Green land optimization method based on urban inland inundation risk assessment model and related device

Abstract

The embodiment of the invention relates to the field of natural disaster risk prediction, and discloses a green land optimization method based on an urban inland inundation risk assessment model, parameters in a maximum entropy model are optimized, and the maximum entropy model after parameter optimization is used as the urban inland inundation risk assessment model, so that model parameters suitable for a city can be optimized; compared with a default parameter method adopted in the prior art, the urban inland inundation risk assessment method has the advantage that the assessment precision can be effectively improved. Meanwhile, according to the method, the obtained urban environment variable test data is pre-evaluated to obtain the urban environment variable evaluation data, so that the operation precision and evaluation accuracy of the urban waterlogging risk model can be improved. In addition, according to the urban inland inundation risk probability distribution diagram and the urban supplementary greenbelt rate output by the urban inland inundation risk assessment model, the greenbelt optimization scheme is determined, and the urban inland inundation risk probability can be effectively reduced.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of natural disaster risk prediction, and in particular to a green space optimization method and related devices based on an urban waterlogging risk assessment model. Background Art

[0002] Global climate change has led to an increase in the frequency of high temperatures, heavy rainfall and urban flooding. With the acceleration of urbanization, population growth and aging of urban drainage systems, the impact of climate change on urban flooding risks has become more significant.

[0003] Urban green spaces, including natural and artificial vegetation, can play an important role in alleviating the risk of urban flooding. The vegetation canopy of green spaces effectively intercepts rainwater, and the soil promotes water infiltration, thereby reducing surface runoff during peak rainfall periods and playing a rain and flood storage function similar to a "sponge". However, the current configuration of urban green spaces is mainly based on landscape beautification, lacking a comprehensive analysis of flood disasters and a comprehensive consideration of the spatial layout of green spaces, which limits the role of flood prevention.

[0004] In recent years, the application of machine learning algorithms in waterlogging research has gradually increased, especially artificial neural networks, random forests, decision trees, and support vector machines, which have been widely used in disaster assessment and prediction. However, the existing waterlogging risk assessment methods based on machine learning have problems such as insufficient data screening and insufficient model parameter adjustment, resulting in low assessment and prediction accuracy. Summary of the invention

[0005] The embodiments of the present application mainly provide a green space optimization method and related devices based on an urban waterlogging risk assessment model, which can optimize the urban green space configuration according to the urban waterlogging risk assessment model to reduce the risk of urban waterlogging.

[0006] To solve the above technical problems, the embodiments of the present application adopt the following technical solutions:

[0007] In a first aspect, the present application provides a green space optimization method based on an urban waterlogging risk assessment model, including:

[0008] Obtaining urban environmental variable test data, which includes at least two of urban topographic data, urban meteorological data, urban underlying surface data, and urban drainage network data;

[0009] Pre-evaluating the urban environmental variable test data and obtaining the urban environmental variable evaluation data, wherein the pre-evaluating the urban environmental variable test data includes using a maximum entropy model to evaluate the permutation importance of the urban environmental variable test data and obtaining the urban environmental variable evaluation data based on the permutation importance;

[0010] Obtain the historical urban waterlogging point data, and obtain urban waterlogging risk assessment data based on the historical urban waterlogging point data and urban environmental variable assessment data;

[0011] According to the urban waterlogging risk assessment data, the maximum entropy model parameters are optimized to obtain the urban waterlogging risk assessment model;

[0012] According to the urban waterlogging risk assessment model, output the urban waterlogging risk probability distribution map;

[0013] Obtain the current data of urban land use and the urban target green space ratio. The urban target green space ratio refers to the proportion of green space that should be achieved in different land use types;

[0014] Obtain the urban supplementary green space rate based on the urban surface cover type data, urban land use status data and urban target green space rate in the urban underlying surface data;

[0015] Determine the green space optimization plan based on the urban waterlogging risk probability distribution map and the urban supplementary green space rate;

[0016] According to the green space optimization plan and urban waterlogging risk assessment model, the waterlogging risk reduction effect after green space optimization is evaluated.

[0017] In some embodiments, the maximum entropy model is optimized according to the urban waterlogging assessment data to obtain an urban waterlogging risk assessment model, including:

[0018] Configure model parameters for the maximum entropy model, including feature type, control coefficient, and regional classification method;

[0019] Combining two or more categories of model parameters to obtain multiple groups of candidate models;

[0020] Perform performance evaluation on multiple groups of candidate models through performance evaluation tools to obtain evaluation results;

[0021] Based on the assessment results, the urban waterlogging risk assessment model is determined.

[0022] In some embodiments, the feature type includes one or more of a linear feature, a linear feature and a quadratic feature, a hinge feature, a linear feature and a quadratic feature, and a hinge feature;

[0023] The control coefficient ranges from 0.5 to 4, and the interval ranges from 0.1 to 0.5; the regional classification methods include longitude and longitude, latitude and latitude, longitude and latitude, and one or more of latitude and longitude.

[0024] In some embodiments, the urban supplementary green space rate is obtained through urban surface cover type data, urban land use status data and urban target green space rate in the urban underlying surface data, including:

[0025] The current green space ratio of different land use types is calculated through the current urban land use data and urban land cover type data;

[0026] The city’s supplementary greening rate is determined based on the city’s target greening rate and the current greening rates of different land use types.

[0027] In some embodiments, determining a green space optimization plan based on the urban waterlogging risk probability distribution map and the urban supplementary green space rate includes:

[0028] The hard ground of different land use types is obtained through the urban waterlogging risk probability distribution map and different land use types in the urban land use status data;

[0029] According to the highest risk areas with the same percentage of hard ground and urban supplementary green space rate of different land use types, the green space optimization priority configuration areas are obtained;

[0030] The green space optimization plan is determined by optimizing the green space priority allocation areas.

[0031] In some embodiments, based on the green space optimization scheme and the urban waterlogging risk assessment model, the waterlogging risk reduction effect after the green space optimization is evaluated, including:

[0032] Obtain urban underlying surface update data based on green space optimization plan;

[0033] The urban underlying surface update data replaces the urban underlying surface data in the urban waterlogging risk assessment data to obtain the urban waterlogging risk prediction data;

[0034] Input the urban waterlogging risk prediction data into the urban waterlogging risk assessment model, and output the urban waterlogging risk probability prediction map after green space optimization;

[0035] The waterlogging risk reduction effect after green space optimization is evaluated based on the urban waterlogging risk probability prediction map after green space optimization.

[0036] In some embodiments, obtaining the urban environmental variable test data includes obtaining two or more of topographic data, urban meteorological data, urban underlying surface data, and urban pipe network data, including:

[0037] The topographic data include one or more data of the city's elevation, slope, and surface roughness;

[0038] The urban meteorological data include one or more of the following data: annual average temperature, maximum temperature of the hottest month, average temperature of the wettest quarter, average temperature of the hottest quarter, annual precipitation, rainfall of the wettest month, rainfall of the wettest quarter, rainfall of the hottest quarter, rainfall of the wettest month, accumulation of rainstorms and accumulation of very heavy rainstorms;

[0039] Urban underlying surface data include one or more of the following data: urban land cover type, normalized vegetation index, normalized building index, and normalized water index;

[0040] Urban drainage network data includes one or more of urban drainage zoning data, pipeline type data, network outlet latitude and longitude information, and network flow data.

[0041] In some embodiments, pre-evaluating the urban environmental variable test data to obtain urban environmental variable evaluation data includes:

[0042] Calculate the Pearson correlation coefficient between the test data of urban environmental variables;

[0043] Maximum entropy model was used to assess the permutation importance of the test data for each urban environmental variable;

[0044] If the absolute value of the Pearson correlation coefficient between two urban environmental variable test data is greater than 0.7, the urban environmental variable test data with smaller replacement importance among the two urban environmental variable test data are removed, and the remaining urban environmental variable test data are combined as urban environmental variable evaluation data.

[0045] In a second aspect, some embodiments of the present application provide an electronic device, including:

[0046] at least one processor, and

[0047] a memory communicatively coupled to at least one processor, wherein:

[0048] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of the first aspect.

[0049] In a third aspect, the present application further provides a non-transitory computer-readable storage medium, which stores computer-executable instructions, and the computer-executable instructions are used to enable an electronic device to execute the method of the first aspect.

[0050] Beneficial effects of the embodiments of the present application: Different from the prior art, the embodiments of the present application provide a green space optimization method based on an urban waterlogging risk assessment model, including: obtaining urban environmental variable test data, the urban environmental variable test data including two or more of urban topographic data, urban meteorological data, urban underlying surface data, and urban drainage network data; pre-evaluating the urban environmental variable test data, obtaining urban environmental variable evaluation data, and pre-evaluating the urban environmental variable test data including using a maximum entropy model to evaluate the replacement importance of urban environmental test variables, and screening urban environmental test variables with relatively large replacement importance as urban environmental variable evaluation data; obtaining urban historical waterlogging point data, and selecting urban environmental test variables with relatively large replacement importance according to the number of historical waterlogging points in the city. According to the urban environmental variable assessment data, the urban waterlogging risk assessment data is obtained; the maximum entropy model is optimized according to the urban waterlogging risk assessment data to obtain the urban waterlogging risk assessment model; according to the urban waterlogging risk assessment model, the urban waterlogging risk probability distribution map is output; the urban land use status data and the urban target green space rate are obtained, and the urban target green space rate refers to the proportion of green space that should be achieved in different land use types; the urban supplementary green space rate is obtained according to the urban surface coverage data, the urban land use status data and the target green space rate in the urban underlying surface data; the green space optimization plan is determined according to the urban waterlogging risk probability distribution map and the urban supplementary green space rate; according to the green space optimization plan and the urban waterlogging risk assessment model, the waterlogging risk reduction effect after green space optimization is evaluated. The urban environmental variable test data such as urban topographic data, urban meteorological data, urban underlying surface data and urban drainage network data are obtained, and pre-evaluated to obtain the urban environmental variable evaluation data. The above steps enable the maximum entropy model to propose urban environmental variable test data with greater relevance and replace the urban environmental variable test data with higher importance according to the actual situation of the city to be evaluated, thereby improving the model operation accuracy and evaluation accuracy. At the same time, the parameters of the maximum entropy model are optimized by using the urban environmental variable evaluation data obtained in the above steps, and the model with the best performance is selected as the urban waterlogging risk assessment model of the city. The model parameters suitable for the city are preferably selected to obtain the urban waterlogging risk assessment model. Compared with the default parameter method used in the prior art, the above method can effectively improve the assessment accuracy of the urban waterlogging risk assessment model. In addition, according to the urban waterlogging risk probability distribution map and the urban supplementary green space rate, the green space optimization plan is determined, which can make up for the problem that the existing green space planning lacks integration with waterlogging risk assessment. When the same area of ​​green space is configured, the rainwater storage function of urban green space can be more effectively exerted to reduce the risk of urban waterlogging. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0052] Figure 1 A schematic diagram of the operating environment of a green space optimization method based on an urban waterlogging risk assessment model provided in one embodiment of the present application;

[0053] Figure 2 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application;

[0054] Figure 3 A flowchart of a green space optimization method based on an urban waterlogging risk assessment model in some embodiments of the present application;

[0055] Figure 4 This is a schematic diagram of the output of the maximum entropy model for parameter optimization in some embodiments of the present application;

[0056] Figure 5 A schematic diagram of the waterlogging risk level corresponding to the urban waterlogging risk probability distribution diagram in some embodiments of the present application;

[0057] Figure 6 A schematic diagram of waterlogging risk levels corresponding to a city waterlogging risk probability prediction map in some embodiments of the present application;

[0058] Figure 7 A schematic diagram of the waterlogging risk level corresponding to the second city waterlogging risk probability prediction map in some embodiments of the present application;

[0059] Figure 8 This is a schematic diagram of the green space priority configuration area in some embodiments of the present application. DETAILED DESCRIPTION

[0060] The present application is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that, for those of ordinary skill in the art, several variations and improvements can also be made without departing from the concept of the present application. These all belong to the protection scope of the present application.

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] It should be noted that, if there is no conflict, the various features in the embodiments of the present application can be combined with each other, all within the scope of protection of the present application. In addition, although the functional module division is performed in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a sequence different from the module division in the device or the flow chart. In addition, the words "first", "second", "third", etc. used herein do not limit the data and execution order, but only distinguish the same items or similar items with basically the same functions and effects.

[0063] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used in this specification and in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The term "and / or" used in this specification includes any and all combinations of one or more of the related listed items.

[0064] In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0065] In recent years, domestic and foreign scholars have conducted extensive and in-depth research on waterlogging risk prediction. The main methods include numerical simulation based on hydrological models, weighted assessment based on geographic information system (GIS) platforms, and risk quantification using machine learning methods. Prediction of climate disaster risks based on machine learning is currently a more commonly used method.

[0066] Among them, unlike algorithms such as random forests and neural networks that rely on positive and negative samples, the maximum entropy model (MaxEnt) only needs positive sample data for high-precision calculations. In view of the continuous development and changes of cities, it is unreasonable to define areas where waterlogging has not occurred as "negative samples". Instead, only known waterlogging areas can be used as "positive samples", and factors related to waterlogging events can be used as environmental variables. Based on the wide application of the MaxEnt model in species distribution research, we can innovatively regard urban waterlogging points as "species" and evaluate potential disaster risk areas by revealing the association mechanism between spatial environment and urban waterlogging. This method provides new research ideas and technical means for urban waterlogging risk prediction.

[0067] The data used in the existing methods for predicting urban waterlogging risk through the maximum entropy model are of low accuracy, so there are certain errors in the waterlogging risk prediction results output by the model. On the other hand, when using the maximum entropy model to predict urban waterlogging risk, the parameters of the maximum entropy model are not adjusted, which makes the model likely to be overfitted, affecting the accuracy of the urban waterlogging risk prediction results output by the model. At the same time, the existing methods for predicting urban waterlogging risk do not optimize the configuration of urban green spaces according to the urban waterlogging risk prediction results output by the model, and there is no specific plan to reduce the risk of urban waterlogging.

[0068] In view of this, before using the maximum entropy model to predict urban waterlogging risk, this scheme optimizes the maximum entropy model and adjusts its parameters through urban waterlogging risk assessment data to improve the accuracy of the maximum entropy model. In addition, the urban green space is optimized according to the urban waterlogging risk prediction results output by the urban waterlogging risk assessment model obtained after the maximum entropy model parameter optimization. At the same time, it can also evaluate the effects of various optimization schemes, determine the optimal green space configuration scheme, and effectively reduce the risk of urban waterlogging.

[0069] Figure 1 It is a schematic diagram of the operating environment of the green space optimization method based on the urban waterlogging risk assessment model provided by an embodiment of the present invention, including an electronic device 10.

[0070] The electronic device 10 is a device that can run according to a program and automatically and quickly process massive amounts of data. It is usually composed of a hardware system and a software system, such as a computer, a server, etc. The electronic device 10 can be a local device or a cloud device, such as a cloud server, a cloud host, a cloud service platform, a cloud computing platform, etc.

[0071] In the above Figure 1 Based on this, another embodiment of the present invention provides a hardware structure diagram of an electronic device 10, specifically, as Figure 2 As shown, the electronic device 10 includes at least one processor 11 and a memory 12 ( Figure 2 (a bus connection and a processor are used as an example).

[0072] Among them, the processor 11 is used to provide computing and control capabilities to control the electronic device 10 to perform corresponding tasks, for example, to control the electronic device 10 to execute any one of the green space optimization methods based on the urban waterlogging risk assessment model provided in the following embodiments of the invention.

[0073] It is understandable that the processor 11 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0074] The memory 12, as a non-transient computer-readable storage medium, can be used to store non-transient software programs, non-transient computer executable programs and modules, such as the program instructions / modules corresponding to the green space optimization method based on the urban waterlogging risk assessment model in the embodiment of the present invention, or the program instructions / modules corresponding to the green space optimization method based on the urban waterlogging risk assessment model in the embodiment of the present invention. The processor 11 can implement the green space optimization method based on the urban waterlogging risk assessment model in any of the following method embodiments by running the non-transient software programs, instructions and modules stored in the memory 12. Specifically, the memory 12 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory 12 may also include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0075] The following is a description of the green space optimization method based on the urban waterlogging risk assessment model provided by the embodiment of the present application in combination with the exemplary application and implementation of the terminal provided by the embodiment of the present application. Figure 3 , Figure 3 It is a flow chart of a green space optimization method based on an urban waterlogging risk assessment model provided in an embodiment of the present application.

[0076] like Figure 3 As shown, the method S100 includes but is not limited to the following steps:

[0077] S110: Acquire urban environmental variable test data, where the urban environmental variable test data includes two or more of urban topographic data, urban meteorological data, urban underlying surface data, and urban drainage network data.

[0078] The urban environmental variable test data is the maximum entropy model input data used for obtaining urban environmental variable evaluation data using the maximum entropy model.

[0079] Among them, urban topography and geomorphology data refers to terrain-related data, including one or more of the city's altitude, slope, and surface roughness. Urban topography and geomorphology data determine the flow path of surface runoff. Therefore, the model can predict the possible areas of urban waterlogging based on topography and geomorphology data. Urban meteorological data refers to the city's rainfall data, including one or more of the city's annual average temperature, the highest temperature in the hottest month, the average temperature in the wettest quarter, the average temperature in the hottest quarter, annual precipitation, rainfall in the wettest month, rainfall in the wettest quarter, rainfall in the hottest quarter, rainfall in the wettest month, rainstorm accumulation, and heavy rainstorm accumulation. Urban meteorological data represents the characteristics and attributes of rainfall events. Rainfall data is the basis for simulating surface runoff. The intensity, duration, and frequency of rainfall data determine the size of runoff. Rainfall data can help the model determine the flow path of surface runoff. By analyzing rainfall data, the model can predict the possibility of waterlogging. Combined with other factors, such as land use type and drainage system conditions, the model can assess the waterlogging risk in different regions. Urban underlying surface data refers to the physical and geographical characteristics of the urban surface, including one or more of the urban surface cover type, normalized vegetation index, normalized building index, and normalized water body index. By analyzing the underlying surface data, the impact of different surface cover types on runoff generation can be evaluated. For example, buildings and roads may increase runoff, while vegetation and grassland may reduce runoff. By using high-quality underlying surface data, the error of model prediction can be reduced because the model can better capture the true situation of surface water flow. Urban drainage network data refers to the detailed information of the drainage system used to collect and discharge rainwater in the city, including one or more of the urban drainage zoning data, pipeline type data, network outlet latitude and longitude information, and network flow data. By analyzing the drainage network data, the drainage capacity of the existing drainage system can be evaluated and its performance under different rainfall conditions can be determined. Combining urban drainage network data with rainfall data, the model can predict areas where waterlogging may occur under specific rainfall conditions.

[0080] The above urban environmental variable test data can help the model simulate the generation, flow and accumulation of surface runoff after precipitation. In addition, by combining urban topographic data, urban meteorological data, urban underlying surface data and urban drainage network data, the model can predict areas where waterlogging may occur and more accurately predict urban waterlogging risks.

[0081] In some embodiments, the above step S110 specifically includes:

[0082] S111: Taking Shenzhen City, Guangdong Province as an example, obtain urban topographic data, urban meteorological data, and urban underlying surface data in the urban environmental variable test data; obtain urban topographic data. The topographic data include three data: urban digital elevation model, surface slope, and surface roughness; obtain urban meteorological data, which include seven data: annual average rainfall, rainstorm accumulation, heavy rainstorm rainfall, extremely heavy rainstorm rainfall, rainstorm frequency, heavy rainstorm frequency, and extremely heavy rainstorm frequency; obtain urban underlying surface data, including four data: urban surface cover type, normalized vegetation index, normalized building index, and normalized water index, a total of 14 urban environmental variable test data.

[0083] Among them, urban surface cover type refers to the different uses and material types of urban surfaces, including natural surfaces and artificial surfaces. Natural surfaces include areas with good permeability such as vegetation and the main sources and confluence areas of surface water such as water bodies. Artificial surfaces include impermeable buildings, roads with poor permeability, squares and parking lots. The urban surface cover type determines the infiltration and runoff of surface water. In the waterlogging model, different types of cover will have different effects on the generation and flow of runoff. The Normalized Difference Vegetation Index (NDVI) is used to quantify the coverage and health of surface vegetation. Since vegetation can absorb precipitation and increase the infiltration rate of soil, vegetation cover can reduce surface runoff. The Normalized Difference Building Index (NDBI) is a data that measures the distribution and density of buildings in the city. According to the Normalized Difference Vegetation Index and the Normalized Difference Building Index, the model can better understand the urban environmental information, making the prediction results of urban waterlogging risk output by the model more accurate. The Normalized Difference Water Index (NDWI) is used to quantify the extent of surface water coverage and is calculated by comparing the difference in water reflection in remote sensing images with the reflection of other surface features. The NDWI can help the maximum entropy model identify surface water bodies, including rivers, lakes, reservoirs, and waterlogged areas; at the same time, the maximum entropy model can better simulate the generation and flow of surface runoff, because water bodies are the main source and confluence area of ​​surface water, improving the accuracy of the urban waterlogging risk prediction results output by the maximum entropy model.

[0084] A digital elevation model (DEM) is a digital representation of surface elevation information that stores the elevation value of each point on the surface in a discrete grid format. A DEM contains two-dimensional geographic spatial information of the surface (usually longitude and latitude coordinates) and corresponding elevation values, which can be actual ground elevations or relative elevations. DEMs can be used to calculate the slope and aspect of the surface, help simulate the flow path of surface water, and predict how runoff flows and gathers on the surface. By analyzing the DEM, the model can assess the risk of waterlogging in different areas of the city.

[0085] The digital elevation model of the city was obtained from the public remote sensing dataset website; ArcGIS software was used to calculate the urban surface slope and surface roughness based on the digital elevation model of the city. Urban rainfall data was obtained through the "real-time grid data table of automatic stations in the city range". According to the intensity of precipitation, rainfall events can be divided into three levels, that is, 24-hour precipitation of 50-99.9 mm is "rainstorm", 100-249.9 mm is "heavy rain", and more than 250 mm is "extremely heavy rain". The annual average rainfall, rainstorm accumulation, heavy rainstorm rainfall, extremely heavy rainstorm rainfall, rainstorm number, heavy rainstorm number and extremely heavy rainstorm number of each station were calculated through the obtained urban rainfall data. The kriging method in the spatial ArcGIS software was used to interpolate the automatic station data into raster data. The Landsat8 series remote sensing image data was downloaded using the USGS Earth Explorer image platform, and the data was processed using the ENVI software. First, the image was subjected to radiation correction, atmospheric correction and geometric correction to complete the image preprocessing. The supervised classification method is used to classify the pixels in the image into three categories: water, impervious ground and vegetation, and the urban surface cover type data is obtained. The band calculation tool is used to calculate the normalized vegetation index, normalized building index and normalized water index.

[0086] S120: Pre-evaluate the urban environmental variable test data to obtain urban environmental variable evaluation data.

[0087] The 14 urban environmental variable test data obtained in step S111 were subjected to Pearson correlation analysis using R software, and the maximum entropy model was used to evaluate the permutation importance of each urban environmental variable test data. If the absolute value of the Pearson correlation coefficient between two urban environmental variable test data is greater than 0.7, the urban environmental variable test data with a smaller permutation importance is removed, and the remaining urban environmental variable test data are combined as environmental variable evaluation data.

[0088] According to the evaluation results, five urban environmental variable test data, including heavy rainstorm rainfall, extremely heavy rainstorm rainfall, number of heavy rainstorms and number of extremely heavy rainstorms in meteorological data and normalized water body index in underlying surface data, were removed.

[0089] The remaining 9 urban environmental variable test data after screening constitute the urban environmental variable evaluation data: digital elevation model, surface slope, surface roughness, average annual rainfall, rainstorm accumulation, number of rainstorms, urban surface cover type, normalized vegetation index and normalized building index.

[0090] By pre-evaluating the urban variable test data and selecting the optimal evaluation data combination according to the Pearson correlation test and the maximum entropy model permutation importance test, the urban environmental variable evaluation data is obtained. Compared with directly obtaining the data required by the maximum entropy model, the above method can propose urban variable test data with higher permutation importance among the urban variable test data with greater correlation according to the actual situation of the city to be evaluated, and input the urban variable test data with higher permutation importance as the urban environmental variable evaluation data into the maximum entropy model, thereby improving the operation precision and evaluation accuracy of the maximum entropy model.

[0091] S130: Acquire historical urban waterlogging point data, and combine it with urban environmental variable assessment data to obtain urban waterlogging risk assessment data.

[0092] The historical urban waterlogging occurrence point data is the specific location of the urban waterlogging occurrence point in the past. Based on the historical waterlogging occurrence point data, the maximum entropy model can accurately simulate the conditions and process of waterlogging. Based on this, the maximum entropy model can perform urban waterlogging risk assessment on the input urban waterlogging risk assessment data and output the corresponding assessment results.

[0093] The original data of historical urban waterlogging occurrence points were obtained. The study area was converted into a 30-meter resolution raster file using GIS software. Data was screened based on the raster file. If the same grid unit (i.e., pixel) contained multiple waterlogging occurrence points, one occurrence point was randomly retained to avoid the influence of spatial autocorrelation on the maximum entropy model results.

[0094] The processed urban waterlogging occurrence point data are combined with the above-mentioned urban environmental variable assessment data to obtain urban waterlogging risk assessment data.

[0095] S140: Optimizing parameters of the maximum entropy model according to the urban waterlogging risk assessment data to obtain an urban waterlogging risk assessment model.

[0096] The Maximum Entropy Model is a probabilistic model that estimates probability distribution by maximizing entropy. In the Maximum Entropy Model, parameters refer to weights or coefficients in the model, which represent the degree of influence of different features on the model output. Parameters (usually expressed as weights) determine the contribution of each feature to the model output, and different parameter values ​​indicate the degree of trust the model has in each feature. Therefore, the choice of maximum entropy model parameters has an impact on its own generalization ability. Appropriate parameter settings can improve the generalization ability of the maximum entropy model, making it perform well on training data while maintaining stable performance on unseen data.

[0097] By optimizing the parameters of the maximum entropy model, the model parameters suitable for the city can be selected to obtain the urban waterlogging risk assessment model. Compared with the default parameter method adopted in the existing technology, the assessment accuracy can be effectively improved.

[0098] In some embodiments, the above step S140 specifically includes:

[0099] S141: Assign model parameters to the maximum entropy model, where the model parameters include feature type, control coefficient, and regional classification method.

[0100] In the maximum entropy model, the feature type usually refers to the attribute or variable type of the input data used for modeling, such as continuous features, discrete features, and ordered features. The feature type of the maximum entropy model defines the structure and nature of the input data, which directly affects the performance and prediction ability of the model. By reasonably selecting the feature type of the maximum entropy model, the prediction accuracy of the model can be improved, while the complexity of the model can be reduced and the generalization ability of the model can be improved.

[0101] The control coefficient of the maximum entropy model is a parameter used to control the feature weight estimation during the model training process. In the maximum entropy model, the control coefficient is usually represented by λ (lambda), which has an important impact on the generalization ability and overfitting risk of the model. The best control coefficient is usually selected through cross-validation. By reasonably selecting the control coefficient of the maximum entropy model, the generalization ability of the model can be improved.

[0102] Regional classification methods refer to techniques for dividing the data set processed by the maximum entropy model into different categories or regions. They are usually used to identify and distinguish patterns or features in the data so that the model can better understand and predict.

[0103] By combining model parameters such as the feature type, control coefficient and regional analysis method of the maximum entropy model, we can obtain maximum entropy models with various parameters. This allows us to optimize the parameters of the maximum entropy model in a more comprehensive range and effectively improve the prediction accuracy and generalization ability of the maximum entropy model.

[0104] In some embodiments, the above step S141 specifically includes:

[0105] S141a: The feature types include one or more of linear features, linear features and quadratic features, hinge features, linear features, quadratic features and hinge features; the control coefficient ranges from 0.5 to 4, and the interval ranges from 0.1 to 0.5; the regional classification methods include one or more of longitude and longitude, latitude and latitude, longitude and latitude, and latitude and longitude.

[0106] Among them, linear features refer to features that are directly linearly related to the output variable. The performance of these features in the model is linear, that is, changes in feature values ​​will affect the probability of the maximum entropy model output in a linear manner. Since linear features usually do not introduce too much complexity and overfitting. Therefore, linear features help improve the generalization ability of the maximum entropy model and help the maximum entropy model avoid overfitting on the training data, thereby reducing the generalization error. In summary, the reasonable selection of linear features in the maximum entropy model helps to improve the predictive performance, generalization ability and interpretability of the model, while reducing computational complexity and overfitting risks.

[0107] Quadratic features are features that have a quadratic relationship with the output variable. The performance of these features in the model is nonlinear, usually involving the square of the eigenvalue or other polynomial forms. Quadratic features can capture nonlinear patterns in the data, such as curvilinear relationships or quadratic curves. By introducing quadratic features, the model can become more complex, so that it can better adapt to the complexity of the data and improve the model's fitting ability. In addition, since quadratic features can capture additional information in the data that is not revealed by the original linear features, quadratic features can also improve the prediction accuracy of the model.

[0108] Hinge Features are a special type of feature used to handle nonlinear boundaries in classification problems. By introducing nonlinear transformations such as step functions or Gaussian functions, the maximum entropy model can capture nonlinear relationships in the data, thereby improving the prediction accuracy of the model. Since hinge features can capture subtle patterns in the data, hinge features can also improve the generalization ability of the maximum entropy model.

[0109] In the maximum entropy model, the value of the control coefficient determines how well the model fits the training data. A smaller control coefficient may cause the model to overfit on the training data, while a larger control coefficient helps reduce overfitting and makes the model perform better on new data.

[0110] It is understandable that if the control coefficient is less than 0.5, it indicates that the maximum entropy model has almost no regularization, and the maximum entropy model is likely to overfit on the training data, showing higher training accuracy and poor generalization ability. If the control coefficient is greater than 4, the maximum entropy model will tend to be smoother, helping to improve the generalization ability of the model, but this situation may cause the maximum entropy model to have a decreased accuracy on the input maximum entropy training data. Therefore, in some embodiments of the present solution, the control coefficient of the maximum entropy model ranges from 0.5 to 4.

[0111] As can be seen from the foregoing, in some embodiments, the feature types of the maximum entropy model include linear features (L), quadratic features (Q) and hinge features (H), and the regulation coefficient of the maximum entropy model ranges from 0.5 to 4. 0.5 is selected as a fixed interval to obtain the value of the specific regulation coefficient within a known range, that is, the specific values ​​of the regulation coefficient of the maximum entropy model are: 0.5, 1, 1.5, 2, 2.5, 3, 3.5 and 4. Among them, the feature type includes multiple combinations, for example, L, LQ, H and LQH. The regional classification method includes one or more of longitude and longitude, latitude and latitude, longitude and latitude, and latitude and longitude. By combining the feature type combination, regulation coefficient and regional classification method of the maximum entropy model, the parameter combination obtained is the model parameter. For example, model parameter (1): the regulation coefficient is 0.5, the feature type combination is L, and the regional classification method is longitude and longitude, or, model parameter (2): the regulation coefficient is 1.5, the feature type is LQ, and the regional classification method is longitude and latitude, etc.

[0112] By using a fixed interval of 0.5 to select the control coefficient within the applicable range, it can not only help the maximum entropy model find a balance between its complexity and generalization ability. Since the fixed interval is relatively small, it can also fine-grainedly adjust the complexity of the model, simplify the cross-validation process, and more accurately find the optimal control coefficient. Compared with larger fixed intervals (such as 1 or 10), choosing 0.5 as the fixed interval avoids sudden changes in model complexity, which helps to avoid drastic fluctuations in model performance.

[0113] S142: Combining two or more types of model parameters to obtain multiple groups of candidate models; performing performance evaluation on the multiple groups of candidate models through a performance evaluation tool to obtain evaluation results; and determining an urban waterlogging risk assessment model based on the evaluation results.

[0114] From the above content, it can be seen that the feature type combination of the above maximum entropy model includes 4 types, the specific values ​​of the control coefficient include 8 types, and the regional classification method includes 4 types. Therefore, there are 128 model parameters obtained by combining the feature type combination, control coefficient and regional classification method of the above maximum entropy model. Therefore, there are 128 maximum entropy models containing different model parameters obtained by the above method. The performance of the maximum entropy model under each of the above model parameters is evaluated according to the performance evaluation tool, and the evaluation results are output. According to the evaluation results, the maximum entropy model under a set of model parameters with higher performance and higher accuracy is selected, and the above maximum entropy model is used as the urban waterlogging risk assessment model.

[0115] Among them, the performance evaluation tools are programming language packages, such as ENMeval, scikit-learn, statsmodels, and dismo. The ENMeval package is used to evaluate and compare the performance of maximum entropy models. The ENMeval package can compare the performance of multiple maximum entropy models in order to select the best model configuration. The ENMeval package also supports cross-validation, which can help determine the generalization ability of the model and provide feature selection methods to help determine the features that have the greatest impact on model performance. The dismo package provides a set of tools and algorithms for training and optimizing distributed machine learning models. At the same time, dismo provides a variety of optimization algorithms, such as gradient descent, stochastic gradient descent, Adam, etc., for optimizing model parameters.

[0116] According to the regional classification method, the urban waterlogging risk assessment data input into the maximum entropy model are divided into four training data sets and validation data sets with different geographical locations, so that the maximum entropy model with different model parameters can be tested spatially independently. Among them, three of the four geographically different data sets are training data sets, and one is a validation data set. The three training data sets are input into the above maximum entropy model for training, and the training is verified based on the maximum entropy model and validation data set. This process is repeated until all data sets are used for verification.

[0117] In the parameter optimization process, according to the input urban waterlogging risk assessment data, the maximum entropy model outputs the corresponding AUC index, such as Figure 4 As shown. Among them, the AUC index (Area Under the ROC Curve) is a data used to evaluate the performance of the classification model, which represents the area under the ROC curve. The ROC curve (Receiver Operating Characteristic Curve) is a graphical tool used to show the performance of the classification model at different thresholds. The AUC index value ranges from 0 to 1, where 1 represents perfect classification performance (that is, the model always predicts correctly), 0.5 represents the performance of random guessing, and less than 0.5 means that the performance of the model is worse than random guessing.

[0118] It can be understood that the AUC index output by the maximum entropy model obtained by training different model parameters is obtained, the maximum entropy model with an AUC index closest to 1 among the above maximum entropy models is selected and its model parameters are obtained, and the urban waterlogging risk assessment model is determined based on the above model parameters and the maximum entropy model.

[0119] By combining the feature type, control coefficient and regional classification method of the maximum entropy model, multiple maximum entropy models with different model parameters are obtained, and the maximum entropy models with multiple different model parameters are trained according to the urban waterlogging risk assessment data, and the maximum entropy model with better performance is selected according to the AUC index output by the above maximum entropy model, and its corresponding model parameters are determined as the model parameters of the urban waterlogging risk assessment model. Through this step, the parameters suitable for the urban waterlogging risk assessment model can be selected to obtain the urban waterlogging risk assessment model, which can effectively improve the assessment accuracy compared with the default parameter method used in the prior art.

[0120] S150: Outputting an urban waterlogging risk probability distribution map according to the urban waterlogging risk assessment model.

[0121] The urban waterlogging risk assessment data is input into the urban waterlogging risk assessment model, and the urban waterlogging risk assessment model outputs an urban waterlogging risk probability distribution map. The above urban waterlogging risk probability distribution map is the urban waterlogging risk probability distribution map before green space optimization.

[0122] S160: Obtain the current urban land use data and the urban target greening rate.

[0123] The current urban land use data specifically include land use types and different types of land use areas. Among them, land use types refer to the classification of land plots used for different purposes, such as industrial land, residential land and public green space. The surface cover characteristics of different land use types (such as hard surface, vegetation cover) will affect the rate and amount of runoff generation. For example, hard surfaces (such as asphalt roads and parking lots) will produce more runoff, while vegetation-covered areas may reduce runoff. Different types of land use areas reflect the distribution of urban surface cover, which can help the model simulate the generation and flow of runoff in different areas, thereby predicting the possibility of waterlogging. By analyzing the current urban land use data, the model can predict the risk of urban waterlogging.

[0124] The urban target green space ratio refers to the proportion of green space that should be achieved in different land use types, that is, the percentage of green space required in the area of ​​different land use types in the city. It can be understood that the urban target green space ratio can be calculated by the ratio of the green space area of ​​different land use types in the city after green space optimization to the area of ​​different land use types in the city. For example, if the urban residential land area S1 = 221.3km 2 , the green area required for urban residential land after optimization is S2 = 66.39km 2 , then the target green area ratio of urban residential land is B = 30%. The specific calculation formula is as follows:

[0125]

[0126] In this embodiment, land use types are divided into 9 categories: logistics and warehousing land, transportation facility land, industrial land, commercial service land, residential land, science and education, medical and health, scenic spots and government agencies and groups land, development reserve land, public facility land, and public green land, and their target greening rates are 20%, 20%, 30%, 20%, 30%, 40%, 50%, 30%, and 90%, respectively.

[0127] S170: Obtain the city's supplementary green space ratio based on the city's surface cover type data, the city's land use status data and the city's target green space ratio.

[0128] Urban land surface cover type data refers to the distribution information of different types of urban surface features, including natural and artificial surfaces; from the above content, it can be seen that urban land use status data includes different land use types and their utilization areas; urban target greening rate is the greening rate that the city needs to achieve; urban supplementary greening rate is the greening rate that the city needs to increase to achieve the urban target greening rate. The urban supplementary greening rate can be calculated through urban land surface cover type data, urban land use status data and urban target greening rate.

[0129] In some embodiments, the above step S170 specifically includes:

[0130] S171: Calculate the current green space ratio of different land use types through the current urban land use data and the urban surface cover type data in the urban waterlogging risk assessment data; determine the urban supplementary green space ratio based on the urban target green space ratio and the current green space ratio of different land use types.

[0131] ArcGIS software was used to spatially overlay the urban land cover types and urban land use status data in the urban underlying surface data to obtain the distribution map of land cover of different land use types. The current green space rate of different land use types was calculated by the zoning statistics function of ArcGIS software.

[0132] From the above content, it can be known that the urban supplementary green space rate can be determined according to the difference between the urban target green space rate and the current green space rate of different land types, as shown in Table 1. For example, if the urban public green space target green space rate B_(1) = 90%, the current green space rate of urban public green space A_(1) = 61%, then the supplementary green space rate of urban public green space C_(1) = 29%. The specific calculation formula is as follows:

[0133] C=BA

[0134] Table 1

[0135]

[0136]

[0137] S180: Determine the green space optimization plan based on the urban waterlogging risk probability distribution map and the urban green space supplement rate.

[0138] From the above content, we can know that the urban supplementary green space rate is the green space rate that needs to be increased in the current city to reach the urban target green space rate. According to the urban waterlogging risk probability distribution map, the probability of waterlogging risk in different areas of the city can be known. The tool can be used to screen the urban non-green space areas, and the green space configuration area can be selected based on the urban supplementary green space rate and the urban non-green space area. For example, the calculated urban supplementary green space rate C_(2) = 10%, and 10% of the urban non-green space area is selected as the green space configuration area.

[0139] Screening urban non-greenland areas and selecting urban non-greenland areas as greenland configuration areas through urban waterlogging risk probability distribution map and urban supplementary greening rate can more effectively improve urban waterlogging risks.

[0140] In some embodiments, the above step S180 specifically includes:

[0141] S181: Obtain hard surfaces of different land use types through the urban waterlogging risk probability distribution map and different land use types in the urban land use status data; obtain green space optimization priority allocation areas based on the hard surfaces of different land use types and the highest risk areas with the same percentage of urban supplementary green space rate; determine the green space optimization plan through the green space optimization priority allocation areas.

[0142] The segmentation tool in ArcGIS software was used to segment the waterlogging risk probability distribution map according to different land use types; the attribute extraction tool in ArcGIS was used to extract the impervious ground in the urban surface cover type and screen the urban non-green area; data was extracted from the urban waterlogging risk probability distribution map to obtain the urban waterlogging risk probability data of the urban non-green area, and the urban non-green area was sorted from high to low according to the urban waterlogging risk probability data, and the green area configuration area was selected in order, that is, the urban non-green area with higher urban waterlogging risk probability data was preferentially selected as the first green area configuration area, that is, if the urban supplementary green area rate is C%, Python software was used to extract the top C% areas with the highest waterlogging risk probability in the impervious ground of different land use types. Ideally, the top C% areas with the highest waterlogging risk probability in the impervious ground of different land use types are the green area priority configuration areas. For example, the calculated urban supplementary green space rate C_(1)=29%. According to the ranking of urban waterlogging risk probability data, the top 29% of urban non-green space areas are selected as the first green space configuration area.

[0143] The non-green areas in the city are sorted according to the urban waterlogging risk probability data. The non-green areas with high urban waterlogging risk probability data are selected as the first green space allocation areas through the city's supplementary greening rate. This can make up for the problem that the existing green space planning lacks integration with waterlogging risk assessment. When allocating the same area of ​​green space, the rainwater storage function of urban green space can be more effectively exerted to reduce the risk of urban waterlogging.

[0144] S190: Based on the green space optimization plan and urban waterlogging risk assessment model, evaluate the waterlogging risk reduction effect after green space optimization.

[0145] The green space optimization scheme is the green space optimization scheme determined by the urban waterlogging risk probability distribution map and the urban green space replenishment rate according to the aforementioned step S180. The waterlogging risk probability after green space optimization is evaluated by the determined green space optimization scheme and the urban waterlogging risk assessment model.

[0146] In some embodiments, the above step S190 specifically includes:

[0147] S190a: According to the green space optimization plan, spatial analysis tools are used to obtain urban underlying surface update data; the urban underlying surface data in the urban waterlogging risk assessment data are replaced by the urban underlying surface update data to obtain urban waterlogging risk prediction data; the urban waterlogging risk prediction data are input into the urban waterlogging risk assessment model to output the urban waterlogging risk probability prediction map after green space optimization; the waterlogging risk reduction effect after green space optimization is evaluated based on the urban waterlogging risk probability prediction map after green space optimization.

[0148] According to the green space optimization scheme determined in the aforementioned step S180, the urban underlying surface data is updated to obtain the updated urban underlying surface data, i.e., the new urban underlying surface data after the green space optimization. The update of the above-mentioned urban underlying surface data includes the update of urban surface cover type data, the update of normalized vegetation index data, and the update of normalized building index data. According to the green space optimization scheme, the impermeable ground at the corresponding position is converted into vegetation to update the urban surface cover type data; according to the green space optimization scheme, the NDVI value at the corresponding position is set to 0.4 (indicating healthy vegetation) to update the normalized vegetation index data; according to the green space optimization scheme, the NDBI value at the corresponding position is set to -0.2 (indicating non-building area) to update the normalized building index data.

[0149] By updating the above-mentioned relevant data, the updated urban underlying surface data can be obtained. The urban underlying surface data in the urban waterlogging risk assessment data is replaced accordingly using the urban underlying surface update data, that is, the urban underlying surface update data is combined with the urban topographic data and urban meteorological data in the urban waterlogging risk assessment data to obtain the urban waterlogging risk prediction data. The urban waterlogging risk prediction data is input into the urban waterlogging risk assessment model to obtain the urban waterlogging risk probability prediction map after green space optimization output by the urban waterlogging risk assessment model.

[0150] The waterlogging risk data of the corresponding urban area in the urban waterlogging risk probability distribution map output by the urban waterlogging risk assessment model is extracted through software tools. The above waterlogging risk data characterizes the probability of waterlogging in the urban area. The range of waterlogging risk data is 0 to 1. The closer the value is to 1, the greater the probability of waterlogging in the corresponding urban area; the closer the value is to 0, the smaller the probability of waterlogging in the corresponding urban area. Based on the above waterlogging risk data, the urban areas in the urban waterlogging risk probability distribution map can be divided into waterlogging risk levels, such as Figure 5 If the waterlogging risk data of the current urban area is less than 0.2, the current urban area belongs to the first-level risk (low risk) area; if the waterlogging risk data of the current urban area is greater than 0.2 and less than 0.4, the current urban area belongs to the second-level risk (lower risk) area; if the waterlogging risk data of the current urban area is greater than 0.4 and less than 0.6, the urban area belongs to the third-level risk (medium risk) area; if the waterlogging risk data of the current urban area is greater than 0.6 and less than 0.8, the current urban area belongs to the fourth-level risk (higher risk) area; if the waterlogging risk data of the current city is greater than 0.8, the current urban area belongs to the fifth-level risk (high risk) area.

[0151] According to the classification of urban area waterlogging risk levels, the urban area areas of different levels are obtained, as shown in Table 2, where R1 is the urban area areas of different waterlogging risk levels obtained according to the urban waterlogging risk probability distribution map, and R2 is the urban area areas of different waterlogging risk levels obtained according to the urban waterlogging risk probability prediction map.

[0152] Table 2

[0153]

[0154]

[0155] According to the urban area with different waterlogging risk levels before green space optimization, Figure 5 As shown in Figure 2, the urban area with different waterlogging risk levels after green space optimization is shown in Figure 2. Figure 6As shown, the reduction rate of the area of ​​the five-level risk area and the absolute reduction of the area of ​​the four-level risk area are calculated. The ratio of the difference between the area of ​​the five-level risk area before green space optimization and the area of ​​the five-level risk area after green space optimization to the area of ​​the five-level risk area before green space optimization is the reduction rate of the area of ​​the five-level risk area. For example, please refer to Table 2. The area of ​​the five-level risk area before green space optimization S1 = 135.22, the area of ​​the five-level risk area after green space optimization S2 = 107.24, and the reduction rate of the area of ​​the five-level risk area P = (135.22-107.24) / 135.22*100% = 20.69%. It can be understood that the absolute reduction in the area of ​​the fourth-level risk area is the difference between the risk area of ​​the fourth-level risk area before green space optimization and the risk area of ​​the fourth-level risk area after green space optimization. For example, please refer to Table 2 again. The area of ​​the fourth-level risk area before green space optimization is S3=499.97, and the area of ​​the fourth-level risk area after green space optimization is S4=496.55. The absolute reduction in the area of ​​the fourth-level risk area is K=S3-S4=3.42.

[0156] Based on the reduction rate of the fifth-level risk area and the absolute reduction in the fourth-level risk area obtained by the above method, the effectiveness of the current green space optimization method in mitigating urban waterlogging risks can be evaluated.

[0157] By calculating the reduction rate of the fifth-level risk area and the absolute reduction of the fourth-level risk area after the optimization of urban green space, the urban waterlogging risk mitigation effect of the current urban green space optimization plan can be evaluated, which can make up for the problem that the existing green space planning lacks integration with waterlogging risk assessment. When allocating the same area of ​​green space, the rainwater storage function of urban green space can be more effectively exerted to reduce the risk of urban waterlogging.

[0158] S190b: The difference in effect between the green space optimization method of the present application and the ordinary method can be verified through the following process.

[0159] Using the python model, randomly select C% of impervious ground as the second green space configuration area to obtain a random green space configuration plan. According to the method of S190a, obtain the second urban underlying surface update data, and use the second urban underlying surface update data to replace the urban underlying surface data in the urban waterlogging risk assessment data accordingly, that is, combine the second urban underlying surface update data with the urban topographic data and urban meteorological data in the urban waterlogging risk assessment data to obtain the second urban waterlogging risk prediction data. Input the second urban waterlogging risk prediction data into the urban waterlogging risk assessment model, and output the second urban waterlogging risk probability prediction map after green space optimization. According to the second urban waterlogging risk probability prediction map, obtain the urban waterlogging risk level map after adding the second green space configuration area, such as Figure 7The area changes of each disaster risk level in the second urban waterlogging risk probability prediction map and the urban waterlogging risk probability distribution map were calculated to evaluate the waterlogging risk reduction effect of random green space configuration and compare it with the waterlogging risk reduction effect in S190a.

[0160] For example, according to the classification of urban area waterlogging risk levels, the urban area areas of different levels are obtained, as shown in Table 3, where R1 is the urban area areas of different waterlogging risk levels obtained according to the urban waterlogging risk probability distribution map; R2 is the urban area areas of different waterlogging risk levels obtained according to the urban waterlogging risk probability prediction map; R3 is the urban area areas of different waterlogging risk levels obtained according to the second urban waterlogging risk probability prediction map.

[0161] Table 3

[0162]

[0163] The area of ​​the five risk areas before green space optimization S1 = 135.22 km 2 , after adding the first green space configuration area, the area of ​​the fifth-level risk area S2 = 107.24km 2 , the reduction rate of the fifth-level risk area P1 = (135.22-107.24) / 135.22*100% = 20.69%. In the case of increasing the same proportion of green space, the area of ​​the zero-risk area after adding the second green space configuration area is S3 = 132.16km 2 , the reduction rate of the fifth-level risk area P2 = (135.22-132.16) / 135.22*100% = 2.2%. It can be understood that the absolute reduction in the area of ​​the fourth-level risk area is the difference between the risk area of ​​the fourth-level area before green space optimization and the risk area of ​​the fourth-level area after green space optimization. For example, please refer to Table 3 again. The area of ​​the fourth-level risk area before green space optimization S4 = 499.97km 2 , after adding the first green space configuration area, the area of ​​the fourth-level risk area S5 = 496.55km 2 The absolute reduction in the area of ​​the fourth-level risk area is K1 = S4-S5 = 3.42 km 2 Under the same circumstances, after adding the second green space configuration area, the area of ​​the fourth-level risk area S6 = 497.25km 2 The absolute reduction in the area of ​​the fourth-level risk area is K2 = S4-S6 = 2.72 km 2 .

[0164] It can be seen that the effect of reducing the risk of urban waterlogging by increasing the first green space configuration area is significantly better than the effect of reducing the risk of urban waterlogging by increasing the second green space configuration area. Based on this, the priority green space configuration area is determined, such as Figure 8The green space optimization method of the present invention is highly effective in reducing the risk of waterlogging.

[0165] In addition, in actual situations, it is possible that the top C% areas with the highest probability of urban flooding risk in impervious surfaces of different land use types are not suitable for conversion into green spaces, such as the presence of important transportation facilities or high-rise buildings that are not suitable for demolition. In this case, appropriate adjustments can be made with reference to the green space priority allocation plan determined by S180 to form a variety of green space optimization plans, and the method of S190 can be repeated to evaluate the effects of different plans, and the urban green space optimization plan with better urban waterlogging risk mitigation effect can be selected, thereby effectively reducing the risk of urban flooding, achieving efficient use of land resources and enhancing urban resilience.

[0166] In summary, this scheme obtains urban environmental variable test data such as urban topographic data, urban meteorological data, urban underlying surface data and urban drainage network data, and pre-evaluates them, and selects the optimal evaluation data combination through Pearson correlation test and maximum entropy model permutation importance test, so as to obtain urban environmental variable evaluation data. The above steps enable the maximum entropy model to propose urban environmental variable test data with higher permutation importance in the urban environmental variable test data with greater correlation according to the actual situation of the city to be evaluated, thereby improving the model operation accuracy and evaluation accuracy. At the same time, the urban environmental variable evaluation data obtained through the above steps optimizes the parameters such as the characteristic type, regulation coefficient and regional classification method of the maximum entropy model, selects the model with the best performance as the urban waterlogging risk assessment model of the city, and preferably selects the model parameters suitable for the city to obtain the urban waterlogging risk assessment model. Compared with the default parameter method adopted in the prior art, the above method can effectively improve the evaluation accuracy of the urban waterlogging risk assessment model. In addition, based on the highest risk areas in the urban waterlogging risk probability distribution map and the urban green space supplementation rate, the green space priority allocation areas, that is, the green space optimization plan, are determined. This can make up for the problem that the existing green space planning lacks integration with waterlogging risk assessment. When allocating the same area of ​​green space, the rainwater storage function of urban green space can be more effectively exerted to reduce the risk of urban waterlogging.

[0167] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] Through the description of the above implementation methods, ordinary technicians in this field can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware. Ordinary technicians in this field can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Under the concept of the present application, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other changes in different aspects of the present application as described above, which are not provided in detail for the sake of simplicity. Although the present application has been described in detail with reference to the aforementioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A green space optimization method based on an urban waterlogging risk assessment model, characterized in that: include: Acquire urban environmental variable test data, wherein the urban environmental variable test data includes at least two of urban topographic data, urban meteorological data, urban underlying surface data, and urban drainage network data; Pre-evaluating the urban environment variable test data to obtain urban environment variable evaluation data, wherein the pre-evaluating the urban environment variable test data includes using a maximum entropy model to evaluate the permutation importance of the urban environment variable test data, and obtaining the urban environment variable evaluation data based on the permutation importance; Acquire historical urban waterlogging point data, and obtain urban waterlogging risk assessment data based on the historical urban waterlogging point data and the urban environmental variable assessment data; Optimizing the parameters of the maximum entropy model according to the urban waterlogging risk assessment data to obtain the urban waterlogging risk assessment model; Outputting an urban waterlogging risk probability distribution map according to the urban waterlogging risk assessment model; Obtaining the current status of urban land use and the urban target greening ratio, where the urban target greening ratio refers to the proportion of green space that should be achieved in different land use types; Obtaining the urban supplementary green space rate according to the urban surface cover type data in the urban underlying surface data, the urban land use status data and the urban target green space rate; Determine a green space optimization plan based on the urban waterlogging risk probability distribution map and the urban supplementary green space rate; According to the green space optimization plan and the urban waterlogging risk assessment model, the waterlogging risk reduction effect after green space optimization is evaluated.

2. The method according to claim 1, characterized in that The step of optimizing the parameters of the maximum entropy model according to the urban waterlogging assessment data to obtain the urban waterlogging risk assessment model includes: Configuring model parameters for the maximum entropy model, wherein the model parameters include feature type, regulation coefficient, and region classification method; Combining two or more of the model parameters to obtain multiple groups of candidate models; Performing performance evaluation on the multiple groups of candidate models using a performance evaluation tool to obtain evaluation results; According to the assessment results, the urban waterlogging risk assessment model is determined.

3. The method according to claim 2, characterized in that The feature type includes one or more of a linear feature, a linear feature and a quadratic feature, a hinge feature, a linear feature and a quadratic feature, and a hinge feature; The control coefficient ranges from 0.5 to 4, and the interval ranges from 0.1 to 0.5; The area classification method includes one or more of longitude and longitude, latitude and latitude, longitude and latitude, and latitude and longitude.

4. The method according to claim 1, characterized in that The urban supplementary green space rate is obtained through the urban surface cover type data in the urban underlying surface data, the urban land use status data and the urban target green space rate, including: Calculate the current green space ratio of different land use types through the urban land use status data and the urban land surface coverage type data; The city's supplementary greening rate is determined based on the city's target greening rate and the current greening rates of the different land use types.

5. The method according to claim 1, characterized in that Determine a green space optimization plan based on the urban waterlogging risk probability distribution map and the urban supplementary green space rate, including: Obtaining hard ground of different land use types through the urban waterlogging risk probability distribution map and different land use types in the urban land use status data; According to the hard ground of different land use types and the urban supplementary green space rate, the green space optimization priority configuration area is obtained; The green space optimization plan is determined through the green space optimization priority configuration area.

6. The method according to claim 1, characterized in that According to the green space optimization plan and the urban waterlogging risk assessment model, the waterlogging risk reduction effect after green space optimization is evaluated, including: Acquire urban underlying surface update data according to the green space optimization plan; Replacing the urban underlying surface data in the urban waterlogging risk assessment data with the urban underlying surface update data to obtain urban waterlogging risk prediction data; Inputting the urban waterlogging risk prediction data into the urban waterlogging risk assessment model, and outputting an urban waterlogging risk probability prediction map after green space optimization; The waterlogging risk reduction effect after green space optimization is evaluated based on the urban waterlogging risk probability prediction map after the green space optimization.

7. The method according to claim 1, characterized in that The obtaining of the urban environmental variable test data includes obtaining two or more of the topographic data, urban meteorological data, urban underlying surface data, and urban pipe network data, including: The topographic data include one or more data of the altitude, slope and surface roughness of the city; The urban meteorological data include one or more of the annual average temperature, the highest temperature of the hottest month, the average temperature of the wettest quarter, the average temperature of the hottest quarter, the annual precipitation, the rainfall of the wettest month, the rainfall of the wettest quarter, the rainfall of the hottest quarter, the rainfall of the wettest month, the accumulated amount of rainstorms and the accumulated amount of heavy rainstorms of the city; The urban underlying surface data includes one or more data of urban surface cover type, normalized vegetation index, normalized building index, and normalized water body index; The urban drainage pipe network data includes one or more of urban drainage zoning data, pipe type data, pipe network outlet latitude and longitude information, and pipe network flow data.

8. The method according to claim 1, characterized in that Preliminary evaluation is performed on the urban environmental variable test data to obtain urban environmental variable evaluation data, including: Calculate the Pearson correlation coefficient between the test data of the urban environmental variables; Using the maximum entropy model to evaluate the permutation importance of each of the urban environmental variable test data; If the absolute value of the Pearson correlation coefficient between the two urban environment variable test data is greater than 0.7, the urban environment variable test data with a smaller replacement importance is removed from the two urban environment variable test data, and the remaining urban environment variable test data are combined as the urban environment variable evaluation data.

9. An electronic device, characterized in that: include: at least one processor, and a memory communicatively coupled to the at least one processor, wherein: The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable an electronic device to execute the method according to any one of claims 1 to 8.

Images