Carbon sink measurement system and method for multi-type urban green space ecosystems

Abstract

The present invention discloses a carbon sink measurement system and method for multi-type urban green space ecosystems, belongs to the technical field of carbon sink measurement, solves the problem of inaccurate evaluation results caused by uneven data distribution or insufficient sample size when the existing method uses a feature selection algorithm to extract characteristic factors. The method comprises: collecting time-series urban green space carbon sink correlation data, determining characteristic factors that affect the carbon sink amount in a carbon sink pretreatment concentration based on a characteristic factor extraction model, analyzing the characteristic factor set in combination with a carbon storage method and a characteristic factor extraction model, and calculating the biomass of trees and shrubs in urban green spaces; the present invention constructs a characteristic factor extraction model by combining a GBR model with a global sensitivity analysis, thereby accurately determining the characteristic factors that affect the carbon sink amount in a carbon sink pretreatment concentration, reducing uncertainty in carbon sink estimation, and ensuring the accuracy of evaluation results when data distribution is uneven or the sample size is insufficient.

Description

Technical Field

[0001] The present invention belongs to the technical field of carbon sink measurement, and in particular relates to a carbon sink measurement system and method for various types of urban green space ecosystems. Background Art

[0002] Carbon sink is an important indicator for evaluating the service function of urban ecosystems. The current focus of urban low-carbon development is to integrate low-carbon goals with ecological concepts, with carbon emission constraints as the premise. Urban green space is an important part of the urban ecosystem and the only direct carbon sink in the city, playing a great role in carbon fixation. Therefore, quantitative analysis of the carbon sink benefits of urban green space can not only promote the development of carbon sink estimation, but also promote the development goals of low-carbon cities.

[0003] Chinese patent CN118643281B discloses an AI-based forest carbon sink prediction system, including a data acquisition module, a carbon sink data enhancement module, a feature selection module, an adaptive environmental factor adjustment module, a forest carbon sink prediction module, and a prediction result analysis module. The data acquisition module collects multi-source data related to forest carbon sinks; the carbon sink data enhancement module enhances the sample data set of multi-source data; the feature selection module extracts features that affect carbon sink prediction; the adaptive environmental factor adjustment module dynamically adjusts real-time environmental factors; the forest carbon sink prediction module realizes the prediction of forest carbon sinks; the prediction result analysis module analyzes the prediction results output by the forest carbon sink prediction model, but the existing method uses feature selection algorithm to extract feature factors. If the data distribution is uneven or the sample size is insufficient, the evaluation results are inaccurate, resulting in insufficient stability and reliability of feature selection. In response to the above problems, we propose a carbon sink measurement system and method for multi-type urban green space ecosystems. Summary of the invention

[0004] The purpose of the present invention is to address the shortcomings of the prior art and provide a carbon sink measurement system and method for multiple types of urban green space ecosystems, thereby solving the problem that when the existing method uses a feature selection algorithm to extract characteristic factors, the evaluation results are inaccurate due to uneven data distribution or insufficient sample size, resulting in insufficient stability and reliability of feature selection.

[0005] The present invention is implemented by a method for measuring carbon sinks in multi-type urban green space ecosystems, and the method for measuring carbon sinks in multi-type urban green space ecosystems comprises:

[0006] Collecting time-series urban green space carbon sink associated data, wherein the carbon sink associated data includes green space ecological data, green space remote sensing data, and urban meteorological data, and preprocessing the carbon sink associated data to obtain a carbon sink preprocessing set;

[0007] Obtain a carbon sink preprocessing set, divide the carbon sink preprocessing set into a training set and a test set, build a characteristic factor extraction model by combining the GBR model and global sensitivity analysis, determine the characteristic factors that affect the carbon sink amount in the carbon sink preprocessing set based on the characteristic factor extraction model, and output the characteristic factor set;

[0008] Load the characteristic factor set, take the characteristic factor set as input, combine the carbon storage method and the characteristic factor extraction model to analyze the characteristic factor set, calculate the biomass of trees and shrubs in urban green spaces, and calculate the total carbon sink in urban green spaces based on the biomass of trees and shrubs in urban green spaces.

[0009] Preferably, the method for preprocessing carbon sink correlation data specifically includes:

[0010] Obtain carbon sink-related data, process missing values ​​and abnormal values ​​of the carbon sink-related data, and obtain a normalized set;

[0011] Extract green space remote sensing data from carbon sink association data, calculate the vegetation index based on hyperspectral influence in green space remote sensing data, and determine the vegetation type corresponding to green space remote sensing data through the vegetation index based on hyperspectral influence;

[0012] Among them, the vegetation index based on hyperspectral influence is calculated by the following formula:

[0013] (1)

[0014] in, represents the reflectivity, The subscript value of represents the wavelength corresponding to the reflectivity. represents the enhancement coefficient;

[0015] Extract green space ecological data from carbon sink-related data, identify vegetation types, vegetation breast diameter, and vegetation tree height in green space ecological data based on vegetation types corresponding to green space remote sensing data, and extract underground stratum depth of green space, soil light group content corresponding to stratum depth, soil recombinant content, light group organic carbon content, and recombinant organic carbon content in combination with carbon sink-related data;

[0016] Based on the soil light group content, soil recombinant content, light group organic carbon content, and recombinant organic carbon content, the light group organic carbon distribution ratio and recombinant organic carbon distribution ratio are determined;

[0017] Obtain the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio, and calculate the soil organic carbon density storage and organic carbon component sensitivity index based on the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio;

[0018] The soil organic carbon density storage is calculated by the following formula:

[0019] (2)

[0020] (3)

[0021] in, represents the soil organic carbon density storage, represents the total organic carbon content of soil. is the soil bulk density, is the soil layer thickness, They are the light group organic carbon distribution ratio and the heavy group organic carbon distribution ratio, It indicates the volume ratio of gravel with diameter greater than 2 mm in the soil, and represents the sensitivity index of organic carbon components, Indicates the minimum value of organic carbon content in various green space types. It indicates the range of total organic carbon and component contents of various green space types;

[0022] The vegetation index based on hyperspectral influence, light group organic carbon allocation ratio, recombinant organic carbon allocation ratio, soil organic carbon density storage and organic carbon component sensitivity index are integrated as the carbon sink pretreatment set.

[0023] Preferably, the method of constructing a characteristic factor extraction model by combining the GBR model and global sensitivity analysis specifically includes:

[0024] The GBR model is used as the initial model of the feature factor extraction model, wherein the initial model includes a model fusion module, a parameter setting module, a regularization module, and an evaluation module;

[0025] Freeze the evaluation module of the initial model, and improve the evaluation module in combination with the global sensitivity analysis to obtain a sensitivity analysis module;

[0026] The model fusion module performs weighted fusion on the sensitivity analysis module and the parameter setting module. The parameter setting module sets the learning rate, number of decision trees, maximum depth of decision trees, leaf nodes of the initial model, and defines the loss function of the initial model.

[0027] Get the training set, use the training set as input, execute the initial model, iteratively train the initial model through the training set, calculate the loss function and gradient of the current model in each round of iterative training, build a new decision tree based on the negative gradient direction, and add the new decision tree to the initial model until the loss function converges;

[0028] Obtain a test set, divide the test set into multiple test subsets, use the test subsets to verify the accuracy of the initial model, and output the feature factor extraction model if the accuracy of the initial model meets the preset accuracy threshold.

[0029] Preferably, the method for determining the characteristic factors affecting the carbon sink quantity in the carbon sink pretreatment concentration based on the characteristic factor extraction model specifically includes:

[0030] Load the carbon sink preprocessing set, extract the initial characteristic factors that affect the carbon sink amount in the carbon sink preprocessing set, and define the initial weights of the initial characteristic factors based on orthogonal initialization;

[0031] Obtain initial characteristic factors, transform and combine the initial characteristic factors based on the orthogonal matrix to generate secondary characteristic factors, calculate the weights of the secondary characteristic factors, and merge the initial characteristic factors and the secondary characteristic factors into a characteristic factor set;

[0032] Among them, the weight of the secondary characteristic factor is calculated by the following formula:

[0033] (4)

[0034] in, is the weight of the secondary characteristic factor, represents the weight of the initial feature factor, are the maximum and minimum weights of the initial characteristic factors associated with the secondary characteristic factors, represents the number of initial characteristic factors;

[0035] Based on the characteristic factor extraction model, the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment set are determined, and the characteristic factor set is output.

[0036] Preferably, the method for determining the characteristic factors affecting the carbon sink amount in the carbon sink pretreatment concentration based on the characteristic factor extraction model specifically further includes:

[0037] Acquire a characteristic factor set, and the characteristic factor extraction model selects characteristic factors whose signal-to-noise ratio coefficients are lower than a preset noise coefficient based on the signal-to-noise ratio coefficients of the characteristic factor set to generate a low signal-to-noise ratio factor set;

[0038] The low signal-to-noise ratio factor set is loaded, and the sensitivity analysis module outputs the global sensitivity index of the characteristic factor based on the low signal-to-noise ratio factor set and the signal-to-noise ratio coefficient of the low signal-to-noise ratio factor;

[0039] A preset sensitivity threshold is used to determine whether the global sensitivity index of the characteristic factor exceeds the preset sensitivity threshold. If it exceeds the sensitivity threshold, the current characteristic factor is determined to be a characteristic factor that affects the carbon sink amount;

[0040] If the preset sensitivity threshold is not exceeded, the current characteristic factor is discarded.

[0041] Preferably, the signal-to-noise ratio coefficient of the characteristic factor set is calculated by the following formula:

[0042] (5)

[0043] (6)

[0044] in, is the signal-to-noise ratio coefficient, Represents the input feature factor data, represents the noise corresponding to the characteristic factor data, is the data volume of the feature factor set, represents the noise factor, represents the urban green space migration coefficient, represents the conversion coefficient of the net ecosystem productivity of urban green space, represents the noise constant, represents the total carbon emissions from urban green spaces;

[0045] The global sensitivity index of the characteristic factor is calculated by the following formula:

[0046] (7)

[0047] in, represents the global sensitivity index of the characteristic factor, represents the variance operator, represents the sensitivity response function, represents the characteristic factor variance;

[0048] (8)

[0049] in, Represents the input feature factor data, Represents the mean of the characteristic factor data.

[0050] Preferably, the method of analyzing the characteristic factor set by combining the carbon storage method and the characteristic factor extraction model specifically includes:

[0051] Obtain the characteristic factor set, and define the biomass allometric growth equation by combining the carbon storage method and the characteristic factor extraction model;

[0052] The biomass expansion factor and biomass root-to-shoot ratio were calculated based on the biomass allometric growth equation;

[0053] Obtain biomass expansion factor and biomass root-to-shoot ratio, and calculate the biomass of trees and shrubs in urban green spaces based on the biomass expansion factor and biomass root-to-shoot ratio;

[0054] Among them, the biomass allometric growth equation is expressed as:

[0055] (9)

[0056] (10)

[0057] in, is the output value of the biomass allometric growth equation, represents the integrated weight of the characteristic factors in the characteristic factor set, are the number of trees and shrubs in green areas, represents the number of characteristic factors, represents the mean value of the global sensitivity index of the characteristic factor, represents the biomass expansion factor;

[0058] (11)

[0059] in, represents the biomass of trees in urban green space, is the root-to-shoot ratio of biomass, Indicates the height of trees in green areas;

[0060] Obtain the biomass of trees and shrubs in urban green spaces, define a biomass coupling model by combining the carbon storage method and the characteristic factor extraction model, and calculate the carbon sink in urban green spaces based on the biomass coupling model;

[0061] The biomass coupling model is defined as:

[0062] (12)

[0063] in, represents the total output value of urban green space carbon sink, is the biomass loss factor of trees and shrubs in urban green space, Represents the kernel function coupling coefficient of the biomass coupling model.

[0064] On the other hand, the present invention also provides a carbon sink measurement system for multi-type urban green space ecosystems, and the carbon sink measurement system for multi-type urban green space ecosystems specifically includes:

[0065] Data collection module, used to collect time-series urban green space carbon sink correlation data;

[0066] The characteristic factor extraction module is used to obtain the carbon sink pretreatment set, divide the carbon sink pretreatment set into a training set and a test set, build a characteristic factor extraction model by combining the GBR model and global sensitivity analysis, determine the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment set based on the characteristic factor extraction model, and output the characteristic factor set;

[0067] The carbon sink measurement module loads the characteristic factor set, takes the characteristic factor set as input, combines the carbon storage method and the characteristic factor extraction model to analyze the characteristic factor set, calculates the biomass of trees and shrubs in urban green spaces, and measures the total carbon sink in urban green spaces based on the biomass of trees and shrubs in urban green spaces.

[0068] Preferably, the data acquisition module specifically includes:

[0069] A normalization unit is used to obtain carbon sink related data, process missing values ​​and abnormal values ​​of carbon sink related data, and obtain a normalized set;

[0070] A remote sensing data processing unit is used to extract green space remote sensing data from carbon sink association data, calculate a vegetation index based on hyperspectral influence in the green space remote sensing data, and determine the vegetation type corresponding to the green space remote sensing data through the vegetation index based on hyperspectral influence;

[0071] The ecological data processing unit is used to extract the green space ecological data from the carbon sink associated data, identify the vegetation type, vegetation breast diameter, and vegetation tree height in the green space ecological data based on the vegetation type corresponding to the green space remote sensing data, and extract the underground stratum depth of the green space, the soil light group content corresponding to the stratum depth, the soil recombinant content, the light group organic carbon content, and the recombinant organic carbon content in combination with the carbon sink associated data;

[0072] The sensitivity index calculation unit determines the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio based on the soil light group content, soil recombinant content, light group organic carbon content and recombinant organic carbon content, and calculates the soil organic carbon density reserve and organic carbon component sensitivity index in combination with the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio;

[0073] The summary integration unit is used to integrate the vegetation index based on hyperspectral influence, the light group organic carbon allocation ratio, the recombinant organic carbon allocation ratio, the soil organic carbon density storage and the organic carbon component sensitivity index into a carbon sink pretreatment set.

[0074] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0075] In an embodiment of the present invention, a characteristic factor extraction model is constructed by combining the GBR model and global sensitivity analysis to accurately determine the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment. On the one hand, it can reduce the uncertainty in carbon sink estimation and ensure the accuracy of the evaluation results when the data is unevenly distributed or the sample size is insufficient. On the other hand, it helps to more accurately calculate the biomass of trees and shrubs in urban green spaces, thereby improving the estimation accuracy of the carbon sink amount.

[0076] In the embodiment of the present invention, the soil organic carbon density reserve and the organic carbon component sensitivity index are calculated in combination with the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio, so that the soil organic carbon density reserve and the organic carbon component sensitivity index can be accurately evaluated and calculated when there are significant differences in carbon density among various green space types, thereby ensuring the effect and accuracy of feature selection of the characteristic factor extraction model.

[0077] In the embodiment of the present invention, a characteristic factor extraction model and a training method are provided. The characteristic factor extraction model uses the GBR model as the initial model, so that the characteristic factor extraction model has high stability. It will not produce large fluctuations due to changes in individual samples, and can resist the influence of noise and outliers, and reflect the overall stability and importance of the characteristics, providing more accurate influencing factor information for carbon sink measurement.

[0078] In the embodiments of the present invention, the carbon storage method and the characteristic factor extraction model are combined to realize the quantitative measurement of the carbon sink of urban green space, which can explore the influence of green space type, green space structure, green space age and human interference degree on the carbon sink capacity of urban green space, and then comprehensively evaluate the carbon storage capacity of urban green space, scientifically evaluate the ecological benefits of urban green space, and further improve the carbon fixation efficiency of urban green space. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a schematic diagram of the implementation process of the carbon sink measurement method for multiple types of urban green space ecosystems provided by the present invention.

[0080] Figure 2 A schematic diagram of the implementation process of the carbon sink correlation data preprocessing method is shown.

[0081] Figure 3 A flowchart of the method for constructing a feature factor extraction model by combining the GBR model and global sensitivity analysis is given.

[0082] Figure 4 A schematic diagram of the implementation process of a method for determining characteristic factors that affect the amount of carbon sinks in carbon sink pretreatment based on a characteristic factor extraction model is shown.

[0083] Figure 5 A schematic diagram of the implementation process of the characteristic factor set method combining the carbon storage method and the characteristic factor extraction model analysis is shown.

[0084] Figure 6 It is a structural schematic diagram of the carbon sink measurement system for multi-type urban green space ecosystems provided by the present invention. DETAILED DESCRIPTION

[0085] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of this application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0086] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0087] When the existing methods use feature selection algorithms to extract characteristic factors, if the data distribution is uneven or the sample size is insufficient, the evaluation results will be inaccurate, resulting in insufficient stability and reliability of feature selection. To address the above problems, we proposed a carbon sink measurement system and method for multi-type urban green space ecosystems. In short, when implementing the carbon sink measurement method for multi-type urban green space ecosystems, we first collect the temporal urban green space carbon sink correlation data, and then preprocess the carbon sink correlation data to obtain a carbon sink preprocessing set. Combined with the GBR model and global sensitivity analysis, a characteristic factor extraction model is constructed. Based on the characteristic factor extraction model, the characteristic factors that affect the carbon sink amount in the carbon sink preprocessing set are determined. Finally, the carbon storage method and the characteristic factor extraction model are combined to analyze the characteristic factor set, calculate the biomass of urban green space trees and shrubs, and calculate the urban green space carbon sink amount based on the biomass of urban green space trees and shrubs. In the embodiment of the present invention, a characteristic factor extraction model is constructed by combining the GBR model and the global sensitivity analysis, so as to accurately determine the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment concentration. On the one hand, it can reduce the uncertainty in the carbon sink estimation, and ensure the accuracy of the evaluation results when the data is unevenly distributed or the sample size is insufficient. On the other hand, it helps to more accurately calculate the biomass of trees and shrubs in urban green spaces, thereby improving the estimation accuracy of the carbon sink amount. This overcomes the problem that the existing method uses a feature selection algorithm to extract characteristic factors, which leads to inaccurate evaluation results if the data is unevenly distributed or the sample size is insufficient, resulting in insufficient stability and reliability of feature selection.

[0088] The embodiment of the present invention provides a carbon sink measurement method for multiple types of urban green space ecosystems. Figure 1 The present invention shows a schematic diagram of the implementation process of a carbon sink measurement method for multiple types of urban green space ecosystems. The carbon sink measurement method for multiple types of urban green space ecosystems specifically includes:

[0089] Step S10, collecting time-series urban green space carbon sink associated data, wherein the carbon sink associated data includes green space ecological data, green space remote sensing data, and urban meteorological data, and pre-processing the carbon sink associated data to obtain a carbon sink pre-processing set;

[0090] It should be noted that the multi-type urban green space includes but is not limited to protective green space, ancillary green space, park green space, and regional green space, and the multi-type urban green space is in accordance with the provisions of the "Urban Green Space Classification Standard" (CJJ / T85-2017). Green space remote sensing data includes but is not limited to high-resolution satellite images, multi-spectral satellite data, and ground measurement data, among which the ground measurement data can be collected by ASD's FieldSpec-HandHeld2 handheld ground feature spectrometer. The handheld ground feature spectrometer has a spectral observation range of 325-1075nm and a spectral resolution of 3nm. The data collection time is 12-13 noon. During data collection, a standard whiteboard calibration is performed after each type of vegetation is measured. When measuring the reflectance spectrum of vegetation, the measurer wears a black coat to avoid interfering with the reflectance spectrum of the measured object.

[0091] During the measurement, the spectrometer was 30cm away from the surface of the vegetation canopy, the lens was vertically facing downward, and the shadow of the measurer was prevented from falling on the observation target. Ten sample points (repeated) were measured for each vegetation, and the distance between each sample point was greater than 2m. During the measurement, the spectrum scanning average was set to 10 times. The data of the 10 sample points of each plant were averaged to obtain the final reflectance of the plant. The data was preprocessed using the ViewSpecpro5.6 software that comes with the spectrometer.

[0092] Step S20, obtaining a carbon sink preprocessing set, dividing the carbon sink preprocessing set into a training set and a test set, building a characteristic factor extraction model in combination with the GBR model and global sensitivity analysis, determining characteristic factors in the carbon sink preprocessing set that affect the carbon sink amount based on the characteristic factor extraction model, and outputting the characteristic factor set;

[0093] Step S30, loading the characteristic factor set, taking the characteristic factor set as input, combining the carbon storage method and the characteristic factor extraction model to analyze the characteristic factor set, calculating the biomass of trees and shrubs in urban green spaces, and measuring the total carbon sink in urban green spaces based on the biomass of trees and shrubs in urban green spaces.

[0094] In an embodiment of the present invention, a characteristic factor extraction model is constructed by combining the GBR model and global sensitivity analysis to accurately determine the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment. On the one hand, it can reduce the uncertainty in carbon sink estimation and ensure the accuracy of the evaluation results when the data is unevenly distributed or the sample size is insufficient. On the other hand, it helps to more accurately calculate the biomass of trees and shrubs in urban green spaces, thereby improving the estimation accuracy of the carbon sink amount.

[0095] The embodiment of the present invention provides a method for preprocessing carbon sink correlation data. Figure 2 The figure shows a flow chart of a method for preprocessing carbon sink-related data. The method for preprocessing carbon sink-related data specifically includes:

[0096] Step S101, obtaining carbon sink associated data, processing missing values ​​and abnormal values ​​of the carbon sink associated data, and obtaining a normalized set;

[0097] It should be noted that the methods for handling missing values ​​and outliers can be filling, marking, replacement, and conversion.

[0098] Step S102, extracting green space remote sensing data from the carbon sink association data, calculating a vegetation index based on hyperspectral influence in the green space remote sensing data, and determining a vegetation type corresponding to the green space remote sensing data through the vegetation index based on hyperspectral influence;

[0099] Among them, the vegetation index based on hyperspectral influence is calculated by the following formula:

[0100] (1)

[0101] in, represents the reflectivity, The subscript value of represents the wavelength corresponding to the reflectivity. represents the enhancement coefficient;

[0102] In this embodiment, the hyperspectral image-based vegetation index (HSVI) is a new vegetation index proposed by Professor Sun Genyun's team at China University of Petroleum. It is mainly used to detect vegetation from hyperspectral images. In formula (1), the enhancement coefficient It can be 4.

[0103] Step S103, extracting green space ecological data from the carbon sink associated data, identifying vegetation types, vegetation breast diameter, and vegetation tree height in the green space ecological data based on the vegetation types corresponding to the green space remote sensing data, and extracting the underground stratum depth of the green space, the soil light group content corresponding to the stratum depth, the soil recombinant content, the light group organic carbon content, and the recombinant organic carbon content in combination with the carbon sink associated data;

[0104] Step S104, determining the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio based on the soil light group content, the soil recombinant content, the light group organic carbon content, and the recombinant organic carbon content;

[0105] Step S105, obtaining the light organic carbon distribution ratio and the recombinant organic carbon distribution ratio, and calculating the soil organic carbon density reserve and the organic carbon component sensitivity index in combination with the light organic carbon distribution ratio and the recombinant organic carbon distribution ratio;

[0106] The soil organic carbon density storage is calculated by the following formula:

[0107] (2)

[0108] (3)

[0109] in, It represents the soil organic carbon density storage (t·hm -2 ), The total organic carbon content of soil (g kg -1 ), is the soil bulk density (g·cm -3 ), is the soil layer thickness (cm), They are the light group organic carbon distribution ratio and the heavy group organic carbon distribution ratio, It indicates the volume ratio of gravel with diameter greater than 2 mm in the soil, and represents the sensitivity index of organic carbon components, Indicates the minimum value of organic carbon content in various green space types. It indicates the range of total organic carbon and component contents of various green space types;

[0110] It should be noted that in the embodiment of the present invention, the soil organic carbon density reserve and the organic carbon component sensitivity index are calculated in combination with the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio, so that the soil organic carbon density reserve and the organic carbon component sensitivity index can be accurately evaluated and calculated when there are significant differences in carbon density among various green space types, thereby ensuring the effect and accuracy of feature selection of the characteristic factor extraction model.

[0111] Step S106, integrating the vegetation index based on hyperspectral influence, the light group organic carbon allocation ratio, the heavy group organic carbon allocation ratio, the soil organic carbon density storage and the organic carbon component sensitivity index into a carbon sink preprocessing set.

[0112] In this embodiment, the sensitivity index of organic carbon components is usually used to analyze the response of soil organic carbon components to changes in different land use patterns. It is generally believed that soil active organic carbon components are more sensitive to land use changes than soil total organic carbon. However, the sensitivity indexes of total organic carbon and light and recombinant organic carbon content in different soil layers to changes in green space types are different. There are differences in the sensitivity indexes of different soil layers to changes in green space types. From the average value of the sensitivity index, light organic carbon (0.546)>recombinant organic carbon (0.333)>total organic carbon (0.308). This shows that the light organic carbon content in the soil is most sensitive to changes in green space types.

[0113] The embodiment of the present invention provides a method for constructing a characteristic factor extraction model by combining the GBR model and global sensitivity analysis. Figure 3The figure shows a schematic diagram of a method for implementing a characteristic factor extraction model by combining a GBR model and a global sensitivity analysis. The method for implementing a characteristic factor extraction model by combining a GBR model and a global sensitivity analysis specifically includes:

[0114] Step S201, using the GBR model as the initial model of the feature factor extraction model, wherein the initial model includes a model fusion module, a parameter setting module, a regularization module, and an evaluation module;

[0115] It should be noted that the setting of the regularization module can prevent overfitting during model training. A regularization term is set in the regularization module. The regularization term penalizes the model parameters and limits the complexity of the model, thereby improving the generalization ability of the model. The loss function of the initial model can be the mean square error (MSE) or absolute error.

[0116] Step S202, freezing the evaluation module of the initial model, and improving the evaluation module in combination with the global sensitivity analysis to obtain a sensitivity analysis module;

[0117] In this embodiment, the sensitivity analysis module is used to determine the input characteristic factors that affect the carbon sink total assessment and the degree of influence. The sensitivity analysis module uses a global sensitivity analysis based on the Sobol′ method when analyzing and determining the characteristic factors.

[0118] Step S203, the model fusion module performs weighted fusion on the sensitivity analysis module and the parameter setting module. The parameter setting module sets the learning rate, number of decision trees, maximum depth of the decision tree, and leaf nodes of the initial model, and defines the loss function of the initial model. In this embodiment, the learning rate of the initial model can be 0.0001-0.0002;

[0119] Step S204, obtaining a training set, taking the training set as input, executing the initial model, iteratively training the initial model through the training set, calculating the loss function and gradient of the current model in each round of iterative training, building a new decision tree based on the negative gradient direction, and adding the new decision tree to the initial model until the loss function converges;

[0120] Step S205, obtaining a test set, dividing the test set into multiple test subsets, and using the test subsets to verify the accuracy of the initial model;

[0121] Step S206: if the accuracy of the initial model meets the preset accuracy threshold, the characteristic factor extraction model is output, and the preset accuracy threshold may be 0.95-0.98.

[0122] If the accuracy of the initial model does not meet the preset accuracy threshold, return to step S204.

[0123] In the embodiment of the present invention, a characteristic factor extraction model and a training method are provided. The characteristic factor extraction model uses the GBR model as the initial model, so that the characteristic factor extraction model has high stability. It will not produce large fluctuations due to changes in individual samples, and can resist the influence of noise and outliers, and reflect the overall stability and importance of the characteristics, providing more accurate influencing factor information for carbon sink measurement.

[0124] The embodiment of the present invention provides a method for determining characteristic factors of carbon sink pretreatment concentration that affect the amount of carbon sink based on a characteristic factor extraction model. Figure 4 The present invention shows a schematic diagram of a method for implementing a characteristic factor extraction model based on which a carbon sink pretreatment concentrates on the amount of carbon sinks. The method for determining the characteristic factor extraction model based on which a carbon sink pretreatment concentrates on the amount of carbon sinks specifically includes:

[0125] Step S301, loading a carbon sink preprocessing set, extracting initial characteristic factors influencing the carbon sink amount in the carbon sink preprocessing set, and defining initial weights of the initial characteristic factors based on orthogonal initialization;

[0126] It should be noted that when defining the initial weights based on orthogonal initialization, an orthogonal matrix of the initial characteristic factors can be constructed. Starting from a non-zero vector, a set of orthogonal vectors are constructed using the Gram-Schmidt orthogonalization process. The matrix composed of these vectors is an orthogonal matrix.

[0127] Step S302, obtaining initial characteristic factors, converting and combining the initial characteristic factors based on an orthogonal matrix to generate secondary characteristic factors, calculating weights of the secondary characteristic factors, and merging the initial characteristic factors and the secondary characteristic factors into a characteristic factor set;

[0128] Among them, the weight of the secondary characteristic factor is calculated by the following formula:

[0129] (4)

[0130] in, is the weight of the secondary characteristic factor, represents the weight of the initial feature factor, are the maximum and minimum weights of the initial characteristic factors associated with the secondary characteristic factors, represents the number of initial characteristic factors. In this embodiment, the number of initial characteristic factors may be 5-15;

[0131] Step S303, determining characteristic factors that affect the carbon sink amount in the carbon sink pretreatment set based on the characteristic factor extraction model, and outputting the characteristic factor set.

[0132] It should be noted that the characteristic factor set includes but is not limited to the following characteristic factors: vegetation type, breast diameter (ground diameter), tree height, calculation of tree and shrub biomass based on biomass allometric growth equation, biomass expansion factor, root-to-stem ratio, etc., organic carbon distribution ratio of each component, vegetation red edge parameters, temperature, precipitation, light, altitude, slope aspect, slope, soil type, vegetation type, vegetation coverage, vegetation age, and industrial emissions.

[0133] Step S304, obtaining a characteristic factor set, and the characteristic factor extraction model selects characteristic factors whose signal-to-noise ratio coefficients are lower than a preset noise coefficient based on the signal-to-noise ratio coefficients of the characteristic factor set to generate a low signal-to-noise ratio factor set;

[0134] Step S305, loading the low signal-to-noise ratio factor set, and the sensitivity analysis module outputs the global sensitivity index of the characteristic factor based on the low signal-to-noise ratio factor set and the signal-to-noise ratio coefficient of the low signal-to-noise ratio factor;

[0135] Step S306, preset a sensitivity threshold, and determine whether the global sensitivity index of the characteristic factor exceeds the preset sensitivity threshold;

[0136] Step S307: if the sensitivity threshold is exceeded, the current characteristic factor is determined to be a characteristic factor affecting the carbon sink amount;

[0137] Step S308: If the sensitivity does not exceed the preset sensitivity threshold, the current characteristic factor is discarded.

[0138] It should be noted that the preset sensitivity threshold setting refers to the first-order sensitivity coefficient setting and can be set to 0.1.

[0139] In this embodiment, the signal-to-noise ratio coefficient of the characteristic factor set is calculated by the following formula:

[0140] (5)

[0141] (6)

[0142] in, is the signal-to-noise ratio coefficient, Represents the input feature factor data, represents the noise corresponding to the characteristic factor data, is the data volume of the feature factor set, represents the noise factor, represents the urban green space migration coefficient, represents the conversion coefficient of the net ecosystem productivity of urban green space, represents the noise constant. In this embodiment, the noise constant may be 2-5. represents the total carbon emissions from urban green spaces;

[0143] The global sensitivity index of the characteristic factor is calculated by the following formula:

[0144] (7)

[0145] in, represents the global sensitivity index of the characteristic factor, represents the variance operator, represents the sensitivity response function, represents the characteristic factor variance;

[0146] (8)

[0147] in, Represents the input feature factor data, Represents the mean of the characteristic factor data.

[0148] The embodiment of the present invention provides a method for analyzing a characteristic factor set by combining a carbon storage method and a characteristic factor extraction model. Figure 5 The figure shows a schematic diagram of the implementation process of the method for analyzing the characteristic factor set by combining the carbon storage method and the characteristic factor extraction model. The method for analyzing the characteristic factor set by combining the carbon storage method and the characteristic factor extraction model specifically includes:

[0149] Step S401, obtaining a characteristic factor set, and defining a biomass allometric growth equation by combining a carbon storage method and a characteristic factor extraction model;

[0150] Step S402, calculating the biomass expansion factor and the biomass root-to-shoot ratio based on the biomass allometric growth equation;

[0151] Step S403, obtaining a biomass expansion factor and a biomass root-to-shoot ratio, and calculating the biomass of trees and shrubs in urban green spaces based on the biomass expansion factor and the biomass root-to-shoot ratio;

[0152] Among them, the biomass allometric growth equation is expressed as:

[0153] (9)

[0154] (10)

[0155] in, is the output value of the biomass allometric growth equation, represents the integrated weight of the characteristic factors in the characteristic factor set, are the number of trees and shrubs in green areas, represents the number of characteristic factors, represents the mean value of the global sensitivity index of the characteristic factor, represents the biomass expansion factor;

[0156] (11)

[0157] in, represents the biomass of trees in urban green space, is the root-to-shoot ratio of biomass, Indicates the height of trees in green space.

[0158] Step S404, obtaining the biomass of trees and shrubs in urban green spaces, defining a biomass coupling model in combination with the carbon storage method and the characteristic factor extraction model, and calculating the carbon sink in urban green spaces based on the biomass coupling model;

[0159] The biomass coupling model is defined as:

[0160] (12)

[0161] in, represents the total output value of urban green space carbon sink, is the biomass loss factor of trees and shrubs in urban green space. In this embodiment, the biomass loss factor may be 0.01-0.15. It represents the kernel function coupling coefficient of the biomass coupling model, and the coupling coefficient can be 0.3-0.4.

[0162] In the embodiments of the present invention, the carbon storage method and the characteristic factor extraction model are combined to realize the quantitative measurement of the carbon sink of urban green space, which can explore the influence of green space type, green space structure, green space age and human interference degree on the carbon sink capacity of urban green space, and then comprehensively evaluate the carbon storage capacity of urban green space, scientifically evaluate the ecological benefits of urban green space, and further improve the carbon fixation efficiency of urban green space.

[0163] The embodiment of the present invention provides a carbon sink measurement system for various types of urban green space ecosystems. Figure 6 The structural schematic diagram of the carbon sink measurement system for multi-type urban green space ecosystems is shown. The carbon sink measurement system for multi-type urban green space ecosystems specifically includes:

[0164] The data collection module 100 is used to collect the time-series urban green space carbon sink related data;

[0165] The characteristic factor extraction module 200 is used to obtain a carbon sink preprocessing set, divide the carbon sink preprocessing set into a training set and a test set, construct a characteristic factor extraction model in combination with the GBR model and global sensitivity analysis, determine the characteristic factors that affect the carbon sink amount in the carbon sink preprocessing set based on the characteristic factor extraction model, and output the characteristic factor set;

[0166] The carbon sink measurement module 300 loads a characteristic factor set, takes the characteristic factor set as input, analyzes the characteristic factor set in combination with the carbon storage method and the characteristic factor extraction model, calculates the biomass of trees and shrubs in urban green spaces, and measures the total carbon sink in urban green spaces based on the biomass of trees and shrubs in urban green spaces.

[0167] In an embodiment of the present invention, data acquisition module 100, characteristic factor extraction module 200, and carbon sink measurement module 300 are connected by 5G or DTU communication connection to realize data interaction, and carbon sink measurement module 300 and characteristic factor extraction module 200 are respectively connected to data acquisition module 100 in a one-to-n manner.

[0168] In this embodiment, the data acquisition module 100 specifically includes:

[0169] A normalization unit 110 is used to obtain carbon sink associated data, process missing values ​​and abnormal values ​​of the carbon sink associated data, and obtain a normalized set;

[0170] The remote sensing data processing unit 120 is used to extract the green space remote sensing data from the carbon sink association data, calculate the vegetation index based on the hyperspectral influence in the green space remote sensing data, and determine the vegetation type corresponding to the green space remote sensing data through the vegetation index based on the hyperspectral influence;

[0171] The ecological data processing unit 130 is used to extract the green space ecological data from the carbon sink associated data, identify the vegetation type, vegetation breast diameter, and vegetation tree height in the green space ecological data based on the vegetation type corresponding to the green space remote sensing data, and extract the underground stratum depth of the green space, the soil light group content corresponding to the stratum depth, the soil recombinant content, the light group organic carbon content, and the recombinant organic carbon content in combination with the carbon sink associated data;

[0172] The sensitivity index calculation unit 140 determines the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio based on the soil light group content, the soil recombinant content, the light group organic carbon content and the recombinant organic carbon content, and calculates the soil organic carbon density reserve and the organic carbon component sensitivity index in combination with the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio;

[0173] The summary integration unit 150 is used to integrate the vegetation index based on hyperspectral influence, the light group organic carbon allocation ratio, the recombinant organic carbon allocation ratio, the soil organic carbon density storage and the organic carbon component sensitivity index into a carbon sink preprocessing set.

[0174] It can be understood that the carbon sink measurement system for multi-type urban green space ecosystems provided by the embodiment of the present invention corresponds to the above-mentioned carbon sink measurement method for multi-type urban green space ecosystems. The explanations, examples, beneficial effects and other parts of the relevant contents can refer to the corresponding contents in the carbon sink measurement method for multi-type urban green space ecosystems, and will not be repeated here.

[0175] On the other hand, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer program instructions, and the computer program instructions can be executed by a processor. When the computer program instructions are executed, the method of any of the above embodiments is implemented.

[0176] Finally, it should be noted that the computer-readable storage medium (e.g., memory) herein may be a volatile memory or a non-volatile memory, or may include both a volatile memory and a non-volatile memory. As an example and not by way of limitation, a non-volatile memory may include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), or a flash memory. A volatile memory may include a random access memory (RAM), which may act as an external cache memory. As an example and not by way of limitation, RAM may be obtained in a variety of forms, such as synchronous RAM (DRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct Rambus RAM (DRRAM). The storage devices of the disclosed aspects are intended to include, but are not limited to, these and other suitable types of memory.

[0177] The various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure herein may be implemented or executed using the following components designed to perform the functions herein: a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination of these components. A general-purpose processor may be a microprocessor, but alternatively, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP, and / or any other such configuration.

[0178] In summary, the present invention provides a carbon sink measurement system and method for multi-type urban green space ecosystems. In an embodiment of the present invention, a characteristic factor extraction model is constructed by combining the GBR model and global sensitivity analysis, so as to accurately determine the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment center. On the one hand, it can reduce the uncertainty in carbon sink estimation and ensure the accuracy of the evaluation results when the data is unevenly distributed or the sample size is insufficient. On the other hand, it helps to more accurately calculate the biomass of trees and shrubs in urban green spaces, thereby improving the estimation accuracy of the carbon sink amount.

[0179] It should be noted that, for the above-mentioned embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, some steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0180] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in this field can still combine, add, delete or make other adjustments to the features in the various embodiments of the present invention according to the circumstances without conflict, without making creative work, so as to obtain different other technical solutions that do not deviate from the concept of the present invention in essence, and these technical solutions also belong to the scope of protection of the present invention.

Claims

1. A carbon sink measurement method for multi-type urban green space ecosystems, characterized in that: The carbon sink measurement method for multi-type urban green space ecosystems includes: Collecting time-series urban green space carbon sink associated data, wherein the carbon sink associated data includes green space ecological data, green space remote sensing data, and urban meteorological data, and preprocessing the carbon sink associated data to obtain a carbon sink preprocessing set; Obtain a carbon sink preprocessing set, divide the carbon sink preprocessing set into a training set and a test set, build a characteristic factor extraction model by combining the GBR model and global sensitivity analysis, determine the characteristic factors that affect the carbon sink amount in the carbon sink preprocessing set based on the characteristic factor extraction model, and output the characteristic factor set; Load the characteristic factor set, take the characteristic factor set as input, combine the carbon storage method and the characteristic factor extraction model to analyze the characteristic factor set, calculate the biomass of trees and shrubs in urban green spaces, and calculate the total carbon sink in urban green spaces based on the biomass of trees and shrubs in urban green spaces; The method for preprocessing carbon sink association data specifically includes: Obtain carbon sink-related data, process missing values ​​and abnormal values ​​of the carbon sink-related data, and obtain a normalized set; Extract green space remote sensing data from carbon sink association data, calculate the vegetation index based on hyperspectral influence in green space remote sensing data, and determine the vegetation type corresponding to green space remote sensing data through the vegetation index based on hyperspectral influence; Among them, the vegetation index based on hyperspectral influence is calculated by the following formula: （1） in, represents the reflectivity, The subscript value of represents the wavelength corresponding to the reflectivity. represents the enhancement coefficient; Extract green space ecological data from carbon sink-related data, identify vegetation types, vegetation breast diameter, and vegetation tree height in green space ecological data based on vegetation types corresponding to green space remote sensing data, and extract underground stratum depth of green space, soil light group content corresponding to stratum depth, soil recombinant content, light group organic carbon content, and recombinant organic carbon content in combination with carbon sink-related data; Based on the soil light group content, soil recombinant content, light group organic carbon content, and recombinant organic carbon content, the light group organic carbon distribution ratio and recombinant organic carbon distribution ratio are determined; Obtain the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio, and calculate the soil organic carbon density storage and organic carbon component sensitivity index based on the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio; The soil organic carbon density storage is calculated by the following formula: （2） （3） in, represents the soil organic carbon density storage, represents the total organic carbon content of soil. is the soil bulk density, is the soil layer thickness, They are the light group organic carbon distribution ratio and the heavy group organic carbon distribution ratio, It indicates the volume ratio of gravel with diameter greater than 2 mm in the soil, and represents the sensitivity index of organic carbon components, Indicates the minimum value of organic carbon content in various green space types. It indicates the range of total organic carbon and component contents of various green space types; The vegetation index based on hyperspectral influence, light group organic carbon allocation ratio, heavy group organic carbon allocation ratio, soil organic carbon density storage and organic carbon component sensitivity index were integrated into the carbon sink pretreatment set; The method for constructing a characteristic factor extraction model by combining the GBR model and global sensitivity analysis specifically includes: The GBR model is used as the initial model of the feature factor extraction model, wherein the initial model includes a model fusion module, a parameter setting module, a regularization module, and an evaluation module; Freeze the evaluation module of the initial model, and improve the evaluation module in combination with the global sensitivity analysis to obtain a sensitivity analysis module; The model fusion module performs weighted fusion on the sensitivity analysis module and the parameter setting module. The parameter setting module sets the learning rate, number of decision trees, maximum depth of decision trees, leaf nodes of the initial model, and defines the loss function of the initial model. Get the training set, use the training set as input, execute the initial model, iteratively train the initial model through the training set, calculate the loss function and gradient of the current model in each round of iterative training, build a new decision tree based on the negative gradient direction, and add the new decision tree to the initial model until the loss function converges; Obtain a test set, divide the test set into multiple test subsets, use the test subsets to verify the accuracy of the initial model, and output the feature factor extraction model if the accuracy of the initial model meets the preset accuracy threshold.

2. The carbon sink measurement method for multi-type urban green space ecosystems according to claim 1 is characterized in that: The method for determining the characteristic factors affecting the carbon sink amount in the carbon sink pretreatment concentration based on the characteristic factor extraction model specifically includes: Load the carbon sink preprocessing set, extract the initial characteristic factors that affect the carbon sink amount in the carbon sink preprocessing set, and define the initial weights of the initial characteristic factors based on orthogonal initialization; Obtain initial characteristic factors, transform and combine the initial characteristic factors based on the orthogonal matrix to generate secondary characteristic factors, calculate the weights of the secondary characteristic factors, and merge the initial characteristic factors and the secondary characteristic factors into a characteristic factor set; Among them, the weight of the secondary characteristic factor is calculated by the following formula: （4） in, is the weight of the secondary characteristic factor, represents the weight of the initial feature factor, are the maximum and minimum weights of the initial characteristic factors associated with the secondary characteristic factors, represents the number of initial characteristic factors; Based on the characteristic factor extraction model, the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment set are determined, and the characteristic factor set is output.

3. The carbon sink measurement method for multi-type urban green space ecosystems according to claim 2 is characterized in that: The method for determining the characteristic factors affecting the carbon sink amount in the carbon sink pretreatment concentration based on the characteristic factor extraction model specifically includes: Acquire a characteristic factor set, and the characteristic factor extraction model selects characteristic factors whose signal-to-noise ratio coefficients are lower than a preset noise coefficient based on the signal-to-noise ratio coefficients of the characteristic factor set to generate a low signal-to-noise ratio factor set; The low signal-to-noise ratio factor set is loaded, and the sensitivity analysis module outputs the global sensitivity index of the characteristic factor based on the low signal-to-noise ratio factor set and the signal-to-noise ratio coefficient of the low signal-to-noise ratio factor; A preset sensitivity threshold is used to determine whether the global sensitivity index of the characteristic factor exceeds the preset sensitivity threshold. If it exceeds the sensitivity threshold, the current characteristic factor is determined to be a characteristic factor that affects the carbon sink amount; If the preset sensitivity threshold is not exceeded, the current characteristic factor is discarded.

4. The carbon sink measurement method for multi-type urban green space ecosystems as claimed in claim 3 is characterized by: The signal-to-noise ratio coefficient of the characteristic factor set is calculated by the following formula: （5） （6） in, is the signal-to-noise ratio coefficient, represents the input feature factor data, represents the noise corresponding to the characteristic factor data, is the data volume of the feature factor set, represents the noise factor, represents the urban green space migration coefficient, represents the conversion coefficient of the net ecosystem productivity of urban green space, represents the noise constant, represents the total carbon emissions from urban green spaces; The global sensitivity index of the characteristic factor is calculated by the following formula: （7） in, represents the global sensitivity index of the characteristic factor, represents the variance operator, represents the sensitivity response function, represents the characteristic factor variance; （8） in, Represents the input feature factor data, Represents the mean of the characteristic factor data.

5. The carbon sink measurement method for multi-type urban green space ecosystems according to claim 1 is characterized in that: The method of analyzing the characteristic factor set by combining the carbon storage method and the characteristic factor extraction model specifically includes: Obtain the characteristic factor set, and define the biomass allometric growth equation by combining the carbon storage method and the characteristic factor extraction model; The biomass expansion factor and biomass root-to-shoot ratio were calculated based on the biomass allometric growth equation; Obtain biomass expansion factor and biomass root-to-shoot ratio, and calculate the biomass of trees and shrubs in urban green spaces based on the biomass expansion factor and biomass root-to-shoot ratio; Among them, the biomass allometric growth equation is expressed as: （9） （10） in, is the output value of the biomass allometric growth equation, represents the integrated weight of the characteristic factors in the characteristic factor set, are the number of trees and shrubs in green areas, represents the number of characteristic factors, represents the mean value of the global sensitivity index of the characteristic factor, represents the biomass expansion factor; （11） in, represents the biomass of trees in urban green space, is the root-to-shoot ratio of biomass, Indicates the height of trees in green space.

6. The carbon sink measurement method for multi-type urban green space ecosystems according to claim 5 is characterized in that: The method of combining the carbon storage method and the characteristic factor extraction model to analyze the characteristic factor set specifically includes: Obtain the biomass of trees and shrubs in urban green spaces, define a biomass coupling model by combining the carbon storage method and the characteristic factor extraction model, and calculate the carbon sink in urban green spaces based on the biomass coupling model; The biomass coupling model is defined as: （12） in, represents the total output value of urban green space carbon sink, is the biomass loss factor of trees and shrubs in urban green space, Represents the kernel function coupling coefficient of the biomass coupling model.

7. A carbon sink measurement system for multi-type urban green space ecosystems, used to implement the carbon sink measurement method for multi-type urban green space ecosystems as claimed in any one of claims 1 to 6, characterized in that: The carbon sink measurement system for multi-type urban green space ecosystems specifically includes: Data collection module, used to collect time-series urban green space carbon sink correlation data; The characteristic factor extraction module is used to obtain the carbon sink pretreatment set, divide the carbon sink pretreatment set into a training set and a test set, build a characteristic factor extraction model by combining the GBR model and global sensitivity analysis, determine the characteristic factors that affect the carbon sink amount in the carbon sink pretreatment set based on the characteristic factor extraction model, and output the characteristic factor set; The carbon sink measurement module loads the characteristic factor set, takes the characteristic factor set as input, combines the carbon storage method and the characteristic factor extraction model to analyze the characteristic factor set, calculates the biomass of trees and shrubs in urban green spaces, and measures the total carbon sink in urban green spaces based on the biomass of trees and shrubs in urban green spaces.

8. The carbon sink measurement system for multi-type urban green space ecosystems as claimed in claim 7 is characterized by: The data acquisition module specifically includes: A normalization unit is used to obtain carbon sink related data, process missing values ​​and abnormal values ​​of carbon sink related data, and obtain a normalized set; A remote sensing data processing unit is used to extract green space remote sensing data from carbon sink association data, calculate a vegetation index based on hyperspectral influence in the green space remote sensing data, and determine the vegetation type corresponding to the green space remote sensing data through the vegetation index based on hyperspectral influence; The ecological data processing unit is used to extract the green space ecological data from the carbon sink associated data, identify the vegetation type, vegetation breast diameter, and vegetation tree height in the green space ecological data based on the vegetation type corresponding to the green space remote sensing data, and extract the underground stratum depth of the green space, the soil light group content corresponding to the stratum depth, the soil recombinant content, the light group organic carbon content, and the recombinant organic carbon content in combination with the carbon sink associated data; The sensitivity index calculation unit determines the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio based on the soil light group content, soil recombinant content, light group organic carbon content and recombinant organic carbon content, and calculates the soil organic carbon density reserve and organic carbon component sensitivity index in combination with the light group organic carbon distribution ratio and the recombinant organic carbon distribution ratio; The summary integration unit is used to integrate the vegetation index based on hyperspectral influence, the light group organic carbon allocation ratio, the recombinant organic carbon allocation ratio, the soil organic carbon density storage and the organic carbon component sensitivity index into a carbon sink pretreatment set.

Images