate in the analysis result exceeds the first preset threshold, recalculate the dynamic boundary feature vector;

[0034] Based on the recalculated dynamic boundary feature vector, update the inundation probability gradient parameters of each grid unit to obtain the updated inundation probability gradient parameters, and calculate the change amplitude of the updated inundation probability gradient parameters;

[0035] Adjust the real-time inundation threat level of the material reserve location according to the change range of the updated inundation probability gradient parameter, input the updated real-time inundation threat level into the dynamic priority scoring function, and dynamically calibrate the weight parameters in the dynamic priority scoring function according to the latest flood flow velocity data;

[0036] When the difference rate of the priority calibration results for consecutive preset times is less than the preset convergence threshold, generate a resource scheduling instruction sequence.

[0037] Optionally, feedback the resource scheduling instruction sequence to the control terminal of the UAV cluster, correct the flight trajectories of the UAVs in the UAV cluster and the data collection area, and generate a closed-loop feedback mechanism, including:

[0038] Analyze the area identifiers greater than the preset priority value in the resource scheduling instruction sequence, extract the longitude and latitude boundaries of each grid cell within the area identifiers greater than the preset priority value and the data collection accuracy requirements, and generate a focused collection task set for the UAV cluster;

[0039] According to the focused collection task set, calculate the trajectory deviation angle and coverage blind area range between the current flight trajectory of the UAV cluster and the target area, and based on the terrain slope attribute of the three-dimensional geographic grid model, generate an obstacle avoidance path correction parameter and a data collection frequency adjustment coefficient for each UAV;

[0040] Inject the obstacle avoidance path correction parameter into the UAV flight control module, dynamically adjust the pitch angle and heading angle of the UAV, so that the flight trajectory of the UAV converges along the terrain contour line towards the area greater than the preset priority value, and at the same time, synchronously modify the sampling interval of the UAV on-board sensor according to the data collection frequency adjustment coefficient;

[0041] After the UAV arrives at the target area, collect and send the latest flood flow velocity dynamic data and the latest terrain elevation data to the cloud server in real time, extract the spatio-temporal consistency verification codes in the latest flood flow velocity dynamic data and the latest terrain elevation data, and verify the geographical coverage coincidence degree between the data collection area and the requirements in the focused collection task set;

[0042] When the geographical coverage coincidence degree is lower than the second preset threshold, recalculate the obstacle avoidance path correction parameter according to the elevation gradient change rate of the missing area, and trigger the secondary trajectory correction of the UAV cluster; if the geographical coverage coincidence degree meets the second preset threshold, mark the latest flood flow velocity dynamic data and the latest terrain elevation data as data passed the closed-loop verification, and update the inundation probability gradient parameter in the spatial overlay map;

[0043] Generate a new resource scheduling instruction sequence according to the updated inundation probability gradient parameter, replace the executed or invalid instruction items in the original instruction sequence, and output the iterated closed-loop control instruction set.

[0044] Optionally, perform multi-dimensional superposition processing on the dynamic flood flow velocity data in the cloud server, and associate and match the superimposed dynamic flood flow velocity data with the inundation paths in historical flood events to generate a superimposed association result, including:

[0045] Perform spatio-temporal feature decomposition on the dynamic flood flow velocity data, extract the extreme values of velocity fluctuations, direction deflection angles, and spatial diffusion continuity indicators of each monitoring point according to a preset time window, and generate a spatio-temporal feature matrix;

[0046] According to the direction deflection angles in the spatio-temporal feature matrix, assign dynamic matching weights to each inundation path in historical flood events;

[0047] Based on the proportional relationship between the extreme values of velocity fluctuations and the peak velocity of the inundation paths recorded in historical flood events, calculate the acceleration factor of the current flood evolution rate relative to historical flood events, and generate a path association intensity coefficient in combination with the dynamic matching weights;

[0048] Perform raster comparison between the spatial diffusion continuity indicators and the spatial coverage density of each inundation path in historical flood events, screen out multiple historical inundation paths with a spatial overlap rate higher than a preset critical value, and use the multiple historical inundation paths as a candidate association path set;

[0049] According to the path association intensity coefficient of each historical inundation path in the candidate association path set and the terrain slope attribute of the three-dimensional geographic grid model, perform coupling correction to obtain a corrected path association intensity coefficient;

[0050] Aggregate the corrected path association intensity coefficients by spatial grid units to generate a superimposed association result covering the flood area, where the superimposed association result includes the spatial matching degree and evolution trend confidence level of each grid unit and historical paths.

[0051] In a second aspect, the present application provides a water conservancy project information management system based on big data, including:

[0052] A receiving module, configured to receive the dynamic flood flow velocity data and terrain elevation data collected by the UAV cluster and the spectral feature data of the flood area transmitted by the satellite remote sensing device, and upload the dynamic flood flow velocity data, terrain elevation data, and spectral feature data to the cloud server;

[0053] A matching module, configured to perform multi-dimensional superposition processing on the dynamic flood flow velocity data in the cloud server, and associate and match the superimposed dynamic flood flow velocity data with the inundation paths in historical flood events to generate a superimposed association result;

[0054] An identification module, configured to construct a three-dimensional geographical model of the flood area based on the terrain elevation data and spectral feature data, identify the boundary features of the water body and terrain mutation area in the three-dimensional geographical model, and generate a spatial superposition map containing the inundation probability gradient in combination with the flood evolution path in the superposition association result;

[0055] A sorting module, configured to perform position association between the spatial superposition map and the material reserve positions in a preset emergency resource distribution map, perform dynamic priority sorting on the associated material reserve positions according to the inundation probability gradient, and generate a resource scheduling instruction sequence linked to the changes in the flood area according to the sorting result;

[0056] A correction module, configured to feedback the resource scheduling instruction sequence to the control terminal of the UAV cluster, correct the flight trajectories and data acquisition areas of the UAVs in the UAV cluster, and generate a closed-loop feedback mechanism.

[0057] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for managing water conservancy project information based on big data as described in the first aspect above.

[0058] In a fourth aspect, an embodiment of the present application provides a computer storage medium, storing a computer program, and when the computer program is executed by a computer, it implements a method for managing water conservancy project information based on big data as described in the first aspect.

[0059] In the embodiment of the present application, by integrating the dynamic data of the flood flow velocity collected by the UAV cluster, the terrain elevation data, and the spectral feature data of the flood area provided by the satellite remote sensing device, and uploading them to the cloud server for processing, accurate prediction and simulation of the flood evolution path are realized; this method can not only generate a spatial superposition map containing the inundation probability gradient, but also dynamically adjust the distribution and scheduling scheme of emergency resources according to this information, greatly improving the scientificity and response speed of disaster prevention and mitigation measures; in addition, by correcting the data acquisition area and flight trajectory of the UAV through the closed-loop feedback mechanism, the accuracy and timeliness of real-time data update are ensured, thereby effectively improving the overall efficiency of water conservancy project information management;

[0060] Furthermore, in the process of constructing the 3D geographical model of the flood area, the multi-scale wavelet decomposition technology is adopted to extract the image texture features of different frequency bands, significantly improving the resolution and clarity of the flood area images, and effectively removing the noise interference at the same time. Based on this, the water body coverage range and its boundary features are identified through a convolutional neural network, and the spatio-temporal weight distribution of each parameter is carried out in combination with the dynamic data of the flood flow velocity, generating a dynamic boundary feature vector representing the flood diffusion trend, realizing the accurate simulation of the flood diffusion process. Finally, the inundation probability weight of each terrain unit is calculated according to the proportional relationship between the historical inundation path and the current flood situation, and mapped into the high-resolution flood area image to form a multi-level inundation probability gradient parameter. This process not only enhances the accuracy of flood warning, but also provides a more intuitive and detailed spatial overlay map for decision-makers, helping to formulate more effective emergency plans and resource scheduling strategies.

[0061] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0063] Figure 1 Shows a flowchart of a water conservancy project information management method based on big data provided by the present application;

[0064] Figure 2 Shows a schematic structural diagram of a water conservancy project information management system based on big data provided by the present application;

[0065] Figure 3 Shows a schematic structural diagram of a computing device provided by the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] In order to enable those skilled in the art to better understand the solutions of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application.

[0067] In some of the processes described in the specification, claims, and the above-mentioned drawings of this application, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0068] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0069] Figure 1 The flowchart of a water conservancy project information management method based on big data provided for the embodiments of the present application is as Figure 1 shown, and the method includes:

[0070] Step 101, receiving the dynamic flood flow velocity data and terrain elevation data collected by the UAV cluster and the spectral feature data of the flood area transmitted by the satellite remote sensing device, and uploading the dynamic flood flow velocity data, terrain elevation data, and spectral feature data to the cloud server;

[0071] In this step, the UAV cluster refers to a group of unmanned aerial vehicles connected through a wireless network and working in cooperation. They can automatically execute data collection tasks according to a preset path; the dynamic flood flow velocity data includes information such as water flow velocity and direction that changes over time, and these data are obtained in real time through sensors on the UAVs; the terrain elevation data is a dataset describing the change in surface height, which reflects the undulation of the terrain in the area and is usually obtained by lidar or photogrammetry technology carried by the UAV; the spectral feature data of the flood area refers to the different frequency information of the electromagnetic waves reflected by the ground objects captured by the satellite remote sensing device, which is used to distinguish water bodies from other ground objects; the cloud server is a group of high-performance computer systems located in a remote data center, which is used to store, process, and analyze massive data, and provide computing resources and services; by uploading the above three types of data to the cloud server, comprehensive monitoring and analysis of flood disasters can be achieved.

[0072] In this embodiment, first, the UAV swarm operates according to a preset flight route, collects dynamic data of flood flow velocity using the speedometers and GPS modules equipped on them, and simultaneously records terrain elevation data through laser rangefinders; secondly, satellite remote sensing equipment scans ground targets according to specific frequency bands and collects spectral feature data of the flood area; then, all the collected data is encoded into a standardized format and sent to the cloud server through a wireless communication link; in the cloud, a distributed file system and a big data processing framework (such as Hadoop or Spark) are used to receive, store, and preliminarily process this data; among them, the dynamic data of flood flow velocity is processed by a filtering algorithm to remove noise interference, the terrain elevation data fills in missing values through interpolation, and the spectral feature data uses image enhancement technology to improve the resolution; finally, these preprocessed data are integrated together, laying a foundation for subsequent multi-dimensional overlay processing and three-dimensional geographical model construction.

[0073] For example, in the face of the upcoming rainy season, the water conservancy management department of a certain area activates a UAV swarm for round-the-clock monitoring; the UAV swarm flies over potential flood risk areas multiple times a day, accurately records the flood flow velocity and its change trend, and simultaneously obtains the latest terrain elevation information; at the same time, a dedicated Earth observation satellite regularly scans this area and transmits detailed spectral feature data; these valuable data are quickly transmitted to the cloud server in the data center located in the suburban area of the city; here, engineers use advanced machine learning algorithms to deeply mine the data, not only effectively improving the quality of the data, but also providing strong support for formulating precise flood control strategies; the whole process ensures the timeliness and accuracy of the data, greatly enhancing the ability to respond to sudden flood events.

[0074] Step 102, perform multi-dimensional overlay processing on the dynamic data of flood flow velocity in the cloud server, associate and match the superimposed dynamic data of flood flow velocity with the inundation paths in historical flood events, and generate a superimposed association result;

[0075] In this step, multi-dimensional overlay processing refers to comprehensively analyzing the dynamic data of flood flow velocity using a variety of algorithms and technologies, including time series analysis, spatial distribution analysis, etc., to comprehensively capture the development trend and characteristics of floods; historical flood events refer to the records of past flood disasters, including information such as inundation paths and affected areas; the inundation path is a trajectory describing the changing area covered by the flood during the diffusion process over time; association and matching is to compare and analyze the current dynamic data of flood flow velocity with historical inundation paths through specific algorithms to find similarities and differences; the superimposed association result is a dataset integrating the current flood situation and historical flood patterns, used to support subsequent inundation prediction and emergency response planning; these technical means work together to help achieve precise assessment of flood risks;

[0076] In this embodiment, first, the spatio-temporal feature decomposition technology is adopted to process the dynamic data of flood flow velocity. The extreme values of velocity fluctuations, the direction deflection angles, and the spatial diffusion continuity indexes of each monitoring point in different time periods are extracted through the sliding window technology to form a spatio-temporal feature matrix. Next, a classification algorithm in machine learning (such as decision tree or random forest) is used to assign dynamic matching weights to each inundation path in historical flood events, and the acceleration factor of the current flood evolution rate relative to historical flood events is calculated based on the extreme values of velocity fluctuations. Then, the path correlation intensity coefficient is generated by combining the dynamic matching weights, and the spatial diffusion situation of the current flood is compared with the historical inundation paths by using the rasterization method, and the historical inundation paths with an overlap rate higher than the preset threshold are selected as the candidate associated path set. Finally, using the Geographic Information System (GIS) tool, the candidate path set is corrected according to the terrain slope attribute of the three-dimensional geographic grid model to obtain the final superimposed association result. The whole process ensures that the current flood situation can be effectively combined with historical data, improving the accuracy of prediction.

[0077] For example, after a certain area encounters heavy rainfall, the cloud server starts to receive the real-time dynamic data of flood flow velocity transmitted by the drone cluster and satellite remote sensing equipment. First, the dynamic data of flood flow velocity is preprocessed to remove noise and fill in missing values. Then, the time series analysis function is used to extract key features, and the random forest algorithm is used to assign dynamic matching weights to the inundation paths in the historical flood event database. Through the simulation analysis of multiple historical events, it is found that the diffusion trend of this flood is highly similar to that of a major flood ten years ago. Subsequently, the QGIS software is used to further optimize and adjust the selected historical inundation paths, and a targeted emergency plan is formulated to continuously update the superimposed association result to ensure that the flood control measures are always in the optimal state.

[0078] Step 103: Based on the terrain elevation data and spectral feature data, construct a three-dimensional geographic model of the flood area, identify the boundary features of the water body and terrain mutation areas in the three-dimensional geographic model, and combine the flood evolution paths in the superimposed association result to generate a spatial superimposed map containing the inundation probability gradient.

[0079] In this step, the three-dimensional geographical model of the flood area is a virtual three-dimensional space constructed based on terrain elevation data and spectral feature data, which can intuitively display the geomorphic features of the flood-affected area; the boundary features of the water body and the terrain mutation area refer to the water body range identified by analyzing the three-dimensional geographical model and the location information of significant terrain changes around it; the flood evolution path describes the specific route of the flood spreading over time; the inundation probability gradient represents the likelihood of inundation at different geographical locations, usually displayed on a map in the form of colors or values; the spatial overlay map is a map that synthesizes information such as terrain, water body distribution, inundation probability, and flood evolution path, and is used to guide flood prevention and disaster relief work; these elements work together to help decision-makers better understand and respond to flood disasters;

[0080] In this embodiment, first, multi-scale wavelet decomposition is performed on the spectral feature data of the flood area to extract image texture features, and a denoised high-resolution flood area image is generated based on the spectral differences between the water body reflectivity and the surface cover; then, the continuous terrain units are divided using the terrain elevation data according to the set elevation mutation threshold, and a three-dimensional geographical grid model with terrain slope attributes is constructed by combining the elevation gradient change rate and the water body coverage range in the denoised flood area image; then, convolutional neural network (CNN) technology is applied in the three-dimensional geographical grid model to identify the boundary features of the water body coverage range and the terrain mutation area, and multiple features including curvature parameters, slope mutation parameters, and water body diffusion direction parameters are extracted; subsequently, the spatio-temporal weight assignment algorithm is used to adjust these feature parameters according to the dynamic data of the flood flow velocity to generate a dynamic boundary feature vector; finally, the dynamic boundary feature vector is spatio-temporally aligned with the historical inundation path in the overlay correlation result, the inundation probability weight of each terrain unit is calculated, and it is mapped into the three-dimensional geographical grid model to generate a spatial overlay map containing the inundation probability gradient, ensuring the accuracy and intuitiveness of flood risk assessment;

[0081] For example, in the face of the upcoming rainy season, unmanned aerial vehicle (UAV) clusters and satellite remote sensing equipment are activated to obtain the latest terrain elevation data and spectral feature data of the flood area; first, MATLAB software is used to perform multi-scale wavelet decomposition on the spectral feature data to generate a clear flood area image; then, ArcGIS software is used to construct a detailed three-dimensional geographical model in combination with the terrain elevation data, and CNN technology is applied in it to accurately identify the water body boundary and the terrain mutation area; by learning historical flood events, a prediction model of the flood evolution path is established and combined with the current flood situation to generate a spatial overlay map containing the inundation probability gradient; as the system continuously updates the overlay map, the effectiveness and timeliness of flood prevention measures are ensured.

[0082] Step 104: Perform location association between the spatial overlay map and the material reserve locations in the preset emergency resource distribution map, conduct dynamic priority ranking on the associated material reserve locations according to the inundation probability gradient, and generate a resource scheduling instruction sequence linked to the changes in the flood area based on the ranking result;

[0083] In this step, the emergency resource distribution map is a map containing information such as the reserve locations of various emergency materials, their types, quantities, etc.; the material reserve locations refer to the specific storage places of these emergency materials; the resource scheduling instruction sequence linked to the changes in the flood area is a material allocation plan dynamically adjusted according to the inundation probability gradient, aiming to ensure that the required materials can be quickly and effectively distributed to the places where they are most needed in case of emergency; by associating the spatial overlay map with the emergency resource distribution map and ranking the material reserve locations based on the inundation probability gradient, effective management and rapid response to resources can be achieved;

[0084] In this embodiment, with the generation of the 3D geographical model and the spatial overlay map, the water conservancy management department obtained a detailed flood risk assessment report; next, engineers used Python scripts combined with the GDAL library to perform geocoding processing on the material reserve locations in the emergency resource distribution map, extracted the longitude and latitude coordinates and material type identifiers of each material reserve point, and generated a material reserve feature set; then, using GIS technology, the inundation probability gradient parameters of each terrain unit in the spatial overlay map were matched with the longitude and latitude coordinates of the material reserve feature set to calculate the real-time inundation threat level of the spatial grid unit where each material reserve location is located; then, according to the material type and the real-time inundation threat level, the dynamic material demand weight coefficient of each material reserve location was calculated, and combined with the terrain slope attribute of the 3D geographical grid model, the cumulative value of the shortest passage path slope from the material reserve location to the high inundation probability gradient area was calculated to generate a terrain passage resistance coefficient; based on the above calculation results, a dynamic priority scoring function for each material reserve location was constructed, and the updated result of the dynamic flood flow velocity data was obtained at preset time intervals, the inundation probability gradient parameters were recalculated, and the priority scoring function of the material reserve location was dynamically adjusted; finally, a resource scheduling instruction sequence was generated according to the latest calculation results to ensure that the resource allocation strategy is synchronized with the development trend of the flood;

[0085] For example, when a certain area is threatened by heavy rainfall, the system first performs geocoding on the received spatial overlay map and the emergency resource distribution map, accurately marking the locations of all material storage points; after successfully generating the spatial overlay map containing the flood probability gradient, it immediately analyzes the risk status of each material storage point; subsequently, it uses a Python script combined with the GDAL library to perform spatial matching analysis on the two sets of data to determine the current flood threat level faced by each storage point; based on this analysis result, it calculates the dynamic material demand weight coefficient for each storage point, and formulates the optimal material allocation route considering terrain factors; then it continuously monitors the flood dynamics, automatically updates the flood probability gradient parameters every 30 minutes, and adjusts the priority ranking of the material storage locations accordingly to generate the latest resource scheduling instructions.

[0086] Step 105: Feed the resource scheduling instruction sequence back to the control terminal of the UAV cluster to correct the flight trajectories of the UAVs in the UAV cluster and the data acquisition area, and generate a closed-loop feedback mechanism.

[0087] In this step, the control terminal of the UAV cluster refers to the central control system used to manage and control the flight tasks of the UAV group, which receives and processes the resource scheduling instruction sequence from the cloud server; the UAV flight trajectory refers to the route along which each UAV flies according to a preset path or a real-time adjusted path; the data acquisition area is the specific geographical range where the UAVs perform data acquisition tasks; by feeding the resource scheduling instructions back to the control terminal of the UAV cluster and dynamically correcting the flight trajectories and data acquisition areas of the UAVs according to the latest flood conditions, a closed-loop feedback mechanism can be formed to ensure the real-time update and efficient utilization of information.

[0088] In this embodiment, first, the system parses the high-priority area identifiers in the resource scheduling instruction sequence, extracts the longitude and latitude boundaries of these areas and the required data acquisition accuracy requirements, and generates a task set for the UAV cluster; then, uses a path planning algorithm to calculate the deviation angle between the current flight trajectory of the UAV and the target area and the coverage blind area range, and generates obstacle avoidance path correction parameters and data acquisition frequency adjustment coefficients for each UAV based on the terrain slope attribute of the three-dimensional geographic grid model; then, injects the obstacle avoidance path correction parameters into the flight control module of the UAV, and dynamically adjusts the pitch angle and heading angle of the UAV through a PID controller to make it converge along the terrain contour line towards the high-priority area; at the same time, modifies the sampling interval of the UAV-borne sensor according to the data acquisition frequency adjustment coefficient to adapt to the data acquisition requirements of different areas; after the UAV arrives at the target area, uses the multi-sensor fusion technology (such as lidar, optical camera, and infrared camera) it carries to collect the latest dynamic data of flood flow velocity and terrain elevation data in real time, and sends them to the cloud server through a wireless communication link; the server side performs spatio-temporal consistency verification on these data to verify whether the area where the data is collected coincides with the geographical coverage required in the task set; if the geographical coverage coincidence degree is lower than the set threshold, recalculate the obstacle avoidance path correction parameters according to the elevation gradient change rate of the missing area, and trigger a secondary trajectory correction; if the coincidence degree meets the standard, mark the verified data as data passed the closed-loop verification, and update the inundation probability gradient parameters in the spatial overlay map; finally, generate a new resource scheduling instruction sequence based on the updated inundation probability gradient parameters, replace the executed or invalid instruction items in the original instruction sequence, and output the iterated closed-loop control instruction set;

[0089] For example, in actual operation, when a certain area is threatened by heavy rainfall, the system generates a spatial overlay map containing the inundation probability gradient and formulates a resource scheduling instruction sequence accordingly; the water conservancy management department quickly feeds back these instructions to the control terminal of the UAV cluster; first, determines the high-risk areas that need to be monitored key points and sets the corresponding data acquisition accuracy requirements; then uses a PID controller to dynamically adjust the flight attitude of the UAVs to ensure that they can fly accurately to the designated area and avoid obstacles; when the UAVs start to execute tasks, uses the various sensors it carries, including lidar and optical cameras, to collect the latest flood flow velocity and terrain data in real time, and transmits this information back to the cloud server through the 4G / 5G network for analysis; once it is found that the data coverage in some areas is insufficient, the system will automatically trigger a secondary trajectory correction to ensure that all high-risk areas are fully monitored.

[0090] In flood disaster monitoring, existing methods are difficult to accurately identify the water body boundary and terrain mutation areas, which results in inaccurate prediction of flood evolution paths. Based on this, in some embodiments, as described in step 103, a three-dimensional geographical model of the flood area is constructed based on the terrain elevation data and spectral feature data, the boundary features of the water body and terrain mutation areas in the three-dimensional geographical model are identified, and a spatial overlay map including the flood inundation probability gradient is generated by combining the flood evolution path in the overlay association result, including:

[0091] Step 201: Perform multi-scale wavelet decomposition on the spectral feature data of the flood area, extract image texture features in different frequency bands, and generate a denoised high-resolution flood area image according to the spectral differences between the water body reflectivity and surface coverings in the image texture features;

[0092] In this step, multi-scale wavelet decomposition is a signal processing technique that decomposes an image through filter banks at different scales to extract information in different frequency bands; the image texture features in different frequency bands refer to the detailed information in different frequency ranges extracted from the image, which reflect the texture characteristics of surface coverings and water bodies; the water body reflectivity refers to the ability of the water body surface to reflect light, which is usually related to factors such as water quality and lighting conditions; surface coverings refer to non-water components such as vegetation and buildings in the flood area; the spectral difference refers to the different degrees of reflection or absorption of electromagnetic waves by different ground objects (such as water bodies and vegetation) at specific wavelengths.

[0093] In this embodiment, first, perform multi-scale wavelet decomposition on the spectral feature data of the flood area to decompose the original image into sub-band images in multiple frequency bands; each sub-band image represents the detailed information at different scales and frequencies; then, use a convolutional neural network (CNN) model to extract features from each sub-band image, identify the image texture features, and classify them according to the spectral differences between the water body reflectivity and surface coverings; through the trained deep learning model, automatically distinguish the water body from other ground objects, remove noise interference and enhance target features; then, use image fusion technology to recombine each sub-band image to generate a denoised high-resolution flood area image; this process not only retains the detailed information of the original image, but also significantly improves the quality and readability of the image; finally, verify its accuracy and reliability by comparing and analyzing the generated image with historical data to ensure that the final result can be used for subsequent three-dimensional geographical modeling and inundation probability prediction.

[0094] Step 202: Divide the terrain elevation data into continuous terrain units according to the elevation mutation threshold, and construct a three-dimensional geographical grid model with terrain slope attributes based on the elevation gradient change rate of the continuous terrain units and the water body coverage range in the denoised high-resolution flood area image;

[0095] In this step, the elevation mutation threshold refers to the height change boundary used to distinguish different terrain features (such as flat ground, hillslopes, etc.); the continuous terrain unit is a geomorphic unit divided according to the elevation mutation threshold, and the terrain within each unit is relatively flat or has a similar slope; the water body coverage range refers to the geographical area occupied by the water body in the flood area image; the three-dimensional geographical grid model with terrain slope attributes is a virtual three-dimensional space constructed based on terrain elevation data and the water body coverage range, where each grid unit contains not only elevation information but also slope and other terrain characteristics; through these parameters and technical means, a detailed three-dimensional geographical model can be constructed to provide a basis for flood simulation;

[0096] In this embodiment, first, the terrain elevation data is processed, and the terrain is divided into multiple continuous terrain units according to the set elevation mutation threshold; by calculating the elevation difference between adjacent points, when the difference exceeds the preset threshold, it is considered that the point belongs to a new terrain unit; then, the ArcGIS software is used to calculate the elevation gradient change rate of each continuous terrain unit, and it is superimposed and analyzed with the water body coverage range in the denoised high-resolution flood area image; on this basis, a three-dimensional modeling software (such as QGIS or GRASS GIS) is used in combination with the triangulated irregular network interpolation method (TIN) to generate a three-dimensional geographical grid model with terrain slope attributes; each grid unit not only records its elevation value but also contains terrain characteristics such as slope and curvature; finally, by comparing and analyzing the generated three-dimensional geographical grid model with historical terrain data, its accuracy and reliability are verified to ensure that the final result can be used for subsequent flood inundation path prediction and emergency response planning.

[0097] Step 203, in the three-dimensional geographical grid model, identify the boundary features between the water body coverage range and adjacent terrain units in the continuous terrain unit through a convolutional neural network, and extract the curvature parameter, slope mutation parameter, and water body diffusion direction parameter in the boundary features;

[0098] In this step, the convolutional neural network (CNN) is a deep learning model commonly used for image recognition and feature extraction; the boundary features between adjacent terrain units refer to the areas where the terrain elevation changes significantly, usually manifested as slope mutations or curvature changes; the curvature parameter is an index describing the degree of terrain bending, reflecting the concave and convex characteristics of the terrain surface; the slope mutation parameter represents the position where the terrain slope changes sharply, usually related to terrain mutation points; the water body diffusion direction parameter refers to the direction in which the flood flows under different terrain conditions;

[0099] In this embodiment, first, a pre-trained convolutional neural network model is used to process the three-dimensional geographical grid model to identify the boundary features between the water body coverage range and the continuous terrain units; the three-dimensional geographical grid model is converted into a two-dimensional slice image and input into the CNN model for feature extraction; then, an edge detection algorithm (such as Canny edge detection) is used to further refine the boundary features, and the curvature parameter, slope mutation parameter, and water body diffusion direction parameter are extracted; then, a mathematical morphology method is used to post-process the extracted features to remove noise and enhance the clarity of the key features; finally, combined with the terrain elevation data and water body coverage range information, a comprehensive terrain analysis result containing all the extracted features is generated.

[0100] Step 204, perform spatio-temporal weight assignment on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the dynamic flood flow velocity data to generate a dynamic boundary feature vector representing the flood diffusion trend;

[0101] In this step, spatio-temporal weight assignment refers to weighting the terrain feature parameters (such as curvature, slope mutation, and water body diffusion direction) according to the time series and spatial distribution characteristics of the dynamic flood flow velocity data; the flood diffusion trend describes the expansion pattern of the flood under different time and space conditions; the dynamic boundary feature vector is a vector that combines time, space, and terrain features and is used to represent the direction and speed of flood diffusion.

[0102] In this embodiment, first, the dynamic flood flow velocity data is decomposed by time series to extract the extreme values of flow velocity fluctuations, direction deflection angles, and spatial diffusion continuity indicators at each monitoring point in different time periods to form a spatio-temporal feature matrix; then, a machine learning algorithm (such as random forest or XGBoost) is used to calculate the spatio-temporal weight factors of each terrain unit based on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter in the three-dimensional geographical grid model; the time weight factors are assigned to the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the extreme values of flow velocity fluctuations, and the spatial weight factors are calculated based on the spatial position relationship of adjacent terrain units; then, the time weight factors and spatial weight factors are coupled and superimposed according to the terrain unit to generate a comprehensive weight; finally, these comprehensive weights are used to perform weighted fusion on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter to obtain the feature components of each terrain unit in the dynamic boundary feature vector.

[0103] Step 205, perform spatio-temporal alignment of the dynamic boundary feature vector with the historical inundation path in the superimposed association result, and calculate the inundation probability weight of each terrain unit in the dynamic boundary feature vector according to the proportional relationship between the flood diffusion rate in the historical inundation path and the current dynamic flood flow velocity data.

[0104] In this step, the historical inundation path refers to the specific path and coverage area of flood spread during past flood events; the flood spread rate is an indicator describing the speed of flood expansion within a specific time period; the inundation probability weight is the likelihood of inundation for each terrain unit calculated based on historical data and current dynamic flood flow velocity data; through these parameters and technical means, the dynamic boundary feature vector can be spatially and temporally aligned with the historical inundation path, and the inundation probability weight of each terrain unit can be calculated to provide support for subsequent inundation prediction;

[0105] In this embodiment, first, the historical inundation path data is processed to extract the spatial coverage density and flood spread rate of each historical inundation path; then, the dynamic time warping (DTW) algorithm is used to spatially and temporally align the dynamic boundary feature vector with the historical inundation path to ensure the corresponding relationship between the two in terms of time and space; then, a scale factor is calculated according to the ratio relationship between the flood spread rate in the historical inundation path and the current dynamic flood flow velocity data; by comparing the historical flood spread rate with the current dynamic flood flow velocity data, the difference between the two is determined, and the inundation probability weight of each terrain unit is adjusted accordingly; finally, combined with the terrain slope attribute in the three-dimensional geographic grid model, multi-level inundation probability gradient parameters covering the flood area are generated and mapped to the high-resolution flood area image to generate a spatial overlay map containing the inundation probability gradient; this process not only improves the accuracy of inundation probability prediction but also enhances the understanding ability of the flood evolution path.

[0106] Step 206: Based on the inundation probability weight and the terrain slope attribute in the three-dimensional geographic grid model, generate inundation probability gradient parameters covering the flood area, and map the inundation probability gradient parameters to the high-resolution flood area image to generate a spatial overlay map containing the inundation probability gradient;

[0107] In this step, the terrain slope attribute refers to the slope information of each terrain unit in the three-dimensional geographic grid model, reflecting the undulation degree of the terrain; the inundation probability gradient parameter is the likelihood distribution of inundation for each area calculated based on the inundation probability weight and the terrain slope attribute; the inundation probability gradient represents the probability difference of inundation at different geographical locations, usually displayed in the form of colors or values; the spatial overlay map is a map integrating information such as terrain, water body distribution, inundation probability, and flood evolution path, used to guide flood prevention and disaster relief work;

[0108] In this embodiment, first, the terrain slope attribute in the three-dimensional geographic grid model is used to perform weighted processing on the inundation probability weight of each terrain unit, generating a multi-level inundation probability gradient parameter; the inundation probability weight is adjusted according to the steepness of the terrain slope, with a lower inundation probability weight assigned to areas with a larger slope and a higher inundation probability weight assigned to flat areas; then, the generated multi-level inundation probability gradient parameter is mapped into the high-resolution flood area image using image processing techniques (such as the OpenCV library); through an interpolation algorithm (such as bilinear interpolation), the smooth transition of the inundation probability gradient parameter in the image is ensured, and a spatial overlay map containing the inundation probability gradient is generated; then, the generated spatial overlay map is verified and optimized to ensure its accuracy and readability.

[0109] Since the time series characteristics of the flood flow velocity dynamic data are crucial for the analysis of the flood diffusion trend, but the prior art often ignores the change rate extreme value and the diffusion direction deflection angle in the time dimension. Based on this, in some embodiments, as described in step 204, spatio-temporal weight allocation is performed on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the flood flow velocity dynamic data, generating a dynamic boundary feature vector characterizing the flood diffusion trend, including:

[0110] Step 301, perform time series decomposition on the flood flow velocity dynamic data, extract the change rate extreme value and the diffusion direction deflection angle of the flood flow velocity within a preset time window, and generate a flow velocity dynamic feature set;

[0111] In this step, time series decomposition is a technique for decomposing time series data into trend, seasonal, and random components; the change rate extreme value refers to the maximum or minimum change rate of the flood flow velocity within a specific time window, reflecting the severity of the change in the flood flow velocity; the diffusion direction deflection angle refers to the change angle of the flood flow direction at different time points, used to describe the direction change of the flood diffusion; the flow velocity dynamic feature set is a data set containing the above features, used for subsequent flood prediction and analysis; through these parameters and technical means, a feature set comprehensively reflecting the dynamic characteristics of the flood flow velocity can be generated, providing support for the prediction of the flood evolution path;

[0112] In this embodiment, first, the dynamic data of flood flow velocity is decomposed by time series to extract the trend component, seasonal component, and random component. Then, the extreme values of the change rate within each time window are calculated using the sliding window technique (such as the rolling function in Pandas). The change rate is determined by calculating the velocity difference between adjacent time points, and the maximum and minimum change rates within each window are found. Next, the trigonometric function is used to calculate the deflection angle of the flood diffusion direction. By comparing the flow direction vectors of adjacent time points, the included angle is calculated as the deflection angle. Finally, the extreme values of the change rate and the deflection angle of the diffusion direction are combined into a set of flow velocity dynamic features and stored in a structured data format for subsequent analysis.

[0113] Step 302: According to the extreme values of the change rate in the flow velocity dynamic feature set, assign time weight factors to the curvature parameter, slope mutation parameter, and water body diffusion direction parameter.

[0114] In this step, the time weight factor refers to the time importance coefficient assigned to the terrain parameters (such as the curvature parameter, slope mutation parameter, and water body diffusion direction parameter) according to the extreme values of the change rate in the flood flow velocity dynamic feature set. These weight factors reflect the degree of change in flood flow velocity in different time periods, thus helping to more accurately evaluate the impact of flood diffusion.

[0115] In this embodiment, first, the flow velocity dynamic feature set is processed to extract the extreme values of the change rate within each time window. Then, the time weight factors are calculated based on the extreme values of the change rate. By setting a reference change rate and determining the time weight factors according to the ratio of the actual change rate to the reference change rate within each time window. For example, if the change rate in a certain period is significantly higher than the reference change rate, the time weight factor for this period is larger; otherwise, it is smaller. Then, machine learning algorithms (such as linear regression or random forest) are used to assign the calculated time weight factors to the curvature parameter, slope mutation parameter, and water body diffusion direction parameter. The specific process includes: using the time weight factor as the input feature and the terrain parameter as the target variable, training the model and generating the weighted terrain parameter. Finally, the weighted terrain parameter is stored in a structured data format for subsequent analysis.

[0116] Step 303: Based on the spatial position relationship of the continuous terrain units in the three-dimensional geographic grid model, calculate the cosine value of the included angle between each terrain unit and the current flood diffusion direction, and assign spatial weight factors to the curvature parameter and slope mutation parameter according to the cosine value.

[0117] In this step, the spatial position relationship of continuous terrain units refers to the relative positions between adjacent terrain units in the three-dimensional geographical grid model; the current flood diffusion direction refers to the main direction of flood flow determined based on real-time monitoring data; the cosine value of the angle is a value obtained by calculating the cosine of the angle between the normal vector of each terrain unit and the flood diffusion direction, which is used to quantify the impact of the terrain on flood diffusion; the spatial weight factor is an importance coefficient assigned to the curvature parameter and the slope mutation parameter based on the cosine value of the angle, reflecting the importance of different terrain units in the flood diffusion path; the combined action of these parameters and technical means can more accurately evaluate the impact of flood diffusion;

[0118] In this embodiment, first, the normal vector of each terrain unit is calculated using the spatial position relationship of continuous terrain units in the three-dimensional geographical grid model; then, the current flood diffusion direction is determined using the dynamic flood flow velocity data, and the cosine value of the angle between the normal vector of each terrain unit and the flood diffusion direction is calculated; for each terrain unit, the normal vector is calculated through its vertex coordinates, and then the cosine value of the angle is calculated using the vector dot product formula; then, spatial weight factors are assigned to the curvature parameter and the slope mutation parameter based on the cosine value of the angle; for example, if the cosine value of the angle of a certain terrain unit is close to 1 (i.e., the normal vector is almost the same as the flood diffusion direction), the spatial weight factor of this unit is higher; conversely, if the cosine value of the angle is close to -1, the spatial weight factor is lower; finally, these spatial weight factors are stored in a structured data format and mapped into the three-dimensional geographical grid model.

[0119] Step 304, couple and superimpose the time weight factor and the spatial weight factor according to the terrain unit to generate a comprehensive curvature weight, a comprehensive slope mutation weight, and a comprehensive water body diffusion direction weight;

[0120] In this step, the comprehensive curvature weight refers to the overall importance coefficient of the curvature parameter after combining the time weight factor and the spatial weight factor; the comprehensive slope mutation weight refers to the overall importance coefficient of the slope mutation parameter after combining the time weight factor and the spatial weight factor; the comprehensive water body diffusion direction weight refers to the overall importance coefficient of the water body diffusion direction parameter after combining the time weight factor and the spatial weight factor; these comprehensive weights reflect the degree of flood impact of different terrain units in the time and space dimensions, providing more accurate data support for subsequent flood simulation and inundation prediction;

[0121] In this embodiment, first, the time weight factor and the spatial weight factor of each terrain unit are coupled and superimposed. The time weight factor and the spatial weight factor are combined in proportion by the weighted average method to generate a comprehensive weight value. Then, using these comprehensive weight values, the comprehensive curvature weight, the comprehensive slope mutation weight, and the comprehensive water body diffusion direction weight are calculated respectively. For example, for each terrain unit, the weighted average of the time weight factor and the spatial weight factor of its curvature parameter is calculated to obtain the comprehensive curvature weight of the unit; similarly, the comprehensive slope mutation weight and the comprehensive water body diffusion direction weight are calculated respectively.

[0122] Step 305: Use the comprehensive curvature weight, the comprehensive slope mutation weight, and the comprehensive water body diffusion direction weight to perform weighted fusion on the curvature parameter, the slope mutation parameter, and the water body diffusion direction parameter, so as to obtain the characteristic component of each terrain unit in the dynamic boundary feature vector;

[0123] In this step, the characteristic component refers to the specific value obtained by performing weighted fusion on the curvature parameter, the slope mutation parameter, and the water body diffusion direction parameter of each terrain unit. These values reflect the importance and influence degree of each terrain feature in the flood diffusion process; through these parameters and technical means, the characteristic component of each terrain unit in the dynamic boundary feature vector can be generated, providing accurate data support for subsequent flood simulation and inundation prediction;

[0124] In this embodiment, first, the comprehensive curvature weight, the comprehensive slope mutation weight, and the comprehensive water body diffusion direction weight of each terrain unit are standardized to ensure that each weight value is in the same order of magnitude; then, the standardized comprehensive weight is fused with the corresponding terrain parameter using a weighted fusion algorithm (such as linear weighted summation); for each terrain unit, calculate the product of its curvature parameter and its comprehensive curvature weight, the product of its slope mutation parameter and its comprehensive slope mutation weight, and the product of its water body diffusion direction parameter and its comprehensive water body diffusion direction weight, and then add these three results to obtain the characteristic component of the unit.

[0125] Step 306: According to the comprehensive water body diffusion direction weight in the characteristic component and the elevation gradient change rate of adjacent terrain units in the continuous terrain unit, correct the continuity of the diffusion trend of the dynamic boundary feature vector, and output the dynamic boundary feature vector that matches the real-time evolution of the flood area;

[0126] In this step, the elevation gradient change rate refers to the change rate of the elevation difference between adjacent terrain units, reflecting the steepness of the terrain and the slope change; the continuity of the diffusion trend refers to whether the diffusion path of the flood between different terrain units smoothly transitions and whether its direction is consistent;

[0127] In this embodiment, first, the elevation gradient change rate between each continuous terrain unit is calculated using a three-dimensional geographic grid model; the elevation gradient change rate is obtained by dividing the elevation difference between adjacent units by the distance between them; then, the diffusion trend of each terrain unit is adjusted using the comprehensive weight of the water body diffusion direction in the generated feature components; the specific process includes: for each terrain unit, a correction factor is calculated based on its comprehensive weight of the water body diffusion direction and the elevation gradient change rate of the adjacent unit; if a certain unit has a relatively high comprehensive weight of the water body diffusion direction and a relatively small elevation gradient change rate, then the correction factor of this unit is larger, indicating that its diffusion trend is more continuous and smooth; otherwise, it is smaller; then, the diffusion trend in the dynamic boundary feature vector is corrected based on the correction factor to ensure its matching with the real-time evolution of the flood area; finally, the corrected dynamic boundary feature vector is output.

[0128] Since the existing emergency resource scheduling system cannot timely adjust the priority ranking of the material reserve locations when facing the dynamically changing flood area, the resource allocation efficiency is low. In some embodiments, according to step 104, the spatial overlay map is associated with the material reserve locations in the preset emergency resource distribution map, and the associated material reserve locations are dynamically prioritized according to the inundation probability gradient, and a resource scheduling instruction sequence linked to the change of the flood area is generated according to the ranking result, including:

[0129] Step 401, perform geographic coding on the material reserve locations in the emergency resource distribution map, extract the longitude and latitude coordinates and the material type identifier of each material reserve location, and generate a material reserve feature set;

[0130] In this step, geographic coding refers to the process of converting an address description into geographic coordinates (such as longitude and latitude); the longitude and latitude coordinates are the specific location identifiers of a point on the earth's surface, used to accurately determine the material reserve location; the material type identifier is the information for classifying and marking different types of emergency materials, such as food, medicine, rescue equipment, etc.; the material reserve feature set is a data set containing all the material reserve locations and their related information, including longitude and latitude coordinates, material type identifiers, etc.;

[0131] In this embodiment, first, geocoding is performed on the material reserve locations in the emergency resource distribution map; the text descriptions (such as addresses or place names) of each material reserve location are input, and they are converted into corresponding latitude and longitude coordinates by calling an online geocoding service; then, the latitude and longitude coordinates and material type identifiers of each material reserve location are extracted and stored in a structured data format (such as a Pandas DataFrame); then, data cleaning techniques (such as removing duplicates, handling missing values, etc.) are used to preprocess the generated dataset to ensure the accuracy and integrity of the data; finally, a material reserve feature set containing all material reserve locations and their related information is generated.

[0132] Step 402: Perform spatial raster matching on the inundation probability gradient parameter of each terrain unit in the spatial overlay map with the latitude and longitude coordinates of the material reserve feature set, calculate the real-time inundation threat level of the spatial raster unit where the material reserve location is located, and assign a dynamic material demand weight coefficient to each material reserve location according to the product relationship between the material type identifier and the real-time inundation threat level.

[0133] In this step, spatial raster matching refers to the process of matching the latitude and longitude coordinates of the material reserve location with the terrain units in the spatial overlay map; the inundation probability gradient parameter is the probability value of inundation for each terrain unit, reflecting the degree of influence of floods on this area; the real-time inundation threat level is the inundation risk assessment of the spatial raster unit where the material reserve location is located calculated based on the inundation probability gradient parameter; the dynamic material demand weight coefficient is the importance coefficient of each material reserve location calculated according to the material type identifier and the real-time inundation threat level, and is used to optimize resource scheduling.

[0134] In this embodiment, first, the spatial overlay map is processed to extract the flood probability gradient parameters of each terrain unit and convert them into a format suitable for spatial grid matching. Then, a Pandas DataFrame is used to store the material reserve feature set, which includes the longitude and latitude coordinates of each material reserve location and the material type identifier. Next, through a spatial grid matching algorithm (such as nearest neighbor interpolation or bilinear interpolation), the longitude and latitude coordinates of the material reserve locations are matched with the terrain units in the spatial overlay map to determine the grid cell where each material reserve location is located and its corresponding flood probability gradient parameter. For each material reserve location, find the grid cell where it is located and obtain the flood probability gradient parameter of this cell as its real-time flood threat level. Next, based on the product relationship between the material type identifier and the real-time flood threat level, calculate the dynamic material demand weight coefficient for each material reserve location. For example, for high-priority materials (such as life-saving equipment), if the flood threat level at its location is high, a higher dynamic material demand weight coefficient is assigned; vice versa. Finally, the calculated dynamic material demand weight coefficient is stored in the material reserve feature set.

[0135] Step 403: Based on the terrain slope attribute of the three-dimensional geographic grid model, calculate the cumulative value of the slopes of the shortest passage paths from the material reserve locations to the high flood probability gradient regions in the spatial overlay map, and generate a terrain passage resistance coefficient according to the cumulative value of the slopes of the shortest passage paths.

[0136] In this step, the cumulative value of the slopes of the shortest passage paths refers to the sum of the slope values of all the terrain units passed through on the shortest path from the material reserve location to the high flood probability gradient region, which reflects the terrain difficulty on the path. The terrain passage resistance coefficient is a value calculated based on the cumulative value of the slopes of the shortest passage paths and is used to quantify the ease of passage of the path.

[0137] In this embodiment, first, the shortest passage paths from each material reserve location to the high flood probability gradient regions in the spatial overlay map are calculated in combination with the three-dimensional geographic grid model. The three-dimensional geographic grid model is converted into an undirected graph, where the nodes represent terrain units, the edges represent the connections between adjacent units, and a weight value is assigned to each edge, which is calculated based on the terrain slope attribute. Then, the Dijkstra algorithm is used to find the shortest paths from the material reserve locations to the high flood probability gradient regions and calculate the cumulative value of the slopes of all the terrain units passed through on these paths. Then, a terrain passage resistance coefficient is generated based on the cumulative value of the slopes of the shortest passage paths. The specific process includes: for each shortest path, calculate the terrain passage resistance coefficient according to its cumulative value of the slopes. The larger the cumulative value of the slopes, the higher the terrain passage resistance coefficient, indicating that the path is more difficult to pass.

[0138] Step 404: Construct a dynamic priority scoring function for each material storage location based on the real-time flooding threat level, material demand weight coefficient, and terrain passage resistance coefficient.

[0139] In this step, the dynamic priority scoring function is a comprehensive evaluation model used to combine the real-time flooding threat level, material demand weight coefficient, and terrain passage resistance coefficient to assign a dynamic priority score to each material storage location. This scoring function can help decision-makers quickly determine which material storage locations need to be prioritized for processing or scheduling, thus more efficiently responding to flood disasters.

[0140] In this embodiment, first, define the logical structure of the dynamic priority scoring function. The scoring function comprehensively considers three key factors: the real-time flooding threat level of the material storage location, the material demand weight coefficient, and the terrain passage resistance coefficient. Among them, the higher the real-time flooding threat level, the greater the risk faced by this location; the higher the material demand weight coefficient, the more important the materials stored at this location; and the higher the terrain passage resistance coefficient, the greater the difficulty in accessing this location. Then read the previously calculated real-time flooding threat level, material demand weight coefficient, and terrain passage resistance coefficient data and integrate them into a unified data table. Then, based on the logical rules of the scoring function, calculate the dynamic priority score for each material storage location. The specific process includes: for each material storage location, assign different weights according to its real-time flooding threat level, material demand weight coefficient, and terrain passage resistance coefficient for comprehensive evaluation. For example, if a certain material storage location has a high real-time flooding threat level and a large material demand weight coefficient, but a low terrain passage resistance coefficient, then the dynamic priority score of this location will be relatively high, indicating that it needs to be prioritized for attention and scheduling.

[0141] Step 405: Obtain the update result of the dynamic data of the flood flow velocity at preset time intervals, trigger the recalculation of the flooding probability gradient parameter according to the update result, dynamically adjust the real-time flooding threat level based on the recalculated flooding probability gradient parameter, trigger the iterative calculation of the priority scoring function, and generate a resource scheduling instruction sequence sorted by urgency according to the iterative calculation result.

[0142] In this step, the dynamic priority scoring function is an evaluation model that is continuously updated according to the real-time flood situation, used to comprehensively consider the real-time flooding threat level, material demand weight coefficient, and terrain passage resistance coefficient of the material storage location, so as to assign a score reflecting its urgency to each material storage location.

[0143] In this embodiment, first, a preset time interval (e.g., every hour) is set, and a timing task scheduling library in Python (such as APScheduler) is used to regularly obtain the update result of the flood flow velocity dynamic data. Then, the updated flood flow velocity dynamic data is processed to extract key features to recalculate the inundation probability gradient parameters. Based on the new flood flow velocity data, the inundation probability of each terrain unit is re-evaluated, and the corresponding inundation probability gradient parameters are updated. Then, according to the recalculated inundation probability gradient parameters, the real-time inundation threat level of each material reserve location is dynamically adjusted. For example, if the inundation probability near a certain location increases significantly, its real-time inundation threat level is correspondingly increased. Next, based on the updated real-time inundation threat level and other relevant parameters (such as the material demand weight coefficient and the terrain passage resistance coefficient), the iterative calculation process of the priority scoring function is triggered. In this process, the dynamic priority scores of all material reserve locations are re-evaluated and sorted from high to low according to the scores. Finally, a resource scheduling instruction sequence sorted by urgency is generated.

[0144] Researchers have found that the development of flood disasters has a high degree of uncertainty and dynamics, and existing methods are difficult to timely adjust the inundation threat level and resource scheduling strategy according to the latest flood flow velocity dynamic data. Based on this, in some embodiments, according to step 405, the update result of the flood flow velocity dynamic data is obtained at a preset time interval, the recalculation of the inundation probability gradient parameters is triggered according to the update result, the real-time inundation threat level is dynamically adjusted based on the recalculated inundation probability gradient parameters, and the iterative calculation of the priority scoring function is triggered. According to the iterative calculation result, a resource scheduling instruction sequence sorted by urgency is generated, including:

[0145] Step 501, divide a preset time window according to the preset flood disaster emergency response level, obtain the latest flood flow velocity dynamic data collected by the UAV cluster at the end of each time window, and extract the time stamp and the extreme value of the flow velocity change rate in the latest flood flow velocity dynamic data.

[0146] In this step, the latest flood flow velocity dynamic data refers to the change information of the flood flow velocity collected by the UAV cluster within a specific time window; the extreme value of the flow velocity change rate refers to the maximum or minimum change rate of the flood flow velocity within a certain time period, which is used to reflect the change of the flood intensity and the diffusion trend; these parameters and technical means can help to monitor the development trend of the flood in real time and provide basic data support for subsequent analysis.

[0147] In this embodiment, first, at the end of each preset time window (for example, every hour), the drone swarm automatically performs data collection tasks to obtain the latest dynamic data of the flood flow velocity. The drone swarm flies along a predetermined path, uses sensor devices to record the flood flow velocity information in real time, and transmits this data back to the data center. Then, these data are processed to extract the timestamps and extreme values of the flow velocity change rate of the dynamic flood flow velocity data within each time window. The specific process includes: importing the data collected by the drones into a Pandas DataFrame, and then using the built-in functions of Pandas to calculate the change rate of the flood flow velocity within each time window, and finding the maximum and minimum change rate values as the extreme values of the flow velocity change rate. For example, if the flood flow velocity increases significantly during a certain period, the extreme value of the flow velocity change rate during that period will be relatively high.

[0148] Step 502: Compare and analyze the extreme value of the flow velocity change rate with the historical data of the same period. When the extreme value of the flow velocity change rate in the analysis result exceeds the first preset threshold, recalculate the dynamic boundary feature vector.

[0149] In this step, the historical data of the same period refers to the dynamic data of the flood flow velocity in the same period in the past. By comparing and analyzing the extreme values of the flow velocity change rate between the current and historical data of the same period, it is possible to evaluate whether the development trend of the current flood is abnormal, and accordingly decide whether to recalculate the dynamic boundary feature vector. When the extreme value of the flow velocity change rate exceeds the first preset threshold, it indicates that the current flood situation may be relatively serious, and it is necessary to re-evaluate the flood diffusion trend.

[0150] In this embodiment, first, compare and analyze the latest dynamic data of the flood flow velocity collected by the drone swarm with the historical data of the same period. Import the two sets of data into a Pandas DataFrame, and calculate the extreme values of the flow velocity change rate within each time window. Then, compare these extreme values of the flow velocity change rate to find out those cases that exceed the first preset threshold. For example, if the flow velocity change rate during a certain period is significantly higher than the historical data of the same period, the data for that period is regarded as abnormal. Once an abnormal situation is found, trigger the process of recalculating the dynamic boundary feature vector.

[0151] Step 503: Based on the recalculated dynamic boundary feature vector, update the flood probability gradient parameter of each grid cell to obtain the updated flood probability gradient parameter, and calculate the change amplitude of the updated flood probability gradient parameter.

[0152] In this step, the updated flood probability gradient parameter refers to the new value obtained by adjusting the flood probability gradient parameter of each grid cell based on the recomputed dynamic boundary feature vector; the change amplitude refers to the degree of difference between the old and new flood probability gradient parameters; by updating the flood probability gradient parameter and calculating its change amplitude, the trend and influence range of flood diffusion can be more accurately reflected;

[0153] In this embodiment, first, the flood probability gradient parameter of each grid cell in the three-dimensional geographic grid model is updated using the recomputed dynamic boundary feature vector; based on the new dynamic boundary feature vector, the flood probability of each grid cell is re-evaluated, and the updated flood probability gradient parameter is generated; then, the change amplitude of the flood probability gradient parameter before and after the update is calculated to quantify the change in the flood diffusion trend; for example, the team found that the flood probability in certain areas increased significantly, indicating that the impact of the flood on these areas is intensifying.

[0154] Step 504, according to the change amplitude of the updated flood probability gradient parameter, adjust the real-time flood threat level of the material reserve location, input the updated real-time flood threat level into the dynamic priority scoring function, and dynamically calibrate the weight parameters in the dynamic priority scoring function according to the latest flood flow velocity data;

[0155] In this step, the real-time flood threat level is the flood risk level faced by the material reserve location calculated based on the updated flood probability gradient parameter; the dynamic priority scoring function is a comprehensive evaluation model used to assign a dynamic priority score to each material reserve location by combining the real-time flood threat level, material demand weight coefficient, and terrain passage resistance coefficient; by adjusting the real-time flood threat level and inputting it into the dynamic priority scoring function, the priority of resource scheduling can be dynamically calibrated;

[0156] In this embodiment, first, according to the change amplitude of the updated flood probability gradient parameter, adjust the real-time flood threat level of each material reserve location; for each material reserve location, re-evaluate its real-time flood threat level according to the change situation of the flood probability gradient parameter in its area; then, input the updated real-time flood threat level into the dynamic priority scoring function, and dynamically calibrate the weight parameters in the dynamic priority scoring function according to the latest flood flow velocity data; for example, if the flood threat level of a certain location increases significantly, then correspondingly increase its weight in the dynamic priority scoring function; finally, generate a new dynamic priority scoring result.

[0157] Step 505, when the difference rate of the priority calibration results for consecutive preset times is less than the preset convergence threshold, generate a resource scheduling instruction sequence;

[0158] In this step, the difference rate of the priority calibration results refers to the degree of difference between consecutive priority calibration results; the resource scheduling instruction sequence is a resource scheduling plan generated by sorting according to the finally determined dynamic priority scores; when the difference rate of the priority calibration results for a continuous preset number of times is less than the preset convergence threshold, it indicates that the resource scheduling priority has tended to be stable, and at this time, a resource scheduling instruction sequence is generated;

[0159] In this embodiment, first, a preset convergence threshold is set (for example, the difference rate is less than 5%), and the results are recorded after each priority calibration; after each recalculation of the dynamic boundary feature vector and update of the dynamic priority scoring function, the current priority calibration result is recorded; then, the difference rate between consecutive priority calibration results is calculated; for example, if the difference rates of the priority calibration results for three consecutive times are all less than 5%, it is considered that the priority has tended to be stable; at this time, a resource scheduling instruction sequence is generated, sorted from high to low according to the dynamic priority scores, and a detailed resource scheduling plan is formulated.

[0160] Since the multi-dimensional superposition processing of the flood flow velocity dynamic data is performed in the cloud server, and the superimposed flood flow velocity dynamic data is associated and matched with the inundation paths in historical flood events to generate a superimposed association result, when the UAV cluster executes the flood disaster monitoring task, there is a lack of an effective closed-loop feedback mechanism, resulting in the data collection area and flight trajectory not being able to adapt to the changes in the development of the flood in a timely manner. Based on this, in some embodiments, according to what is described in step XXX, the resource scheduling instruction sequence is fed back to the control terminal of the UAV cluster to correct the UAV flight trajectories and data collection areas in the UAV cluster, and a closed-loop feedback mechanism is generated, including:

[0161] Step 601, parse the area identifiers greater than the preset priority value in the resource scheduling instruction sequence, extract the longitude and latitude boundaries of each grid cell within the area identifiers greater than the preset priority value and the data collection accuracy requirements, and generate a focused collection task set for the UAV cluster;

[0162] In this step, the area identifier greater than the preset priority value refers to the high-risk areas that need to be focused on marked in the resource scheduling instruction sequence; the longitude and latitude boundaries are the specific geographical ranges of these areas; the data collection accuracy requirement refers to the accuracy standard for data collection within these areas; by parsing this information, a focused collection task set for the UAV cluster can be generated to guide the UAVs to efficiently execute the data collection task;

[0163] In this embodiment, first, read the resource scheduling instruction sequence file and extract all region identifiers greater than the preset priority value, as well as their corresponding longitude and latitude boundaries and data acquisition accuracy requirements. For example, read and parse the geographical coordinate ranges and data acquisition accuracies of each high-risk region from a CSV file. Secondly, based on the extracted longitude and latitude boundaries, combined with the three-dimensional geographical grid model, calculate the coverage ranges of each grid cell to ensure that the geographical boundaries of each grid cell are accurately identified. Then, associate the longitude and latitude boundaries and data acquisition accuracy requirements of each high-risk region with the corresponding grid cells to form a task list containing geographical boundaries and accuracy requirements. Finally, generate a focused acquisition task set for the UAV cluster according to the task list and send it to the UAV control center to provide detailed guidance for the subsequent path planning and data acquisition of the UAVs.

[0164] Step 602, according to the focused acquisition task set, calculate the trajectory deviation angle between the current flight trajectory of the UAV cluster and the target area and the coverage blind area range, and generate an obstacle avoidance path correction parameter and a data acquisition frequency adjustment coefficient for each UAV based on the terrain slope attribute of the three-dimensional geographical grid model.

[0165] In this step, the trajectory deviation angle refers to the included angle between the current flight trajectory of the UAV and the target area; the coverage blind area range refers to the area not covered during the flight of the UAV; the obstacle avoidance path correction parameter is a parameter for adjusting the flight path of the UAV calculated according to the terrain slope attribute; the data acquisition frequency adjustment coefficient is a coefficient for adjusting the sampling interval of the UAV sensor determined according to the terrain complexity. By calculating these parameters, the flight path and data acquisition efficiency of the UAV can be optimized.

[0166] In this embodiment, first, use the relative position between the current position of the UAV and the target area to calculate the trajectory deviation angle through trigonometric functions. For example, if the UAV is currently located at 35.6900°N, 139.6920°E, and the center point of the target area is at 35.6950°N, 139.6940°E, then calculate the included angle between the two. Secondly, analyze the coverage blind area range on the flight route of the UAV based on the three-dimensional geographical grid model to find the uncovered areas. Then, according to the terrain slope attribute, use GIS tools to calculate the obstacle avoidance path correction parameter to ensure that the UAV can safely avoid obstacles. Then, determine the data acquisition frequency adjustment coefficient according to the terrain complexity, and increase the sampling frequency for complex terrain areas. Finally, send the obstacle avoidance path correction parameter and the data acquisition frequency adjustment coefficient to the UAV flight control module to dynamically adjust the flight trajectory and sensor sampling interval of the UAV to ensure efficient data acquisition.

[0167] Step 603: Inject the obstacle avoidance path correction parameters into the UAV flight control module, dynamically adjust the pitch angle and heading angle of the UAV, so that the flight trajectory of the UAV converges along the terrain contour towards the area greater than the preset priority value, and at the same time, synchronously modify the sampling interval of the UAV's on-board sensors according to the data acquisition frequency adjustment coefficient;

[0168] In this step, the obstacle avoidance path correction parameters are used to adjust the flight path of the UAV to avoid obstacles and fly along the contour line; the pitch angle and heading angle are key parameters of the UAV flight attitude; the data acquisition frequency adjustment coefficient is used to modify the sampling interval of the UAV's on-board sensors; by injecting these parameters, dynamic adjustment of the UAV flight path and synchronous modification of the data acquisition frequency can be achieved;

[0169] In this embodiment, first, inject the obstacle avoidance path correction parameters into the UAV flight control module, and use the PID controller to adjust the pitch angle and heading angle of the UAV in real time, so that the flight trajectory of the UAV converges along the terrain contour towards the area greater than the preset priority value; for example, when the UAV approaches the mountainous area, the system will automatically adjust its flight path to avoid the peaks; second, according to the data acquisition frequency adjustment coefficient, synchronously modify the sampling interval of the on-board sensors through the UAV's flight control system; if the terrain of a certain flight path is relatively complex, the flight speed will be appropriately reduced and the sampling frequency will be increased to ensure data quality; finally, the UAV continues to execute the task according to the new flight path and sampling frequency to ensure the effectiveness and accuracy of data acquisition.

[0170] Step 604: After the UAV arrives at the target area, collect and send the latest dynamic flood flow velocity data and the latest terrain elevation data to the cloud server in real time, extract the spatio-temporal consistency check code from the latest dynamic flood flow velocity data and the latest terrain elevation data, and verify the geographical coverage overlap between the data collection area and the geographical coverage required in the focused collection task set;

[0171] In this step, the spatio-temporal consistency check code is a code used to verify the time and space consistency of data collection; the geographical coverage overlap refers to the matching degree between the actual data collection area of the UAV and the geographical coverage range required in the focused collection task set; by collecting and sending the latest dynamic flood flow velocity data and terrain elevation data in real time, and performing spatio-temporal consistency check and geographical coverage overlap verification, the accuracy and integrity of the data can be ensured;

[0172] In this embodiment, first, when the drone arrives at the target area, it activates the on-board sensors to collect the latest dynamic data of flood flow velocity and terrain elevation data in real time, and sends these data to the cloud server through the wireless communication module; second, extract the spatio-temporal consistency check codes from the received data to ensure the time and space consistency of the data; for example, check whether the data collection timestamps are continuous and without repetition, and whether the geographical locations are consistent with the assigned areas; then, based on the drone flight records and the geographical boundaries assigned for the task, calculate the geographical coverage overlap degree to evaluate the matching situation between the actual collected area and the task requirements; finally, feedback the verification result to the drone control center, and if data loss or inconsistency is found, trigger re-collection or adjustment of the flight path.

[0173] Step 605, when the geographical coverage overlap degree is lower than the second preset threshold, recalculate the obstacle avoidance path correction parameters according to the elevation gradient change rate of the missing area, and trigger the secondary trajectory correction of the drone swarm; if the geographical coverage overlap degree meets the second preset threshold, mark the latest dynamic data of flood flow velocity and the latest terrain elevation data as data passed in the closed-loop verification, and update the flood probability gradient parameters in the spatial overlay map.

[0174] In this embodiment, first, check whether the geographical coverage overlap degree is lower than the second preset threshold (such as 85%). If it is lower than the threshold, recalculate the obstacle avoidance path correction parameters according to the elevation gradient change rate of the missing area. For example, use differential GPS data and topographic maps to calculate the elevation gradient change rate of the missing area, and adjust the drone flight path accordingly. Second, trigger the secondary trajectory correction of the drone swarm, re-plan the flight route to ensure coverage of the missing area. Then, the drone continues to execute the task according to the new flight path until the geographical coverage overlap degree meets the standard. Finally, mark the latest dynamic data of flood flow velocity and terrain elevation data as data passed in the closed-loop verification, and update the flood probability gradient parameters in the spatial overlay map to ensure the consistency and accuracy of the data.

[0175] Step 606, generate a new resource scheduling instruction sequence according to the updated flood probability gradient parameters, replace the executed or invalid instruction items in the original instruction sequence, and output the iterated closed-loop control instruction set.

[0176] In this embodiment, first, based on the latest dynamic data of flood flow velocity and terrain elevation data, re-evaluate the priority of each material reserve location and generate new resource scheduling instructions. For example, if the flood threat in certain areas increases significantly, adjust the resource scheduling priority of these areas accordingly. Second, output the newly generated sequence of resource scheduling instructions as a closed-loop control instruction set and send it to the water conservancy management department for reference and execution. Finally, the system will automatically repeat the above process to ensure that the resource scheduling is always up-to-date and accurate. For example, during this response process, due to the generation of an accurate closed-loop control instruction set, the water conservancy management department was able to quickly respond to flood changes, effectively allocate emergency materials, and protect the lives and property of local residents.

[0177] In summary, through the above steps, the refined management and optimization of the data collection task of the UAV cluster are achieved, ensuring the accuracy and integrity of data collection. By dynamically adjusting the UAV flight path, data collection frequency, and resource scheduling instructions, not only the scientificity and accuracy of emergency response are improved, but also the rapid response ability to flood changes is enhanced. This method provides strong support for flood control and disaster relief work, greatly improving the scientificity and efficiency of water conservancy project information management.

[0178] Since the existing methods for analyzing dynamic flood flow velocity data are difficult to comprehensively capture their spatio-temporal characteristics, the correlation between the inundation paths in historical flood events and the current flood situation is not close enough. Based on this, in some embodiments, according to step 102, perform multi-dimensional superposition processing on the dynamic flood flow velocity data in the cloud server, and correlate and match the superimposed dynamic flood flow velocity data with the inundation paths in historical flood events to generate a superimposed correlation result, including:

[0179] Step 701, decompose the spatio-temporal characteristics of the dynamic flood flow velocity data, extract the velocity fluctuation extreme value, direction deflection angle, and spatial diffusion continuity index of each monitoring point according to a preset time window, and generate a spatio-temporal characteristic matrix.

[0180] In this step, spatio-temporal characteristic decomposition refers to the analysis of dynamic flood flow velocity data in the time and space dimensions. The velocity fluctuation extreme value refers to the maximum or minimum velocity change value of each monitoring point within a preset time window. The direction deflection angle refers to the change angle of the flood flow direction. The spatial diffusion continuity index measures the continuity and consistency of flood diffusion between different regions. Through these parameters, a matrix containing various spatio-temporal characteristics can be generated, providing a basis for subsequent analysis.

[0181] In this embodiment, first, dynamic data of flood flow velocity is read, and the data is segmented according to a preset time window (such as every hour); for example, the data obtained from the sensor is divided into one time slice per hour. Secondly, the extreme value of velocity fluctuation, the angle of direction deflection, and the spatial diffusion continuity index at each monitoring point within each time window are calculated; for the data within each time window, the maximum and minimum change values of the flow velocity are calculated as the extreme value of velocity fluctuation, the angle change of the water flow direction is calculated as the angle of direction deflection, and the continuity of flood diffusion is evaluated. Then, these extracted features are integrated into a spatio-temporal feature matrix, where each row represents a monitoring point and each column represents a feature (such as the extreme value of velocity fluctuation, the angle of direction deflection, etc.).

[0182] Step 702: According to the angle of direction deflection in the spatio-temporal feature matrix, assign dynamic matching weights to each inundation path in historical flood events.

[0183] In this embodiment, first, the angle of direction deflection of the current flood is compared with the angle distribution of historical events; for example, if the angles of direction deflection of two paths are very close, a higher dynamic matching weight is assigned. Secondly, the calculated dynamic matching weights are stored in the database and associated with each inundation path. Then, SQL queries are used to bind these weights to the data of historical inundation paths to form a list of historical inundation paths containing dynamic matching weights.

[0184] Step 703: Based on the proportional relationship between the extreme value of velocity fluctuation and the peak velocity of the inundation path recorded in historical flood events, calculate the acceleration factor of the current flood evolution rate relative to historical flood events, and generate a path association intensity coefficient in combination with the dynamic matching weight.

[0185] In this embodiment, first, based on the proportional relationship between the extreme value of velocity fluctuation and the peak velocity of the inundation path recorded in historical flood events, calculate the acceleration factor of the current flood evolution rate relative to historical flood events; for example, if the extreme value of velocity fluctuation of the current flood is significantly higher than the historical record, a higher acceleration factor is calculated. Secondly, for each historical inundation path, calculate the proportional relationship between its extreme value of velocity fluctuation and the extreme value of velocity fluctuation of the current flood, and multiply it by the corresponding dynamic matching weight to obtain the path association intensity coefficient.

[0186] Step 704: Perform raster comparison between the spatial diffusion continuity index and the spatial coverage density of each inundation path in historical flood events, screen out multiple historical inundation paths with a spatial overlap rate higher than a preset critical value, and use the multiple historical inundation paths as a candidate associated path set.

[0187] In this embodiment, first, the three-dimensional geographic grid model is converted into a raster format, and the flood coverage density of each raster cell is calculated; secondly, based on a spatial overlap rate threshold (such as 80%), historical inundation paths with a spatial overlap rate higher than the preset critical value are screened out; for example, if the spatial coverage density of a certain historical path highly coincides with the spatial diffusion continuity index of the current flood on multiple raster cells, it is included in the candidate associated path set.

[0188] Step 705, perform coupling correction on the path association strength coefficient of each historical inundation path in the candidate associated path set and the terrain slope attribute of the three-dimensional geographic grid model to obtain the corrected path association strength coefficient;

[0189] In this embodiment, first, calculate the terrain slope passed by each path, and adjust the path association strength coefficient according to the steepness of the slope; secondly, combine the terrain slope adjustment coefficient with the original path association strength coefficient to generate the corrected path association strength coefficient; for example, if a certain path passes through relatively flat terrain, the corrected path association strength coefficient is higher; otherwise, it is lower.

[0190] Step 706, aggregate the corrected path association strength coefficients according to spatial raster cells to generate a superimposed association result covering the flood area, where the superimposed association result includes the spatial matching degree and evolution trend confidence of each raster cell with the historical path;

[0191] In this embodiment, first, aggregate the corrected path association strength coefficients according to spatial raster cells to generate a superimposed association result covering the flood area, then accumulate the path association strength coefficients of each raster cell to form a two-dimensional array representing the entire flood area; then, based on the superimposed association result, calculate the spatial matching degree and evolution trend confidence of each raster cell with the historical path; for example, if the path association strength coefficient of a certain raster cell is higher, its spatial matching degree and evolution trend confidence are also higher; then, export the superimposed association result as a visualization map.

[0192] Figure 2 The following is a schematic structural diagram of a water conservancy project information management system based on big data provided by an embodiment of the present application, as Figure 2 shown, the system includes:

[0193] A receiving module 21, configured to receive the dynamic flood flow velocity data and terrain elevation data collected by the UAV cluster and the spectral feature data of the flood area transmitted by the satellite remote sensing device, and upload the dynamic flood flow velocity data, terrain elevation data, and spectral feature data to the cloud server;

[0194] A matching module 22, configured to perform multi-dimensional superposition processing on the dynamic flood flow velocity data in the cloud server, associate and match the superimposed dynamic flood flow velocity data with the inundation paths in historical flood events, and generate a superimposed association result;

[0195] An identification module 23, configured to construct a three-dimensional geographical model of the flood area based on the terrain elevation data and spectral feature data, identify the boundary features of the water body and terrain mutation areas in the three-dimensional geographical model, and generate a spatial superposition map including the inundation probability gradient in combination with the flood evolution path in the superimposed association result;

[0196] A sorting module 24, configured to perform position association between the spatial superposition map and the material storage positions in a preset emergency resource distribution map, perform dynamic priority sorting on the associated material storage positions according to the inundation probability gradient, and generate a resource scheduling instruction sequence linked to the changes in the flood area according to the sorting result;

[0197] A correction module 25, configured to feedback the resource scheduling instruction sequence to the control terminal of the UAV cluster, correct the flight trajectories and data acquisition areas of the UAVs in the UAV cluster, and generate a closed-loop feedback mechanism.

[0198] Figure 2 The described water conservancy project information management system based on big data can execute Figure 1 The described water conservancy project information management method based on big data in the illustrated embodiment, and its implementation principle and technical effects will not be elaborated further. For the water conservancy project information management system based on big data in the above embodiment, the specific manners in which each module and unit perform operations have been described in detail in the embodiment of the method, and will not be elaborated here.

[0199] In a possible design, Figure 2 The water conservancy project information management system based on big data in the illustrated embodiment can be implemented as a computing device, such as Figure 3 shown, and this computing device can include a storage component 31 and a processing component 32;

[0200] The storage component 31 stores one or more computer instructions, and among them, the one or more computer instructions are called and executed by the processing component 32.

[0201] The processing component 32 is used for the Figure 1 described water conservancy project information management method based on big data in the above embodiment.

[0202] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A water conservancy project information management method based on big data, characterized in that, Including: Receiving the dynamic data of flood flow velocity and terrain elevation data collected by the UAV swarm, as well as the spectral feature data of the flood area transmitted by the satellite remote sensing equipment, and uploading the dynamic data of flood flow velocity, terrain elevation data, and spectral feature data to the cloud server; Performing multi-dimensional superposition processing on the dynamic data of flood flow velocity in the cloud server, associating and matching the superimposed dynamic data of flood flow velocity with the inundation paths in historical flood events to generate a superimposed association result; Constructing a three-dimensional geographic model of the flood area based on the terrain elevation data and spectral feature data, identifying the boundary features of the water body and terrain mutation areas in the three-dimensional geographic model, and combining the flood evolution path in the superimposed association result to generate a spatial superimposed map including the inundation probability gradient; Associating the position of the spatial superimposed map with the material reserve positions in the preset emergency resource distribution map, dynamically prioritizing the associated material reserve positions according to the inundation probability gradient, and generating a resource scheduling instruction sequence linked to the changes in the flood area according to the sorting result; Feeding back the resource scheduling instruction sequence to the control terminal of the UAV swarm, correcting the flight trajectories of the UAVs in the UAV swarm and the data collection area, and generating a closed-loop feedback mechanism.

2. The method according to claim 1, wherein Constructing a three-dimensional geographic model of the flood area based on the terrain elevation data and spectral feature data, identifying the boundary features of the water body and terrain mutation areas in the three-dimensional geographic model, and combining the flood evolution path in the superimposed association result to generate a spatial superimposed map including the inundation probability gradient, including: Performing multi-scale wavelet decomposition on the spectral feature data of the flood area, extracting the image texture features of different frequency bands, and generating a denoised high-resolution flood area image according to the spectral differences between the water body reflectivity and the surface cover in the image texture features; Dividing the terrain elevation data into continuous terrain units according to the elevation mutation threshold, and constructing a three-dimensional geographic grid model with terrain slope attributes based on the elevation gradient change rate of the continuous terrain units and the water body coverage range in the denoised high-resolution flood area image; In the three-dimensional geographic grid model, identifying the boundary features between the water body coverage range and the adjacent terrain units in the continuous terrain units through a convolutional neural network, and extracting the curvature parameter, slope mutation parameter, and water body diffusion direction parameter in the boundary features; Performing spatio-temporal weight allocation on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the dynamic data of flood flow velocity to generate a dynamic boundary feature vector representing the flood diffusion trend; Performing spatio-temporal alignment on the dynamic boundary feature vector and the historical inundation path in the superimposed association result, and calculating the inundation probability weight of each terrain unit in the dynamic boundary feature vector according to the proportional relationship between the flood diffusion rate in the historical inundation path and the current dynamic data of flood flow velocity; Based on the inundation probability weights and the terrain slope attributes in the three-dimensional geographic grid model, generate inundation probability gradient parameters covering the flood area, and map the inundation probability gradient parameters to the high-resolution flood area image to generate a spatial overlay map containing the inundation probability gradient.

3. The method according to claim 2, characterized in that, Perform spatio-temporal weight allocation on the curvature parameter, slope mutation parameter, and water body diffusion direction parameter according to the dynamic flood flow velocity data to generate a dynamic boundary feature vector representing the flood diffusion trend, including: Perform time series decomposition on the dynamic flood flow velocity data, extract the extreme value of the change rate of the flood flow velocity within a preset time window and the deflection angle of the diffusion direction, and generate a velocity dynamic feature set; According to the extreme value of the change rate in the velocity dynamic feature set, allocate time weight factors to the curvature parameter, slope mutation parameter, and water body diffusion direction parameter; Based on the spatial position relationship of continuous terrain units in the three-dimensional geographic grid model, calculate the cosine value of the angle between each terrain unit and the current flood diffusion direction, and allocate spatial weight factors to the curvature parameter and slope mutation parameter according to the cosine value of the angle; Couple and overlay the time weight factor and the spatial weight factor according to the terrain unit to generate a comprehensive curvature weight, a comprehensive slope mutation weight, and a comprehensive water body diffusion direction weight; Use the comprehensive curvature weight, the comprehensive slope mutation weight, and the comprehensive water body diffusion direction weight to perform weighted fusion on the curvature parameter, the slope mutation parameter, and the water body diffusion direction parameter to obtain the feature component of each terrain unit in the dynamic boundary feature vector; According to the comprehensive weight of the water body diffusion direction in the feature component and the elevation gradient change rate of adjacent terrain units in the continuous terrain unit, correct the continuity of the diffusion trend of the dynamic boundary feature vector, and output a dynamic boundary feature vector matching the real-time evolution of the flood area.

4. The method according to claim 1, wherein Associate the position of the spatial overlay map with the material reserve positions in the preset emergency resource distribution map, perform dynamic priority sorting on the associated material reserve positions according to the inundation probability gradient, and generate a resource scheduling instruction sequence linked to the change of the flood area according to the sorting result, including: Perform geographic coding on the material reserve positions in the emergency resource distribution map, extract the longitude and latitude coordinates and material type identifiers of each material reserve position, and generate a material reserve feature set; Perform spatial grid matching on the inundation probability gradient parameters of each terrain unit in the spatial overlay map and the longitude and latitude coordinates of the material reserve feature set, calculate the real-time inundation threat level of the spatial grid unit where the material reserve position is located, and allocate a dynamic material demand weight coefficient to each material reserve position according to the product relationship between the material type identifier and the real-time inundation threat level; Based on the terrain slope attribute of the three-dimensional geographic grid model, calculate the cumulative value of the shortest passage path slope from the material reserve position to the high inundation probability gradient area in the spatial overlay map, and generate a terrain passage resistance coefficient; Construct a dynamic priority scoring function for each material storage location according to the real-time inundation threat level, material demand weight coefficient, and terrain passage resistance coefficient; Obtain the update result of the dynamic flood velocity data at preset time intervals, trigger the recalculation of the inundation probability gradient parameter according to the update result, dynamically adjust the real-time inundation threat level based on the recalculated inundation probability gradient parameter, trigger the iterative calculation of the priority scoring function, and generate a resource scheduling instruction sequence sorted by urgency according to the iterative calculation result.

5. The method according to claim 4, characterized in that, Obtain the update result of the dynamic flood velocity data at preset time intervals, trigger the recalculation of the inundation probability gradient parameter according to the update result, dynamically adjust the real-time inundation threat level based on the recalculated inundation probability gradient parameter, trigger the iterative calculation of the priority scoring function, and generate a resource scheduling instruction sequence sorted by urgency, including: Divide a preset time window according to the preset flood disaster emergency response level, obtain the latest dynamic flood velocity data collected by the UAV cluster at the end of each time window, and extract the timestamp and the extreme value of the flow velocity change rate from the latest dynamic flood velocity data; Compare and analyze the extreme value of the flow velocity change rate with the historical data of the same period. When the extreme value of the flow velocity change rate in the analysis result exceeds the first preset threshold, recalculate the dynamic boundary feature vector; Based on the recalculated dynamic boundary feature vector, update the inundation probability gradient parameter of each grid cell to obtain the updated inundation probability gradient parameter, and calculate the change amplitude of the updated inundation probability gradient parameter; Adjust the real-time inundation threat level of the material storage location according to the change amplitude of the updated inundation probability gradient parameter, input the updated real-time inundation threat level into the dynamic priority scoring function, and dynamically calibrate the weight parameters in the dynamic priority scoring function according to the latest flood velocity data; When the difference rate of the priority calibration results for consecutive preset times is less than the preset convergence threshold, generate a resource scheduling instruction sequence.

6. The method according to claim 1, wherein Feed back the resource scheduling instruction sequence to the control terminal of the UAV cluster to correct the flight trajectories and data collection areas of the UAVs in the UAV cluster, and generate a closed-loop feedback mechanism, including: Parse the area identifier greater than the preset priority value in the resource scheduling instruction sequence, extract the longitude and latitude boundaries and data collection accuracy requirements of each grid cell within the area identifier greater than the preset priority value, and generate a focused collection task set for the UAV cluster; According to the focused collection task set, calculate the trajectory deviation angle and coverage blind area range between the current flight trajectory of the UAV cluster and the target area, and generate an obstacle avoidance path correction parameter and a data collection frequency adjustment coefficient for each UAV based on the terrain slope attribute of the three-dimensional geographic grid model; Inject the obstacle avoidance path correction parameter into the UAV flight control module to dynamically adjust the pitch angle and heading angle of the UAV, so that the flight trajectory of the UAV converges along the terrain contour line towards the area greater than the preset priority value, and at the same time, synchronously modify the sampling interval of the UAV on-board sensor according to the data collection frequency adjustment coefficient; After the UAV arrives at the target area, it collects and transmits the latest dynamic data of flood flow velocity and the latest terrain elevation data to the cloud server in real time, extracts the spatio-temporal consistency check codes in the latest dynamic data of flood flow velocity and the latest terrain elevation data, and verifies the coincidence degree of the data collection area with the geographical coverage required in the focused collection task; When the geographical coverage coincidence degree is lower than the second preset threshold, recalculate the obstacle avoidance path correction parameters according to the elevation gradient change rate of the missing area, and trigger the secondary trajectory correction of the UAV cluster; if the geographical coverage coincidence degree meets the second preset threshold, mark the latest dynamic data of flood flow velocity and the latest terrain elevation data as data passed the closed-loop verification, and update the inundation probability gradient parameters in the spatial overlay map; Generate a new resource scheduling instruction sequence according to the updated inundation probability gradient parameters, replace the executed or invalid instruction items in the original instruction sequence, and output the iterated closed-loop control instruction set.

7. The method according to claim 1, wherein Perform multi-dimensional overlay processing on the flood flow velocity dynamic data in the cloud server, associate and match the overlaid flood flow velocity dynamic data with the inundation paths in historical flood events, and generate an overlay association result, including: Decompose the spatio-temporal characteristics of the flood flow velocity dynamic data, extract the flow velocity fluctuation extreme value, direction deflection angle and spatial diffusion continuity index of each monitoring point according to the preset time window, and generate a spatio-temporal characteristic matrix; According to the direction deflection angle in the spatio-temporal characteristic matrix, assign dynamic matching weights to each inundation path in historical flood events; Based on the ratio relationship between the flow velocity fluctuation extreme value and the peak flow velocity of the inundation path recorded in historical flood events, calculate the acceleration factor of the current flood evolution rate relative to historical flood events, and generate a path association intensity coefficient in combination with the dynamic matching weight; Perform raster comparison between the spatial diffusion continuity index and the spatial coverage density of each inundation path in historical flood events, screen out multiple historical inundation paths with a spatial overlap rate higher than the preset critical value, and use the multiple historical inundation paths as the candidate association path set; Perform coupling correction on the path association intensity coefficient of each historical inundation path in the candidate association path set with the terrain slope attribute of the three-dimensional geographical grid model to obtain the corrected path association intensity coefficient; Aggregate the corrected path association intensity coefficients according to the spatial grid unit to generate an overlay association result covering the flood area, where the overlay association result includes the spatial matching degree and evolution trend confidence degree of each grid unit with the historical path.

8. A water conservancy project information management system based on big data, characterized in that, Including: A receiving module, configured to receive the dynamic data of flood flow velocity and terrain elevation data collected by the UAV cluster and the spectral characteristic data of the flood area transmitted by the satellite remote sensing device, and upload the dynamic data of flood flow velocity, terrain elevation data and spectral characteristic data to the cloud server; A matching module, configured to perform multi-dimensional overlay processing on the flood flow velocity dynamic data in the cloud server, associate and match the overlaid flood flow velocity dynamic data with the inundation paths in historical flood events, and generate an overlay association result; An identification module, configured to construct a three-dimensional geographical model of the flood area based on the terrain elevation data and spectral feature data, identify the boundary features of the water body and terrain mutation area in the three-dimensional geographical model, and generate a spatial overlay map containing the flood probability gradient in combination with the flood evolution path in the overlay association result; A sorting module, configured to perform position association between the spatial overlay map and the material reserve positions in a preset emergency resource distribution map, dynamically prioritize the associated material reserve positions according to the flood probability gradient, and generate a resource scheduling instruction sequence linked to the changes in the flood area according to the sorting result; A correction module, configured to feedback the resource scheduling instruction sequence to the control terminal of the UAV cluster, correct the flight trajectories of the UAVs and the data acquisition areas in the UAV cluster, and generate a closed-loop feedback mechanism.

9. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a big data-based water conservancy project information management method according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that, There is a computer program stored, and when the computer program is executed by a computer, it implements a big data-based water conservancy project information management method according to any one of claims 1 to 7.

Images