A drainage construction management system for preventing and controlling urban waterlogging in high-density urban areas

Abstract

The present invention relates to the technical field of drainage construction management, and discloses a drainage construction management system for preventing waterlogging in high-density urban areas, including a catchment area division module for collecting hydrological environment data of a target city and dividing the target city based on the hydrological environment data to obtain a number of catchment units; a data collection module for obtaining the built environment data of the target city and obtaining an internal waterlogging factor set for each catchment unit based on the built environment data; an internal waterlogging statistics module for obtaining the internal waterlogging data set of the target city and obtaining the internal waterlogging density of each catchment unit based on the internal waterlogging data set; and a stationarity correlation module for correlating the internal waterlogging factor set and the internal waterlogging density of each catchment unit to obtain a number of correlation sequences, and calculating the single-line interpretation value of a single internal waterlogging factor in the internal waterlogging factor set of each correlation sequence for the internal waterlogging density and the collinear interpretation value of multiple internal waterlogging factors for the internal waterlogging density.

Description

Technical Field

[0001] The present invention relates to the technical field of drainage construction management, and more specifically, it relates to a drainage construction management system for preventing waterlogging in high-density urban areas. Background Art

[0002] With the acceleration of global climate change and urbanization, the waterlogging disasters in high-density urban areas are becoming increasingly severe, seriously threatening the safety of residents' lives and the sustainable development of cities. Traditional waterlogging prevention and control measures only rely on the expansion construction of drainage pipe networks, without finding the key factors causing waterlogging in the target city through scientific methods, and then carrying out low-cost drainage construction based on the key factors. Therefore, there are problems such as high cost and poor effect in the existing urban waterlogging prevention and control. Summary of the Invention

[0003] The present invention provides a drainage construction management system for preventing waterlogging in high-density urban areas to solve the technical problems raised in the background art.

[0004] The present invention provides a drainage construction management system for preventing waterlogging in high-density urban areas, including:

[0005] A catchment area division module, configured to collect hydrological environment data of the target city, and divide the target city based on the hydrological environment data to obtain a number of catchment units;

[0006] A data collection module, configured to obtain built environment data of the target city, and obtain an internal waterlogging factor set for each catchment unit based on the built environment data;

[0007] An internal waterlogging statistics module, configured to obtain an internal waterlogging data set of the target city, and obtain the internal waterlogging density of each catchment unit based on the internal waterlogging data set;

[0008] A stationarity correlation module, configured to correlate the internal waterlogging factor set and the internal waterlogging density of each catchment unit to obtain a number of correlation sequences, and calculate the single-line interpretation value of a single internal waterlogging factor in the internal waterlogging factor set of each correlation sequence on the internal waterlogging density and the collinear interpretation value of multiple internal waterlogging factors on the internal waterlogging density;

[0009] A geographical correlation module, configured to calculate the geographical action scale of each internal waterlogging factor based on a multi-scale geographically weighted regression model, and weight the corresponding single-line interpretation value and collinear interpretation value by the geographical action scale of the internal waterlogging factor to obtain a first updated interpretation and a second updated interpretation;

[0010] An internal waterlogging management module, configured to determine characteristic internal waterlogging factors based on the first updated interpretation and the second updated interpretation, and perform corresponding management on the drainage construction of the target city based on the characteristic internal waterlogging factors.

[0011] Further, a number of catchment units are obtained, including:

[0012] Obtain a planar image of the target city, rasterize the planar image to obtain a number of grid cells; collect the hydrological environment data of the grid cell in the x-th row and y-th column, where the hydrological environment data includes: elevation data, surface water flow direction, surface roughness, impervious area ratio, stormwater pipe network density, and drainage pipe slope;

[0013] Calculate the fluidity of the grid cell in the x-th row and y-th column based on the hydrological environment data, and the fluidity calculation formula is as follows:

[0014] ;

[0015] Wherein, represents the fluidity of the grid cell in the x-th row and y-th column, represents the normalization coefficient, and respectively represent the first weight parameter and the second weight parameter, represents the elevation data of the grid cell in the x-th row and y-th column, represents the surface water flow direction of the grid cell in the x-th row and y-th column, represents the surface roughness of the grid cell in the x-th row and y-th column, represents the surface roughness influence weight, represents the impervious area ratio of the grid cell in the x-th row and y-th column, represents the stormwater pipe network density of the grid cell in the x-th row and y-th column, represents the drainage pipe slope of the grid cell in the x-th row and y-th column;

[0016] Based on the fluidity of a number of grid cells, obtain a fluidity matrix; process the fluidity matrix through a Sobel operator, extract a number of edge features, and combine the number of edge features to obtain a water catchment network; fit the water catchment network to the planar image, and segment the planar image through the water catchment network to obtain a number of sub-blocks, and each sub-block is used as a catchment unit.

[0017] Further, obtain the waterlogging factor set of each catchment unit, including:

[0018] The built environment data of the target city includes: drainage system data and drainage system related data;

[0019] The drainage system data includes: stormwater pipe network density, drainage pipe slope, stormwater inlet drainage capacity, and rainwater storage capacity;

[0020] The drainage system related data includes: terrain slope of the target city, building density, and green space coverage rate;

[0021] Obtain the drainage system data and the data related to the drainage system of each catchment unit, and take each item of data in the drainage system data and the data related to the drainage system as the waterlogging factors of the catchment unit to obtain the waterlogging factor set of the catchment unit.

[0022] Further, the waterlogging density of the catchment unit includes:

[0023] Obtain the first waterlogging data of the private platform, where the first waterlogging data includes: waterlogging time and waterlogging coordinates;

[0024] Obtain the second waterlogging data of the pan-media platform, specifically including:

[0025] Set preset keywords to screen the text of the pan-media platform to obtain the target text;

[0026] Process the target text based on the NLP natural language model to obtain the second waterlogging data, where the second waterlogging data includes: waterlogging time and waterlogging coordinates;

[0027] Merge the first waterlogging data and the second waterlogging data to obtain the complete waterlogging data;

[0028] Calculate the waterlogging density of each catchment unit based on the waterlogging events and waterlogging areas in the complete waterlogging data. The calculation formula of the waterlogging density is as follows:

[0029] ;

[0030] Where, represents the waterlogging density of the jth catchment unit, represents the quantity of the first waterlogging data, represents the index of represents the waterlogging catchment area of the ath first waterlogging data, represents the influence weight of the ath first waterlogging data, represents the quantity of the second waterlogging data, represents the index of represents the waterlogging coordinates of the bth second waterlogging data, represents the coordinate range of the ith catchment unit, represents the first indicator function. If , then the value is 1, otherwise the value is 0, represents the second indicator function. If , then the value is 1, otherwise the value is 0, represents the confidence coefficient of the pan-media platform, represents the bias constant term of the waterlogging;

[0031] Among them, the confidence coefficient of the pan-media platform is obtained based on the order of magnitude of the number of users of the corresponding pan-media platform, including: obtaining the order of magnitude of the number of users of the pan-media platform, and weighting the preset benchmark confidence coefficient based on the order of magnitude to obtain .

[0032] Furthermore, the single-line interpretation value includes:

[0033] Establish a single-line interpretation framework for each waterlogging factor in the waterlogging factor set. The single-line interpretation framework includes: the discrete classification of the corresponding waterlogging factor, and the waterlogging density of the catchment unit of the corresponding waterlogging factor;

[0034] Among them, the discrete classification of the waterlogging factor includes: generating S clustering centers for the s-th waterlogging factor, clustering the s-th waterlogging factor based on K-means clustering to obtain S discrete classifications of the s-th waterlogging factor, and assigning discrete classification weights to each discrete classification;

[0035] Calculate the single-line interpretation value of the s-th waterlogging factor for the S discrete classifications of the s-th waterlogging factor through an improved geographical factor detector model. The calculation formula of the improved factor detector model is as follows:

[0036] ;

[0037] Among them, represents the single-line interpretation value of the s-th waterlogging factor, represents the S discrete classifications of the s-th waterlogging factor, represents the index of, represents the average waterlogging density of the catchment unit corresponding to the h-th discrete classification of the s-th waterlogging factor, represents the variance of the waterlogging density of the catchment unit corresponding to the h-th discrete classification of the s-th waterlogging factor, represents the average waterlogging density of several catchment units in the target city, represents the variance of the waterlogging density of several catchment units in the target city.

[0038] Furthermore, the collinear interpretation value includes:

[0039] Calculate the collinear interpretation value for any multiple waterlogging factors through an improved geographical interaction detector model. The calculation formula of the improved geographical interaction detector model is as follows:

[0040] ;

[0041] ;

[0042] Among them, Represents the collinear interpretation value of m waterlogging factors, Represents the number of waterlogging factors, Represents The index of, Represents the weighted waterlogging density of the catchment unit corresponding to the j-th waterlogging factor, Represents the variance of the single-line interpretation values of m waterlogging factors; Represents the number of catchment units, Represents The index of, Represents the waterlogging density of the n-th catchment unit, Represents the discrete classification weight of the j-th waterlogging factor corresponding to the n-th catchment unit.

[0043] Furthermore, obtaining the first updated interpretation and the second updated interpretation includes:

[0044] J catchment units in the target city, each catchment unit having a waterlogging density And Z waterlogging factors. For each catchment unit, a regression catchment unit is established to obtain an initial multi-scale geographically weighted regression model. The calculation formula of the initial multi-scale geographically weighted regression model is as follows:

[0045] ;

[0046] Among them, Represents the waterlogging density of the j-th catchment unit, And Respectively represent the first local regression coefficient and the second local regression coefficient, Represents the bias coefficient, Represents the regression catchment unit of the j-th catchment unit , Represents the number of waterlogging factors, Represents The index of, Represents the discrete classification weight of the z-th waterlogging factor of the j-th catchment unit;

[0047] By using spatial weighted least squares method for And The values are estimated, including: obtaining the spatial weighted weights based on the Gaussian kernel function. The calculation formula of the Gaussian kernel function is as follows:

[0048] ;

[0049] Among them, Represents the scale weight of the z-th waterlogging factor at the target regression point For the j-th catchment unit, Represents the bandwidth of the z-th waterlogging factor, represents the Euclidean distance between the j-th catchment unit and the corresponding regressed catchment unit represents the natural exponent

[0050] wherein, the bandwidth of the waterlogging factor z is optimized and obtained based on the AICc locally weighted regression evaluation index, and the AICc locally weighted regression evaluation index is as follows:

[0051] ;

[0052] Based on the AICc locally weighted regression evaluation index, the bandwidth of each waterlogging factor is obtained, and the Z-term waterlogging factors are processed based on the maximum-minimum value normalization to obtain the geographical action scale of each waterlogging factor; the single-line interpretation value and the collinear interpretation value are weighted by the geographical action scale to obtain the first updated interpretation value and the second updated interpretation value.

[0053] Furthermore, the characteristic waterlogging factors include:

[0054] Set the first significance influence threshold, compare the first significance influence threshold with the first updated interpretation value of each waterlogging factor, and take the waterlogging factors with the first updated interpretation value greater than the first significance influence threshold as the dominant significant factors to obtain the dominant significant set;

[0055] Set the second significance influence threshold, traverse all combinations of waterlogging factors, obtain the second updated interpretation value corresponding to each combination, and if the second updated interpretation value is greater than the second significance influence threshold, then take the waterlogging factors in the combination corresponding to the second updated interpretation value as the latent significant factors to obtain the latent significant set;

[0056] Obtain the intersection of the dominant significant set and the latent significant set, and take all the waterlogging factors in the intersection as the characteristic waterlogging factors.

[0057] Furthermore, corresponding management of the drainage construction of the target city is carried out based on the characteristic waterlogging factors, including:

[0058] Based on the expert determination of the correlation between the built environment data of the target city corresponding to the characteristic waterlogging factors and the drainage construction, the correlation includes: positive correlation and negative correlation;

[0059] If the characteristic waterlogging factor is positively correlated, the built environment data of the target city corresponding to the characteristic waterlogging factor is increased through drainage construction;

[0060] If the characteristic waterlogging factor is negatively correlated, the built environment data of the target city corresponding to the characteristic waterlogging factor is reduced through drainage construction.

[0061] The beneficial effects of the present invention are as follows:

[0062] 1. By integrating multi-source data from the government and the pan-media, the present invention realizes the comprehensive collection and accurate assessment of urban waterlogging risks, ensuring the breadth and accuracy of data sources, thereby effectively making up for the deficiencies of traditional single data collection methods.

[0063] 2. The present invention adopts a multi-scale geographically weighted regression model and an improved factor detector to dynamically update the statistical explanatory power and spatial influence range of each waterlogging factor, effectively capturing the spatial heterogeneity within the city and providing a solid technical support for accurately identifying characteristic waterlogging factors.

[0064] 3. By innovatively integrating multi-source data collection, geographic information processing, and statistical models, the refined and quantitative assessment of urban waterlogging risks in high-density urban areas is realized. By means of sub-catchment zoning, waterlogging factor screening, and dynamic update of explanatory values in multi-scale geographically weighted regression, scientific, accurate, and cost-effective decision-making support for drainage construction and waterlogging prevention is provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a module diagram of a drainage construction management system for urban waterlogging prevention in high-density urban areas of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] Now, the subject matter described herein will be discussed with reference to exemplary embodiments. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. Without departing from the scope of protection of the content of this specification, changes can be made to the functions and arrangements of the elements discussed. Each example can omit, substitute, or add various processes or components as needed. Additionally, the features described relative to some examples can also be combined in other examples.

[0067] As Figure 1 shown, a drainage construction management system for urban waterlogging prevention in high-density urban areas includes:

[0068] A sub-catchment zoning module, configured to collect hydrological environment data of a target city and divide the target city based on the hydrological environment data to obtain a number of sub-catchment units;

[0069] A data collection module, configured to obtain built environment data of the target city and obtain a set of waterlogging factors for each sub-catchment unit based on the built environment data;

[0070] A waterlogging statistics module, configured to obtain a waterlogging data set of the target city and obtain the waterlogging density of each sub-catchment unit based on the waterlogging data set;

[0071] A stationarity correlation module, configured to correlate the waterlogging factor set and waterlogging density of each catchment unit to obtain a number of correlation sequences, and calculate the single-line interpretation value of a single waterlogging factor in the waterlogging factor set of each correlation sequence for the waterlogging density and the collinear interpretation value of multiple waterlogging factors for the waterlogging density;

[0072] A geographical correlation module, configured to calculate the geographical action scale of each waterlogging factor based on a multi-scale geographically weighted regression model, and weight the corresponding single-line interpretation value and collinear interpretation value by the geographical action scale of the waterlogging factor to obtain a first updated interpretation and a second updated interpretation;

[0073] A waterlogging management module, configured to determine characteristic waterlogging factors based on the first updated interpretation and the second updated interpretation, and perform corresponding management on the drainage construction of the target city based on the characteristic waterlogging factors.

[0074] In an embodiment of the present invention, obtaining a number of catchment units includes:

[0075] Obtaining a planar image of the target city, rasterizing the planar image to obtain a number of grid units; collecting hydrological environment data of the grid unit in the x-th row and y-th column, where the hydrological environment data includes: elevation data, surface water flow direction, surface roughness, impervious area ratio, rainwater pipe network density, and drainage pipe slope;

[0076] Calculating the circulation degree of the grid unit in the x-th row and y-th column based on the hydrological environment data, and the circulation degree calculation formula is as follows:

[0077] ;

[0078] Wherein, represents the circulation degree of the grid unit in the x-th row and y-th column, represents the normalization coefficient, and respectively represent the first weight parameter and the second weight parameter, represents the elevation data of the grid unit in the x-th row and y-th column, represents the surface water flow direction of the grid unit in the x-th row and y-th column, represents the surface roughness of the grid unit in the x-th row and y-th column, represents the surface roughness influence weight, represents the impervious area ratio of the grid unit in the x-th row and y-th column, represents the rainwater pipe network density of the grid unit in the x-th row and y-th column, represents the drainage pipe slope of the grid unit in the x-th row and y-th column;

[0079] Based on the fluidity of several grid cells, a fluidity matrix is obtained; the fluidity matrix is processed by the Sobel operator to extract several edge features, and a water sink network is combined based on the several edge features; the water sink network is fitted to the planar image, and the planar image is segmented by the water sink network to obtain several sub-blocks, and each sub-block is used as a water sink unit.

[0080] Specifically, first, obtain the planar image of the target city and rasterize it, dividing the entire city into multiple grid cells of a fixed size. Each grid cell serves as an independent spatial sampling unit, and its size is carefully designed to ensure that it can fully reflect local hydrological and topographical features, providing a balanced and accurate spatial basis for subsequent data collection and calculation. Subsequently, collect multiple key hydrological environment data within each grid cell, including elevation, surface water flow direction, surface roughness, impervious area ratio, rainwater pipe network density, and drainage pipe slope. The above data respectively reveal factors such as terrain undulation, water flow direction, surface friction resistance, hardening degree, and drainage system distribution, all of which are important variables affecting the smoothness of water flow. To eliminate the dimensional differences between different data, all collected data are normalized, and by setting the first and second weight parameters, different importance is assigned to key indicators such as elevation and roughness, so as to accurately calculate the fluidity value of each grid cell. Based on the fluidity values of all grid cells, a fluidity matrix is formed, which comprehensively reflects the comprehensive ability of water flow smoothness in each region of the city. Using the Sobel operator to perform edge detection on the fluidity matrix can extract edge features representing water flow convergence or segmentation, and these features are further used to combine and generate a water sink network, and the network is fitted to the original planar image to achieve precise segmentation of the urban space. Finally, each sub-block segmented by the water sink network is defined as an independent water sink unit. This method fully integrates spatial image processing and edge detection technologies, and calculates the fluidity through comprehensive hydrological environment data, realizing a detailed division of the hydrological characteristics and drainage conditions within the city, providing accurate and reliable spatial units for subsequent waterlogging risk assessment and drainage construction management.

[0081] In an embodiment of the present invention, an urban waterlogging factor set for each water sink unit is obtained, including:

[0082] The built environment data of the target city includes: drainage system data and drainage system-related data;

[0083] The drainage system data includes: rainwater pipe network density, drainage pipe slope, rainwater inlet drainage capacity, and rainwater storage capacity;

[0084] The drainage system-related data includes: terrain slope of the target city, building density, and green space coverage rate;

[0085] Obtain the drainage system data and the data related to the drainage system for each catchment unit, and use each item of data in the drainage system data and the data related to the drainage system as the waterlogging factors of the catchment unit to obtain the waterlogging factor set of the catchment unit.

[0086] Specifically, in order to fully reveal the internal mechanism of the formation of urban waterlogging risk in the target city, this claim discloses a detailed method for obtaining the waterlogging factor set of each catchment unit from the built environment data. Specifically, the built environment data of the target city is divided into two major categories: one is the drainage system data, and the other is the data related to the drainage system. First, the drainage system data mainly reflects the composition and performance parameters of the drainage infrastructure itself, including the density of rainwater pipe networks, the slope of drainage pipes, the drainage capacity of rainwater inlets, and the rainwater storage capacity. These parameters directly determine the drainage efficiency and operation stability of the city under rainstorm conditions. Second, the data related to the drainage system covers the external environmental factors that affect the operation of the drainage system, such as the terrain slope, building density, and green space coverage rate. The terrain slope affects the flow direction and confluence of rainwater, the building density usually increases with the increase of impervious area, while the green space coverage rate is conducive to the natural infiltration and absorption of rainwater. By collecting the above two types of data in each catchment unit and regarding each item of data as an independent waterlogging factor, the present invention constructs a multi-dimensional waterlogging factor set. This factor set not only comprehensively reflects the actual situation of the drainage system and its operating environment, but also provides a detailed data basis for subsequent statistical analysis of waterlogging risk, stationarity correlation, and optimization of drainage construction, so as to realize the refined assessment of urban waterlogging risk.

[0087] In an embodiment of the present invention, the waterlogging density of the catchment unit includes:

[0088] Obtain the first waterlogging data of the private platform, and the first waterlogging data includes: waterlogging time and waterlogging coordinates;

[0089] Obtain the second waterlogging data of the pan-media platform, specifically including:

[0090] Set preset keywords to screen the text of the pan-media platform to obtain the target text;

[0091] Process the target text based on the NLP natural language model to obtain the second waterlogging data, and the second waterlogging data includes: waterlogging time and waterlogging coordinates;

[0092] Merge the first waterlogging data and the second waterlogging data to obtain the complete waterlogging data;

[0093] Calculate the waterlogging density of each catchment unit based on the waterlogging events and waterlogging areas in the complete waterlogging data, and the calculation formula of the waterlogging density is as follows:

[0094] ;

[0095] Among them, represents the waterlogging density of the j-th catchment unit, represents the quantity of the first waterlogging data, represents the index of represents the waterlogging catchment area of the a-th first waterlogging data, represents the influence weight of the a-th first waterlogging data, represents the quantity of the second waterlogging data, represents the index of represents the waterlogging coordinates of the b-th second waterlogging data, represents the coordinate range of the i-th catchment unit, represents the first indicator function. If , then the value is 1, otherwise the value is 0. represents the second indicator function. If , then the value is 1, otherwise the value is 0. represents the confidence coefficient of the pan-media platform, represents the bias constant term of waterlogging;

[0096] Among them, the confidence coefficient of the pan-media platform is obtained by taking values based on the order of magnitude of the number of users of the corresponding pan-media platform, including: obtaining the order of magnitude of the number of users of the pan-media platform, and weighting the preset benchmark confidence coefficient based on the order of magnitude to obtain .

[0097] It should be noted that in order to avoid the situation where the second waterlogging data of the pan-media platform cannot be obtained or the second waterlogging data is empty, therefore, the bias constant term of waterlogging is set, and the bias constant term of waterlogging is a parameter based on artificial customization.

[0098] Specifically, the waterlogging information is obtained from two data sources, and these information are merged and used to accurately quantify the waterlogging risk of each catchment unit. First, the present invention collects the first waterlogging data from a private platform. These data include the occurrence time and geographical coordinates of waterlogging events, usually from the government or professional monitoring institutions, and have high data credibility. At the same time, the second waterlogging data is also obtained from the pan-media platform. When collecting the second waterlogging data, the text of the pan-media platform is screened through preset keywords to obtain the target text, and then a natural language processing (NLP) model is used to parse the target text to extract the time and coordinate information of the waterlogging event. These two parts of data are strictly merged to eliminate duplicates and noise, forming a complete waterlogging data set. Next, for each catchment unit, the waterlogging density is calculated based on this complete data set. The specific process is as follows: First, for each event in the first waterlogging data, statistics are made according to its waterlogging catchment area and the predefined influence weight. At the same time, for each event in the second waterlogging data, the indicator function is used to judge whether it falls within the coordinate range of the current catchment unit, and the corresponding weight is assigned in combination with the confidence coefficient of the pan-media platform (this confidence coefficient is obtained by weighting the preset reference value according to the order of magnitude of the platform user volume). Finally, the weighted statistical results of the two types of data are added together to form the waterlogging density of the catchment unit.

[0099] In an embodiment of the present invention, the single-line interpretation value includes:

[0100] A single-line interpretation framework is established for each waterlogging factor in the waterlogging factor set. The single-line interpretation framework includes: the discrete classification of the corresponding waterlogging factor, and the waterlogging density of the catchment unit of the corresponding waterlogging factor;

[0101] Among them, the discrete classification of the waterlogging factor includes: generating S clustering centers for the s-th waterlogging factor, clustering the s-th waterlogging factor based on K clustering to obtain S discrete classifications of the s-th waterlogging factor, and assigning discrete classification weights to each discrete classification respectively;

[0102] The single-line interpretation value of the s-th waterlogging factor is calculated for the S discrete classifications of the s-th waterlogging factor through an improved geographical factor detector model. The calculation formula of the improved factor detector model is as follows:

[0103] ;

[0104] Among them, represents the single-line interpretation value of the s-th waterlogging factor, represents the S discrete classifications of the s-th waterlogging factor, represents the index of, represents the average waterlogging density of the catchment unit corresponding to the h-th discrete classification of the s-th waterlogging factor, Denotes the variance of the waterlogging density of the catchment unit corresponding to the h-th discrete classification of the s-th waterlogging factor. Denotes the average waterlogging density of several catchment units in the target city. Denotes the variance of the waterlogging density of several catchment units in the target city.

[0105] Specifically, the waterlogging factors are discretely classified, and the single-line interpretation value is calculated using an improved geographical factor detector model to quantify the explanatory ability of each waterlogging factor on the waterlogging density of each catchment unit. First, to reflect the influence characteristics of each waterlogging factor in different regions, in this embodiment, each waterlogging factor is discretized using a clustering method. Specifically, for the s-th waterlogging factor, S clustering centers are generated by loading, and the factor data is divided into S discrete categories based on the K-clustering algorithm. Each category not only represents the performance of the factor under a certain degree of influence but is also assigned corresponding discrete classification weights, which reflect the concentration and volatility of the waterlogging density of the catchment units within each category. Next, for each discrete classification, the present invention uses an improved geographical factor detector model to calculate the single-line interpretation value of the s-th waterlogging factor. The calculation formula of this model includes: the average waterlogging density of the catchment units corresponding to this discrete category, which is used to reflect the waterlogging risk level under this category; the variance of the waterlogging density of the catchment units within this discrete category, which reflects the fluctuation of the waterlogging density within this category; the average value and variance of the waterlogging density of the catchment units throughout the target city are used as the global reference. By combining the statistical indicators of each discrete category and performing weighted summation, this model can obtain the single-line interpretation value, which reflects the explanatory power of the s-th waterlogging factor on the waterlogging density when acting alone. In other words, the single-line interpretation value not only reflects the local statistical characteristics of the factor data but also fully considers the unevenness of the spatial distribution, thus providing a solid basis for the subsequent quantitative analysis of the waterlogging risk.

[0106] It should be noted that the factor detector is a statistical tool for quantifying the explanatory power of a single waterlogging factor on the spatial distribution of the target variable. Its basic idea is to compare the ratio between the local variance of the waterlogging density of each catchment unit and the overall variance of the waterlogging density of the entire city under the factor division, so as to obtain an interpretation value. The value range of the interpretation value is usually from 0 to 1. The higher the value, the greater the spatial distribution difference that the waterlogging factor can explain, and thus the stronger the individual influence on the occurrence of waterlogging. The factor detector provides a quantitative basis for the subsequent screening and key attention of waterlogging factors.

[0107] In an embodiment of the present invention, the collinear interpretation value includes:

[0108] The collinear interpretation value is calculated for any multiple waterlogging factors using an improved geographical interaction detector model. The calculation formula of the improved geographical interaction detector model is as follows:

[0109] ;

[0110] ;

[0111] Among them, represents the collinear interpretation value of m waterlogging factors, represents the number of waterlogging factors, represents the index of represents the weighted waterlogging density of the catchment unit corresponding to the j-th waterlogging factor, represents the variance of the single-line interpretation values of m waterlogging factors; represents the number of catchment units, represents the index of represents the waterlogging density of the n-th catchment unit, represents the discrete classification weight of the j-th waterlogging factor corresponding to the n-th catchment unit.

[0112] Specifically, the collinear interpretation values of multiple waterlogging factors are calculated by an improved geographical interaction detector model to quantify the joint interpretation ability of multiple waterlogging factor combinations on the waterlogging density of catchment units. First, for any m waterlogging factors, the present invention adopts an improved geographical interaction detector model, and the core of this model lies in comparing the weighted waterlogging density of each waterlogging factor in each catchment unit with the statistical fluctuations of the single-line interpretation values of these factors.

[0113] Specifically, the model realizes the calculation of the collinear interpretation value through the following steps: For each waterlogging factor, corresponding classification weights have been assigned in each catchment unit according to the discrete classification method. These weights reflect the performance of waterlogging density at different discrete levels and also provide a basis for subsequent multi-factor combination analysis. For the j-th waterlogging factor, statistics are carried out in each catchment unit according to its weighted waterlogging density, and the overall weighted mean and the degree of fluctuation (i.e., the variance of the single-line interpretation value) are calculated, which are used as a measure of the imbalance of the influence of a single factor on waterlogging density.

[0114] The weighted waterlogging density of the above-mentioned multiple factors is combined. In the formula, the following are comprehensively considered: the weighted waterlogging density corresponding to m waterlogging factors together, which reflects the combined effect of the factor combination on the waterlogging density in each catchment unit; at the same time, the statistical variance of the single-line interpretation values of the m waterlogging factors is taken into account as an index to measure the synergistic effect between factors; combined with the number of each catchment unit and its corresponding discrete classification weight, a quantitative value reflecting the overall combined interpretation ability is obtained, that is, the collinear interpretation value. The improved geographical interaction detector model can not only reveal the local influence of a single waterlogging factor but also reflect the interaction and non-linear synergistic effect between multiple factors, thus more accurately describing the influence of the multi-factor combination on the waterlogging density.

[0115] It should be noted that the interaction detector is an extension based on the factor detector method and is used to evaluate the combined effect of multiple waterlogging factor combinations in explaining the spatial distribution of the target variable. It judges whether there is a synergistic or inhibitory effect between these factors by comparing the fluctuation of the weighted waterlogging density under the combined action of multiple factors with the fluctuation when each factor acts alone. If the explanatory power of the factor combination is significantly higher than the simple superposition of the explanatory powers of each individual factor, it indicates a positive interaction or non-linear enhancement effect; otherwise, there may be a negative or complementary effect. The results of the interaction detector help to identify which factor combinations have a decisive impact on the waterlogging risk in a complex urban environment, providing a more comprehensive basis for system optimization and decision-making.

[0116] In an embodiment of the present invention, obtaining the first updated interpretation and the second updated interpretation includes:

[0117] There are J catchment units in the target city, and each catchment unit has a waterlogging density and Z waterlogging factors. For each catchment unit, a regression catchment unit is established to obtain an initial multi-scale geographically weighted regression model. The calculation formula of the initial multi-scale geographically weighted regression model is as follows:

[0118] ;

[0119] Among them, represents the waterlogging density of the jth catchment unit, and respectively represent the first local regression coefficient and the second local regression coefficient, represents the bias coefficient, represents the regression catchment unit of the jth catchment unit , represents the number of waterlogging factors, represents the index of, represents the discrete classification weight of the zth waterlogging factor of the jth catchment unit;

[0120] Estimate the values of and by spatial weighted least squares method, including: obtaining the spatial weighted weights based on the Gaussian kernel function, and the calculation formula of the Gaussian kernel function is as follows:

[0121] ;

[0122] Among them, represents the scale weight of the z-th waterlogging factor at the target regression point for the j-th catchment unit, represents the bandwidth of the z-th waterlogging factor, represents the Euclidean distance between the j-th catchment unit and the corresponding regression catchment unit, represents the natural exponent;

[0123] Among them, the bandwidth of the waterlogging factor z is optimized and obtained based on the AICc locally weighted regression evaluation index, and the AICc locally weighted regression evaluation index is as follows:

[0124] ;

[0125] Obtain the bandwidth of each waterlogging factor based on the AICc locally weighted regression evaluation index, and process the z waterlogging factors based on the maximum-minimum value normalization to obtain the geographical action scale of each waterlogging factor; weight the single-line interpretation value and the collinear interpretation value through the geographical action scale to obtain the first updated interpretation value and the second updated interpretation value.

[0126] Specifically, by performing local regression analysis within each catchment unit and adopting the data processing method of weighted within the region, the propagation range of each waterlogging factor in space can be finely measured. By using strict locally weighted regression technology and parameter optimization strategies, the statistical performance of each factor in different regions can be matched with its actual spatial influence range, realizing the accurate extraction of the geographical action scale of the waterlogging factor. On this basis, the present invention further normalizes the obtained spatial influence scale and uses it as a weighting factor to fuse and correct the single-item and combined waterlogging factor interpretation values obtained in traditional statistical analysis. The updated interpretation value in this way not only retains the essential connotation of the factor's statistical interpretation of waterlogging density, but also fully reflects its spatial inhomogeneity and diffusion characteristics, thus providing a more scientific and targeted decision-making basis for waterlogging risk assessment and drainage construction management.

[0127] It should be noted that the multi-scale geographically weighted regression (MGWR) model is an advanced spatial regression method. Based on local weighted regression, it allows each explanatory variable to have its own independent spatial range of action, that is, its own bandwidth parameter. This feature enables the model to more precisely reveal the differential impacts of different waterlogging factors in various urban regions, thus providing a more reliable basis for the spatial quantitative analysis of waterlogging risks. In the implementation process, the MGWR model conducts local regression analysis on each catchment unit, and uses spatial weighting techniques to assign higher weights to data that are closer, thereby obtaining local regression coefficients. Different from traditional geographically weighted regression, MGWR determines the optimal bandwidth for each waterlogging factor separately. This process usually evaluates a series of candidate bandwidths using information criteria (such as the modified AICc), and selects the parameter that makes the local model have the best fitting effect and reasonable complexity. The bandwidth obtained in this way not only reflects the statistical relationship of each waterlogging factor, but also intuitively characterizes its actual diffusion range and scale of action in space. Through the normalization process of the spatial action scale of each factor, the model further combines this information with the original statistical explanatory values to form updated explanatory indicators, thus achieving a comprehensive evaluation that takes into account both statistical explanatory power and spatial influence range.

[0128] It should be noted that is an improved information criterion used to evaluate the balance between the goodness of fit and complexity of the model in the local weighted regression model. Its core idea is to correct the traditional Akaike information criterion (AIC) to adapt to the possible biases in the case of a small sample size or local regression environment. By introducing a penalty term in the local model, this evaluation index can effectively avoid overfitting, thus helping to determine the optimal bandwidth parameter.

[0129] In an embodiment of the present invention, the characteristic waterlogging factors include:

[0130] Set a first significance impact threshold, compare the first significance impact threshold with the first updated explanatory value of each waterlogging factor, and take the waterlogging factors with the first updated explanatory value greater than the first significance impact threshold as dominant significant factors to obtain a dominant significant set;

[0131] Set a second significance impact threshold, traverse all combinations of waterlogging factors, obtain the second updated explanatory value corresponding to each combination, and if the second updated explanatory value is greater than the second significance impact threshold, then take the waterlogging factors in the combination corresponding to the second updated explanatory value as latent significant factors to obtain a latent significant set;

[0132] Obtain the intersection of the dominant significant set and the latent significant set, and take all the waterlogging factors in the intersection as characteristic waterlogging factors.

[0133] Specifically, for the strategy of identifying and determining characteristic waterlogging factors, a double-threshold screening method is adopted to ensure full consideration in both statistical interpretability and factor interaction. Specifically, first, a first significance influence threshold is defined to compare the first updated interpretation values of each waterlogging factor; any factor whose first updated interpretation value exceeds this threshold is considered to have a significant influence at the single-action level and is included in the dominant significant set. At the same time, a second significance influence threshold is also set, and by traversing and screening the second updated interpretation values calculated for each combination of waterlogging factors, any factor whose value exceeds this threshold is considered to have potential significance in the factor combination effect and thus constitutes a latent significant set. Finally, the intersection of the above two sets is taken to obtain the final characteristic waterlogging factors.

[0134] In an embodiment of the present invention, corresponding management of the drainage construction of the target city is performed based on the characteristic waterlogging factors, including:

[0135] Based on expert judgment of the correlation between the built environment data of the target city corresponding to the characteristic waterlogging factors and the drainage construction, the correlation includes: positive correlation and negative correlation;

[0136] If the characteristic waterlogging factor is positively correlated, the built environment data of the target city corresponding to the characteristic waterlogging factor is increased through drainage construction;

[0137] If the characteristic waterlogging factor is negatively correlated, the built environment data of the target city corresponding to the characteristic waterlogging factor is reduced through drainage construction.

[0138] Specifically, through expert judgment and data association, the positive and negative correlations between the built environment data of the target city and the drainage construction measures are clarified, so as to achieve targeted adjustment. Specifically, this solution first relies on the characteristic waterlogging factors obtained from the foregoing module, and combines the comprehensive evaluation of the expert on the urban drainage system and its operating conditions to determine the correlation between each characteristic factor and the drainage construction parameters: if it is found through evaluation that a certain characteristic factor is positively correlated with the drainage construction index, that is, an increase in the value of this factor tends to lead to an increase in the waterlogging risk, then the management strategy requires corresponding increase of this index in the drainage construction (for example, increasing the pipe network density) to enhance the drainage capacity; conversely, if a certain characteristic factor is negatively correlated with the drainage construction index, the waterlogging risk is alleviated by reducing this index (such as reducing the impervious area). In this way, through expert judgment and data feedback, the drainage construction management module can dynamically adjust the urban built environment according to the specific action direction of the characteristic waterlogging factors, so as to achieve precise and targeted waterlogging prevention and control.

[0139] The above has described the embodiments of this example, but this example is not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative rather than restrictive. Under the inspiration of this example, those of ordinary skill in the art can also make many forms, all of which fall within the protection scope of this example.

Claims

1. A drainage construction management system for waterlogging prevention and control in high-density urban areas, characterized in that: include: The water catchment zoning module is used to collect the hydrological environment data of the target city and divide the target city based on the hydrological environment data to obtain several water catchment units; The data collection module is used to obtain the built environment data of the target city and obtain the waterlogging factor set of each catchment unit based on the built environment data; The waterlogging statistics module is used to obtain the waterlogging dataset of the target city and obtain the waterlogging density of each catchment unit based on the waterlogging dataset; The stationary association module is used to associate the waterlogging factor set of each watershed unit with the waterlogging density to obtain several association sequences, and calculate the single-line explanation value of the single waterlogging factor in the waterlogging factor set of each association sequence to the waterlogging density and the collinear explanation value of the multiple waterlogging factors to the waterlogging density; A geographical association module is used to calculate the geographical effect scale of each waterlogging factor based on a multi-scale geographical weighted regression model, and weight the corresponding single-line explanation value and collinear explanation value of the geographical effect scale of the waterlogging factor to obtain a first updated explanation value and a second updated explanation value; The waterlogging management module is used to determine the characteristic waterlogging factor based on the first updated interpretation value and the second updated interpretation value, and to perform corresponding management on the drainage construction of the target city based on the characteristic waterlogging factor.

2. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 1, characterized in that: Several water catchment units are obtained, including: Obtain a plane image of the target city and rasterize the plane image to obtain a number of grid cells; collect the hydrological environment data of the grid cell in the xth row and yth column, the hydrological environment data including: elevation data, surface water flow direction, surface roughness, impervious area ratio, rainwater pipe network density and drainage pipe slope; The flow rate of the grid cell in the xth row and yth column is calculated based on the hydrological environment data. The flow rate calculation formula is as follows: ； in, represents the flow rate of the grid cell at row x and column y, represents the normalization coefficient, and denote the first weight parameter and the second weight parameter respectively, Represents the elevation data of the grid cell at row x and column y. represents the surface water flow direction of the grid cell in row x and column y, represents the surface roughness of the grid cell in row x and column y, Indicates the influence weight of surface roughness, represents the impervious area ratio of the grid cell in row x and column y, represents the stormwater pipe network density of the grid cell in row x and column y, represents the slope of the drainage pipe of the grid cell at row x and column y; Based on the circulation of several grid units, a circulation matrix is ​​obtained; the circulation matrix is ​​processed by the Sobel operator to extract several edge features, and a water sink network is obtained based on the combination of several edge features; the water sink network is fitted to the plane image, and the plane image is segmented by the water sink network to obtain several sub-blocks, each of which is regarded as a water sink unit.

3. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 2, characterized in that: The waterlogging factor set of each catchment unit is obtained, including: Built environment data of target cities include: drainage system data and drainage system-related data; Drainage system data include: stormwater pipe network density, drainage pipe slope, stormwater inlet drainage capacity and stormwater storage capacity; Drainage system-related data include: terrain slope, building density and green space coverage of the target city; The drainage system data and drainage system related data of each watershed unit are obtained, and each data in the drainage system data and drainage system related data is used as the waterlogging factor of the watershed unit to obtain the waterlogging factor set of the watershed unit.

4. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 3, characterized in that: Obtain the waterlogging density of each catchment unit, including: Obtain the first flood data from the private platform, the first flood data including: flood time and flood coordinates; Obtain the second flood data of the pan-media platform, including: Set preset keywords to filter the texts on the pan-media platform and obtain the target texts; The target text is processed based on the NLP natural language model to obtain the second waterlogging data, which includes: waterlogging time and waterlogging coordinates; The first waterlogging data and the second waterlogging data are merged to obtain complete waterlogging data; The waterlogging density of each catchment unit is calculated based on the waterlogging events and waterlogging areas in the complete waterlogging data. The calculation formula of waterlogging density is as follows: ； in, represents the waterlogging density of the j-th catchment unit, Indicates the number of first flood data, express The index of represents the waterlogging catchment area of ​​the ath first waterlogging data, represents the impact weight of the a-th first flood data, Indicates the number of the second flood data, express The index of represents the waterlogging coordinates of the bth second waterlogging data, represents the coordinate range of the ith catchment unit, represents the first indicator function, if , then the value is 1, otherwise the value is 0. represents the second indicator function, if , then the value is 1, otherwise the value is 0. represents the confidence coefficient of the pan-media platform, represents the bias constant term for waterlogging; The confidence coefficient of the pan-media platform is determined based on the order of magnitude of the number of users of the corresponding pan-media platform, including: obtaining the order of magnitude of the number of users of the pan-media platform, weighting the preset benchmark confidence coefficient based on the order of magnitude, and obtaining .

5. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 4, characterized in that: Computes single-line interpretation values, including: A single-line interpretation framework is established for each waterlogging factor in the waterlogging factor set. The single-line interpretation framework includes: discrete classification of the corresponding waterlogging factor and waterlogging density of the catchment unit of the corresponding waterlogging factor; The discrete classification of waterlogging factors includes: generating S cluster centers by loading the s-th waterlogging factor, clustering the s-th waterlogging factor based on K clustering, obtaining S discrete classifications of the s-th waterlogging factor, and assigning discrete classification weights to each discrete classification; The single-line explanation value of the s-th waterlogging factor is calculated for the S discrete classifications of the s-th waterlogging factor through the improved geographic factor detector model. The calculation formula of the improved factor detector model is as follows: ； in, represents the single-line interpretation value of the sth waterlogging factor, represents the S discrete categories of the s-th waterlogging factor, express The index of represents the average waterlogging density of the catchment unit corresponding to the h-th discrete classification of the s-th waterlogging factor, represents the variance of waterlogging density of the catchment unit corresponding to the h-th discrete classification of the s-th waterlogging factor, represents the average waterlogging density of several catchment units in the target city, Represents the variance of waterlogging density of several catchment units in the target city.

6. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 5, characterized in that: Compute collinearity explanation values, including: The collinearity explanation value is calculated for any number of waterlogging factors through the improved geographic interactive detector model. The calculation formula of the improved geographic interactive detector model is as follows: ； ； in, represents the collinearity explanation value of m waterlogging factors, represents the number of waterlogging factors, express The index of represents the weighted waterlogging density of the catchment unit corresponding to the jth waterlogging factor, It represents the variance of the single-line explanation value of m waterlogging factors; represents the number of water catchment units, express The index of represents the waterlogging density of the nth catchment unit, Represents the discrete classification weight of the jth waterlogging factor corresponding to the nth catchment unit.

7. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 6, characterized in that: Obtaining a first updated interpretation value and a second updated interpretation value, including: There are J catchment units in the target city, each of which has a waterlogging density as well as The waterlogging factor is taken into account. A regression model is established for each watershed unit to obtain the initial multi-scale geographically weighted regression model. The calculation formula of the initial multi-scale geographically weighted regression model is as follows: ； in, represents the waterlogging density of the j-th catchment unit, and denote the first local regression coefficient and the second local regression coefficient, respectively. represents the bias coefficient, The watershed unit representing the regression of the j-th watershed unit , represents the number of waterlogging factors, express The index of represents the first Discrete classification weights of waterlogging factors; By using spatial weighted least squares and The value is estimated, including: obtaining the spatial weighted weight based on the Gaussian kernel function, and the calculation formula of the Gaussian kernel function is as follows: ； in, Indicates The waterlogging factor is at the target regression point The scale weight of the j-th catchment unit is: Indicates The bandwidth of the waterlogging factor, represents the Euclidean distance between the jth catchment unit and the corresponding regressed catchment unit, represents the natural index; Among them, the bandwidth of the waterlogging factor The optimization is performed based on the AICc local weighted regression evaluation index, and the AICc local weighted regression evaluation index is as follows: ； The bandwidth of each waterlogging factor was obtained based on the AICc local weighted regression evaluation index, and the maximum and minimum values ​​were normalized. The waterlogging factors are processed to obtain the geographical effect scale of each waterlogging factor; the single-line explanation value and the collinear explanation value are weighted by the geographical effect scale to obtain the first updated explanation value and the second updated explanation value.

8. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 7, characterized in that: Determine characteristic waterlogging factors, including: The first significant impact threshold is set, and the first significant impact threshold is compared with the first updated explanation value of each waterlogging factor. The waterlogging factor whose first updated explanation value is greater than the first significant impact threshold is taken as a dominant significant factor to obtain a dominant significant set. Set the second significant impact threshold, traverse all combinations of waterlogging factors, and obtain the second updated explanation value corresponding to each combination. If the second updated explanation value is greater than the second significant impact threshold, take the waterlogging factor in the combination corresponding to the second updated explanation value as a potential significant factor to obtain a potential significant set. The intersection of the dominant significant set and the latent significant set is obtained, and the waterlogging factors in the intersection are taken as characteristic waterlogging factors.

9. A drainage construction management system for waterlogging prevention and control in high-density urban areas according to claim 8, characterized in that: Based on the characteristic waterlogging factors, the drainage construction of the target city is managed accordingly, including: The correlation between the built environment data and drainage construction of the target city corresponding to the characteristic waterlogging factors determined by experts, including positive correlation and negative correlation; If the characteristic waterlogging factor is positively correlated, the built environment data of the target city corresponding to the characteristic waterlogging factor will be increased through drainage construction; If the characteristic waterlogging factor is negatively correlated, the built environment data of the target city corresponding to the characteristic waterlogging factor will be reduced through drainage construction.

Images