A flood transfer path optimization method based on flood loss assessment

Abstract

The application discloses a flood-avoiding transfer path optimization method based on flood loss evaluation, relates to the technical field of flood disaster emergency management, and comprises the following steps: collecting basic data and resource data in advance, and pre-processing the data, wherein the basic data comprises real-time rainfall, water level, social and economic data and satellite remote sensing data, and the resource data comprises refuge capacity, material reserve quantity and traffic capacity data; and the waterlogged range is extracted through an SWI-OTSU algorithm. The application deeply fuses path optimization and cross-regional emergency resource scheduling, constructs a resource scheduling model by using a linear programming algorithm, is used for realizing the matching of paths, resources and demands, solves the problem that resources are disconnected with paths in traditional methods, significantly improves the overall collaborative efficiency of emergency response, and realizes the real-time response of disaster loss evaluation, resource state monitoring and path optimization by means of a high-frequency data updating mechanism, and completely gets rid of the hysteresis of a traditional static model.

Description

Technical Field

[0001] This invention relates to the field of flood disaster emergency management technology, and more specifically, to a method for optimizing flood evacuation routes based on flood loss assessment. Background Technology

[0002] Floods are frequent natural disasters worldwide, posing a serious threat to people's lives and property due to their suddenness and destructiveness. Traditional flood evacuation route planning methods mainly rely on historical flood data, static topographic information, and empirical rules, which have significant limitations: poor timeliness of assessment, as they are mostly based on pre-generated flood risk maps and cannot respond to dynamically changing flood scenarios in real time. For example, in sudden heavy rainfall events, static models cannot accurately reflect the rapid expansion of the flooded area; insufficient accuracy of loss assessment, usually only considering geographical distance or simple flood depth, ignoring economic losses, such as the impact of damage to buildings, farmland, and infrastructure on route selection, leading to a disconnect between evacuation plans and the actual disaster situation; single data source, with most methods relying on single data sources such as topographic elevation data, lacking the integrated application of multi-source data such as remote sensing, real-time rainfall, and socio-economic data, limiting the comprehensiveness and accuracy of the assessment; and lack of dynamic adaptability in route optimization, as traditional route planning algorithms, such as Dijkstra's algorithm, are not combined with dynamic flood loss assessment and cannot adjust routes in real time during the evolution of the disaster to minimize potential losses.

[0003] In recent years, although some studies have begun to incorporate remote sensing technology and multi-source data, such as extracting flooded areas through radar imagery or using deep learning for land cover classification, these technologies have not yet been deeply integrated with route optimization and cross-regional resource allocation. Meanwhile, flood loss assessments often focus on post-disaster statistics, lacking real-time quantification capabilities during disasters, and resource allocation is mostly reactive, failing to deeply link flood evacuation routes with cross-regional emergency resources. This leads to problems such as accessible routes but insufficient resources, or sufficient resources but blocked routes. Specifically, some evacuation routes reach shelters that are already saturated, but haven't been promptly redeployed to backup shelters, or lack corresponding transportation capacity during evacuation, resulting in low evacuation efficiency and disorder.

[0004] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0005] In response to the problems in related technologies, this invention proposes a flood avoidance and evacuation route optimization method based on flood loss assessment to overcome the aforementioned technical problems existing in the existing related technologies.

[0006] The technical solution of this invention is implemented as follows:

[0007] A method for optimizing flood evacuation routes based on flood loss assessment includes the following steps:

[0008] Basic data and resource data are collected in advance and the data is preprocessed. The basic data includes real-time rainfall, water level, socio-economic data and satellite remote sensing data. The resource data includes shelter capacity, material reserves and transportation capacity data.

[0009] The flooded area was extracted using the SWI-OTSU algorithm, and the RDDA network model was used to classify land features, quantify their value and economic loss coefficient, and calculate the total economic loss of the region.

[0010] Construct a resource supply and demand balance model to assess the degree of supply and demand matching, use a linear programming algorithm to dynamically allocate cross-regional emergency resources, and establish a real-time resource data update strategy;

[0011] A multi-dimensional weighted graph is constructed based on the loss assessment and resource status results. An improved A* algorithm is used to generate dynamic transfer paths with the goals of minimizing loss and optimizing resource matching.

[0012] The step of extracting the flooded area using the SWI-OTSU algorithm includes the following steps:

[0013] The water body characteristics, pre-defined using normalized water body indices, are expressed as follows:

[0014] ;

[0015] In the formula, For the green band reflectance of remote sensing images, Reflectivity in the near-infrared band;

[0016] The OTSU algorithm is used to achieve threshold adaptive segmentation for accurate extraction of flooded areas, including: assuming the image grayscale level is... The threshold T divides the image into foreground water bodies and background non-water bodies, and the objective function is to maximize the inter-class variance. , is represented as:

[0017] ;

[0018] In the formula, Foreground pixel ratio The percentage of background pixels. The average value of the foreground pixels. This represents the average pixel value of the background.

[0019] The method of classifying land features and quantifying their value and economic loss coefficient using the RDDA network model includes: quantifying land feature value, i.e., the benchmark value per unit area of ​​a certain type of land feature. , is represented as:

[0020] ;

[0021] In the formula, k represents the land feature category. Let k be the area of ​​the k-th type of land cover within the i-th sample region. Let be the asset value per unit area of ​​the k-th type of land feature within the i-th sample region, and n be the number of sample regions;

[0022] Among them, the economic loss coefficient of land features , is represented as:

[0023] ;

[0024] In the formula, h is the flood depth. This represents the maximum loss coefficient for the k-th type of land cover, with a value range of [0,1]. The loss rate parameter is obtained by fitting historical data.

[0025] The total economic loss of the calculation area is expressed as:

[0026] ;

[0027] Where m represents the total number of land cover categories, and p represents the number of raster cells within the flooded area. Let k be the area of ​​the k-th type of land feature within the j-th grid cell. Let the flood depth be the j-th grid cell. is the rainfall correction factor, and R is the real-time rainfall, fitted from historical rainfall loss data.

[0028] The process of constructing a resource supply and demand balance model to assess the degree of supply and demand matching includes the following steps:

[0029] Shelter supply and demand matching , is represented as:

[0030] ;

[0031] In the formula, Let be the maximum capacity of the s-th shelter. Let be the current number of residents in the s-th shelter. The number of people planned to be transferred to the s-th shelter;

[0032] Transportation capacity supply and demand matching degree , is represented as:

[0033] ;

[0034] In the formula, Let r be the unit time capacity of the r-th transportation route. To allow for transfer time, This refers to the number of passengers that the line needs to accommodate for transfers.

[0035] The method of dynamically allocating cross-regional emergency resources using a linear programming algorithm includes defining the resource scheduling objective function, i.e., minimizing the total scheduling cost. , is represented as:

[0036] ;

[0037] In the formula, S represents the number of shelters, and T represents the number of affected areas. Let be the number of people transferred from the t-th disaster area to the s-th shelter. For the cost of transfer per unit of employees, The quantity of supplies to be allocated from the s-th shelter to the t-th area. Cost of material allocation for the unit.

[0038] The multi-dimensional weight graph is represented as follows:

[0039] ;

[0040] In the formula, These are the normalized coefficients for the loss weight, resource weight, and traffic weight, respectively. The loss weight for the j-th grid is expressed as: , This represents the potential loss of the raster. The resource weight of the j-th grid is expressed as: , The nearest shelter to grid j; The traffic weight of the j-th grid is expressed as: , This represents the road traffic speed corresponding to this grid.

[0041] Wherein, the improved A* algorithm heuristic function , is represented as:

[0042] ;

[0043] In the formula, The cumulative cost from the starting point to the current node n is expressed as:

[0044] ;

[0045] In the formula, Let j be the side length of the grid cell. This is the set of grid cells traversed by the path from the starting point to node n;

[0046] in, The heuristic cost from the current node n to the destination shelter is expressed as:

[0047] ;

[0048] In the formula, Let n be the coordinates of node n. The coordinates of the endpoint The average weight of the region The resource fit coefficient for the final refuge.

[0049] The beneficial effects of this invention are:

[0050] This invention deeply integrates path optimization with cross-regional emergency resource scheduling, and uses a linear programming algorithm to construct a resource scheduling model to achieve the matching of paths, resources and demands. This solves the problem of resource and path disconnect in traditional methods, significantly improves the overall coordination efficiency of emergency response, and achieves real-time response of disaster loss assessment, resource status monitoring and path optimization by means of a high-frequency data update mechanism, completely getting rid of the lag of traditional static models.

[0051] Meanwhile, with minimizing economic losses as the core objective, and combining a multi-dimensional weight model and an improved A* algorithm, the system considers resource utilization efficiency when optimizing relocation paths. This effectively avoids high-loss areas and prevents resource waste or shortages, ensuring a smooth and orderly relocation process. Its path optimization, loss assessment, and resource scheduling form a closed-loop linkage, which can respond synchronously to the dynamic changes in flood evolution, resource consumption, and relocation needs. Furthermore, the various parameters in the model can be adapted to the actual conditions of different regions using historical data, supporting application scenarios at different scales, such as urban blocks, watersheds, and cross-regional areas. It has strong generalization ability and scalability. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart illustrating a flood evacuation route optimization method based on flood loss assessment according to an embodiment of the present invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0055] According to an embodiment of the present invention, a method for optimizing flood evacuation routes based on flood loss assessment is provided.

[0056] like Figure 1 As shown, the flood evacuation route optimization method based on flood loss assessment according to an embodiment of the present invention includes the following steps:

[0057] Step S1: Collect basic data and resource data in advance, and preprocess the data. The basic data includes real-time rainfall, water level, socio-economic data and satellite remote sensing data. The resource data includes shelter capacity, material reserves and transportation capacity data.

[0058] Specifically, it collects basic data by using web crawling technology to obtain real-time rainfall, water level, and socio-economic data, such as population density and asset distribution, from public platforms, such as meteorological and hydrological stations; and obtains surface observation data before and after flood events, including water body extent and land cover types, such as buildings, farmland, and roads, through satellite remote sensing data, such as Sentinel-1 radar images and multispectral data.

[0059] Among them, resource data can be collected through cross-regional emergency management platforms, transportation department databases, shelter management systems, etc. Real-time emergency resource data can be obtained, including the current capacity, remaining capacity, and material reserves of each shelter, such as food, drinking water, medical supplies, and transportation capacity data, such as the number of buses, dispatch status, and road traffic capacity.

[0060] This technical solution preprocesses the data, including standardizing and denoising the data, such as using morphological filtering and interpolation to eliminate image noise, and standardizing the format of resource data and removing outliers to ensure data quality and consistency.

[0061] Step S2 involves extracting the flooded area using the SWI-OTSU algorithm, classifying land features and quantifying their value and economic loss coefficients using the RDDA network model, and calculating the total regional economic loss based on the UFIITC model. This includes the following steps:

[0062] Step S201, the water body characteristics are expressed as follows through the normalized water body index:

[0063]

[0064] In the formula, For the green band reflectance of remote sensing images, This refers to the reflectivity in the near-infrared band.

[0065] Step S202, using the OTSU algorithm to achieve threshold adaptive segmentation for accurate extraction of the flooded area, includes: setting the image grayscale level as follows: The threshold T divides the image into foreground water bodies and background non-water bodies, and the objective function is to maximize the inter-class variance. , is represented as:

[0066] ;

[0067] In the formula, Foreground pixel ratio The percentage of background pixels. The average value of the foreground pixels. This represents the average pixel value of the background.

[0068] Step S203: Use the RDDA (Residual Deep Discriminative Analysis) network model to classify land features in multispectral remote sensing data, identify categories such as buildings, farmland, and infrastructure, and assign economic loss coefficients to each type of land feature based on its value.

[0069] Among them, quantified land feature value refers to the benchmark value per unit area of ​​a certain type of land feature. , is represented as:

[0070] ;

[0071] In the formula, k represents the type of land feature, such as buildings or farmland. Let k be the area of ​​the k-th type of land cover within the i-th sample region. Let be the asset value per unit area of ​​the k-th type of land feature within the i-th sample region, and n be the number of sample regions;

[0072] Among them, the economic loss coefficient of land features , is represented as:

[0073] ;

[0074] In the formula, h is the flood depth. This represents the maximum loss coefficient for the k-th type of land cover, with a value range of [0,1]. The loss rate parameter is obtained by fitting historical data.

[0075] Step S204: Calculate the total regional economic loss. , is represented as:

[0076] ;

[0077] Where m represents the total number of land cover categories, and p represents the number of raster cells within the flooded area. Let k be the area of ​​the k-th type of land feature within the j-th grid cell. Let the flood depth be the j-th grid cell. is the rainfall correction factor, and R is the real-time rainfall, fitted from historical rainfall loss data.

[0078] Step S3 involves constructing a resource supply and demand balance model to assess the degree of supply and demand matching, dynamically allocating cross-regional emergency resources using a linear programming algorithm, and establishing a real-time resource data update strategy; this includes the following steps:

[0079] Step S301: Based on real-time transfer demand data and collected resource data, construct a resource supply and demand balance model to assess the supply and demand matching degree of shelters and transportation capacity in various regions.

[0080] Among them, the supply and demand matching degree of shelters , is represented as:

[0081] ;

[0082] In the formula, Let be the maximum capacity of the s-th shelter. Let be the current number of residents in the s-th shelter. This represents the number of people planned to be transferred to the s-th shelter.

[0083] Among them, the degree of matching between supply and demand of transportation capacity , is represented as:

[0084] ;

[0085] In the formula, Let r be the unit time capacity of the r-th transportation route. To allow for transfer time, This refers to the number of passengers that the line needs to accommodate for transfers.

[0086] Step S302: A linear programming algorithm is used to construct a resource scheduling objective function, with the goal of maximizing resource utilization efficiency and minimizing transfer costs, to dynamically allocate cross-regional emergency resources. The constraints include the upper limit of shelter capacity, the timeliness of material transportation, road capacity, and vehicle scheduling costs.

[0087] The resource scheduling objective function is to minimize the total scheduling cost. , is represented as:

[0088] ;

[0089] The shelter capacity constraint is expressed as follows:

[0090] ;

[0091] The transfer demand constraint is expressed as:

[0092] ;

[0093] The material matching constraint is expressed as follows:

[0094] ;

[0095] Among them, the transportation timeliness constraint is expressed as:

[0096] ;

[0097] The nonnegativity constraint is represented as follows:

[0098] ;

[0099] In the formula, S represents the number of shelters, and T represents the number of affected areas. Let be the number of people transferred from the t-th disaster area to the s-th shelter. For the cost of transfer per unit of employees, The quantity of supplies to be allocated from the s-th shelter to the t-th area. For the unit's material allocation cost, Let t be the number of people to be relocated in the t-th disaster-stricken area. Let be the total amount of supplies needed in the t-th disaster-stricken area. Let be the transportation time from the t-th region to the s-th shelter. This is the maximum permissible transportation time.

[0100] Step S303: Establish a dynamic resource data update strategy, and synchronize the data on shelter capacity and transportation capacity changes from various data sources every 15 minutes to ensure the real-time performance and accuracy of resource scheduling.

[0101] Among them, the real-time remaining capacity of the shelter , is represented as:

[0102] ;

[0103] In the formula, t is the current time, and t−1 is the previous update time. minute, for The number of new residents admitted to the s-th shelter within a given time period for The amount of supplies allocated to the shelter within a given timeframe. for The amount of supplies consumed by the shelter within a given time period.

[0104] Step S4: Construct a multi-dimensional weighted graph based on the loss assessment and resource status results, and use the improved A* algorithm to generate a dynamic transfer path with the goal of minimizing loss and optimizing resource matching;

[0105] Specifically, the multi-dimensional weighted graph integrates flood loss assessment results and resource supply and demand status assessment results, transforming them into a raster-style multi-dimensional weighted graph, i.e., the comprehensive weight of raster cells, expressed as:

[0106] ;

[0107] In the formula, These are the normalized coefficients for the loss weight, resource weight, and traffic weight, respectively. The loss weight for the j-th grid is expressed as: , This represents the potential loss of the raster. The resource weight of the j-th grid is expressed as: , The nearest shelter to grid j; The traffic weight of the j-th grid is expressed as: , This represents the road traffic speed corresponding to this grid.

[0108] Among them, the generation of dynamic transfer paths is achieved by using an improved A* algorithm with the dual objective functions of minimizing losses and optimizing resource matching, and by dynamically generating flood avoidance transfer paths based on real-time flood evolution data and resource scheduling results.

[0109] Specifically, improve the heuristic function of the A* algorithm. , is represented as:

[0110] ;

[0111] In the formula, The cumulative cost from the starting point to the current node n is expressed as:

[0112] ;

[0113] In the formula, Let j be the side length of the grid cell. This is the set of grid cells traversed by the path from the starting point to node n;

[0114] in, The heuristic cost from the current node n to the destination shelter is expressed as:

[0115] ;

[0116] In the formula, Let n be the coordinates of node n. The coordinates of the endpoint The average weight of the region The resource fit coefficient for the final refuge.

[0117] Meanwhile, the optimization process considers the following constraints, as detailed below:

[0118] The safety of the transfer route, i.e., avoiding deep water areas, is represented as: , The safe flood depth threshold;

[0119] Time efficiency, or the shortest reachable time, is expressed as: , Maximum allowable transfer time;

[0120] Loss aversion, which means prioritizing avoiding high-loss areas, is expressed as: , This is the maximum allowable loss threshold;

[0121] Resource matching, which prioritizes routes to shelters with sufficient remaining capacity, adequate supplies, and guaranteed transportation, is represented as: , This represents the minimum acceptable supply-demand match.

[0122] In addition, during implementation, it reassesses the loss and resource status through real-time data streams, such as remote sensing images updated every 30 minutes and resource data updated every 15 minutes, and adjusts the path planning scheme in sync to ensure that the scheme adapts to the dynamic changes of disasters and fluctuations in resource supply and demand as a dynamic update mechanism.

[0123] Furthermore, during implementation, it can also be integrated with a cross-regional emergency command system to generate a visualized comprehensive solution map and a comprehensive evaluation report.

[0124] Specifically, during implementation, it can be directly connected to the cross-regional emergency command system to support decision-making and execution. The system generates a visualized comprehensive solution map including recommended evacuation routes, alternative routes, risk warnings, and resource status annotations for each shelter, such as remaining capacity, material availability, and transportation capacity scheduling. Simultaneously, it outputs a comprehensive assessment report, including evacuation time, resource allocation plans, and cross-regional collaboration suggestions.

[0125] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Those skilled in the art, upon considering the disclosure in the specification and embodiments, will readily conceive of other embodiments of this disclosure. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0126] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for optimizing flood evacuation routes based on flood loss assessment, characterized in that, Includes the following steps: Basic data and resource data are collected in advance and the data is preprocessed. The basic data includes real-time rainfall, water level, socio-economic data and satellite remote sensing data. The resource data includes shelter capacity, material reserves and transportation capacity data. The flooded area was extracted using the SWI-OTSU algorithm, and the RDDA network model was used to classify land features, quantify their value and economic loss coefficient, and calculate the total economic loss of the region. Construct a resource supply and demand balance model to assess the degree of supply and demand matching, use a linear programming algorithm to dynamically allocate cross-regional emergency resources, and establish a real-time resource data update strategy; A multi-dimensional weighted graph is constructed based on the loss assessment and resource status results. An improved A* algorithm is used to generate dynamic transfer paths with the goals of minimizing loss and optimizing resource matching.

2. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 1, characterized in that, The extraction of the flooded area using the SWI-OTSU algorithm includes the following steps: The water body characteristics, pre-defined using normalized water body indices, are expressed as follows: ； In the formula, For the green band reflectance of remote sensing images, Reflectivity in the near-infrared band; The OTSU algorithm is used to achieve threshold adaptive segmentation for accurate extraction of flooded areas, including: assuming the image grayscale level is... The threshold T divides the image into foreground water bodies and background non-water bodies, and the objective function is to maximize the inter-class variance. , is represented as: ； In the formula, Foreground pixel ratio This represents the percentage of background pixels. The average value of the foreground pixels. This represents the average value of the background pixels.

3. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 2, characterized in that, The method of using the RDDA network model to classify land features and quantify their value and economic loss coefficient includes: quantifying the value of land features, i.e., the benchmark value per unit area of ​​a certain type of land feature. , is represented as: ； In the formula, k represents the land feature category. Let k be the area of ​​the k-th type of land cover within the i-th sample region. Let be the asset value per unit area of ​​the k-th type of land feature within the i-th sample region, and n be the number of sample regions; Among them, the economic loss coefficient of land features , is represented as: ； In the formula, h is the flood depth. This represents the maximum loss coefficient for the k-th type of land cover, with a value range of [0,1]. The loss rate parameter is obtained by fitting historical data.

4. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 3, characterized in that, The total economic loss of the calculation area is expressed as: ； Where m represents the total number of land cover categories, and p represents the number of raster cells within the flooded area. Let k be the area of ​​the k-th type of land feature within the j-th grid cell. Let the flood depth be the j-th grid cell. is the rainfall correction factor, and R is the real-time rainfall, fitted from historical rainfall loss data.

5. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 1, characterized in that, Constructing a resource supply and demand balance model to assess the degree of supply and demand matching includes the following steps: Shelter supply and demand matching , is represented as: ； In the formula, Let be the maximum capacity of the s-th shelter. Let be the current number of residents in the s-th shelter. The number of people planned to be transferred to the s-th shelter; Transportation capacity supply and demand matching degree , is represented as: ； In the formula, Let r be the unit time capacity of the r-th transportation route. To allow for transfer time, This refers to the number of passengers that the line needs to accommodate for transfers.

6. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 5, characterized in that, The method of dynamically allocating cross-regional emergency resources using a linear programming algorithm includes defining the resource scheduling objective function, namely, minimizing the total scheduling cost. , is represented as: ； In the formula, S represents the number of shelters, and T represents the number of affected areas. Let be the number of people transferred from the t-th disaster area to the s-th shelter. For the cost of transfer per unit of employees, The quantity of supplies to be allocated from the s-th shelter to the t-th area. Cost of material allocation for the unit.

7. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 1, characterized in that, The multi-dimensional weight graph is represented as follows: ； In the formula, These are the normalized coefficients for the loss weight, resource weight, and traffic weight, respectively. The loss weight for the j-th grid is expressed as: , This represents the potential loss of the raster. The resource weight of the j-th grid is expressed as: , The nearest shelter to grid j; The traffic weight of the j-th grid is expressed as: , This represents the road traffic speed corresponding to this grid.

8. The flood avoidance and evacuation route optimization method based on flood loss assessment according to claim 7, characterized in that, The improved A* algorithm heuristic function , is represented as: ； In the formula, The cumulative cost from the starting point to the current node n is expressed as: ； In the formula, Let j be the side length of the grid cell. This is the set of grid cells traversed by the path from the starting point to node n; in, The heuristic cost from the current node n to the destination shelter is expressed as: ； In the formula, Let n be the coordinates of node n. The coordinates of the endpoint The average weight of the region The resource fit coefficient for the final refuge.

Images