Urban blue-green space layout optimization method based on form-ecological response model
Through the urban blue-green space layout optimization method based on the morphology-ecological response model, the problem of low optimization efficiency in the existing technology is solved, and a more scientific and comprehensive urban blue-green space layout optimization is achieved, improving optimization efficiency and overall ecology are improved. Development ability.
Patent Information
- Application Number
- CN202411235452.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-12-06
AI Technical Summary
The existing urban blue-green space layout optimization method lacks holistic research on multiple variables and cannot accurately reflect the correlation between form and ecology. It is affected by human factors, interest factors and cognitive limitations, resulting in low optimization efficiency.
The urban blue-green space layout optimization method based on the morphology-ecological response model is adopted. By constructing the morphological characteristic index set and the ecological benefit index set of urban blue-green space, the morphology-ecological response model is trained using the machine-learning random forest method to carry out optimization cycles. Improve the optimization efficiency of blue and green space.
This method reduces the impact of human factors and cognitive limitations on the optimization method, can comprehensively consider the impact of multiple morphological characteristics on ecological benefits, improves the optimization efficiency of blue-green space, and ensures the comprehensiveness and scientific nature of the overall ecological development of the city. sex.
Smart Images

Figure CN119090082A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of urban ecological space transformation, and specifically relates to an urban blue-green space layout optimization method based on a morphological-ecological response model. Background Art
[0002] Urban blue-green space has multiple ecological functions, is the main carrier of ecological benefits, and is also an important focus of high-quality urban development. At present, my country's cities are shifting from incremental development to stock development, and the overall optimization of layout in limited space is particularly critical. How to efficiently exert the ecological efficiency of urban blue-green space has become an urgent problem to be solved. In previous urban construction, there was a separation between spatial form and ecological benefits, while the two are closely related and influence each other. Existing methods for morphological optimization of urban blue-green space are mostly limited to regression analysis of simple mathematical models, or analysis of influencing factors of a single ecological benefit. They lack holistic research on multiple variables and cannot accurately reflect the correlation between form and ecology. Affected by human factors, interest factors and cognitive limitations, the optimization of blue-green space layout has limitations and ambiguity. Summary of the invention
[0003] To solve the above problems, the present invention discloses a method for optimizing the layout of urban blue-green spaces based on a morphological-ecological response model. The morphological-ecological response model is used for optimization cycles. Due to its iterative characteristics, it can adapt to most application scenarios and provide a variety of optimization results, greatly improving the optimization efficiency of blue-green spaces.
[0004] To achieve the above object, the technical solution of the present invention is as follows:
[0005] The urban blue-green space layout optimization method based on the morphological-ecological response model includes the following steps:
[0006] S1, construct the urban blue-green space morphological characteristic index set and ecological benefit index set, and calculate them;
[0007] S2, using the random forest method of machine learning, the urban blue-green space "morphology-ecology" response model is obtained by training the urban blue-green space morphological characteristic index set and ecological benefit index set of the target area;
[0008] S3, obtain the planning scheme of the target area, translate its land use data, and calculate the internal morphological characteristic indicators;
[0009] S4, based on the results of the morphological characteristic indicators of S3 and the urban blue-green space "morphology-ecology" response model of S2, identify typical areas where ecological benefits need to be improved and obtain the ecological benefit optimization target;
[0010] S5, construct an optimization iterative cycle based on the “morphology-ecology” response model of urban blue-green space, output the optimized urban blue-green space morphological feature series, and screen them, and then use the Landscape Generator platform (Landscape Generator, referred to as LG) to obtain the optimized urban blue-green space distribution.
[0011] Furthermore, in S1, the urban blue-green space morphological characteristic index set is composed of urban green space morphological characteristic index and urban blue space morphological characteristic index. Both types of indexes include area index, shape index and aggregation index. Among them, the area index is the total three-dimensional area, the proportion of the three-dimensional area of the largest patch, and the three-dimensional area ratio of the core patch; the shape index is the average perimeter area ratio, the average fractal dimension, and the landscape shape index; the aggregation index is the degree of aggregation, the degree of spread, the degree of connection, the degree of separation, the degree of splitting, the number of patches, and the density of patches.
[0012] Furthermore, in S1, the target layer of the ecological benefit evaluation index set is ecological benefit; the criterion layer includes stormwater resilience, cooling benefit, carbon sink benefit, and biodiversity; and the indicator layer includes surface runoff, waterlogging point density, urban surface temperature, net primary productivity, carbon storage, and habitat quality.
[0013] Furthermore, in S1, the calculation steps of the blue-green space morphological characteristic index set and the ecological benefit index set are as follows:
[0014] S11, obtaining multispectral remote sensing data, land use data, NDVI data (Normalized Difference Vegetation Index), vegetation type data, daily radiation data, temperature data, precipitation data, urban soil type data, and urban land cover data in the target area;
[0015] S12, identification of urban blue-green spaces using supervised classification based on multispectral remote sensing data;
[0016] S13, using Fragstats software to calculate the shape index and concentration index of urban blue space and urban green space;
[0017] The average perimeter area ratio is calculated as follows:
[0018]
[0019] Among them, RARA MN is the average perimeter-to-area ratio, N is the total number of samples, and p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample;
[0020] Among them, the calculation formula of average fractal dimension is:
[0021]
[0022] Among them, FRAC MN is the average fractal dimension, N is the total number of samples, p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample;
[0023] Secondly, the landscape shape index calculation formula is:
[0024]
[0025] Where LSI is the landscape shape index, e ik It refers to the total length of the landscape edge between the i-th sample and the k-th sample, and A is the area of the entire calculation area;
[0026] Again, the aggregation calculation formula is:
[0027]
[0028] Among them, AI is the degree of aggregation, g ij is the number of adjacent samples of the same type as the ijth sample, max(g ij ) is the maximum number of neighbors of the ijth sample;
[0029] Then, the spread calculation formula is:
[0030]
[0031] Among them, COHESION is the degree of spread, p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample, and Z is the number of all samples in the calculation area;
[0032] Secondly, the connectivity calculation formula is:
[0033]
[0034] Among them, CONNECT is the connectivity, c ijk is the number of possible connections between the ijth sample and the surrounding landscapes of the same type, n i is the total number of samples of this type of landscape in the entire calculation area;
[0035] Then, the separation degree is calculated as:
[0036]
[0037] Among them, DIVISION is the separation degree, a ijis the area of the ijth sample, and A is the area of the entire calculation area;
[0038] Secondly, the splitting degree calculation formula is:
[0039]
[0040] Among them, SPLIT is the splitting degree, a ij is the area of the ijth sample, and A is the area of the entire calculation area;
[0041] Again, the plaque number calculation formula is:
[0042] NP=n i
[0043] Where NP is the number of plaques, n i is the number of samples of a certain type of landscape in the calculation area;
[0044] Then, the plaque density is calculated as:
[0045]
[0046] Where PD is the plaque density, n i is the number of samples of a certain type of landscape in the calculation area, and A is the area of the entire calculation area;
[0047] Finally, before calculating the shape index and convergence index, the semivariogram function is used to calculate the most appropriate moving window size. The formula is as follows:
[0048]
[0049] Among them, X(i) and X(i+h) represent the values of the landscape pattern index X at position i and position i+h, h is the moving window size, N(h) represents the number of samples when the moving window size is h, and λ(h) represents the semivariance of X. Generally speaking, as the moving window increases, λ(h) will first increase and then decrease. When the trend of λ(h) tends to be flat, it means that the moving window size is more appropriate;
[0050] S14, the area index of urban blue space and urban green space is calculated by correcting the plane projection area used for the patch type area, the proportion of patch to landscape area, and the proportion of core area to the entire landscape area in the landscape pattern index to the three-dimensional surface area calculated based on DEM data (Digital Elevation Model, referred to as DEM). The specific calculation formula is as follows:
[0051] First, the total three-dimensional area is calculated as:
[0052]
[0053] Where VCA is the total three-dimensional area, va ij is the three-dimensional surface area of the ijth sample;
[0054] Secondly, the calculation formula for the proportion of the largest plaque three-dimensional area is:
[0055]
[0056] Among them, VLPI is the proportion of the largest plaque three-dimensional area, va ij is the three-dimensional surface area of the ijth sample, VA is the three-dimensional surface area of the entire calculation area;
[0057] Finally, the calculation formula for the core plaque three-dimensional area ratio is:
[0058]
[0059] Among them, VCPLAND is the three-dimensional area ratio of the core plaque, The three-dimensional surface area of the core sample in the region is calculated for the ijth sample, and VA is the three-dimensional surface area of the entire calculation region. Use a moving window of the same size as the shape index and the convergence index;
[0060] S15, calculating the total ecological benefit of the target area according to the ecological benefit indicator set;
[0061] S16, based on the calculation results of each ecological benefit indicator, normalize and standardize each indicator;
[0062] For positive indicators:
[0063]
[0064] Among them, X i ' is the normalized value of the i-th ecological benefit index, X i is the original value of the i-th ecological benefit index, X imin is the minimum value of the i-th ecological benefit index, and X imax is the maximum value of the i-th ecological benefit index;
[0065] For negative indicators:
[0066]
[0067] S17, the weights of the four types of ecological benefit indicators, namely, stormwater resilience, cooling benefit, carbon sequestration benefit, and biodiversity, are considered to be equally weighted. For benefits indicated by two indicators, such as carbon sequestration benefit and stormwater management benefit, it is necessary to equally weight the two indicators and then equally weight them again to calculate the total ecological benefit.
[0068] Further, S15, the total ecological benefit of the target area is calculated, and the specific steps are as follows:
[0069] S151, based on the multispectral remote sensing data in S11, uses the Radiative TransferEquation (RTE) to invert the LST value;
[0070] S152, based on the multispectral remote sensing data, daily radiation data, temperature data, and precipitation data obtained in S11, the CASA model is used to calculate the net primary productivity;
[0071] S153, use the Carbon Storage and Sequestration module provided by the InVEST (Integrated Valuation of Ecosystem Services and Trade-offs) platform to estimate the carbon storage in the target area;
[0072] S154, calculating the surface runoff volume of the target area according to the SCS-CN model (Soil Conservation Service Curve Number, runoff curve model);
[0073] S155, based on the remote sensing images of continuous sunny days and after rain in S11 multispectral remote sensing data, mosaic the flood inundation range in GIS software. Use the "point density analysis" tool to obtain the density of waterlogging points to characterize the waterlogging risk in the target area;
[0074] S156, based on the urban land cover data and multispectral remote sensing data of S11, the habitat quality of the target area was calculated using the HabitatQuality module of the InVEST platform.
[0075] Furthermore, in S2, the specific steps for training and obtaining the “morphology-ecology” response model of urban blue-green space are as follows:
[0076] S21, in SPSS software (Statistical Product and Service Solutions), the Spearman correlation coefficient is used to calculate the correlation between morphological characteristic indicators. The specific formula is:
[0077]
[0078] Among them, P S is the correlation coefficient between the two morphological indicators, d i is the sequence difference of the i-th data, and n is the number of calculated samples;
[0079] S22, use Bootstrap sampling method to randomly sample the training set to construct the training set, and the data that does not appear in the training set is the test set. Calculate the average test error of all decision trees and calculate the average rate of decrease in accuracy:
[0080]
[0081] Where, MDA is the average decrease rate of accuracy, which is the main factor to measure the importance of the indicator. n is the total number of decision trees, error t is the out-of-bag error of decision tree t when a feature is not added, error' t is the out-of-bag error of decision tree t after adding a certain feature;
[0082] S23, write a program in Matlab software (MATrix LABoratory) to perform random forest regression and obtain the results of the importance of morphological characteristic indicators; screen out the indicators with an importance less than 0.3. Then, among the remaining indicators, screen out the indicators with correlation coefficients greater than 0.8 and less importance, and the remaining indicators are the key morphological characteristic indicators;
[0083] S24, based on the calculation results of S23, select the appropriate number of decision trees through the three model features of the deterministic coefficient R2, the mean absolute error MAE, and the model running time T. Adjust the three parameters of the maximum depth of the decision tree, the minimum cotyledon tree, and the number of randomly generated seeds to fit the response model.
[0084] S25, based on the fitting results of S24, the model accuracy is verified by R2 and mean absolute error.
[0085] Furthermore, in S3, the specific steps of calculating the morphological characteristic index in the target area planning scheme are:
[0086] S31, identifying blue-green spaces, translating the planned land types based on land use / land cover classification, classifying forest land, grassland, and cultivated land as urban green spaces, and water bodies as urban blue spaces;
[0087] S32, calculating the morphological characteristic indicators in the target area, including the total three-dimensional area of the blue space, the spread, and the proportion of the three-dimensional area of the largest patch of green space, the degree of aggregation, and the landscape shape index.
[0088] Furthermore, in S4, the steps for identifying typical areas where ecological benefits need to be improved are:
[0089] S41, predict the ecological benefits of the target area based on the calculation results of the urban blue-green space morphological characteristic indicators in S3 and the morphological-ecological response model;
[0090] S42, the ecological benefits of the target area are evaluated using the natural breakpoint method and are divided into five ecological levels: excellent, relatively good, good, average, and poor;
[0091] S43, based on the ecological benefit classification of the target areas, areas where the proportion of areas with poor ecological benefits exceeds 90% are areas to be optimized.
[0092] Furthermore, in S5, the logic of constructing the morphology optimization loop is:
[0093] S51, write a program in the Matlab software platform, perform optimization iteration based on the morphological-ecological response model trained in S2, set the maximum number of iterative adjustments of the morphological characteristic indicators to 5000 times, and the output standard is to evaluate whether the new morphological characteristic series is better than the ecological benefit of the original planning scheme after each adjustment. If the ecological benefit is better, output the corresponding morphological characteristic series to obtain the morphological optimization result that meets the ecological goal improvement;
[0094] S52, before outputting the final optimized morphological feature series, it is necessary to make trade-offs among various options, that is, retaining those with higher ecological benefits and retaining those that are closer to the morphological features before optimization.
[0095] The beneficial effects of the present invention are:
[0096] 1. The urban blue-green space layout optimization method based on the morphological-ecological response model provided by the present invention constructs a set of urban blue-green space morphological characteristic indexes and an ecological benefit index set for the specific ecosystem of urban blue-green space, and trains a morphological-ecological response model through the random forest method. The morphological-ecological response model trained in this way minimizes the limitations of the blue-green space optimization method due to human factors, interest factors and cognitive limitations, ensuring that the blue-green space can promote the overall ecological development of the city. And the constructed morphological-ecological response model can comprehensively consider the impact of various morphological characteristics on ecological benefits, and has strong complexity, comprehensiveness and scientificity.
[0097] 2. The urban blue-green space layout optimization method based on the morphological-ecological response model provided by the present invention performs optimization cycles using the morphological-ecological response model. Due to its iterative characteristics, it can adapt to most application scenarios and provide a variety of optimization results, greatly improving the optimization efficiency of the blue-green space. BRIEF DESCRIPTION OF THE DRAWINGS
[0098] Figure 1 is a flow chart of the urban blue-green space optimization method of the present invention;
[0099] Figure 2 is a blue-green space distribution map of an urban area according to an implementation example of the present invention;
[0100] Figure 3 is the LST inversion result of the implementation case of the present invention;
[0101] Figure 4 is the NPP value of the embodiment of the present invention;
[0102] Figure 5 is the Ctotal value of the embodiment of the present invention;
[0103] Figure 6 is the runoff calculation result of the implementation case of the present invention;
[0104] Figure 7 It is the density of waterlogging points in the implementation case of the present invention;
[0105] Figure 8 is the runoff inversion result of the implementation case of the present invention;
[0106] Fig. 9 is the total ecological benefit of the implementation case of the present invention;
[0107] Fig.10 is the correlation coefficient of the morphological index of the implementation case of the present invention;
[0108] Fig.11 It is the importance of the embodiment index of the present invention;
[0109] Fig.12 is a curve diagram showing the change in the number of decision trees and model characteristics in the implementation cases of the present invention;
[0110] Fig.13 This is a scatter plot of actual value-predicted value of the implementation case of the present invention
[0111] Fig.14 It is a land classification map of the planning scheme for the implementation case of the present invention;
[0112] Fig.15 It is a blue-green space distribution map of the implementation case planning scheme of the present invention;
[0113] Fig.16 It is a blue-green space morphological characteristic diagram of the planning scheme of the implementation case of the present invention;
[0114] Fig.17 It is an ecological benefit diagram of the planning scheme for the implementation case of the present invention;
[0115] Fig.18 It is an ecological benefit zoning map of the planning scheme for the implementation case of the present invention;
[0116] Fig.19 It is a typical optimized area map of blue-green space of the planning scheme for the implementation case of the present invention;
[0117] Fig. 20It is an optimized cycle diagram based on the response model of the implementation case of the present invention;
[0118] Fig.21 It is an optimization scheme diagram of a typical area of the planning scheme for the implementation case of the present invention. DETAILED DESCRIPTION
[0119] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.
[0120] Taking Nanjing as an example, Figure 1 As shown, the method for optimizing the layout of urban blue-green space based on the morphological-ecological response model of the present invention specifically comprises the following steps:
[0121] S1, construct the urban blue-green space morphological characteristic index set and ecological benefit index set, and calculate them. The urban blue-green space morphological characteristic index set is shown in Table 1, and the urban blue-green space ecological benefit index set is shown in Table 2.
[0122] Table 1. Urban blue-green space morphological characteristic index set
[0123]
[0124] Table 2 Ecological benefit index set of urban blue-green space
[0125]
[0126]
[0127] S11, obtain multispectral remote sensing data, land use data, NDVI data, vegetation type data, daily radiation data, temperature data, precipitation data, urban soil type data, and urban surface cover data in the target area. The multispectral remote sensing data comes from the Landsat-8 satellite remote sensing image set (path-row: 120-38) published on the official website of the United States Geological Survey (https: / / www.usgs.gov / ). It is used for identifying urban land types, inverting surface temperature, simulating carbon sink data, identifying waterlogged areas, etc. The DEM data comes from the SRTM v3.1 (Shuttle Radar Topography Mission) data set measured and collected by the National Aeronautics and Space Administration (NASA, https: / / www.earthdata.nasa.gov / ), which is used for the calculation of three-dimensional morphological indicators. Urban land cover data (GLC_FCS30-2020) comes from the 30-resolution fine-scale land feature dataset (http: / / data.casearth.cn / sdo / detail) released by the Institute of Space Information Innovation, Chinese Academy of Sciences, and is used to simulate carbon sink data, calculate carbon storage, simulate runoff, calculate habitat quality, etc. Urban soil type data comes from the Chinese Soil Science Database (http: / / vdb3.soil.csdb.cn / ), including soil type distribution, soil moisture content and other data, which are used to simulate runoff. In addition, Nanjing meteorological data is also required when simulating carbon sink data and runoff. The data of Nanjing meteorological monitoring stations are mainly used and obtained through Kriging interpolation calculation.
[0128] S12, using supervised classification method to identify urban blue-green space based on multispectral remote sensing data. The forest land, grassland and cultivated land in the LUCC system are regarded as urban green space, water area (including lakes, rivers, wetlands, etc.) is classified as urban blue space, and construction land and unused land are gray space. The results are as follows: Figure 2 .
[0129] S13, using Fragstats software to calculate the shape index and concentration index of urban blue space and urban green space;
[0130] The average perimeter area ratio is calculated as follows:
[0131]
[0132] Among them, FRAC MN is the average perimeter-to-area ratio, N is the total number of samples, and p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample;
[0133] Among them, the calculation formula of average fractal dimension is:
[0134]
[0135] Among them, FRAC MN is the average fractal dimension, N is the total number of samples, p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample;
[0136] Secondly, the landscape shape index calculation formula is:
[0137]
[0138] Where LSI is the landscape shape index, e ik It refers to the total length of the landscape edge between the i-th sample and the k-th sample, and A is the area of the entire calculation area;
[0139] Again, the aggregation calculation formula is:
[0140]
[0141] Among them, AI is the degree of aggregation, g ij is the number of adjacent samples of the same type as the ijth sample, max(g ij ) is the maximum number of neighbors of the ijth sample;
[0142] Then, the spread calculation formula is:
[0143]
[0144] Among them, COHESION is the degree of spread, p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample, and Z is the number of all samples in the calculation area;
[0145] Secondly, the connectivity calculation formula is:
[0146]
[0147] Among them, CONNECT is the connectivity, c ijk is the number of possible connections between the ijth sample and the surrounding landscapes of the same type, n i is the total number of samples of this type of landscape in the entire calculation area;
[0148] Then, the separation degree is calculated as:
[0149]
[0150] Among them, DIVISION is the separation degree, a ij is the area of the ijth sample, and A is the area of the entire calculation area;
[0151] Secondly, the splitting degree calculation formula is:
[0152]
[0153] Among them, SPLIT is the splitting degree, a ij is the area of the ijth sample, and A is the area of the entire calculation area;
[0154] Again, the plaque number calculation formula is:
[0155] NP=n i
[0156] Where NP is the number of plaques, n i is the number of samples of a certain type of landscape in the calculation area;
[0157] Then, the plaque density is calculated as:
[0158]
[0159] Where PD is the plaque density, n i is the number of samples of a certain type of landscape in the calculation area, and A is the area of the entire calculation area;
[0160] Finally, before calculating the shape index and convergence index, the semivariogram function is used to calculate the most appropriate moving window size. The formula is as follows:
[0161]
[0162] Among them, X(i) and X(i+h) represent the values of the landscape pattern index X at position i and position i+j, h is the moving window size, N(h) represents the number of samples when the moving window size is h, and λ(h) represents the semivariance of X. Generally speaking, as the moving window increases, λ(h) will first increase and then decrease. When the trend of λ(h) tends to be flat, it means that the moving window size is more appropriate;
[0163] S14, the area index of urban blue space and urban green space is calculated by correcting the plane projection area used in the landscape pattern index, the proportion of patch type area in the landscape area, and the proportion of core area in the entire landscape area into a three-dimensional surface area calculated based on DEM data. The specific calculation formula is as follows:
[0164] First, the total three-dimensional area is calculated as:
[0165]
[0166] Where VCA is the total three-dimensional area, va ij is the three-dimensional surface area of the ijth sample;
[0167] Secondly, the calculation formula for the proportion of the largest plaque three-dimensional area is:
[0168]
[0169] Among them, VLPI is the proportion of the largest plaque three-dimensional area, va ij is the three-dimensional surface area of the ijth sample, VA is the three-dimensional surface area of the entire calculation area;
[0170] Finally, the calculation formula for the core plaque three-dimensional area ratio is:
[0171]
[0172] Among them, VCPLAND is the three-dimensional area ratio of the core plaque, The three-dimensional surface area of the core sample in the region is calculated for the ijth sample, and VA is the three-dimensional surface area of the entire calculation region. Use a moving window of the same size as the shape index and the convergence index;
[0173] S15, calculate the ecological benefits of the study area according to the ecological benefit indicator set, as follows
[0174] S151, based on the multispectral remote sensing data in S11, uses the Radiative Transfer Equation (RTE) to invert the LST value, such as Figure 3 .
[0175] The core principle is to subtract the brightness of the atmospheric radiation reflected upward and downward from the thermal infrared radiation brightness received by the Landsat-8 satellite thermal infrared sensor, and then convert the radiation brightness into temperature based on the Planck constant.
[0176]
[0177] Among them, B(T s ) represents the ground blackbody radiation, ε is the surface reflectivity, τ is the atmospheric transmittance, L λ is the thermal infrared radiation brightness, L u is the upward atmospheric radiance, L d is the downward atmospheric radiation brightness. When using remote sensing data for calculation, τ, L u and L dThe values of can be obtained by inputting the imaging time and the central longitude and latitude in the Atmospheric Correction Parameter Calculator (ACPC) provided by NASA. In addition, ε in the formula can be calculated by NDVI:
[0178]
[0179] Among them, ρ NIR is the near-infrared reflectivity, ρ RED is the red band reflectance. For the Landsat-8 satellite, the near infrared band is the 5th band and the red band is the 4th band. Finally, the value of LST can be obtained by the following formula:
[0180]
[0181] Among them, T s is the value of LST, K1 and K2 are pre-launch calibration constants
[103] . The corresponding values for the Landsat-8 satellite are: K1 = 774.89 W / (m 2 μmsr),K2=1321.08K.
[0182] S152, based on the multispectral remote sensing data, daily radiation data, temperature data, and precipitation data obtained in S11, the CASA model is used to calculate the net primary productivity (NPP), such as Figure 4 .
[0183] The principle is to consider the impact of solar radiation, temperature, and moisture on vegetation, calculate the effective radiation absorbed by vegetation that can be used for photosynthesis, and then convert it through the vegetation light energy utilization rate to estimate NPP. The effective radiation absorbed by vegetation can be obtained by the following formula:
[0184] APAR(x,t)=SOL(x,t)*FPAR(x,t)
[0185] Among them, APAR(x, t) represents the effective radiation absorbed by sample x in month t, and SOL(x, t) represents the total solar radiation received by sample x in month t, which is obtained through the weather station data of S11. FPAR(x, t) is the radiation component of photosynthesis of sample x in month t, which is calculated by the Normalized Difference Vegetation Index (NDVI) or Vegetation Cover (SR) in S151. FPAR is calculated by combining NDVI and SR:
[0186] FPAR(x, t) = 0.5FPAR NDVI (x,t)+0.5FPAR SR (x, t)
[0187]
[0188] In addition, the light energy utilization rate can be obtained by the following formula:
[0189] ε(x, t) = T ε1 (x,t)*T ε2 (x,t)*W ε (t, t)*ε max
[0190] Among them, ε(x, t) is the actual light energy utilization rate of sample x in month t. ε1 (x, t) and T ε2 (x, t) are the stress coefficients of high temperature and low temperature on the utilization of vegetation light energy, W ε (x, t) is the stress coefficient of water on vegetation light energy utilization, ε max It is the maximum light energy utilization of vegetation under ideal conditions without restrictions.
[0191] Finally, the monthly NPP is obtained by multiplying the effective radiation APAR with the vegetation light energy utilization rate ε. To obtain the NPP data for one year, it is necessary to accumulate the results month by month:
[0192] NPP(x,t)=APAR(x,t)*ε(x,t)
[0193] S153, use the Carbon Storage and Sequestration module provided by the InVEST (Integrated Valuation of Ecosystem Services and Trade-offs) platform to estimate the carbon storage in the target area, such as Figure 5 .
[0194] The calculation formula is as follows:
[0195]
[0196] Among them, C total is the total carbon storage, C i is the total carbon density of the i-th type of land, S i is the total area of the i-th type of land. i It can be calculated by the following formula:
[0197] C i =(C above +C below +C soil +C dead )*S
[0198] Among them, C above , C below , C soil and C dead It is divided into aboveground carbon density, underground carbon density, soil carbon density and dead organic matter carbon density per unit area. This case refers to the carbon density values of Shanghai and Wuhan, and corrects the carbon density values based on the temperature difference between the two cities and Nanjing:
[0199] C BT =28*T+398
[0200]
[0201] Where T is the average annual temperature of the region, K BT is the carbon density correction factor based on temperature difference, C' BT is the carbon density data of the central urban area of Nanjing, C' B ' T is the carbon density data of Wuhan or Shanghai, and the annual average temperature of the region is obtained through the urban meteorological station. Finally, the carbon density parameters of the central urban area of Nanjing are shown in Table 3.
[0202] Table 3 Calculation parameters of carbon density in the central urban area of Nanjing
[0203]
[0204]
[0205] S154, the surface runoff volume of the target area is calculated according to the SCS-CN runoff model (Soil Conservation Service Curve Number, runoff curve model), such as Figure 6 .
[0206] Among them, the SCS-CN model was developed by the United States Soil Protection Agency (US EPA). The core principle is that under the premise of water balance, it is assumed that the ratio of runoff to maximum runoff is equal to the ratio of infiltration to maximum possible retention, and it is assumed that the initial rainfall loss is proportional to the potential maximum precipitation loss, and the following SCS-CN classic formula is obtained:
[0207]
[0208] Among them, Q is the runoff, P is the average annual rainfall, and λS represents the initial rainfall loss. λ is a parameter mainly affected by regional geological and climatic conditions. According to previous studies, the λ value of Nanjing's central urban area is 0.2. The calculation formula of S is:
[0209]
[0210] Among them, S is the maximum possible retention capacity, and CN is a dimensionless parameter determined by soil type, land use, surface conditions and previous soil moisture content. The runoff or infiltration of rainwater during rainfall is determined by the soil characteristics before rainfall. The CN values of this case are shown in Table 4. In the table, soil type A represents sandy soil, loamy sandy soil and sandy loam, soil type B represents loam and silt loam, soil type C is sandy clay loam, and soil type D represents clay loam, silt clay loam, sandy clay, silt clay and clay.
[0211] Table 4 CN values in the central urban area of Nanjing
[0212]
[0213] S155, based on the remote sensing images of continuous sunny days and after rain in S11 multispectral remote sensing data, remove the permanent water area in the target area in GIS software to obtain seasonal water bodies, and then obtain the waterlogged area after subtraction. Use the "point density analysis" tool to obtain the waterlogging point density to characterize the waterlogging risk in the target area, such as Figure 7 .
[0214] S156, based on the urban land cover data and multispectral remote sensing data of S11, the habitat quality of the target area is calculated using the HabitatQuality module of the InVEST platform, such as Figure 8 .
[0215] The core principle is to use the original habitat suitability coefficient of the current land use situation, minus factors such as topography, soil, and human activities that may threaten biodiversity, to calculate habitat quality:
[0216]
[0217] Among them, HQ xj is the habitat quality of the jth sample of the jth land use type, H j is the habitat suitability coefficient of the jth land use type, and the value is referenced in Table 5. z is the scale constant, with a value of 2.5, and k is the half-saturation constant. Finally, D xy It represents the degree of habitat degradation and is the main factor affecting habitat quality. The calculation formula is:
[0218]
[0219] Among them, R is the number of factors threatening biodiversity, Y r is the number of grids of threat factor r in the calculation area, ω ris the weight of threat factor r, refer to Table 6 for the value, r y is the threat level of grid y, i rxy is r y Threat level to habitat grid x, β x represents the accessibility of habitat grid x, S jr is the sensitivity of the jth land use type to the threat factor r, and the value is given in Table 5. rxy It can be calculated by the distance between grids and the distance affected by the threat factor:
[0220]
[0221] Among them, d xy is the straight-line distance between grid x and grid y, d rmax is the maximum impact distance of threat factor r, and its value is referenced in Table 6. rxy When calculating, there are two calculation formulas: linear attenuation and exponential attenuation. When calculating the model, the calculation formula is selected according to the characteristics of the threat factor r, and the value is referenced in Table 6.
[0222] Table 5 Habitat characteristics of various land use types in the central urban area of Nanjing
[0223]
[0224] Table 6 Characteristics of threat factors in the central urban area of Nanjing
[0225]
[0226] S16, based on the calculation results of each ecological benefit, normalize and standardize each indicator.
[0227] For positive indicators:
[0228]
[0229] Among them, X i ' is the normalized value of the i-th ecological benefit index, X i is the original value of the i-th ecological benefit index, X imin is the minimum value of the i-th ecological benefit index, and X imax It is the maximum value of the i-th ecological benefit indicator.
[0230] For negative indicators:
[0231]
[0232] S17, the weights of the four ecological benefit indicators of stormwater resilience, cooling benefit, carbon sink benefit, and biodiversity are considered to be the same value and weighted equally. For the benefits indicated by two indicators, such as carbon sink benefit and stormwater management benefit, it is necessary to weight the two indicators equally and then weight them equally again to calculate the total ecological benefit (ECO), such as Fig. 9 .
[0233] The calculation result is:
[0234] ECO=0.25LST'+0.25(0.5NPP'+0.5C total ')+0.25(0.5Q'+0.5F n ')+0.25HQ'
[0235] S2, using the random forest method of machine learning, the urban blue-green space "morphology-ecology" response model is obtained by training the urban blue-green space morphological characteristic index set and ecological benefit index set in the target area.
[0236] S21, in SPSS software, the Spearman correlation coefficient was used to calculate the correlation between morphological characteristic indicators. The specific formula is:
[0237]
[0238] Among them, P S is the correlation coefficient between the two morphological indicators, d i is the sequence difference of the ith data, and n is the number of samples for calculation. The calculation results are as follows Fig.10 Generally speaking, when the absolute value of the correlation coefficient is less than 0.3, it is considered that there is no correlation between the indicators; when the absolute value of the correlation coefficient is greater than 0.3 and less than 0.5, it is considered that there is a weak correlation between the indicators; when the absolute value of the correlation coefficient is greater than 0.5 and less than 0.8, it indicates that there is a moderate correlation between the indicators; and when the absolute value of the correlation coefficient is greater than 0.8, it indicates that there is a strong correlation between the indicators.
[0239] S22, use Bootstrap sampling method to randomly sample the training set to construct the training set, and the data that does not appear in the training set is the test set. Calculate the average test error of all decision trees and calculate the average rate of decrease in accuracy:
[0240]
[0241] Where, MDA is the average decrease rate of accuracy, which is the main factor to measure the importance of the indicator. n is the total number of decision trees, error t is the out-of-bag error of decision tree t when a feature is not added, error' t is the out-of-bag error of decision tree t after adding a certain feature.
[0242] S23, write a program in the Matlab software platform to perform random forest regression and obtain the results of the importance of morphological characteristic indicators. Screen out the indicators with an importance less than 0.3. Then, among the remaining indicators, screen out the indicators with correlation coefficients greater than 0.8 and less importance, and the remaining indicators are the key morphological characteristic indicators.
[0243] S24, based on the calculation results of S23, select the appropriate number of decision trees through the three model features of the deterministic coefficient R2, the mean absolute error MAE, and the model running time T. Adjust the three parameters of the maximum depth of the decision tree (MaxNumSplits), the minimum leaf tree (MinLeafSize), and the number of randomly generated seeds (Random Number Generator) to fit the response model.
[0244] First, we select the appropriate number of decision trees by determining the coefficient R2, mean absolute error MAE, and model running time T. The results are as follows: Fig.12 .
[0245] Among them, when the number of decision trees is less than 50, R2 is small and does not exceed 0.5, indicating that the model is underfitting. As numTrees increases, R2 tends to stabilize after the number of decision trees reaches 50, basically maintaining at around 0.7, indicating that the model has reached a certain accuracy. When numTrees is 100, R2 reaches its peak. MAE is the mean absolute error of the model. The smaller the MAE, the more accurate the model. The curve of the number of decision trees and MAE shows that MAE has a relatively stable value only after numTrees is greater than 50. Before that, MAE shows a gradual downward trend. In addition, in terms of computing time, the larger the numTrees, the longer the computing time, and the lower the model efficiency. After numTrees is greater than 100, the computing time growth trend increases significantly. Taking all factors into consideration, this paper determines that the optimal number of decision trees is 100.
[0246] Secondly, MaxNumSplits is the maximum depth of the decision tree. It limits the maximum depth to prevent overfitting of the model. The sample size in this case exceeds 30,000, so MaxNumSplits is set to 10.
[0247] Again, MinLeafSize is the minimum cotyledon tree. Selecting a moderate MinLeafSize can enable the model to capture subtle features of the data and be more stable. After analysis, MinLeafSize is set to 8 in this case.
[0248] Finally, RNG (Random Number Generator) is a randomly generated seed number that does not affect the accuracy of the model. Generally, an integer can be set. In this case, an uncommon prime number 43 is used as the value of rng.
[0249] S25, based on the fitting results of S24, the model accuracy is verified by R2 and Mean Absolute Error. The results are as follows Fig.13 .
[0250] S3, selected the Nanjing City Master Plan in the Nanjing City Land and Space Master Plan (2021-2035) and extracted the Nanjing Central City Master Plan (such as Fig.14 ) as the basic data, and calculate the morphological characteristic indicators in the target area. The specific steps include:
[0251] S31, in the central urban planning scheme of Nanjing, land is divided into 29 types, which correspond one by one to the six major land use types in the LUCC classification based on the meaning of each type of land use. Among them, 21 types of land, including first and second type residential land, administrative office land, cultural facilities land, general education facilities land, sports land, medical and health land, etc., all belong to the category of construction land. Park green space, protective green space, suburban green space, and ecological green space correspond to forest land and grassland. Agricultural land corresponds to cultivated land in LUCC, and water area corresponds to water body in LUCC. It is worth noting that there is no land type corresponding to unused land in LUCC in future urban planning, because LUCC mainly starts from the current status of land use. Some undeveloped bare land is called unused land, and urban planning itself is to guide land development, so there is no unused space. Finally, the same as the blue-green space identification method of the current land use, forest land, grassland, and cultivated land are classified as urban green space, water bodies are regarded as urban blue space, and construction space is classified as gray space. The results are as follows Fig.15 .
[0252] S32, calculate key morphological characteristic indicators in GIS software and Fragstats software. In order to ensure comparability with the existing blue-green space morphological characteristics, a 500m moving window is also used for calculation, including the total three-dimensional area of the blue space, the spread, and the three-dimensional area proportion of the largest patch of green space, the aggregation, and the landscape shape index. The results are as follows Fig.16 .
[0253] S4, based on the morphological characteristic index results of S3 and the urban blue-green space "morphology-ecology" response model of S2, identifies typical areas where ecological benefits need to be improved and obtains ecological benefit optimization targets.
[0254] S41, based on the calculation results of the urban blue-green space morphological characteristic indicators in S3 and the morphological-ecological response model, predict the ecological benefits of the target area. Fig.17 The color changes in the figure represent the level of ecological benefits. Green colors indicate excellent ecological benefits, yellow colors indicate good ecological benefits, and red or light pink areas may have low ecological benefits.
[0255] S42, use the natural breakpoint method to evaluate the ecological benefits of the planning scheme and identify the areas to be optimized. The ecological benefits after planning are divided into five ecological zones: excellent, relatively good, good, average, and poor. Fig.18 .
[0256] S43, based on the ecological benefit classification of the target area, the area with poor ecological benefit accounting for more than 90% is the area to be optimized. In this case, we identified the following three areas according to the above two standards, namely Xinjiekou area, Baijia Lake area, and the development zone on the west side of Jiangbei New District as typical areas for subsequent optimization of blue-green space forms, such as Figure 7 There is no area or shape restriction for the area to be optimized, but in order to take the surrounding conditions into consideration, a 500m buffer zone needs to be set after the range is selected.
[0257] S5, construct an optimization iteration cycle based on the “morphology-ecology” response model of urban blue-green space, output the optimized urban blue-green space morphological feature series, and screen them, such as Fig.19 Then the Landscape Generator (LG) platform was used to obtain the optimized urban blue-green space distribution. The results are as follows Fig. 20 .
[0258] In actual cases, the morphological characteristic index adjustment criteria should be formulated to save computing efficiency. For urban green space, when the VLPI value of green space is less than 16%, the proportion of urban green space area should be increased as much as possible; when the VLPI value of green space is between 16% and 36%, the AI value of green space should be increased. In addition, the LSI value of green space should be minimized.
[0259] S51, write a program in the Matlab software platform, perform optimization iterations based on the morphological-ecological response model trained in S2, set the maximum number of iterative adjustments of the morphological characteristic indicators to 5000 times, and the output standard is to evaluate whether the new morphological characteristic series is better than the ecological benefits of the original planning scheme after each adjustment. If the ecological benefits are better, output the corresponding morphological characteristic series to obtain the morphological optimization results that meet the ecological goal improvement.
[0260] S52, before outputting the final optimized morphological feature sequence, it is necessary to make a choice among the various schemes: retain those with higher ecological benefits, and retain those that are closer to the morphological features before optimization. Since the morphological feature sequence of the pixel points included in each optimization scheme is too large, only sample data is shown here, as shown in Table 7.
[0261] Table 7 Example of blue-green spatial morphological feature series for the optimization scheme
[0262]
[0263] It should be noted that the above content only illustrates the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. For ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications all fall within the protection scope of the claims of the present invention.
Claims
1. Urban blue-green space layout optimization method based on morphological-ecological response model, It is characterized in that The following steps are involved: S1, construct the urban blue-green space morphological characteristic index set and ecological benefit index set, and calculate them; S2, using the random forest method of machine learning, the urban blue-green space "morphology-ecology" response model is obtained by training the urban blue-green space morphological characteristic index set and ecological benefit index set in the target area; S3, obtain the planning scheme of the target area, translate its land use data, and calculate the internal morphological characteristic indicators; S4, based on the results of the morphological characteristic indicators of S3 and the urban blue-green space "morphology-ecology" response model of S2, identify typical areas where ecological benefits need to be improved and obtain the ecological benefit optimization target; S5, construct an optimization iterative cycle based on the "morphology-ecology" response model of urban blue-green space, output the optimized urban blue-green space morphological feature series, and screen them, and then use the landscape generator platform to obtain the optimized urban blue-green space distribution.
2. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that In S1, the urban blue-green space morphological characteristic indicator set is composed of urban green space morphological characteristic indicators and urban blue space morphological characteristic indicators. Both types of indicators include area indicators, shape indicators, and aggregation indicators; among them, the area indicators are the total three-dimensional area, the proportion of the three-dimensional area of the largest patch, and the three-dimensional area ratio of the core patch; the shape indicators are the average perimeter area ratio, the average fractal dimension, and the landscape shape index; the aggregation indicators are aggregation degree, spread degree, connectivity, separation degree, split degree, number of patches, and patch density.
3. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that In S1, the target layer of the ecological benefit evaluation index set is ecological benefit; the criterion layer includes stormwater resilience, cooling benefit, carbon sink benefit, and biodiversity; the indicator layer includes surface runoff, waterlogging point density, urban surface temperature, net primary productivity, carbon storage, and habitat quality.
4. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that In S1, the calculation steps of the urban blue-green space morphological characteristic index set and the ecological benefit index set are as follows: S11, obtaining multispectral remote sensing data, land use data, NDVI data, vegetation type data, daily radiation data, temperature data, precipitation data, urban soil type data, and urban land cover data in the target area; S12, identification of urban blue-green spaces using supervised classification based on multispectral remote sensing data; S13, using Fragstats software to calculate the shape index and concentration index of urban blue space and urban green space; The average perimeter area ratio is calculated as follows: Among them, PARA MN is the average perimeter-to-area ratio, N is the total number of samples, and p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample; Among them, the calculation formula of average fractal dimension is: Among them, FRAC MN is the average fractal dimension, N is the total number of samples, p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample; Secondly, the landscape shape index calculation formula is: Where LSI is the landscape shape index, e ik It refers to the total length of the landscape edge between the i-th sample and the k-th sample, and A is the area of the entire calculation area; Again, the aggregation calculation formula is: Among them, AI is the degree of aggregation, g ij is the number of adjacent samples of the same type as the ijth sample, max(g ij ) is the maximum number of neighbors of the ijth sample; Then, the spread calculation formula is: Among them, COHESION is the degree of spread, p ij is the perimeter of the ijth sample, a ij is the area of the ijth sample, and Z is the number of all samples in the calculation area; Secondly, the connectivity calculation formula is: Among them, CONNECT is the connectivity, c ijk is the number of possible connections between the ijth sample and the surrounding landscapes of the same type, n i is the total number of samples of this type of landscape in the entire calculation area; Then, the resolution is calculated as: Among them, DIVISION is the separation degree, a ij is the area of the ijth sample, and A is the area of the entire calculation area; Secondly, the splitting degree calculation formula is: Among them, SPLIT is the splitting degree, a ij is the area of the ijth sample, and A is the area of the entire calculation area; Again, the plaque number calculation formula is: NP=n i Where NP is the number of plaques, n i is the number of samples of a certain type of landscape in the calculation area; Then, the plaque density is calculated as: Where PD is the plaque density, n i is the number of samples of a certain type of landscape in the calculation area, and A is the area of the entire calculation area; Finally, before calculating the shape index and convergence index, the semivariogram function is used to calculate the most appropriate moving window size. The formula is as follows: Among them, X(i) and X(i+h) represent the values of the landscape pattern index X at position i and position i+h, h is the moving window size, N(h) represents the number of samples when the moving window size is h, and λ(h) represents the semivariance of X. Generally speaking, as the moving window increases, λ(h) will first increase and then decrease. When the trend of λ(h) tends to be flat, it means that the moving window size is more appropriate; S14, the area index of urban blue space and urban green space is calculated by correcting the plane projection area used in the landscape pattern index, the proportion of patch type area in the landscape area, and the proportion of core area in the entire landscape area into a three-dimensional surface area calculated based on DEM data. The specific calculation formula is as follows: First, the total three-dimensional area is calculated as: Where VCA is the total three-dimensional area, va ij is the three-dimensional surface area of the ijth sample; Secondly, the calculation formula for the proportion of the largest plaque three-dimensional area is: Among them, VLPI is the proportion of the largest plaque three-dimensional area, va ij is the three-dimensional surface area of the ijth sample, VA is the three-dimensional surface area of the entire calculation area; Finally, the calculation formula for the core plaque three-dimensional area ratio is: Among them, VCPLAND is the three-dimensional area ratio of the core plaque, The three-dimensional surface area of the core sample in the region is calculated for the ijth sample, and VA is the three-dimensional surface area of the entire calculation region. Use a moving window of the same size as the shape index and the convergence index; S15, calculating the ecological benefits of the target area according to the ecological benefit indicator set; S16, based on the calculation results of each ecological benefit indicator, normalize and standardize each indicator; For positive indicators: Among them, X i ' is the normalized value of the i-th ecological benefit index, X i is the original value of the i-th ecological benefit index, X imin is the minimum value of the i-th ecological benefit index, and X imax is the maximum value of the i-th ecological benefit index; For negative indicators: S17, the weights of the four types of ecological benefit indicators, namely, stormwater resilience, cooling benefit, carbon sequestration benefit, and biodiversity, are considered to be equally weighted with the same values; among them, the two indicators indicating carbon sequestration benefit and stormwater management benefit are equally weighted, and then the total ecological benefit is calculated by equally weighting again.
5. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 4, It is characterized in that S15 calculates the total ecological benefits of the target area. The specific steps are: S151, based on the multispectral remote sensing data in S11, uses the transmission equation to invert the LST value; S152, based on the multispectral remote sensing data, daily radiation data, temperature data, and precipitation data obtained in S11, the CASA model is used to calculate the net primary productivity; S153, estimate the carbon storage in the target area using the Carbon Storage and Sequestration module provided by the InVEST platform; S154, calculate the surface runoff volume of the target area according to the SCS-CN runoff generation model; S155, based on the remote sensing images of continuous sunny days and after rain in S11 multispectral remote sensing data, mosaic the flood inundation range in GIS software; use the "point density analysis" tool to obtain the density of waterlogging points to characterize the waterlogging risk in the target area; S156, based on the urban land cover data and multispectral remote sensing data of S11, the habitat quality of the target area was calculated using the HabitatQuality module of the InVEST platform.
6. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that The specific steps of training and obtaining the urban blue-green space "morphology-ecology" response model in S2 are: S21, in SPSS software, the Spearman correlation coefficient was used to calculate the correlation between morphological characteristic indicators. The specific formula is: Among them, P S is the correlation coefficient between the two morphological indicators, d i is the sequence difference of the i-th data, and n is the number of calculated samples; S22, use Bootstrap sampling method to randomly sample the training set to construct the training set, and the data that does not appear in the training set is the test set. Calculate the average test error of all decision trees and calculate the average rate of decrease in accuracy: Where MDA is the average decrease rate of accuracy, which is the main factor to measure the importance of the indicator; n is the total number of decision trees, error t is the out-of-bag error of decision tree t when a feature is not added, error' t is the out-of-bag error of decision tree t after adding a certain feature; S23, write a program in Matlab software to perform random forest regression and obtain the results of the importance of morphological characteristic indicators; screen out the indicators with an importance less than 0.3; then screen out the indicators with correlation coefficients greater than 0.8 and less importance from the remaining indicators, and the remaining indicators are key morphological characteristic indicators; S24, based on the calculation results of S23, select the appropriate number of decision trees through the three model characteristics of the determination coefficient R2, the mean absolute error MAE, and the model running time T; adjust the three parameters of the maximum depth of the decision tree, the minimum cotyledon tree, and the number of randomly generated seeds to fit the response model; S25, based on the fitting results of S24, the model accuracy is verified by R2 and mean absolute error.
7. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that The specific steps for calculating the morphological characteristic indicators in the target area planning scheme in S3 are: S31, identifying blue-green space, based on land use / land cover classification, translating the planned land types, merging forest land, grassland, and cultivated land into urban green space, and taking water bodies as urban blue space; S32, calculating the morphological characteristic indicators in the target area, including the total three-dimensional area of the blue space, the spread, and the proportion of the three-dimensional area of the largest patch of green space, the degree of aggregation, and the landscape shape index.
8. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that The steps for identifying typical areas for ecological benefit improvement in S4 are: S41, predict the ecological benefits of the target area based on the calculation results of the urban blue-green space morphological characteristic indicators in S3 and the morphological-ecological response model; S42, using the natural breakpoint method to classify the ecological benefits of the target area into five levels: excellent, relatively good, good, average, and poor; S43, based on the ecological benefit classification of the target areas, areas where the proportion of areas with poor ecological benefits exceeds 90% are areas to be optimized.
9. The urban blue-green space layout optimization method based on the morphological-ecological response model according to claim 1, It is characterized in that In S5, the steps to construct the morphology optimization loop are: S51, write a program in Matlab software, perform optimization iteration based on the morphological-ecological response model trained in S2, set the maximum number of iterative adjustments of morphological characteristic indicators to 5000 times, and the output standard is to evaluate whether the new morphological characteristic series is better than the original planning scheme in ecological benefits after each adjustment. If the ecological benefits are better, output the corresponding morphological characteristic series to obtain the morphological optimization results that meet the ecological goal improvement; S52, before outputting the final optimized morphological feature series, it is necessary to make trade-offs among the various options: retain those with higher ecological benefits and retain those that are closer to the morphological features before optimization.
Citation Information
Cited By
GBDT model-based reduction method for aerosol pollution caused by high-score urban green land form
CN119692866A
Method and system for measuring and calculating coupling degree of green space structure and built environment
CN119989165A
Intelligent evaluation method and system for green space supplement cost based on vacancy probability
CN119990516A
Basin ecological health threshold evaluation method based on landscape fragmentation and microbial network modularity
CN120015131A
Assessment Method for Watershed Ecological Health Threshold Based on Landscape Fragmentation and Microbial Network Modularity
CN120015131B
