An urban waterlogging state analysis method and system based on spatial heterogeneity

Abstract

The application belongs to the technical field of urban waterlogging analysis, and more particularly relates to a method and system for analyzing the state of urban waterlogging based on spatial heterogeneity. The method of the application screens out key driving factors affecting the formation of waterlogging by combining the benchmark role of shadow variables and the importance evaluation mechanism of multiple models through the Boruta algorithm, and can obtain accurate dominant driving factors. By using the spatial distribution data of each dominant driving factor in the waterlogging unit, each waterlogging unit is recursively divided into multiple waterlogging subgroups, and then the quantitative calculation ability of the factor contribution value of the geographic detector model can be used to effectively utilize the spatial heterogeneity between each waterlogging subgroup, accurately calculate the factor contribution value of the corresponding target dominant factor of each waterlogging subgroup, and thus accurately obtain the multi-factor synergistic results between each dominant driving factor through the factor contribution value of each target dominant factor, thereby providing data support for generating accurate prevention and control strategies for urban waterlogging through each dominant driving factor.

Description

Technical Field

[0001] This application belongs to the field of urban waterlogging analysis technology, and more specifically, relates to a method and system for analyzing urban waterlogging status based on spatial heterogeneity. Background Technology

[0002] Urban flooding due to torrential rain has become a major disaster threatening the safety of modern cities. Its causes are complex, influenced by the coupling of multiple factors such as rainfall intensity, topography, land use, drainage capacity, and vegetation cover. As a typical complex adaptive system, the mechanism of urban flooding is far from a linear superposition of single factors, but rather stems from the nonlinear interaction of multiple factors. Especially in highly urbanized areas, due to intensive human intervention and rapid reconstruction of landscape elements, the spatial heterogeneity and functional dynamism of these driving factors are significantly enhanced, leading to typical spatial nonstationarity characteristics in the flooding driving mechanism: the degree of influence of driving factors on flood formation varies significantly with spatial location.

[0003] Existing methods for analyzing urban flooding generally suffer from the following problems: 1. They rely on a single model to screen for flood-related driving factors. The screening process is easily affected by the characteristics of the model and it is difficult to stably identify the dominant driving factors in urban flooding scenarios of varying complexity. 2. They ignore the spatial heterogeneity between different flooded areas in the city and only analyze the dominant factors of each flooded area. As a result, they cannot effectively quantitatively assess the interaction of multiple factors, which seriously affects the efficiency of subsequent governance of urban flooded areas. Summary of the Invention

[0004] To address the aforementioned deficiencies in existing technologies, this application provides a method for analyzing urban flooding status based on spatial heterogeneity. This method aims to accurately obtain the dominant driving factors among the various flooding driving factors in the target area, and effectively obtain the multi-factor synergistic results among the dominant driving factors based on the spatial heterogeneity among the flooding units in the urban target area.

[0005] Firstly, this application provides a method for analyzing urban flooding status based on spatial heterogeneity, including: S1. Preprocess the spatial distribution data of each waterlogging driving factor in the target area and construct a spatial dataset. Each sub-dimension of the spatial dataset corresponds to the spatial distribution data of a waterlogging driving factor. S2. Based on the target feature selection model, obtain the comprehensive feature ranking of each waterlogging driving factor in the spatial dataset, and take the top waterlogging driving factors in the comprehensive feature ranking as the dominant driving factors to construct a set of dominant driving factors. S3. Obtain the waterlogging density distribution map corresponding to the target area, and divide each watershed unit in the target area into waterlogged units and non-waterlogged units; S4. Based on the spatial distribution data of each dominant driving factor in the waterlogging unit, each waterlogging unit is recursively segmented to obtain multiple waterlogging subgroups. Each waterlogging subgroup corresponds to a dominant driving factor as the target dominant factor. S5. Based on the geographic detector model, the factor contribution values ​​of each waterlogging subgroup are quantitatively calculated to obtain the factor contribution values ​​of the target dominant factors corresponding to each waterlogging subgroup. Based on the factor contribution values ​​of each target dominant factor, the multi-factor synergy results among the dominant driving factors are obtained.

[0006] Furthermore, the driving factors of waterlogging include at least rainfall, elevation standard deviation, population density, slope, shadow variable threshold, water system distance difference, slope standard deviation, vegetation cover, land use type, and elevation.

[0007] Furthermore, the spatially distributed data are preprocessed, including outlier removal, missing value imputation, and data normalization for the waterlogging driving factors at each location point within the target area.

[0008] Furthermore, the target feature selection model is the Boruta algorithm model optimized by multiple classification models, including at least XGBoost, LightGBM, and SVM.

[0009] Furthermore, based on the target feature selection model, a comprehensive feature ranking of each waterlogging driving factor in the spatial dataset is obtained, including: S21. Based on the spatial distribution data of each waterlogging driving factor, construct the original feature matrix, and randomly arrange each column of the original feature matrix to generate shadow variables. Based on each shadow variable and the original feature matrix, construct the extended feature matrix. S22. Based on the trained classification model, the extended feature matrix is ​​predicted to obtain feature scores of shadow variables corresponding to each waterlogging driving factor. S23. Weighted fusion of the feature scores corresponding to each waterlogging driving factor is performed to obtain the comprehensive feature score of each waterlogging driving factor, and the comprehensive feature scores are ranked.

[0010] Among them, the target feature selection model belongs to the improved Boruta algorithm. It is an integrated feature selection algorithm based on shadow variable statistical testing and multi-model weighted voting, which can effectively reduce the collinearity interference of high-dimensional feature elements and the risk of misselection of noisy features. The algorithm constructs randomly arranged shadow variables as a benchmark, combines multi-model (XGBoost, LightGBM, SVM) heterogeneity evaluation and weighted voting mechanism, quantifies the statistical significance of feature importance, dynamically eliminates redundant variables, and ultimately effectively retains the flood driving factors that are globally representative of urban flooding.

[0011] Furthermore, the watershed units in the target area are divided into flood-prone units and non-flood-prone units, including: The waterlogging density characteristics of each watershed unit in the waterlogging density distribution map are obtained based on the CatBoost model, and each watershed unit is classified into two categories based on the preset waterlogging density discrimination threshold.

[0012] Furthermore, based on the spatial distribution data of each dominant driving factor in the waterlogging unit, each waterlogging unit is recursively segmented to obtain multiple waterlogging subgroups, including: Based on the Cubist regression tree algorithm and the preset segmentation thresholds corresponding to each dominant driving factor, multiple waterlogging subgroups are obtained. Local regression operations are then performed on each dominant driving factor corresponding to each waterlogging subgroup to obtain the target variable variance ratio of each dominant driving factor. The dominant driving factor with the highest target variable variance ratio is taken as the target dominant factor of the waterlogging subgroup.

[0013] The Cubist regression tree algorithm can use a large number of binary tree structures, where each branch can represent the threshold segmentation result of a certain dominant driving factor. Therefore, by setting a large number of preset segmentation thresholds, a large number of waterlogging subgroups can be obtained, and the threshold segmentation size of each branch directly affects the target dominant factor of each waterlogging subgroup.

[0014] Furthermore, the multi-factor synergistic results include nonlinear weakening, single-factor nonlinear weakening, binary enhancement, independent and nonlinear enhancement.

[0015] Secondly, this application also provides an urban flooding status analysis system based on spatial heterogeneity, for performing any of the methods in the first aspect, including: The dataset construction module is used to preprocess the spatial distribution data of each waterlogging driver in the target area and construct a spatial dataset. The dominant driving factor partitioning module is used to obtain the comprehensive characteristic ranking of each waterlogging driving factor in the spatial dataset, and to construct the dominant driving factor set by taking the top waterlogging driving factors in the comprehensive characteristic ranking as the dominant driving factors. Watershed unit division is used to obtain the waterlogging density distribution map corresponding to the target area, and to divide each watershed unit in the target area into waterlogged units and non-waterlogged units. The waterlogging subgroup segmentation unit is used to recursively segment each waterlogging unit based on the spatial distribution data of each dominant driving factor in the waterlogging unit, thereby obtaining multiple waterlogging subgroups. The multi-factor synergy result acquisition unit is used to quantify the factor contribution values ​​of each waterlogging subgroup based on the geographic detector model, obtain the factor contribution values ​​of the target dominant factors corresponding to each waterlogging subgroup, and obtain the multi-factor synergy results among the dominant driving factors based on the factor contribution values ​​of each target dominant factor.

[0016] Thirdly, this application also provides an electronic device, comprising: at least one memory for storing a program; and at least one processor for executing the program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to execute any of the methods of the first aspect.

[0017] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art: This application's method, through preprocessing operations such as outlier removal, missing value imputation, and data normalization on the spatial distribution data of waterlogging driving factors in the target area, effectively ensures the authenticity of the data input to the target feature selection model. By employing the Boruta algorithm combined with the benchmark role of shadow variables and a multi-model importance assessment mechanism, it identifies key driving factors influencing waterlogging formation, effectively improving the robustness and generalization ability of the target feature selection model, thus obtaining accurate dominant driving factors. Using the spatial distribution data of each dominant driving factor in waterlogging units, each waterlogging unit is recursively divided into multiple waterlogging subgroups. The quantitative calculation capability of the factor contribution value of the geographic detector model can then be utilized to effectively leverage the spatial heterogeneity between the subgroups, achieving accurate calculation of the factor contribution value of the target dominant factor corresponding to each subgroup. This allows for accurate acquisition of the multi-factor synergistic results among the dominant driving factors through their factor contribution values, providing data support for subsequent generation of precise urban waterlogging prevention and control strategies based on these dominant driving factors. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the urban flooding status analysis method provided in the embodiments of this application.

[0020] Figure 2 This is another flowchart illustrating the urban flooding status analysis method provided in this application embodiment.

[0021] Figure 3 This is a schematic diagram illustrating the recursive partitioning principle of the waterlogging subgroup provided in this application embodiment.

[0022] Figure 4 This is a schematic diagram illustrating the calculation principle of factor contribution value provided in the embodiments of this application.

[0023] Figure 5 This is a schematic diagram of the multi-factor synergistic results provided in the embodiments of this application.

[0024] Figure 6 This is a schematic diagram of the urban flooding status analysis system provided in the embodiments of this application.

[0025] Figure 7 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0026] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0027] In the following description, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The following description provides multiple embodiments of this application, which can be substituted or combined with each other. Therefore, this application can also be considered to include all possible combinations of the same and / or different embodiments described. Thus, if one embodiment includes features A, B, and C, and another embodiment includes features B and D, then this application should also be considered to include embodiments containing one or more other possible combinations of A, B, C, and D, even if such embodiments are not explicitly described in the following text.

[0028] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the function and arrangement of the described elements without departing from the scope of this application. Various processes or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

[0029] Figure 1 This is a flowchart illustrating the urban flooding status analysis method provided in this application embodiment. The method includes at least the following steps: S1. Preprocess the spatial distribution data of each waterlogging driving factor in the target area and construct a spatial dataset. Each sub-dimension of the spatial dataset corresponds to the spatial distribution data of a waterlogging driving factor.

[0030] In this embodiment, the implementing entity can be the central controller of an urban municipal management system. The formation mechanism of urban flooding is a complex system process involving multi-dimensional and non-linear coupling of natural geographical elements and human activities. Flooding can be analyzed from four sets of driving factors: meteorological and hydrological data, topography, land use, and human intervention. In this embodiment, rainfall, elevation standard deviation, population density, slope, shadow variable threshold, water system distance difference, slope standard deviation, vegetation cover, land use type, and elevation are selected as examples. The implementing object of this application is all location points within the target area of ​​the city. The spatial distribution data of each flooding driving factor is similar to a layer of a thermal remote sensing image covering this target area. The two-dimensional reference image of this thermal remote sensing image can be the flooding density distribution map of this application.

[0031] In addition, since the spatial distribution data may be damaged during the collection or transmission process, it is necessary to perform preprocessing processes such as outlier removal, missing value imputation and data normalization on each spatial distribution data according to the coordinate range of the target area before constructing the spatial dataset. This is to ensure that the subsequent construction of the feature matrix based on the spatial dataset can proceed smoothly. The normalization process unifies the data of various waterlogging driving factors with different units, which also ensures the construction of the feature matrix.

[0032] S2. Based on the target feature selection model, obtain the comprehensive feature ranking of each waterlogging driving factor in the spatial dataset, and take the top waterlogging driving factors as the dominant driving factors to construct a set of dominant driving factors.

[0033] In this embodiment, the target feature selection model is a Boruta algorithm optimized by multiple classification models, including at least XGBoost, LightGBM, and SVM. The Boruta algorithm is a wrapper-type feature selection algorithm based on Random Forest (RF). Compared to traditional wrapper algorithms that only select a subset of features that minimize classifier error, this algorithm identifies all features with predictive utility for the target variable through a global search mechanism. Therefore, it is suitable for the application scenario of this application, which includes a large amount of spatially distributed data on waterlogging drivers. Traditional Boruta relies on the split gain of random forests to evaluate feature importance. Its limitations are: tree models are sensitive to nonlinear relationships and may ignore linearly dominant features; single models are susceptible to parameter settings or noise interference, leading to unstable importance evaluation. Therefore, this application uses multi-model weighted feature selection to replace the single random forest framework of the traditional Boruta algorithm, aiming to overcome model bias and enhance the robustness of feature selection.

[0034] XGBoost captures complex nonlinear interactions through gradient boosting tree structures, and its split gain importance score is robust to high-dimensional sparse data; LightGBM uses histogram optimization and deep growth strategies to efficiently identify the local contribution of feature combinations in multidimensional data; SVM quantifies the strength of linear associations based on the absolute value of regression coefficients, supplementing the blind spot of tree models for global linear patterns.

[0035] In one possible implementation, based on a target feature selection model, a comprehensive feature ranking of each waterlogging driving factor in the spatial dataset is obtained, including: S21. Based on the spatial distribution data of each waterlogging driving factor, construct the original feature matrix, and randomly arrange each column of the original feature matrix to generate shadow variables. Based on each shadow variable and the original feature matrix, construct the extended feature matrix. S22. Based on the trained classification model, the extended feature matrix is ​​predicted to obtain feature scores of shadow variables corresponding to each waterlogging driving factor. S23. Weighted fusion of the feature scores corresponding to each waterlogging driving factor is performed to obtain the comprehensive feature score of each waterlogging driving factor, and the comprehensive feature scores are ranked.

[0036] In the embodiments of this application, such as Figure 2 As shown, step S2 combines the Boruta algorithm with the benchmark role of shadow variables and a multi-model importance assessment mechanism. The weights of the feature scores corresponding to each flooding driving factor for each classification model can be preset, for example, based on the validation set performance metrics of each model on the validation set for each flooding driving factor classification, or based on some evaluation metrics. By constructing a multi-model weighted voting Boruta algorithm, the robustness and generalization ability of feature selection are improved, making it suitable for scenarios with high requirements for noise filtering in high-dimensional data, multi-model collaborative decision-making, and interpretability.

[0037] S3. Obtain the waterlogging density distribution map corresponding to the target area, and divide each watershed unit in the target area into waterlogged units and non-waterlogged units.

[0038] In one possible implementation, the watershed units in the target area are divided into flood-prone units and non-flood-prone units, including: The waterlogging density features of each watershed unit in the waterlogging density distribution map are obtained based on the CatBoost model, and each watershed unit is classified into two categories based on the preset waterlogging density discrimination threshold. In this embodiment, CatBoost uses oblivious trees as the base learner. Compared to XGBoost and LightGBM, which use ordinary decision trees, the different decision conditions for left and right nodes at each layer can lead to overfitting at the same depth. CatBoost, however, uses the same splitting criterion at each layer of the tree. The waterlogging density distribution map is used to characterize the waterlogging density distribution in each watershed unit of the target area (the watershed unit is divided using the natural breakpoint method based on the number of waterlogged points and the watershed area in the waterlogging density distribution map). It effectively reflects the waterlogging density at each location in the target area and has a mapping relationship with the spatial distribution data of each dominant driving factor. Therefore, a preset waterlogging density discrimination threshold can be set, and the CatBoost model can effectively divide the watershed units according to their waterlogging distribution.

[0039] S4. Based on the spatial distribution data of each dominant driving factor in the waterlogging unit, each waterlogging unit is recursively segmented to obtain multiple waterlogging subgroups. Each waterlogging subgroup corresponds to a dominant driving factor as the target dominant factor.

[0040] In one possible implementation, based on the spatial distribution data of each dominant driving factor in the waterlogging unit, each waterlogging unit is recursively segmented to obtain multiple waterlogging subgroups, including: Based on the Cubist regression tree algorithm and the preset segmentation thresholds corresponding to each dominant driving factor, multiple waterlogging subgroups are obtained. Local regression operations are then performed on each dominant driving factor corresponding to each waterlogging subgroup to obtain the target variable variance ratio of each dominant driving factor. The dominant driving factor with the highest target variable variance ratio is taken as the target dominant factor of the waterlogging subgroup.

[0041] In this embodiment of the application, the Cubist regression tree algorithm can reveal the non-stationary relationships in the space of waterlogging, such as... Figure 3 As shown, the Cubist regression tree divides urban flooding space into six heterogeneous subgroups using rules to clearly define local driving forces spatially. There are six rules in total, with one rule corresponding to each subgroup. The constructed set of local driving force rules indicates significant differences in the dominant driving factor parameters for flooding formation across different geographical subgroups: in watershed units with FVC ≥ 0.42 (normalized value) and Land_use ≥ 0.65 (Rule 1 subgroup), the variation in flooding density is mainly driven by the synergistic effect of vegetation cover, elevation, and land use type; while in watershed units with DEM > 0.21 and FVC > 0.42 (Rule 6 subgroup), the disaster-causing pattern is dominated by topographic elevation.

[0042] The purpose of performing local regression calculations on the dominant driving factors corresponding to each subgroup of waterlogging is to determine the proportion of the target variable variance in each subgroup of waterlogging, that is, to determine whether the dominant driving factor exhibits a non-stationary relationship within the corresponding subgroup of waterlogging. The larger the proportion of the target variable variance, the more stable the dominant driving factor is within the corresponding subgroup of waterlogging, and the more likely it is to serve as the target dominant factor for that subgroup of waterlogging.

[0043] S5. Based on the geographic detector model, the factor contribution values ​​of each waterlogging subgroup are quantitatively calculated to obtain the factor contribution values ​​of the target dominant factors corresponding to each waterlogging subgroup. Based on the factor contribution values ​​of each target dominant factor, the multi-factor synergy results among the dominant driving factors are obtained.

[0044] In the embodiments of this application, such as Figure 4 As shown, the Geographical Detector Model (GDM) is a spatial statistical analysis model that can effectively diagnose spatial heterogeneity among driving factors and reveal their driving forces. Because it does not assume linearity, this method can effectively avoid the limitations of multivariate collinearity. GDM typically includes four detections: factor detection, interaction detection, risk detection, and ecological monitoring. This embodiment mainly studies the dominant spatial factors and driving forces.

[0045] Among them, the geographic detector model uses a factor detector to detect the explanatory power of each representative driving factor on the spatial differentiation of urban flooding. The degree of explanation here This is the factor contribution value of this application, calculated using the following formula:

[0046]

[0047]

[0048]

[0049] in, The number of units, and the layers are as follows: =1,2,…, where L is the number of layers. , For unit The waterlogging density value in the overall and the first Values ​​in the layer This represents the average waterlogging density value for the entire area. In the first Number of units per layer SSW is the variance of each unit layer. SSW is the sum of the variances within each layer, and SST is the total variance of the entire region. The value is between 0 and 1. =0 indicates that there is no correlation between the independent variable and the dependent variable. The larger the value, the stronger the explanatory power of the representative factors of urban flooding. The flooding density value is obtained from the flooding density distribution map corresponding to the target area. If the factor contribution value of the dominant driving factor in a certain flooding subgroup exceeds a certain value, which can be 0.6, it can be considered a strong dominant unit; if it is lower than this value, it is a weak dominant unit.

[0050] like Figure 5 As shown, after obtaining the factor contribution values ​​of the target dominant factors corresponding to each waterlogging subgroup, it is necessary to evaluate the multi-factor synergy results among the dominant driving factors in order to determine the synergistic effect of each factor on waterlogging. In this application, the synergistic effect of factors is determined by the interaction detector of the geographic detector model.

[0051] Interaction detection is a key module in the analysis of multi-factor coupling effects among the four core functions of a geographic detector. Its core objective is to identify the explanatory power of two or more independent variables on the spatial differentiation of the dependent variable Y when they act together, determining whether the effects of multiple factors are synergistic enhancement, antagonistic weakening, or independent. The calculation is still based on the q-statistic of factor detection. The core logic is to define the type of interaction by quantitatively comparing the q-values ​​of single factors with those of combined factors. In the method of this application, each dominant driving factor within each waterlogging unit in the target area is randomly paired. Since the preceding S1 process has already performed normalization, the values ​​of the two paired dominant driving factors can be superimposed. Then, based on the superimposed factor combination value, the explanatory power is calculated, yielding a comprehensive factor contribution value for comparing and analyzing the combined effects of two factors.

[0052] The multi-factor synergistic results in the embodiments of this application include nonlinear weakening, single-factor nonlinear weakening, binary enhancement, independent enhancement, and nonlinear enhancement. (See also...) Figure 5 As shown in the figure x 1 and x 2 represents different dominant driving factors. When the contribution value of the comprehensive factor is greater than the sum of the contributions of the two single factors, the multi-factor synergy result is characterized by nonlinear enhancement; when the contribution value of the comprehensive factor is equal to the sum of the contributions of the two single factors, the multi-factor synergy result is characterized by each factor being independent; when the contribution value of the comprehensive factor is greater than the maximum value among the two single factors, the multi-factor synergy result is characterized by binary enhancement; when the contribution value of the comprehensive factor is less than the minimum value among the two single factors, the multi-factor synergy result is characterized by nonlinear weakening; when the contribution value of the comprehensive factor is between the contributions of the two single factors, the multi-factor synergy result is characterized by single-factor nonlinear weakening, that is, the factor with the smallest contribution value has an inhibitory effect on the other factor.

[0053] Figure 6 An urban flooding status analysis system based on spatial heterogeneity is provided for embodiments of this application, such as... Figure 6 As shown, the system includes at least: The dataset construction module is used to preprocess the spatial distribution data of each waterlogging driver in the target area and construct a spatial dataset. The dominant driving factor partitioning module is used to obtain the comprehensive characteristic ranking of each waterlogging driving factor in the spatial dataset, and to construct the dominant driving factor set by taking the top waterlogging driving factors in the comprehensive characteristic ranking as the dominant driving factors. Watershed unit division is used to obtain the waterlogging density distribution map corresponding to the target area, and to divide each watershed unit in the target area into waterlogged units and non-waterlogged units. The waterlogging subgroup segmentation unit is used to recursively segment each waterlogging unit based on the spatial distribution data of each dominant driving factor in the waterlogging unit, thereby obtaining multiple waterlogging subgroups. The multi-factor synergy result acquisition unit is used to quantify the factor contribution values ​​of each waterlogging subgroup based on the geographic detector model, obtain the factor contribution values ​​of the target dominant factors corresponding to each waterlogging subgroup, and obtain the multi-factor synergy results among the dominant driving factors based on the factor contribution values ​​of each target dominant factor.

[0054] like Figure 7 As shown, Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: a processor 701, a communications interface 702, a memory 703, and a communication bus 704. The processor 701, communications interface 702, and memory 703 communicate with each other via the communication bus 704. The processor 701 can call software instructions in the memory 703 to execute the methods described in the above embodiments.

[0055] Furthermore, the logical instructions in the aforementioned memory 703 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application.

[0056] Based on the methods in the above embodiments, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0057] Based on the methods in the above embodiments, this application provides a computer program product that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0058] It is understood that the processor in the embodiments of this application can be a CPU (Central Processing Unit), or other general-purpose processors, DSPs (Digital Signal Processors), ASICs (Application Specific Integrated Circuits), FPGAs (Field Programmable Gate Arrays), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0059] The method steps in this application embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, ROM (Read-only Memory), PROM (Programmable ROM), EPROM (Erasable PROM), EEPROM (Electrically Erasable EPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0060] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line DSL) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., SSD (Solid State Disk)).

[0061] It is understood that the various numerical designations used in the embodiments of this application are merely for the convenience of description and are not intended to limit the scope of the embodiments of this application.

[0062] Those skilled in the art will readily understand that the above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for analyzing urban flooding status based on spatial heterogeneity, characterized in that, include: S1. Preprocess the spatial distribution data of each waterlogging driving factor in the target area and construct a spatial dataset, wherein each sub-dimension of the spatial dataset corresponds to the spatial distribution data of a waterlogging driving factor. S2. Based on the target feature selection model, obtain the comprehensive feature ranking of each of the waterlogging driving factors in the spatial dataset, and take the waterlogging driving factors with the highest comprehensive feature ranking as the dominant driving factors to construct a set of dominant driving factors. S3. Obtain the waterlogging density distribution map corresponding to the target area, and divide each watershed unit in the target area into waterlogged units and non-waterlogged units; S4. Based on the spatial distribution data of each dominant driving factor in the waterlogging unit, each waterlogging unit is recursively segmented to obtain multiple waterlogging subgroups. Each waterlogging subgroup corresponds to a dominant driving factor as a target dominant factor. S5. Based on the geographic detector model, the factor contribution values ​​of each of the waterlogging subgroups are quantitatively calculated to obtain the factor contribution values ​​of the target dominant factors corresponding to each of the waterlogging subgroups. Based on the factor contribution values ​​of each of the target dominant factors, the multi-factor synergy results among the dominant driving factors are obtained.

2. The urban flooding status analysis method according to claim 1, characterized in that, The waterlogging drivers include at least rainfall, elevation standard deviation, population density, slope, shadow variable threshold, water system distance difference, slope standard deviation, vegetation cover, land use type, and elevation.

3. The urban flooding status analysis method according to claim 1, characterized in that, The preprocessing of the spatial distribution data includes: performing outlier removal, missing value imputation, and data normalization on the waterlogging driving factors at each location point within the target area.

4. The urban flooding status analysis method according to claim 3, characterized in that, The target feature selection model is the Boruta algorithm model optimized by multiple classification models, including at least XGBoost, LightGBM, and SVM.

5. The urban flooding status analysis method according to claim 4, characterized in that, The method for obtaining a comprehensive feature ranking of each of the waterlogging driving factors in the spatial dataset based on the target feature selection model includes: S21. Based on the spatial distribution data of each of the aforementioned waterlogging driving factors, construct an original feature matrix, and randomly arrange each column of the original feature matrix to generate shadow variables. Based on each of the shadow variables and the original feature matrix, construct an extended feature matrix. S22. Based on the trained classification model, the extended feature matrix is ​​predicted to obtain feature scores of shadow variables corresponding to each of the waterlogging driving factors. S23. The feature scores corresponding to each of the waterlogging driving factors are weighted and fused to obtain the comprehensive feature score of each of the waterlogging driving factors, and the comprehensive feature scores are sorted.

6. The urban flooding status analysis method according to claim 1, characterized in that, The step of dividing each watershed unit in the target area into flood-prone units and non-flood-prone units includes: The waterlogging density characteristics of each watershed unit in the waterlogging density distribution map are obtained based on the CatBoost model, and each watershed unit is subjected to binary classification based on a preset waterlogging density discrimination threshold.

7. The urban flooding status analysis method according to claim 1, characterized in that, Based on the spatial distribution data of each of the dominant driving factors in the waterlogging unit, the waterlogging unit is recursively segmented to obtain multiple waterlogging subgroups, including: Based on the Cubist regression tree algorithm and the preset segmentation threshold corresponding to each of the dominant driving factors, multiple waterlogging subgroups are obtained. Local regression operations are performed on each of the dominant driving factors corresponding to each of the waterlogging subgroups to obtain the target variable variance ratio of each of the dominant driving factors. The dominant driving factor with the highest target variable variance ratio is taken as the target dominant factor of the waterlogging subgroup.

8. The urban flooding status analysis method according to claim 1, characterized in that, The multi-factor synergistic results include nonlinear weakening, single-factor nonlinear weakening, binary enhancement, independent and nonlinear enhancement.

9. A system for analyzing urban flooding status based on spatial heterogeneity, used to perform the method as described in any one of claims 1-8, characterized in that, include: The dataset construction module is used to preprocess the spatial distribution data of each waterlogging driver in the target area and construct a spatial dataset. The dominant driving factor partitioning module is used to obtain the comprehensive characteristic ranking of each waterlogging driving factor in the spatial dataset, and to construct a dominant driving factor set by taking the top waterlogging driving factors in the comprehensive characteristic ranking as dominant driving factors. The watershed unit division unit is used to obtain the waterlogging density distribution map corresponding to the target area and divide each watershed unit in the target area into waterlogged units and non-waterlogged units. The waterlogging subgroup segmentation unit is used to recursively segment each waterlogging unit based on the spatial distribution data of each dominant driving factor in the waterlogging unit to obtain multiple waterlogging subgroups. The multi-factor synergy result acquisition unit is used to quantify the factor contribution values ​​of each of the waterlogging subgroups based on the geographic detector model, obtain the factor contribution values ​​of the target dominant factors corresponding to each of the waterlogging subgroups, and obtain the multi-factor synergy results among the dominant driving factors based on the factor contribution values ​​of each of the target dominant factors.

10. An electronic device, characterized in that, include: At least one memory for storing computer programs; At least one processor is configured to execute a program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to perform the method as described in any one of claims 1-8.

Images