Fodder storage and detention area resident evacuation path dynamic planning method based on crowd sensing

Abstract

The invention relates to a flood storage and detention area resident evacuation path dynamic planning method based on crowd sensing, and belongs to the technical field of evacuation path planning. Multi-source heterogeneous data in a flood storage and detention area is collected, data cleaning is conducted on the multi-source heterogeneous data in the flood storage and detention area, and the evacuation path of residents in the flood storage and detention area is obtained; and performing data fusion on the multi-source heterogeneous data after data cleaning, updating the fused data, predicting flood condition data of a preset coordinate point, constructing a dynamic road network based on the fused data and the flood condition data of the preset coordinate point, and sending the dynamic road network to a preset terminal according to a preset mode. And finally, evaluating real-time state information of the dynamic road network, and dynamically updating the dynamic road network based on an evaluation result. According to the method, the multi-source heterogeneous data and the ant colony algorithm are fused, the multi-source heterogeneous data serve as the constraint condition of the ant colony algorithm, the evacuation path in the flood storage and detention area can be optimized, and the evacuation path is more reasonable.

Description

Technical Field

[0001] The present invention relates to the technical field of evacuation route planning, and particularly to a dynamic planning method for the evacuation routes of residents in flood detention areas based on crowd-sourced perception. Background Art

[0002] In the face of natural disasters such as floods, the rapid and safe evacuation of residents in flood detention areas is of crucial importance. Traditional evacuation route planning methods often have problems such as information lag and fixed routes, making it difficult to meet the needs of real-time changing disaster scenarios. Moreover, starting from the individual, the carrying capacity of resettlement points is ignored, and the group optimum cannot be achieved. With the development of technologies such as mobile Internet and intelligent transportation, by integrating multi-source information such as mobile phone GPS signals and traffic camera data, constructing a real-time flood evolution and forecast and early warning model, and combining advanced ant colony algorithms and AR navigation technologies, the dynamic planning of evacuation routes for residents in flood detention areas can be realized. By generating a real-time three-dimensional map and overlaying virtual elements such as direction indicators and real-time remaining capacity of resettlement points on the visualization interface, the efficient combination of centralized and decentralized transfers has great practical significance for residents in flood detention areas. Summary of the Invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a dynamic planning method for the evacuation routes of residents in flood detention areas based on crowd-sourced perception.

[0004] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] The first aspect of the present invention provides a dynamic planning method for the evacuation routes of residents in flood detention areas based on crowd-sourced perception, including the following steps:

[0006] Collect multi-source heterogeneous data in the flood detention area and perform data cleaning on the multi-source heterogeneous data in the flood detention area;

[0007] Perform data fusion on the multi-source heterogeneous data after data cleaning, update the fused data, and predict the flood situation data at preset coordinate points;

[0008] Construct a dynamic road network based on the fused data and the flood situation data at preset coordinate points, and send the dynamic road network to a preset terminal in a preset manner;

[0009] Evaluate the real-time status information of the dynamic road network and perform dynamic updates on the dynamic road network based on the evaluation results.

[0010] In the dynamic planning method for the evacuation routes of residents in flood detention areas based on crowd-sourced perception, collecting multi-source heterogeneous data in the flood detention area and performing data cleaning on the multi-source heterogeneous data in the flood detention area is specifically as follows:

[0011] Obtain the GPS signal data of residents' mobile phones or in-vehicle devices in the flood detention area, and install traffic cameras at preset roads, intersections, and preset key nodes in the flood detention area to capture traffic data information in real time through the traffic cameras;

[0012] Construct multi-source heterogeneous data in the flood detention area based on the traffic data information and GPS signal data, and by setting a data threshold range, propose data points that deviate from the data threshold range based on the data threshold range;

[0013] Use AI image recognition and video analysis technologies to remove misidentified data caused by video jitter, light changes, and occlusion factors, and unify the data display formats collected by traffic cameras of different brands and models.

[0014] In the dynamic planning method for the evacuation routes of residents in the flood detention area based on crowd-sensing, perform data fusion on the multi-source heterogeneous data after data cleaning and update the fused data. Specifically:

[0015] Introduce the pd.merge_asof function, use the pd.merge_asof function to match the multi-source heterogeneous data after data cleaning according to the timestamp, set the tolerance range, and perform approximate matching according to the tolerance range to align the multi-source heterogeneous data after cleaning;

[0016] After data alignment, use a data fusion algorithm to fuse the multi-source heterogeneous data after data alignment, determine the reliability weights of each type of multi-source heterogeneous data and perform scoring processing to obtain reliability weight values;

[0017] Calculate the total scores of each data source based on the reliability weight values, align the two data sources according to the timestamp, and define column mapping. Finally, fuse the data. Through data fusion, obtain a unified data set containing timestamps, longitudes, latitudes, and traffic data information.

[0018] In the dynamic planning method for the evacuation routes of residents in the flood detention area based on crowd-sensing, predict the flood situation data at preset coordinate points. Specifically:

[0019] Collect historical data collected by rain gauges, satellite remote sensing, and hydrological stations. Based on the historical data collected by the rain gauges, satellite remote sensing, and hydrological stations, construct all-element monitoring information, and build a flood feature data prediction model based on a deep neural network;

[0020] Use the all-element monitoring information as the model input of the flood feature data prediction model, perform model parameter setting and verification, pre-play the flood scenario, and dynamically correct the model parameters in combination with the all-element monitoring information and the flood scenario;

[0021] Obtain the all-element monitoring information of the preset coordinate points within the preset time, and input the all-element monitoring information into the flood characteristic data prediction model for prediction to obtain the flood situation data of the preset coordinate points.

[0022] In the dynamic planning method for the evacuation routes of residents in flood detention areas based on crowdsensing, construct a dynamic road network based on the fused data and the flood situation data of the preset coordinate points, specifically:

[0023] Introduce the ant colony algorithm, determine the nodes and edges of the flood detention area based on the fused data and the flood situation data of the preset coordinate points, and initialize the number of ants, pheromone evaporation coefficient, pheromone importance factor, and heuristic factor of the ant colony algorithm;

[0024] Obtain the number of people to be transferred, use the number of people to be transferred as the number of ants in the ant colony algorithm, use the real-time water depth, path length, planning time, and road capacity as heuristic factors, start from the starting node based on the nodes and edges of the flood detention area, and select the next path according to the pheromone concentration and heuristic information;

[0025] Gradually construct the evacuation route from the starting node to the target node to generate an initial path, set constraint conditions based on the flood situation data of the preset coordinate points, and determine whether the transfer path meets the constraint conditions;

[0026] When the transfer path meets the constraint conditions, the current transfer path iteration ends. At the end of each iteration, adjust the pheromone concentration according to the path length of the ants and the flood risk. When the transfer path does not meet the constraint conditions, the iteration continues, and the pheromone concentration on the path is updated in real time during the movement of the ants to reduce pheromone evaporation;

[0027] Repeat the process of path construction and pheromone update. After a preset number of iterations, obtain the optimal evacuation path, and construct a dynamic road network according to the optimal evacuation path.

[0028] In the dynamic planning method for the evacuation routes of residents in flood detention areas based on crowdsensing, send the dynamic road network to the preset terminal in a preset manner, specifically including:

[0029] Obtain the flood situation data of the preset coordinate points, and obtain the boundary of the risk area according to the flood situation data of the preset coordinate points. Take the boundary of the risk area as the benchmark, set an early warning buffer zone, and construct a polygonal electronic fence;

[0030] Render the warning circle layer through the mobile map SDK based on the polygonal electronic fence, obtain the geographical location information of the user, and determine whether the geographical location information of the user is within the warning circle;

[0031] When the geographical location information of the user is within the warning circle, send the dynamic road network to a preset terminal in a preset manner, generate a prompt message, and determine whether there is a response based on the prompt message within a preset time;

[0032] When there is no response within the preset time, include the continuously unresponsive users in the rescue priority calculation model.

[0033] In the method for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing, evaluate the real-time status information of the dynamic road network, and dynamically update the dynamic road network based on the evaluation results. Specifically:

[0034] Obtain the user quantity information of the resettlement points in each evacuation route from the dynamic road network, set the user quantity constraint conditions for the resettlement points, and determine whether the user quantity information of the resettlement points in the evacuation route is greater than the user quantity constraint conditions for the resettlement points;

[0035] When the user quantity information of the resettlement points in the evacuation route is greater than the user quantity constraint conditions for the resettlement points, re-plan the evacuation routes for the users without assigned resettlement points and update the dynamic road network;

[0036] When the user quantity information of the resettlement points in the evacuation route is not greater than the user quantity constraint conditions for the resettlement points, conduct evacuation planning according to the current dynamic road network.

[0037] In the method for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing, it further includes:

[0038] After planning the optimal evacuation routes, generate relevant hazard avoidance information, and input the relevant hazard avoidance information into the AR navigation function through a preset terminal;

[0039] Display the relevant hazard avoidance information and the evacuation routes in a virtual scene, construct a real-time virtual scene map, and represent the positions of hazard avoidance points and scene navigation instructions in the real-time virtual scene map through the representation of the virtual scene.

[0040] The second aspect of the present invention provides a system for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing, including a memory and a processor. The memory includes a program for the method for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing. When the program for the method for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing is executed by the processor, the steps of any one of the methods for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing are implemented.

[0041] The third aspect of the present invention provides a computer-readable storage medium, including a program for the dynamic planning method of the evacuation route of residents in flood detention areas based on crowd-sourced sensing. When the program for the dynamic planning method of the evacuation route of residents in flood detention areas based on crowd-sourced sensing is executed by a processor, the steps of any of the dynamic planning methods of the evacuation route of residents in flood detention areas based on crowd-sourced sensing are implemented.

[0042] The present invention solves the defects existing in the background art and has the following beneficial effects:

[0043] The present invention collects multi-source heterogeneous data in flood detention areas, cleans the multi-source heterogeneous data in flood detention areas, and then fuses the multi-source heterogeneous data after data cleaning, updates the fused data, predicts the flood situation data at preset coordinate points, thereby constructs a dynamic road network based on the fused data and the flood situation data at preset coordinate points, sends the dynamic road network to a preset terminal in a preset manner, finally evaluates the real-time status information of the dynamic road network, and dynamically updates the dynamic road network based on the evaluation results. By integrating multi-source heterogeneous data and the ant colony algorithm, and using the multi-source heterogeneous data as the constraint conditions of the ant colony algorithm, the present invention can optimize the evacuation routes in flood detention areas and make the evacuation routes more reasonable. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0045] Figure 1 Shows the overall flowchart of the dynamic planning method of the evacuation route of residents in flood detention areas based on crowd-sourced sensing;

[0046] Figure 2 Shows the system architecture diagram of the dynamic planning method of the evacuation route of residents in flood detention areas based on crowd-sourced sensing;

[0047] Figure 3 Shows the data collection and fusion method flowchart of the dynamic planning method of the evacuation route of residents in flood detention areas based on crowd-sourced sensing;

[0048] Figure 4 Shows the flowchart of the improved ant colony algorithm module;

[0049] Figure 5 Shows the system block diagram of the dynamic planning system of the evacuation route of residents in flood detention areas based on crowd-sourced sensing. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] In order to more clearly understand the above-mentioned objects, features, and advantages of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0051] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0052] The first aspect of the present invention provides a method for dynamically planning the evacuation routes of residents in flood detention areas based on crowd-sourced sensing, including the following steps:

[0053] S102: Collect multi-source heterogeneous data in the flood detention area and perform data cleaning on the multi-source heterogeneous data in the flood detention area;

[0054] S104: Through data fusion of the multi-source heterogeneous data after data cleaning, update the fused data, and predict the flood situation data of preset coordinate points;

[0055] S106: Based on the fused data and the flood situation data of the preset coordinate points, construct a dynamic road network and send the dynamic road network to a preset terminal in a preset manner;

[0056] S108: Evaluate the real-time status information of the dynamic road network and perform dynamic updates on the dynamic road network based on the evaluation results.

[0057] It should be noted that by fusing multi-source heterogeneous data and the ant colony algorithm, and using the multi-source heterogeneous data as the constraint conditions of the ant colony algorithm, the present invention can optimize the evacuation routes in the flood detention area and make the evacuation routes more reasonable.

[0058] As Figures 1 to 4 shown, among them, in step S102, by cooperating with mobile communication operators, GPS signal data of residents' mobile phones or in-vehicle devices in the flood detention area is obtained. These data include information such as the longitude and latitude coordinates and timestamps of mobile phones, and can reflect the location distribution, moving speed, and staying duration of residents in real time. The data collection frequency is set according to actual needs, generally once every few minutes, to ensure the real-time nature of the data.

[0059] Install traffic cameras at the main roads, intersections and key nodes in the flood detention area to capture basic information such as traffic flow, vehicle speed, road congestion, etc., as well as abnormal information such as road waterlogging and accident status in real time. The cameras transmit video data to the data processing center through the network, and the data processing center uses AI image recognition and video analysis technologies to identify the number of vehicles, pedestrian flow, waterlogging depth, water level information, etc. Construct multi-source heterogeneous data in the flood detention area based on traffic data information and GPS signal data.

[0060] Among them, when performing data cleaning, clean the received mobile phone GPS signals and traffic camera data to remove outliers and incorrect data. For example, for mobile phone GPS signals, there may be inaccurate positioning due to signal occlusion or interference. By setting reasonable thresholds, data points that deviate significantly from the normal range are excluded. Use AI image recognition and video analysis technologies to remove misidentified data caused by factors such as video jitter, light changes, and occlusion. Unify the data display formats collected by traffic cameras of different brands and models for subsequent data processing.

[0061] Among them, in step S104, fuse the cleaned mobile phone GPS signals and traffic camera data to obtain more comprehensive and accurate traffic and personnel distribution information in the flood detention area.

[0062] Data fusion adopts the method of weighted average and assigns different weights according to the reliability and timeliness of the data. For example, for the evacuation route planning with high requirements for timeliness, higher weights are given to the latest data. During the data fusion process, it is first necessary to align the data from different data sources. This usually involves the alignment of timestamps. For example, GPS data and camera data are matched according to timestamps. We used the pd.merge_asof function, which can perform approximate matching according to timestamps and allows a certain tolerance range (such as 1 minute). After data alignment, it is necessary to select an appropriate fusion algorithm. Data fusion can be achieved through various methods. Common methods include the weighted average method, Bayesian estimation method, Kalman filtering method, etc. For example, using the weighted average method, the weights are first determined. The weights can be determined according to the reliability, timeliness, richness of details, and redundancy of the data sources. Suppose there are two data sources now, each data source is scored, and the weight of reliability is set to 0.3, timeliness is 0.2, richness of details is 0.4, and redundancy is 0.1. According to the scores, use the code to calculate the total scores of each data source respectively, align the two data sources according to timestamps, define a function to fuse the data, define column mapping, and finally fuse the data. For example, the score corresponding to the reliability of the mobile phone GPS signal is 4, the score of timeliness is 5, the richness of details is 3, and the redundancy is 2; the score corresponding to the reliability of the traffic camera data is 5, the score of timeliness is 3, the richness of details is 5, and the redundancy is 2. After data fusion, we obtain a unified data set containing timestamps, longitude, latitude, and traffic condition information. This data set can be used for subsequent analysis and decision-making, such as real-time optimizing evacuation routes, pushing hazard avoidance information, etc. The positioning accuracy reaches 1 meter, the recognition error of the water accumulation depth is <10 cm, and the road condition changes are responded within 5 minutes. At the same time, the fused data is updated in real time to reflect the latest situation of the flood detention area and construct a dynamic road network. The update frequency is dynamically adjusted according to the development of the disaster and the evacuation progress to ensure timely and accurate data support for subsequent route planning.

[0063] Among them, in step S104, the flood situation data of the preset coordinate point can be predicted, and a flood characteristic data prediction model can be constructed through a deep neural network. The rainfall monitoring accuracy and coverage can be improved by integrating multi-source data such as rain radar, satellite remote sensing, and hydrological stations. Collect all-factor (water level, flow, soil moisture, etc.) monitoring information to provide real-time input conditions for the model, and use historical flood inundation data to calibrate and verify model parameters. Set up a pre-drilled flood scenario (such as the flood diversion process under different scheduling strategies). Before the flood storage area is activated, the model is used to predict the flooding range, water level peak and evolution time after flood diversion (flood situation data) to provide a basis for personnel transfer and engineering scheduling, and dynamically correct the model parameters in combination with emergency monitoring data (such as breach flow, dike danger) to improve the forecast accuracy, so as to obtain the full-factor monitoring information of the preset coordinate point within the preset time, and input the flood characteristic data prediction model according to the full-factor monitoring information for prediction, and obtain the flood situation data of the preset coordinate point.

[0064] It should be noted that the following steps may also be included:

[0065] According to the flood situation data, flood risk status is divided to obtain flood risk status classification, a dynamic Bayesian network is introduced to obtain flood risk status classification of several timestamps, and a flood risk status classification state matrix is ​​constructed based on the flood risk status classification of the timestamps;

[0066] Inputting the flood risk state classification state matrix into the dynamic Bayesian network, taking the flood risk state classification of each timestamp as an observation vector, and calculating the state transition probability value of the observation vector transformed to another level of observation vector;

[0067] When the state transition probability value is greater than a preset state transition probability value, the current observation vector is transformed into an observation vector of another level, and the flood risk state classification state matrix is ​​updated;

[0068] Different multi-source heterogeneous data acquisition frequencies are divided according to the flood risk status classification, and the flood risk status classification state of the current timestamp is obtained according to the flood risk status classification state matrix, and the data acquisition frequency of the flood risk status classification state of the current timestamp is adjusted.

[0069] It should be noted that the flood risk status grading includes high-risk level, medium-risk level, low-risk level, etc. When the state transition probability value is greater than the preset state transition probability value, the current observation vector is transformed into an observation vector of another level, and the flood risk status grading state matrix is updated, enabling rapid monitoring and updating through a remote terminal in advance, improving the monitoring accuracy of the flood risk status grading; moreover, different multi-source heterogeneous data collection frequencies are divided according to the flood risk status grading, and the flood risk status grading at the current timestamp is obtained according to the flood risk status grading state matrix, and the data collection frequency is adjusted for the flood risk status grading at the current timestamp, which can improve the rationality of data collection. In a particularly urgent situation, the data collection frequency can be adjusted, and updates can be made based on real-time data; secondly, in a non-urgent situation, the data collection frequency can be reduced, and the monitoring cost can be reduced.

[0070] Among them, in step S106, the ant colony algorithm is an optimization algorithm that simulates the foraging behavior of ants. In nature, ants can find the shortest path from the ant nest to the food source because ants release pheromones during the foraging process, and other ants will choose paths according to the concentration of pheromones. The path with a higher pheromone concentration is considered a better path, and more ants will choose to take such a path, thus further enhancing the pheromones and forming a positive feedback mechanism.

[0071] The ant colony algorithm abstracts this natural phenomenon into a mathematical model for solving various optimization problems, such as path planning. When a flood occurs in a flood detention area, residents need to be quickly evacuated. These "ants" are like the avatars of residents. A high pheromone concentration means that this road has been chosen more times or this road is more suitable for evacuation. According to the fused data, the nodes and edges of the flood detention area are determined. Nodes represent key locations such as road intersections and landmark buildings, and edges represent the roads connecting these nodes. In the evacuation path planning, each iteration is equivalent to searching for the optimal path again. (Optimal can be understood as the shortest evacuation time or the largest number of transferred people)

[0072] Initialize the parameters of the ant colony algorithm, including the number of ants (the number of people to be transferred), the pheromone evaporation coefficient, the importance factor of pheromone, the heuristic factor, etc. For the evacuation route planning of flood detention areas, assuming there are 100 key nodes in the flood detention area, the number of ants can be set from 10 to 20; the importance factor of pheromone (α) is usually set from 0.5 to 5, and the simulation results show that within this range, the convergence speed of the algorithm is relatively stable; the size of the heuristic factor (β) has a greater impact on the number of cycles of the ant colony algorithm, and generally β ∈ [1.2, 5.0]; the choice of the amount of pheromone (Q) can generally be taken as Q ∈ [3, 6]; ρ represents the pheromone volatility, 1 - ρ is the pheromone residue coefficient (0 ≤ ρ ≤ 1), and ρ ∈ [0.3, 0.9].

[0073] The judgment conditions for transfer path selection are: the predicted inundation range is not in the path, and the path will not be flooded within the effective travel time. Then assume that each ant starts from the starting node and selects the next path according to the pheromone concentration and heuristic information, gradually constructing an evacuation path from the starting node to the target node. The heuristic information can include factors such as real-time water depth, path length, planned travel time (traffic congestion degree), road traffic capacity, etc., to guide the ants to select a better path.

[0074] The calculation formula for the heuristic factor is:

[0075]

[0076] Path selection probability:

[0077]

[0078] Among them, α and β are the importance weights of pheromone and heuristic factor respectively, and P ij represents the probability value of the selection of path ij, A ij represents the pheromone in path ij, B ij represents the heuristic factor in path ij; A ik represents the pheromone in path ik, B ik represents the heuristic factor in path ik.

[0079] Among them, in the process of path planning by the ant colony algorithm, during the movement of ants, the pheromone concentration on the path is updated in real time, reducing pheromone evaporation, so as to perform local update; and at the end of each iteration, the pheromone concentration is adjusted according to the path length of the ants and the flood risk, so as to perform global update, satisfying the following relationship:

[0080] A ij =(1 - ρ)×A ij +ΔA ij

[0081] where ρ is the pheromone evaporation rate, and ΔA ij According to the path fitness calculation, the processes of repeated path construction and pheromone update are carried out. After multiple iterations, the ant colony algorithm gradually converges to the optimal or approximately optimal evacuation path. The number of iterations is set according to the actual requirements and the algorithm convergence speed, generally ranging from dozens to hundreds of times.

[0082] Among them, in step S106, the dynamic road network is sent to a preset terminal in a preset manner, specifically including:

[0083] Obtain the flood situation data of the preset coordinate point, and obtain the boundary of the risk area based on the flood situation data of the preset coordinate point. Taking the boundary of the risk area as the benchmark, set an early warning buffer zone and construct a polygonal electronic fence;

[0084] Based on the polygonal electronic fence, render the warning circle layer through the mobile map SDK, obtain the geographical location information of the user, and determine whether the geographical location information of the user is within the warning circle;

[0085] When the geographical location information of the user is within the warning circle, send the dynamic road network to the preset terminal in a preset manner, generate a prompt message, and determine whether there is a response based on the prompt message within a preset time;

[0086] When there is no response within the preset time, include the continuously unresponsive users in the rescue priority calculation model.

[0087] It should be noted that according to the flood storage and detention area flood forecasting and early warning system, the boundary of the risk area is dynamically delimited. The buffer expansion mechanism is adopted. Taking the flood prediction boundary as the benchmark, expand 500 meters outward as the early warning buffer zone to generate a polygonal electronic fence. Render the warning circle layer through the mobile map SDK, and use the red - orange double - color gradient to distinguish the emergency evacuation area (the area directly threatened by floods) and the preparatory evacuation area (the buffer zone). After the user enters the warning circle, the system pops up an application interface of "whether to request centralized transfer". According to the application data, the optimal path is recommended for self - transferred users preferentially. Establish a low - latency communication protocol to ensure that the user - selected data (user ID, GPS coordinates, selection type, timestamp, etc.) is synchronized to the government command platform within 3 seconds, and allocate rescue vehicle resources in real time. For unresponsive personnel, a secondary reminder (vibration + voice broadcast) is triggered after 5 minutes, and continuously unresponsive ones are included in the rescue priority calculation model.

[0088] In the method for dynamically planning the evacuation path of residents in the flood storage and detention area based on crowd - sourced perception, evaluate the real - time status information of the dynamic road network, and dynamically update the dynamic road network based on the evaluation results, specifically as follows:

[0089] Obtain the user quantity information of the resettlement points in each evacuation route from the dynamic road network, set the user quantity constraint conditions for the resettlement points, and determine whether the user quantity information of the resettlement points in the evacuation route is greater than the user quantity constraint conditions for the resettlement points;

[0090] When the user quantity information of the resettlement points in the evacuation route is greater than the user quantity constraint conditions for the resettlement points, re-plan the evacuation routes for the users without assigned resettlement points and update the dynamic road network;

[0091] When the user quantity information of the resettlement points in the evacuation route is not greater than the user quantity constraint conditions for the resettlement points, conduct the evacuation plan according to the current dynamic road network.

[0092] It should be noted that when the user's GPS enters the electronic fence of the resettlement point (such as setting a radius of 300 meters) and stays continuously for more than 5 minutes; a dynamic QR code scanning interface pops up and is linked with the registration terminal (PDA device) at the resettlement point site to verify and ensure the physical presence authenticity. When the real-time capacity of a certain resettlement point > 95% (it can also be the number threshold), the following operations are automatically executed: push a red warning (including the over-limit location and the list of resettlement points that can be diverted) to the government platform. The front-end navigation system shields this resettlement point and re-plans the evacuation routes of the residents who have not arrived through the ant colony algorithm, triggers the government resource scheduling interface, and requests to allocate additional temporary resettlement facilities, so as to re-plan the dynamic routes and make the evacuation route planning more reasonable.

[0093] Among them, this method also includes: after planning the optimal evacuation route, generate relevant hazard avoidance information, and input the relevant hazard avoidance information into the AR navigation function through a preset terminal;

[0094] Display the relevant hazard avoidance information and the evacuation route in the virtual scene, construct a real-time virtual scene map, and represent the location of the hazard avoidance points and the scene navigation instructions in the real-time virtual scene map through the representation of the virtual scene.

[0095] It should be noted that according to the optimal evacuation route planned by the ant colony algorithm, corresponding hazard avoidance information is generated. The hazard avoidance information includes route guidance, safety tips, estimated arrival time, resettlement point capacity, etc., to help residents evacuate to a safe area quickly and safely. Through the AR navigation function of mobile phones or other mobile terminals, the hazard avoidance information is pushed to residents in the form of intuitive images and texts. AR navigation uses the camera and sensors of the mobile phone to superimpose the virtual evacuation route and hazard avoidance information on the real scene. The AR navigation interface superimposes the location of the hazard avoidance points and multilingual voice prompts, and residents only need to act according to the navigation instructions, which improves the convenience of evacuation.

[0096] In addition, this method also includes:

[0097] Obtain the size information of the roads in each evacuation route, and conduct a preview through virtual reality technology based on the size information of the roads in the evacuation route to obtain the maximum evacuation capacity threshold, and obtain the real-time number of people flow information of each evacuation route;

[0098] Judge whether the real-time number of people flow information of the evacuation route is greater than the maximum evacuation capacity threshold;

[0099] When the real-time number of people flow information of each evacuation route is greater than the maximum evacuation capacity threshold, after allocating the maximum number of people corresponding to the maximum evacuation capacity threshold according to the dynamic road network, re-plan the paths for the remaining people on the unallocated evacuation routes, and update the dynamic road network;

[0100] When the real-time number of people flow information of each evacuation route is not greater than the maximum evacuation capacity threshold, allocate evacuation routes according to the current dynamic road network.

[0101] It should be noted that this method continues to fully consider the size of the paths to estimate the maximum evacuation capacity threshold. Thus, when the real-time number of people flow information of each evacuation route is greater than the maximum evacuation capacity threshold, after allocating the maximum number of people corresponding to the maximum evacuation capacity threshold according to the dynamic road network, re-plan the paths for the remaining people on the unallocated evacuation routes, and update the dynamic road network to improve the planning rationality of the dynamic road network.

[0102] In addition, this method also includes:

[0103] Obtain the information collection delay characteristic data of the multi-source heterogeneous data collection nodes in each flood detention area within a preset time;

[0104] Set the information collection delay characteristic data threshold, and judge whether the information collection delay characteristic data of the multi-source heterogeneous data collection nodes in the flood detention area within a preset time is greater than the information collection delay characteristic data threshold;

[0105] Obtain the data collection nodes where the information collection delay characteristic data within a preset time is greater than the information collection delay characteristic data threshold, and judge whether the dynamic road network passes through the data collection nodes where the information collection delay characteristic data within a preset time is greater than the information collection delay characteristic data threshold;

[0106] When the dynamic road network passes through the data collection nodes where the information collection delay characteristic data within a preset time is greater than the information collection delay characteristic data threshold, re-plan the evacuation route until the dynamic road network no longer passes through the data collection nodes where the information collection delay characteristic data within a preset time is greater than the information collection delay characteristic data threshold.

[0107] It should be noted that since the data acquisition nodes may be temporarily damaged due to abnormal conditions such as bad weather and floods, when the information collection delay characteristic data within the preset time in the dynamic road network is greater than the data acquisition node where the information collection delay characteristic data threshold is located, it indicates that this route is also infeasible. Through this method, abnormal data acquisition nodes can be identified, so as to re-plan the evacuation route and improve the planning rationality of the dynamic road network.

[0108] As Figure 5 shown, the second aspect of the present invention provides a dynamic planning system 4 for the evacuation routes of residents in flood detention areas based on crowd-sourced sensing, including a memory 41 and a processor 42. The memory 41 includes a program for the dynamic planning method of the evacuation routes of residents in flood detention areas based on crowd-sourced sensing. When the program for the dynamic planning method of the evacuation routes of residents in flood detention areas based on crowd-sourced sensing is executed by the processor 42, the steps of any of the dynamic planning methods of the evacuation routes of residents in flood detention areas based on crowd-sourced sensing are implemented.

[0109] The third aspect of the present invention provides a computer-readable storage medium, including a program for the dynamic planning method of the evacuation routes of residents in flood detention areas based on crowd-sourced sensing. When the program for the dynamic planning method of the evacuation routes of residents in flood detention areas based on crowd-sourced sensing is executed by a processor, the steps of any of the dynamic planning methods of the evacuation routes of residents in flood detention areas based on crowd-sourced sensing are implemented.

[0110] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed with each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be electrical, mechanical, or other forms.

[0111] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units; they can be located in one place or distributed to multiple network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] In addition, in each embodiment of the present invention, each functional unit can be entirely integrated into one processing unit, or each unit can be separately regarded as one unit, or two or more units can be integrated into one unit; the above-mentioned integrated unit can be implemented in the form of hardware, or in the form of a hardware plus software functional unit.

[0113] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including the above method embodiments; and the foregoing storage medium includes: removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks or optical discs and other various media that can store program codes.

[0114] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of the various embodiments of the present invention. And the foregoing storage medium includes: removable storage devices, ROM, RAM, magnetic disks or optical discs and other various media that can store program codes.

[0115] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A dynamic planning method for evacuation routes of residents in flood storage areas based on crowd intelligence perception is characterized by: The following steps are involved: Collecting multi-source heterogeneous data in a flood storage and detention area, and performing data cleaning on the multi-source heterogeneous data in the flood storage and detention area; By fusing the cleaned multi-source heterogeneous data, updating the fused data, and predicting the flood situation data of the preset coordinate points; Building a dynamic road network based on the fused data and flood situation data of preset coordinate points, and sending the dynamic road network to a preset terminal in a preset manner; The real-time status information of the dynamic road network is evaluated, and the dynamic road network is dynamically updated based on the evaluation results.

2. The method for dynamic planning of evacuation paths for residents in flood storage areas based on crowd intelligence perception according to claim 1 is characterized in that: Collect multi-source heterogeneous data in the flood storage and detention area, and perform data cleaning on the multi-source heterogeneous data in the flood storage and detention area, specifically: Obtain GPS signal data from mobile phones or vehicle-mounted devices of residents in the flood storage and detention area, and install traffic cameras on preset roads, intersections and preset key nodes in the flood storage and detention area to capture traffic data information in real time through traffic cameras; Constructing multi-source heterogeneous data in the flood storage area according to traffic data information and GPS signal data, and setting a data threshold range, and proposing data points that deviate from the data threshold range based on the data threshold range; AI image recognition and video analysis technologies are used to remove misidentified data caused by video jitter, light changes, and occlusion factors, and to unify the display format of data collected by traffic cameras of different brands and models.

3. The method for dynamic planning of evacuation paths for residents in flood storage areas based on crowd intelligence perception according to claim 1 is characterized in that: By performing data fusion on the multi-source heterogeneous data after data cleaning, the fused data is updated, specifically: Introduce the pd.merge_asof function, use the pd.merge_asof function to match the cleaned multi-source heterogeneous data according to the timestamp, set the tolerance range, and approximately match according to the tolerance range, so that the cleaned multi-source heterogeneous data is aligned; After data alignment, the data fusion algorithm is used to fuse the aligned multi-source heterogeneous data, determine the reliability weight of each type of multi-source heterogeneous data and perform scoring processing to obtain the reliability weight value; The total score of each data source is calculated based on the reliability weight value, the two data sources are aligned according to the timestamp, and the column mapping is defined. Finally, the data is fused. Through data fusion, a unified data set containing timestamp, longitude, latitude and traffic data information is obtained.

4. The method for dynamic planning of evacuation paths for residents in flood storage areas based on crowd intelligence perception according to claim 1 is characterized in that: Predict flood situation data at preset coordinate points, specifically: Collect historical data collected by rainfall radar, satellite remote sensing, and hydrological stations, build full-factor monitoring information based on the historical data collected by rainfall radar, satellite remote sensing, and hydrological stations, and build a flood characteristic data prediction model based on a deep neural network; The full-factor monitoring information is used as a model input of the flood characteristic data prediction model to set and verify model parameters, preview flood scenarios, and dynamically correct model parameters in combination with the full-factor monitoring information and flood scenarios; The full-factor monitoring information of the preset coordinate point within the preset time is obtained, and the full-factor monitoring information is input into the flood characteristic data prediction model for prediction to obtain the flood situation data of the preset coordinate point.

5. The method for dynamic planning of evacuation paths for residents in flood storage areas based on crowd intelligence perception according to claim 1 is characterized in that: A dynamic road network is constructed based on the fused data and the flood situation data of the preset coordinate points, specifically: Introducing an ant colony algorithm, determining nodes and edges of the flood storage area based on the fused data and flood situation data of preset coordinate points, and initializing the number of ants, pheromone volatility coefficient, pheromone importance factor, and heuristic factor of the ant colony algorithm; The number of people to be transferred is obtained, and the number of people to be transferred is used as the number of ants in the ant colony algorithm. The real-time water depth, path length, planning time, and road capacity are used as heuristic factors. Based on the nodes and edges of the flood storage area, the next path is selected from the starting node according to the pheromone concentration and heuristic information. Step by step, construct an evacuation path from a starting node to a target node, generate an initialization path, set constraints based on flood condition data of the preset coordinate points, and determine whether the transfer path meets the constraints; When the transfer path meets the constraint condition, the current transfer path iteration ends. At the end of each iteration, the pheromone concentration is adjusted according to the path length of the ants and the flood risk. When the transfer path does not meet the constraint condition, the iteration continues. During the movement of the ants, the pheromone concentration on the path is updated in real time to reduce the volatilization of pheromones. The process of path construction and pheromone updating is repeated, and after a preset number of iterations, an optimal evacuation path is obtained, and a dynamic road network is constructed according to the optimal evacuation path.

6. The method for dynamic planning of evacuation paths for residents in flood storage areas based on crowd intelligence perception according to claim 1 is characterized in that: Sending the dynamic road network to a preset terminal in a preset manner specifically includes: Obtain flood situation data of preset coordinate points, and obtain risk area boundaries according to the flood situation data of the preset coordinate points, set a warning buffer zone based on the risk area boundaries, and construct a polygonal electronic fence; Based on the polygonal electronic fence, a warning circle layer is rendered through a mobile map SDK to obtain the geographical location information of the user, and determine whether the geographical location information of the user is within the warning circle; When the geographical location information of the user is in the warning circle, the dynamic road network is sent to a preset terminal in a preset manner, a prompt message is generated, and it is determined whether there is a response based on the prompt message within a preset time; When there is no response within a preset time, users who continue to be unresponsive are included in the rescue priority calculation model.

7. The method for dynamic planning of evacuation routes for residents in flood storage and detention areas based on crowd intelligence perception according to claim 1 is characterized in that: Evaluate the real-time status information of the dynamic road network, and dynamically update the dynamic road network based on the evaluation results, specifically: Acquire user quantity information of the placement points in each evacuation route from the dynamic road network, set user quantity constraint conditions of the placement points, and determine whether the user quantity information of the placement points in the evacuation route is greater than the user quantity constraint conditions of the placement points; When the number of users at the placement points in the evacuation path is greater than the number of users at the placement points, the evacuation path is replanned for users who are not assigned placement points, and the dynamic road network is updated; When the user quantity information of the placement point in the evacuation path is not greater than the user quantity constraint condition of the placement point, evacuation planning is performed according to the current dynamic road network.

8. The method for dynamic planning of evacuation routes for residents in flood storage and detention areas based on crowd intelligence perception according to claim 1 is characterized in that: Also includes: After the optimal evacuation route is planned, relevant risk avoidance information is generated, and the relevant risk avoidance information is input into the AR navigation function through a preset terminal; The relevant evacuation information and evacuation routes are displayed in a virtual scene, a real-time virtual scene graph is constructed, and the locations of evacuation points and scene navigation instructions are represented in the real-time virtual scene graph through the representation of the virtual scene.

9. A dynamic planning system for evacuation routes of residents in flood storage areas based on crowd intelligence perception, characterized in that: It comprises a memory and a processor, wherein the memory comprises a method program for dynamically planning the evacuation path of residents in a flood storage and detention area based on crowd intelligence perception, and when the method program for dynamically planning the evacuation path of residents in a flood storage and detention area based on crowd intelligence perception is executed by the processor, the steps of the method for dynamically planning the evacuation path of residents in a flood storage and detention area based on crowd intelligence perception as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that: It includes a method program for dynamically planning the evacuation path of residents in a flood storage and detention area based on crowd intelligence perception. When the method program for dynamically planning the evacuation path of residents in a flood storage and detention area based on crowd intelligence perception is executed by a processor, the steps of the method for dynamically planning the evacuation path of residents in a flood storage and detention area based on crowd intelligence perception as described in any one of claims 1 to 8 are implemented.

Images