Urban green land carbon storage and sink quantification method based on remote sensing and field investigation

Abstract

The invention provides an urban green land carbon storage and sink quantification method based on remote sensing and field investigation, and relates to the technical field of environmental science and engineering, remote sensing technology and ecological assessment, and the method comprises the steps: obtaining a high-definition remote sensing image, combining a green land planning file, determining the distribution of various types of green lands, calculating an NDVI, and extracting an effective green land space; and generating a sampling point scheme according to the area proportion. Growing parameters of arbors and shrubs in the quadrat are investigated on site, biomass is calculated and converted into carbon reserves, an NDVI and carbon density model is established, and a carbon density spatial distribution diagram is generated. And constructing a carbon sink model based on the relationship between the age of the arbor and the carbon reserves, calculating the annual carbon sink, establishing a relationship model between the NDVI and the carbon sink, and generating a carbon sink spatial distribution diagram. And integrating the two diagrams to output carbon storage and sink capability evaluation results of different green land types, thereby providing a scientific basis for urban ecological planning.

Description

Technical Field

[0001] This application belongs to the field of environmental science and engineering, remote sensing technology and ecological assessment technology, and specifically relates to a method for quantifying carbon storage in urban green spaces based on remote sensing and field surveys. Background Art

[0002] With the acceleration of urbanization, urban green spaces, as important ecosystem components, play a key role in regulating climate, improving air quality, and enhancing the quality of the human living environment. In recent years, the carbon sequestration function of urban green spaces has garnered increasing attention. Vegetation absorbs atmospheric carbon dioxide through photosynthesis and fixes it as organic carbon, creating carbon reserves (carbon storage) and carbon sinks (carbon absorption rate), making it a crucial carbon sink in urban ecosystems.

[0003] Currently, assessments of vegetation carbon storage capacity, both domestically and internationally, often rely on field surveys combined with remote sensing inversion. Traditional methods typically rely on surveys of fixed plots to obtain biomass data, which are then converted to carbon storage. However, these methods suffer from limited survey coverage, high costs, and difficulty in scaling up to large scales. Some studies have attempted to use vegetation indices (such as the Normalized Difference Vegetation Index (NDVI)) from remote sensing imagery to establish empirical models for spatial extrapolation of carbon storage. However, these studies often fail to fully consider the impact of different urban green space types (such as parks, sheltered green spaces, and ancillary green spaces) on carbon storage capacity, resulting in low model accuracy and poor applicability.

[0004] Furthermore, existing carbon sink estimation techniques rely heavily on static carbon storage data and lack modeling of vegetation growth dynamics (such as changes in tree age), making it difficult to accurately reflect annual carbon absorption capacity. Furthermore, most methods fail to integrate and analyze the spatial distribution of carbon density and carbon sinks, failing to provide comprehensive spatial decision-making support for refined urban greening management. Summary of the Invention

[0005] This application provides a method for quantifying carbon storage in urban green spaces based on remote sensing and field surveys to solve one of the above technical problems.

[0006] The technical solutions adopted in this application are: The present invention provides a method for quantifying carbon storage in urban green spaces based on remote sensing and field surveys, including: S1. Obtain high-definition remote sensing images of the study area and, in combination with urban green space planning documents, determine the spatial distribution range of various types of green space within the study area. Calculate the normalized vegetation index based on the high-definition remote sensing images, extract grids with normalized vegetation index values ​​within a preset threshold range as urban green space, and generate corresponding sampling point layout plans based on the area proportions of various types of green space. S2. Conduct a field survey of the sampling points according to the sampling point layout plan, record the growth parameters of the trees and shrubs in the sample plot, including the diameter at breast height, plant height, and age of the trees, and the coverage area and type of the shrubs. Based on the growth parameters, calculate the biomass of the trees and shrubs in the sample plot, convert it into carbon storage in combination with the carbon content parameter, and summarize it to form the total carbon storage of the sample plot; S3. extracting the normalized vegetation index value and the corresponding carbon storage data of the sample point, establishing a relationship model between the normalized vegetation index and carbon density, and applying the relationship model to perform spatial calculations on the study area to generate a carbon density spatial distribution map; S4. Based on the relationship between the tree age and carbon storage of the trees in the sample plot, a carbon sequestration calculation model is constructed, and the annual carbon sequestration of the sample plot is calculated in combination with the carbon sequestration calculation model; S5. Extract the normalized vegetation index value and the corresponding carbon sink data of the sample point, establish a relationship model between the normalized vegetation index and the carbon sink, and apply the model to perform spatial calculations on the study area to generate a carbon sink spatial distribution map. Integrate the carbon density spatial distribution map and the carbon sink spatial distribution map to output the carbon storage capacity assessment results of different green space types in the study area.

[0007] According to one embodiment of the present application, the step S1 is specifically as follows: The resolution of the high-definition remote sensing image is 10 meters, and the boundaries of the study area are delineated by mask processing technology; The urban green space planning document includes vector data of park green space, protective green space, square land, ancillary green space and regional green space, which is used to assist in confirming the spatial distribution range of various types of green space; The calculation of the normalized vegetation index is achieved through a band operation method, and the preset threshold range is [0.2, 1]. The sampling point layout plan is generated through a GIS tool, and the number of sampling points is adjusted according to the area proportion of each type of green space. At the same time, the distribution of sampling points is optimized through spatial uniformity analysis to ensure that key areas of different green space types are covered.

[0008] According to one embodiment of the present application, the step S2 is specifically as follows: The field sample survey used a standard sample area of ​​20m×20m; The tree age is determined by comparing remote sensing images of different years, observing the number of branches, measuring the number of annual rings of felled tree stumps, or by expert experience and judgment; The coverage area of ​​the shrubs is measured by visual estimation or grid method, and the species are confirmed by field identification or specimen comparison; The tree biomass is calculated using an allometric growth model, and the shrub biomass is calculated based on the biomass per unit area parameters corresponding to the coverage area and type; The carbon content parameter is determined according to plant species or local standards, and the total carbon storage of the sample is generated through the conversion relationship between biomass and carbon storage.

[0009] According to one embodiment of the present application, the step S3 is specifically as follows: The relationship model between the normalized vegetation index and carbon density is fitted to the sample point data using a linear, exponential, power function or logarithmic model using R language software, and the model with the highest goodness of fit is selected as the final model; The spatial calculation is achieved through the raster calculator of GIS software to generate a carbon density spatial pattern map that distinguishes different green space types and annotates the regional total carbon storage and mean value.

[0010] According to one embodiment of the present application, the step S4 is specifically as follows: The carbon sequestration calculation model is grouped and fitted based on tree species or growth characteristics, and the carbon sequestration calculation model type includes a logistic model, a Gompertz model or other nonlinear models; The annual carbon sequestration is determined by calculating the difference between the carbon storage in the current year and the previous year, and is corrected based on the tree growth rate parameters.

[0011] According to one embodiment of the present application, the step S5 is specifically as follows: The relationship model between the normalized difference vegetation index and carbon sequestration was fitted to the sample point data using R language software, and the model with the highest goodness of fit was selected as the final model; The spatial calculation is achieved through a raster calculator in GIS software to generate a carbon sink spatial pattern map and mark the spatial distribution characteristics of high carbon sink areas and low carbon sink areas.

[0012] According to one embodiment of the present application, it further includes: The carbon storage capacity assessment results include the total regional carbon storage, average carbon density, spatial agglomeration characteristics and the contribution of different green space types; The decision support content includes recommendations for priority protection of high-carbon sink areas, recommendations for optimization and transformation of low-carbon sink areas, and scientific guidance on urban greening layout.

[0013] According to one embodiment of the present application, the carbon sequestration calculation model further includes: In view of the fact that tree growth rate is affected by climate change and soil fertility environmental factors, a correction coefficient K is introduced, K=1±Δ, where Δ is the comprehensive weight value of the influence of environmental factors; The carbon sequestration calculation model was retrospectively verified using the sample carbon storage data from the past five years to ensure that the deviation rate between the predicted values ​​and the measured values ​​of the carbon sequestration calculation model was less than 10%.

[0014] A second aspect of the present application provides a computer-readable storage medium having a program stored thereon, which implements the steps in the method when executed by a processor.

[0015] In a third aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor implements the steps in the method described above when executing the program.

[0016] Due to the adoption of the above technical solution, the beneficial effects achieved by this application are as follows: This application clarifies the spatial distribution of various types of green spaces by combining urban green space planning documents, calculates NDVI based on high-definition remote sensing images, and extracts grids that meet the threshold as green space, ensuring the accuracy of the data source. Furthermore, a sampling point layout plan is generated based on the area proportion of each type of green space, achieving differentiated and representative sampling of different types of green spaces, laying the foundation for subsequent classification modeling. This method overcomes the defects of the traditional method of mixing green space types, and significantly improves the discrimination and reliability of the carbon storage capacity assessment of different green space types.

[0017] This application calculates carbon storage based on the growth parameters of trees within the sample plot, such as diameter at breast height, plant height, and tree age, combined with carbon content parameters. Furthermore, a carbon sink calculation model is constructed based on the relationship between tree age and carbon storage, enabling accurate estimates of annual carbon sinks. This method breaks through the limitations of static carbon storage assessments and introduces dynamic factors of vegetation growth, making carbon sink capacity assessments more timely and scientific. By establishing a relationship model between NDVI and carbon sinks and performing spatial calculations, spatial visualization of carbon sinks at the urban scale has been achieved for the first time, filling the technical gap in dynamic spatial measurement of carbon sinks.

[0018] This application extracts NDVI values ​​from sample points and measured carbon storage / sink data to establish relationship models between NDVI and carbon density, and between NDVI and carbon sink, respectively. These models are then applied to spatial calculations across the entire study area, achieving high-precision expansion from point to surface. This method leverages the wide coverage of remote sensing data and the accuracy of field surveys, improving the accuracy of spatial distribution maps of carbon density and sinks through a model-driven approach and overcoming the poor generalization capabilities of traditional empirical models.

[0019] This application integrates the generated spatial distribution maps of carbon density and carbon sinks to output carbon storage and sequestration capacity assessment results for different green space types within the study area. This not only visually demonstrates the spatial pattern of carbon storage and sequestration, but also identifies areas of high carbon sink potential and low carbon sink potential, providing a scientific basis for optimizing urban green space layout, tree species selection, and maintenance management. This integrated output mechanism enhances the readability and application value of the assessment results, supporting urban low-carbon development and ecological planning decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 A schematic diagram of a process for quantifying urban green space carbon storage based on remote sensing and field surveys provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0021] Reference numerals: 810 , processor; 820 , communication interface; 830 , memory; 840 , communication bus. DETAILED DESCRIPTION

[0022] In order to more clearly illustrate the overall concept of the present application, a detailed description is given below in an illustrative manner in conjunction with the accompanying drawings.

[0023] The following description sets forth many specific details to facilitate a thorough understanding of the present application. However, the present application may also be implemented in other ways than those described herein, and therefore, the scope of protection of the present application is not limited by the specific embodiments disclosed below. It should be noted that the embodiments of the present application and the features of each embodiment may be combined with each other unless there is a conflict.

[0024] In this application, unless otherwise expressly specified and limited, a first feature "above" or "below" a second feature may be that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in an appropriate manner in any one or more embodiments or examples.

[0025] Example 1 like Figure 1 As shown in the figure, a method for quantifying carbon storage in urban green spaces based on remote sensing and field surveys includes: S1. Obtain high-definition remote sensing images of the study area, and combine them with urban green space planning documents to determine the spatial distribution range of various types of green spaces in the study area. Calculate the normalized vegetation index based on the high-definition remote sensing images, extract the grids with normalized vegetation index values ​​within the preset threshold range as urban green space, and generate corresponding sampling point layout plans based on the area proportions of various types of green spaces.

[0026] As mentioned above, first, obtain high-definition remote sensing images of the study area. This typically involves satellite imagery with a resolution of 10 meters or higher, so that different types of vegetation cover can be clearly distinguished. Next, in conjunction with urban green space planning documents (such as urban green space system planning maps), clarify the spatial distribution of various types of green space, including park green space, protective green space, square land, ancillary green space, and surrounding area green space. Based on these high-definition remote sensing images, calculate the Normalized Difference Vegetation Index (NDVI), which is an important indicator for assessing vegetation health and density. By setting a reasonable threshold range (for example, [0.2, 1]), grid cells that meet the definition of urban green space can be screened out as preliminary urban green space spaces. Finally, generate a corresponding sampling point layout plan based on the proportion of each type of green space area to ensure representative and uniform sampling.

[0027] For example, let's assume that the core urban area of ​​a city is selected as the study area, using high-definition remote sensing images with a resolution of 10 meters and referring to local urban green space system planning documents. First, the high-definition imagery is interpreted to identify different types of green spaces. Then, the NDVI is calculated and the grids within the range of [0.2, 1] are determined to be potential urban green space spaces. For example, if park green space accounts for 32% of the total green space area, when formulating the sampling plan, it is necessary to ensure that approximately 32% of the sampling points are located in park green space. For each green space type, sampling points are randomly distributed according to a certain ratio, and the location of the sampling points is optimized using geographic information system software such as ArcGIS to ensure that all important green space patches are covered.

[0028] It should be noted that, in specific implementation scenarios, the above-mentioned approach can be further expanded to include more refined green space classification and analysis. For example, in addition to the main green space types, specific functional areas (such as leisure and entertainment areas and ecological protection areas) can also be refined to conduct separate carbon storage capacity assessments. In addition, the impact of seasonal changes on NDVI values ​​can be considered, and the NDVI threshold can be adjusted based on high-definition remote sensing imagery in different seasons to obtain a more accurate spatial distribution of urban green spaces. At the same time, in terms of sampling point layout, the sampling density can be increased in high-priority areas (such as areas with greater carbon sequestration potential) based on historical data or existing research results, thereby improving the accuracy and efficiency of the entire assessment process. In addition, considering the possible changes in green spaces brought about by future urban development, high-definition remote sensing images and green space planning documents can be regularly updated, and green space classification and sampling strategies can be dynamically adjusted to ensure the effectiveness and reliability of long-term monitoring.

[0029] S2. Conduct field surveys of the sampling points according to the sampling point layout plan, and record the growth parameters of the trees and shrubs in the sample plots, including the diameter at breast height, plant height, and age of the trees, and the coverage area and type of the shrubs. Calculate the biomass of the trees and shrubs in the sample plots based on the growth parameters, and convert them into carbon storage in combination with the carbon content parameter, and summarize them to form the total carbon storage of the sample plots.

[0030] As mentioned above, a sample plot survey is carried out in the field according to the pre-set sampling point layout plan. Each sample plot is usually 20m×20m in size to ensure that different types of green spaces in the study area are fully covered. In each sample plot, detailed growth parameters of trees and shrubs are recorded, including but not limited to the diameter at breast height, plant height, and age of trees, as well as the coverage area and species of shrubs. These data are the basis for calculating biomass. Subsequently, a specific allometric growth model (such as ) Calculate the biomass of individual trees based on their diameter at breast height (DBH), height, and age. This is then converted into carbon storage using the carbon content (typically between 0.43 and 0.52). For shrubs, refer to local standards (such as DB11 / T953-2013) based on their cover area and type to obtain biomass per unit area. Carbon storage is then calculated using the carbon content (typically 0.45). Finally, the carbon storage of all trees and shrubs is aggregated to form the total carbon storage for the sample plot.

[0031] For example, let's assume a sample plot survey is conducted in a park (G1) in the core urban area of ​​a city. A 20m x 20m sample plot was selected, and field measurements revealed 10 trees and several shrubs. For each tree, we measured its diameter at breast height (DBH) using a caliper and its height using a stadiometer. We also estimated its age using various methods (such as comparing historical remote sensing imagery, observing branch counts, and expert judgment). For example, a 20-year-old Chinese pine tree with a diameter of 20cm and a height of 8 meters has a biomass of approximately 0.5 tons, calculated using its allometric growth model. Combined with a carbon content of 0.45, this yields a carbon storage of approximately 0.225 tons. For shrubs, we measured their cover area and, based on the species (for example, broadleaf shrubs), found a biomass per unit area of ​​0.3 tons per square meter. If the cover area is 5 square meters, the carbon storage of the shrubs in this sample plot is 0.675 tons (0.3 tons per square meter * 5 square meters * carbon content of 0.45). The carbon storage of all trees and shrubs was added together to obtain the total carbon storage of the sample.

[0032] It should be noted that, in specific implementation scenarios, the above-mentioned solutions can be further expanded to include more detailed vegetation classification and assessment. Drones or LiDAR technology can be introduced to assist in field surveys to improve the accuracy and efficiency of data collection. For example, drones can be used to capture high-resolution images, combined with deep learning algorithms to automatically identify and measure the diameter at breast height, plant height, and crown width of trees, thereby reducing the workload and errors of manual measurement. Furthermore, for the critical link of tree age estimation, a machine learning-based tree age prediction model can be developed, using a large amount of historical data to train the model and improve the accuracy of tree age estimation. At the same time, taking into account the differences in vegetation characteristics across cities or regions, a national or international database can be established to collect allometric growth models and carbon content parameters of vegetation in various regions, so that parameters can be quickly adjusted when applied in different regions, thereby improving the versatility and adaptability of the method.

[0033] S3. Extract the normalized vegetation index value and the corresponding carbon storage data of the sample point, establish a relationship model between the normalized vegetation index and carbon density, and use the relationship model to perform spatial calculations on the study area to generate a carbon density spatial distribution map.

[0034] As mentioned above, after completing the sample survey and calculating the carbon storage of each sample, the next step is to extract the Normalized Difference Vegetation Index (NDVI) values ​​from these sample points and correlate them with the corresponding carbon storage data. Using statistical analysis software (such as R language), various mathematical models (such as linear, exponential, power function, or logarithmic models) are tried to establish a relationship model between NDVI and carbon density. The model with the highest fit is selected as the final model to ensure that the model can accurately describe the relationship between the two. Subsequently, using geographic information system (GIS) tools (such as the raster calculator in ArcGIS), the selected relationship model is applied to the NDVI layer of the entire study area, thereby achieving spatial expansion from point to surface and generating a spatial distribution map of carbon density. This step not only reveals the differences in carbon density between different regions but also provides a scientific basis for subsequent optimization of urban green spaces.

[0035] For example, suppose a field survey of multiple plots is completed in the core urban area of ​​a city, and the NDVI value and carbon storage of each plot are recorded. Next, the data are processed using R language software, and different regression models (such as linear, exponential, power function, etc.) are tried to fit the relationship between NDVI and carbon density. Assume that the power function model has the highest goodness of fit ( The model with the largest value was selected as the final model. The model was then applied to the NDVI layer for the entire study area using the ArcGIS Raster Calculator function to generate a spatial distribution map of carbon density. For example, in the central area of ​​the parkland (G1), due to its dense and healthy vegetation, the NDVI values ​​were high, resulting in a higher carbon density according to the model. Meanwhile, at the edge of the attached green space (XG), due to its low vegetation cover and relatively low NDVI values, the carbon density was also lower. The resulting spatial distribution map of carbon density visually demonstrates the differences in carbon storage capacity between regions.

[0036] It should be noted that in specific implementation scenarios, in addition to NDVI, the above scheme can also be combined with other remote sensing-derived indicators (such as leaf area index (LAI) and fractional vegetation cover (FVC)) and topographic factors (such as altitude, slope, and aspect) to establish a multi-indicator comprehensive model to improve the accuracy of carbon density estimation. For example, although the NDVI value of vegetation in certain high-altitude areas is not high, its carbon density may be higher than that of low-altitude areas due to special climatic conditions.

[0037] S4. Based on the relationship between the tree age and carbon storage of the trees in the sample plot, a carbon sequestration calculation model is constructed, and the annual carbon sequestration of the sample plot is calculated in combination with the carbon sequestration calculation model.

[0038] As mentioned above, to construct a carbon sequestration calculation model based on the relationship between the tree age and carbon storage in the sample plot, it is first necessary to classify the trees in the sample plot and determine whether to model them individually or in groups based on the number of trees of each type. For tree species with a large number (such as more than 100 trees), the relationship model between their tree age and carbon storage can be established independently; for species with a small number, they can be grouped according to the similarity of their growth characteristics and then modeled. Then, different mathematical models (such as linear, power function, exponential model, etc.) are used to fit these relationships and select The model with the highest value was selected as the final model. This model was then combined with tree carbon storage data from the previous year and the current year to calculate the annual carbon sink. This was done by comparing the difference in carbon storage between the two time points to determine how much carbon dioxide the trees absorbed and fixed during that time. Finally, the annual carbon sinks for trees in all plots were summarized to assess the carbon sequestration capacity of the entire study area.

[0039] For example, suppose that in all the urban green space plots surveyed, there are a large number of poplar trees (more than 100 trees) and a small number of pine trees and other miscellaneous trees. For the poplar trees, according to their tree age and corresponding carbon storage data, we try to use multiple mathematical models for fitting analysis and find that the exponential model can best describe the relationship between tree age and carbon storage (with the highest value). Therefore, an exponential model was selected as the carbon sequestration calculation model for poplar trees. At the same time, for those pine trees and other miscellaneous woods with smaller numbers, they were grouped according to their growth characteristics and corresponding models were established for each type of tree. Next, these models were used, combined with the tree carbon storage data from the previous year and the current year, to calculate the carbon sequestration of each type of tree during this period. For example, if the average carbon storage of poplar trees was 20 tons / hectare last year and increased to 25 tons / hectare this year, this means that in the past year, an additional 5 tons of carbon were fixed per hectare of poplar forest. By accumulating the annual carbon sequestration of trees in all sample plots, we can obtain the total carbon sequestration for the entire study area, which is crucial for assessing the carbon sequestration potential of cities.

[0040] It should be noted that in specific implementation scenarios, building on the above approach, the model for the relationship between tree age and carbon storage can also incorporate the influence of environmental factors such as soil fertility, precipitation, and temperature, in addition to tree age. By comprehensively analyzing how these factors interact to influence tree growth and carbon storage capacity, it is possible to more accurately predict the carbon sequestration of trees under different conditions, thereby optimizing urban greening strategies.

[0041] S5. Extract the normalized vegetation index value and the corresponding carbon sink data of the sample point, establish a relationship model between the normalized vegetation index and the carbon sink, and apply the model to perform spatial calculations on the study area to generate a carbon sink spatial distribution map. Integrate the carbon density spatial distribution map and the carbon sink spatial distribution map to output the carbon storage capacity assessment results of different green space types in the study area.

[0042] As mentioned above, after completing the sample survey and calculating the carbon sequestration for each plot, the next step is to extract Normalized Difference Vegetation Index (NDVI) values ​​from these plot points and correlate them with the corresponding carbon sequestration data. Using statistical analysis software (such as R), various mathematical models (e.g., linear, exponential, power function, or logarithmic models) are tested to establish a relationship between NDVI and carbon sequestration. The model with the best fit is selected as the final model to ensure that it accurately describes the relationship between the two. Next, using Geographic Information System (GIS) tools (such as the Raster Calculator in ArcGIS), the selected relationship model is applied to the NDVI layer for the entire study area, achieving spatial expansion from point to surface and generating a spatial distribution map of carbon sequestration. Finally, the spatial distribution maps of carbon density and carbon sequestration are integrated to produce an assessment of the carbon storage capacity of different green space types within the study area. This step not only reveals differences in carbon sequestration potential across regions but also provides a scientific basis for urban planning and ecological management.

[0043] For example, suppose that a field survey of multiple plots is completed in the core urban area of ​​a city, and the NDVI value and carbon sequestration amount of each plot are recorded. Next, the data are processed using R language software, and different regression models (such as linear, exponential, power function, etc.) are tried to fit the relationship between NDVI and carbon sequestration amount. Assume that the power function model has the highest goodness of fit ( The model with the largest value is used as the final model. Then, using the raster calculator function of ArcGIS, the model is applied to the NDVI layer of the entire study area to generate a spatial distribution map of carbon sequestration. For example, in the central area of ​​the park green space (G1), due to the dense vegetation and good health, the NDVI value is high, so the carbon sequestration predicted by the model is also high; while in the edge area of ​​the attached green space (XG), the vegetation coverage is low, the NDVI value is relatively low, and the corresponding carbon sequestration is also low. Finally, the generated carbon density spatial distribution map and carbon sequestration spatial distribution map are integrated to output a comprehensive evaluation result map, which intuitively shows the differences in carbon storage and sequestration capacity of different green space types.

[0044] It should be noted that in specific implementation scenarios, the above approach can be combined with other remote sensing-derived indicators (such as leaf area index (LAI) and fractional vegetation cover (FVC)) and topographic factors (such as altitude, slope, and aspect) in addition to NDVI to develop a multi-indicator integrated model to improve the accuracy of carbon sink estimation. For example, although vegetation in certain high-altitude areas may have lower NDVI values, their carbon sinks may be higher than those in lower altitudes due to unique climatic conditions. By incorporating more relevant factors, a more comprehensive reflection of actual carbon sinks can be achieved.

[0045] According to one embodiment of the present application, the step S1 is specifically as follows: The resolution of the high-definition remote sensing image is 10 meters, and the boundaries of the study area are delineated by mask processing technology; The urban green space planning document includes vector data of park green space, protective green space, square land, ancillary green space and regional green space, which is used to assist in confirming the spatial distribution range of various types of green space; The calculation of the normalized vegetation index is achieved through a band operation method, and the preset threshold range is [0.2, 1]. The sampling point layout plan is generated through a GIS tool, and the number of sampling points is adjusted according to the area proportion of each type of green space. At the same time, the distribution of sampling points is optimized through spatial uniformity analysis to ensure that key areas of different green space types are covered.

[0046] As mentioned above, the first step is to acquire high-definition remote sensing imagery with a 10-meter resolution. Using mask processing techniques, we can precisely define the boundaries of the study area, eliminating data interference outside the study area. This helps us focus on the actual research object and improves the accuracy of subsequent analysis.

[0047] Next, the spatial distribution of various green space types was determined based on the urban green space planning document. This document includes vector data for park green space (G1), protective green space (G2), plaza land (G3), ancillary green space (XG), and regional green space (EG). This data is crucial for identifying and distinguishing different types of green space, as it not only reflects the greening layout in urban planning but also directly influences the assessment of carbon sequestration capacity. Using this vector data, the specific location and coverage of each type of green space can be determined.

[0048] Next, the Normalized Difference Vegetation Index (NDVI) was calculated using a band-wise operation. The NDVI is an important indicator of vegetation growth, typically ranging from 0.2 to 1. Within this range, higher values ​​indicate more lush vegetation, while lower values ​​may indicate sparse vegetation or non-vegetated surfaces. Using this preset threshold range to select grid cells that meet the definition of urban green space effectively distinguishes true green space from other land uses.

[0049] Finally, to ensure accurate sampling of different green space types, a rational sampling point layout plan must be designed. This plan is generated using GIS tools, and the number of sampling points is adjusted based on the area proportion of each type of green space. For example, if a particular type of green space is large, the number of sampling points within that type of green space should be increased accordingly. Furthermore, a spatial uniformity analysis is required to optimize the distribution of sampling points to ensure that key areas of different green space types within the study area are adequately covered. This is done to minimize data errors caused by sample selection bias and ensure that the resulting carbon storage quantification results are highly representative and reliable. In this way, a comprehensive and accurate understanding of the carbon storage status of various green space types within the study area can be achieved.

[0050] According to one embodiment of the present application, the step S2 is specifically as follows: The field sample survey used a standard sample area of ​​20m×20m; The tree age is determined by comparing remote sensing images of different years, observing the number of branches, measuring the number of annual rings of felled tree stumps, or by expert experience and judgment; The coverage area of ​​the shrubs is measured by visual estimation or grid method, and the species are confirmed by field identification or specimen comparison; The tree biomass is calculated using an allometric growth model, and the shrub biomass is calculated based on the biomass per unit area parameters corresponding to the coverage area and type; The carbon content parameter is determined according to plant species or local standards, and the total carbon storage of the sample is generated through the conversion relationship between biomass and carbon storage.

[0051] As mentioned above, the first step is to conduct a field survey using a pre-defined standard sample size. Each sample size is set at 20m x 20m to ensure consistency and comparability of data collection while also covering sufficient vegetation coverage to provide reliable data support.

[0052] To determine the age of trees, a variety of methods are used for comprehensive judgment: Comparing remote sensing images from different years: By viewing historical remote sensing images, we can observe the changes in the crown width of target trees in different years and estimate their age based on the growth rate model.

[0053] Observe the number of branches: For some types of trees, you can infer the tree's age by observing the number of branches on the main trunk. For example, if a certain type of tree adds a major branch every year, you can roughly estimate the tree's age based on the number of branches.

[0054] Measuring the number of growth rings on felled tree stumps: If there are felled trees in the sample plot, the number of growth rings on the stumps can be directly measured to determine the tree age.

[0055] Expert judgment: The results of the preliminary estimate will be submitted to a review panel composed of multiple forestry experts. Through cross-comparison and on-site verification, the estimated values ​​with large errors will be corrected to improve the accuracy of tree age estimation.

[0056] The relevant parameters of the shrubs were measured and confirmed by the following methods: Cover: This can be measured using either a visual estimate or a grid method. The visual estimate is a direct estimate of the ground area covered by shrubs. The grid method divides the sample into smaller grids, counts the number of grids covered by shrubs, and calculates the total cover.

[0057] Species Identification: Confirm the species of shrubs through field identification or specimen comparison. Field staff identify the plant based on morphological characteristics (such as leaf shape, flower and fruit characteristics, etc.), and if necessary, collect specimens for further laboratory identification.

[0058] Next, the biomass is calculated based on the obtained growth parameters: Tree biomass: Using allometric growth models (e.g. ), where D is the diameter at breast height (DBH), H is the plant height, and a and b are constants. These constants are usually determined based on existing research results or local standards for specific tree species.

[0059] Shrub biomass: Calculated based on the biomass per unit area parameters corresponding to the shrub cover area and type. For example, broadleaf shrubs, coniferous shrubs, and mixed broadleaf and coniferous shrubs may have different biomass per unit area parameters. These parameters can be obtained from local standards (such as DB11 / T 953-2013).

[0060] Finally, the carbon content parameter is used to convert biomass into carbon storage: Carbon content parameters are determined based on plant species or local standards. For example, the carbon content of trees generally ranges from 0.43 to 0.52, while that of shrubs is 0.45. By multiplying the biomass by the corresponding carbon content, the carbon storage of each tree and shrub can be calculated. The carbon storage of all trees and shrubs is then summed to form the total carbon storage of the quadrat.

[0061] According to one embodiment of the present application, the step S3 is specifically as follows: The relationship model between the normalized vegetation index and carbon density is fitted to the sample point data using a linear, exponential, power function or logarithmic model using R language software, and the model with the highest goodness of fit is selected as the final model; The spatial calculation is achieved through the raster calculator of GIS software to generate a carbon density spatial pattern map that distinguishes different green space types and annotates the regional total carbon storage and mean value.

[0062] As mentioned above, we first need to extract the Normalized Difference Vegetation Index (NDVI) values ​​and corresponding carbon storage data from the sample point data. These data will be used to establish a relationship model between NDVI and carbon density.

[0063] Data preparation: Collect NDVI values ​​and corresponding carbon storage data for each plot point. These data come from previous field surveys and calculations.

[0064] Model fitting: Use statistical analysis software (such as R language) to process the sample point data. Try various mathematical models to fit the relationship between NDVI and carbon density, including but not limited to linear, exponential, power function or logarithmic models. Each model will generate a fitting result and calculate the corresponding goodness of fit index (such as value).

[0065] Select the best model: Select the best model as the final model based on the goodness of fit index. Usually choose The model with the highest value is selected because it means that this model best describes the relationship between NDVI and carbon density.

[0066] Apply the model to the study area: Using the raster calculator function in geographic information system (GIS) software (such as ArcGIS), the selected optimal model is applied to the NDVI layer of the entire study area. This allows spatial expansion from point data to surface data, generating a spatial pattern map of carbon density that distinguishes different green space types.

[0067] Mark key information: Mark the total amount and mean of regional carbon storage in the generated carbon density spatial pattern map. Specifically: Total carbon storage: Sum up the carbon storage of all sample plots to obtain the total carbon storage of the entire study area.

[0068] Mean carbon density: Calculate the average carbon density of all sample plots to reflect the overall level of carbon density in the study area.

[0069] These annotations help to visually demonstrate the differences in carbon storage capacity between different regions and provide a scientific basis for urban planners.

[0070] According to one embodiment of the present application, the step S4 is specifically as follows: The carbon sequestration calculation model is grouped and fitted based on tree species or growth characteristics, and the carbon sequestration calculation model type includes a logistic model, a Gompertz model or other nonlinear models; The annual carbon sequestration is determined by calculating the difference between the carbon storage in the current year and the previous year, and is corrected based on the tree growth rate parameters.

[0071] As mentioned above, a mathematical model for estimating carbon sequestration is first constructed based on tree age and carbon storage data obtained from sample plot surveys. Because different tree species or trees with similar growth characteristics differ in growth rate, biomass accumulation pattern, and carbon fixation capacity, the trees need to be classified. For tree species with a large population and statistical significance, separate carbon sequestration calculation models are developed. For tree species with a small population or similar growth characteristics, they are grouped according to their common biological characteristics and a unified model is fitted. This classification or group modeling approach can more accurately reflect the carbon accumulation patterns of different tree types, improving the model's applicability and prediction accuracy.

[0072] When choosing a model, we used mathematical models that can describe the nonlinear characteristics of vegetation growth, including logistic models, Gompertz models, or other nonlinear models applicable to biological growth processes. These models can better depict the changing trends in carbon storage of trees at different growth stages (e.g., slow growth in juvenile stages, rapid accumulation in middle stages, and stabilization in old age), and are more scientific and reasonable than linear models. By fitting the measured tree age and carbon storage data into these models, the model with the highest goodness of fit was selected as the carbon sequestration model for that tree species or group.

[0073] After the model is constructed, it is used to calculate the carbon storage of trees in the current year and the previous year, respectively. The difference between the two is the carbon sink for that year. This difference reflects the net amount of carbon fixed by trees in the sample plot through photosynthesis over a one-year period and is a core indicator for measuring the carbon absorption capacity of green spaces. To improve the accuracy of the calculation results, the preliminary calculation results need to be corrected in combination with the growth rate parameters of the trees. Growth rate parameters can be derived from long-term observation data, literature, or local standards, and are used to reflect the impact of environmental factors (such as climate conditions, soil quality, and management level) on tree growth rate. For example, in years with low precipitation or a significant urban heat island effect, tree growth may be inhibited. In this case, the carbon sink estimate can be appropriately lowered by introducing a negative correction coefficient. Conversely, in years with excellent growth conditions, a positive correction can be made.

[0074] According to one embodiment of the present application, the step S5 is specifically as follows: The relationship model between the normalized difference vegetation index and carbon sequestration was fitted to the sample point data using R language software, and the model with the highest goodness of fit was selected as the final model; The spatial calculation is achieved through a raster calculator in GIS software to generate a carbon sink spatial pattern map and mark the spatial distribution characteristics of high carbon sink areas and low carbon sink areas.

[0075] As mentioned above, we first need to use the sample point data obtained from preliminary surveys and calculations to establish a relationship model between the Normalized Difference Vegetation Index (NDVI) and carbon sequestration. Each sample point contains two key data points: the NDVI value extracted from high-definition remote sensing imagery, which reflects the vegetation cover and growth activity at that location; and the annual carbon sequestration value calculated in step S4, which represents the net amount of carbon sequestered by the sample point in a year. These two types of data form the basis for model fitting.

[0076] To find the optimal mathematical relationship between NDVI and carbon sequestration, R language software was used to perform fitting analysis on the sample point data using various models, including but not limited to linear models, exponential models, power function models, logarithmic models, etc. Each model generates a corresponding fitting curve and calculates the goodness of fit indicators, such as the coefficient of determination ( ), root mean square error (RMSE), etc. By comparing these indicators, the model with the highest goodness of fit is selected as the final relationship model. For example, if the power function model If the value is significantly higher than that of other models, it is determined to be the NDVI-carbon sink conversion model for the study area. This model can effectively reflect the nonlinear or saturated relationship between vegetation index and carbon absorption capacity, ensuring the accuracy of subsequent spatial expansion.

[0077] Once the model is finalized, the spatial calculation phase begins. Using the raster calculator function within the Geographic Information System (GIS) software, the selected relationship model is applied to the NDVI raster data for the entire study area. The NDVI data covers the entire region in pixels, and the model calculates the corresponding carbon sink value pixel by pixel, thus expanding the scope from a limited number of sample points to a continuous spatial surface. Ultimately, a complete spatial pattern map of carbon sinks is generated, with each pixel value representing the estimated carbon sink at that location, forming a continuous spatial distribution pattern.

[0078] To further enhance the map's readability and applicability, the generated carbon sink spatial pattern map clearly identifies the spatial distribution characteristics of high- and low-carbon sink areas. High-carbon sink areas typically correspond to well-managed green spaces with high vegetation cover, vigorous tree growth, and good management, such as large parks or mature shelterbelts. Low-carbon sink areas tend to occur in areas with sparse vegetation, dominated by shrubs or lawns, or frequent human disturbance, such as some ancillary green spaces or green spaces along transportation routes. By setting appropriate thresholds or employing spatial cluster analysis methods, areas with carbon sinks significantly above or below the regional average are identified and marked with different colors or symbols in the map, visually demonstrating the spatial differences in carbon sink capacity.

[0079] According to one embodiment of the present application, it further includes: The carbon storage capacity assessment results include the total regional carbon storage, average carbon density, spatial agglomeration characteristics and the contribution of different green space types; The decision support content includes recommendations for priority protection of high-carbon sink areas, recommendations for optimization and transformation of low-carbon sink areas, and scientific guidance on urban greening layout.

[0080] As mentioned above, regional total carbon storage refers to the cumulative carbon storage of all green space plots within the study area. It reflects the total amount of organic carbon sequestered by the region's vegetation ecosystem and is a key indicator of the carbon sequestration capacity of urban green spaces. This value, obtained by aggregating measured or model-estimated carbon storage from all plots, serves as baseline data for urban carbon storage, enabling long-term monitoring and comparative analysis.

[0081] Mean carbon density refers to the average carbon storage per unit area of ​​green space. It is calculated by dividing the total regional carbon storage by the total green space area. This metric is used to evaluate the carbon storage efficiency of different green space types or regions, helping to identify green space types and management methods with higher carbon storage efficiency.

[0082] Spatial clustering refers to the spatial distribution of carbon density and carbon sinks. Spatial analysis methods can be used to identify high-value clusters (e.g., hotspots) and low-value clusters (e.g., coldspots). For example, certain large parks or ecological corridors may form distinct high-value carbon sink clusters, while urban fringes or newly constructed green spaces may exhibit dispersed or low-value distributions. This characteristic helps reveal the spatial pattern of carbon functions in urban green spaces and the factors influencing them.

[0083] The contribution of different green space types refers to the proportion of each type of green space (such as park green space, protective green space, ancillary green space, and plaza land) in total carbon storage and annual carbon sinks. This categorized statistical analysis can clarify which green space types play a dominant role in carbon storage and sinking, providing a quantitative basis for optimizing green space structure. For example, although ancillary green space has a large area but a low unit carbon density, park green space, which is smaller in area, contributes a higher proportion of carbon sinks. This information is of great guiding significance for green space planning.

[0084] Based on the above assessment results, corresponding decision support content is generated to guide urban green space planning and management practices. Specifically, it includes: Prioritize the protection of high-carbon sink areas, such as avoiding encroachment on high-carbon sink green spaces during urban development, limiting high-intensity human interference, and strengthening conservation and management to maintain their high carbon sink capacity. These areas can also be included in urban ecological control lines or key ecological function zones for long-term protection.

[0085] Recommendations for optimizing and renovating low-carbon sink areas are proposed, such as increasing tree planting density, replacing them with high-carbon-sequestering species, improving soil conditions, or enhancing maintenance levels to enhance their carbon sequestration potential. For plazas or ancillary green spaces primarily composed of lawns, it is recommended to gradually increase the proportion of trees and shrubs, creating a multi-layered greening structure to improve carbon sequestration capacity per unit area.

[0086] Provide scientific guidance for urban greening layout, and propose green space system optimization plans based on urban master plans, current land use status, and future development directions. For example, prioritize the layout of green spaces with high carbon sink potential in urban expansion areas, build ecological corridors in areas with poor connectivity to enhance carbon sink network effects, or concentrate high-carbon sink vegetation in areas with significant urban heat island effects to synergistically improve microclimates.

[0087] According to one embodiment of the present application, the carbon sequestration calculation model further includes: In view of the fact that tree growth rate is affected by climate change and soil fertility environmental factors, a correction coefficient K is introduced, K=1±Δ, where Δ is the comprehensive weight value of the influence of environmental factors; The carbon sequestration calculation model was retrospectively verified using the sample carbon storage data from the past five years to ensure that the deviation rate between the predicted values ​​and the measured values ​​of the carbon sequestration calculation model was less than 10%.

[0088] As mentioned above, to improve the applicability and prediction accuracy of the carbon sequestration calculation model under different environmental conditions, an environmental correction mechanism is further introduced. Since tree growth rates depend not only on tree age and species but are also significantly influenced by external environmental factors, such as annual mean temperature fluctuations, precipitation fluctuations, air pollution levels, differences in soil fertility, and the urban heat island effect, these factors can cause trees of the same age to exhibit different carbon accumulation rates in different regions or years. Therefore, a correction factor, K, is introduced into the existing carbon sequestration calculation model to dynamically adjust the model output. This correction factor, K, is calculated as K = 1 ± Δ, where Δ represents the combined weight of environmental factors and is determined based on actual monitoring data or regional environmental assessment results. When environmental conditions are superior to the baseline (e.g., adequate precipitation, fertile soil, and suitable temperature), Δ is positive (K > 1), indicating accelerated growth and a corresponding increase in carbon sequestration. When environmental conditions are adverse (e.g., drought, poor soil, or high temperature stress), Δ is negative (K < 1), indicating growth inhibition and a moderate decrease in carbon sequestration. The determination of Δ can be based on a multi-factor weighted scoring method, combining information such as meteorological data, soil testing reports, and urban environmental quality bulletins to quantitatively evaluate various environmental factors and assign corresponding weights, ultimately synthesizing a comprehensive impact value.

[0089] In order to further verify the reliability and stability of the carbon sequestration calculation model, historical data were used for retrospective verification. The specific method is to use the measured carbon storage data accumulated in the same or adjacent sample plots in the past five years as the verification benchmark. The carbon storage results predicted by the model after inputting the corresponding tree age and environmental parameters in each year are compared one by one with the carbon storage obtained from the actual survey of that year, and the deviation rate between the predicted value and the measured value is calculated. If the average deviation rate of a certain type of tree or a certain type of green space exceeds 10%, it is determined that the model needs to be optimized, and the model structure or correction coefficient parameters need to be readjusted until the deviation rate is controlled within 10%. This verification process ensures that the model is not only valid in the current year, but also has a certain degree of temporal continuity and long-term prediction capabilities, and can reflect the dynamic trend of vegetation carbon sequestration.

[0090] An embodiment of a second aspect of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of any one of the embodiments of the first aspect when executing the program.

[0091] Figure 2 An example of a physical structure diagram of an electronic device is shown below. Figure 2 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the method in any embodiment of the first aspect above, the method including: S1. Obtain high-definition remote sensing images of the study area and, in combination with urban green space planning documents, determine the spatial distribution range of various types of green space within the study area. Calculate the normalized vegetation index based on the high-definition remote sensing images, extract grids with normalized vegetation index values ​​within a preset threshold range as urban green space, and generate corresponding sampling point layout plans based on the area proportions of various types of green space. S2. Conduct a field survey of the sampling points according to the sampling point layout plan, record the growth parameters of the trees and shrubs in the sample plot, including the diameter at breast height, plant height, and age of the trees, and the coverage area and type of the shrubs. Based on the growth parameters, calculate the biomass of the trees and shrubs in the sample plot, convert it into carbon storage in combination with the carbon content parameter, and summarize it to form the total carbon storage of the sample plot; S3. extracting the normalized vegetation index value and the corresponding carbon storage data of the sample point, establishing a relationship model between the normalized vegetation index and carbon density, and applying the relationship model to perform spatial calculations on the study area to generate a carbon density spatial distribution map; S4. Based on the relationship between the tree age and carbon storage of the trees in the sample plot, a carbon sequestration calculation model is constructed, and the annual carbon sequestration of the sample plot is calculated in combination with the carbon sequestration calculation model; S5. Extract the normalized vegetation index value and the corresponding carbon sink data of the sample point, establish a relationship model between the normalized vegetation index and the carbon sink, and apply the model to perform spatial calculations on the study area to generate a carbon sink spatial distribution map. Integrate the carbon density spatial distribution map and the carbon sink spatial distribution map to output the carbon storage capacity assessment results of different green space types in the study area.

[0092] Furthermore, the logic instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as standalone products, stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for causing a computer device (such as a personal computer, server, or network device) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories, random access memories, magnetic disks, or optical disks.

[0093] On the other hand, the present invention further provides a computer program product, comprising a computer program. The computer program may be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the method provided by each of the above methods, including: S1. Obtain high-definition remote sensing images of the study area and, in combination with urban green space planning documents, determine the spatial distribution range of various types of green space within the study area. Calculate the normalized vegetation index based on the high-definition remote sensing images, extract grids with normalized vegetation index values ​​within a preset threshold range as urban green space, and generate corresponding sampling point layout plans based on the area proportions of various types of green space. S2. Conduct a field survey of the sampling points according to the sampling point layout plan, record the growth parameters of the trees and shrubs in the sample plot, including the diameter at breast height, plant height, and age of the trees, and the coverage area and type of the shrubs. Based on the growth parameters, calculate the biomass of the trees and shrubs in the sample plot, convert it into carbon storage in combination with the carbon content parameter, and summarize it to form the total carbon storage of the sample plot; S3. extracting the normalized vegetation index value and the corresponding carbon storage data of the sample point, establishing a relationship model between the normalized vegetation index and carbon density, and applying the relationship model to perform spatial calculations on the study area to generate a carbon density spatial distribution map; S4. Based on the relationship between the tree age and carbon storage of the trees in the sample plot, a carbon sequestration calculation model is constructed, and the annual carbon sequestration of the sample plot is calculated in combination with the carbon sequestration calculation model; S5. Extract the normalized vegetation index value and the corresponding carbon sink data of the sample point, establish a relationship model between the normalized vegetation index and the carbon sink, and apply the model to perform spatial calculations on the study area to generate a carbon sink spatial distribution map. Integrate the carbon density spatial distribution map and the carbon sink spatial distribution map to output the carbon storage capacity assessment results of different green space types in the study area.

[0094] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for cigarette box image recognition provided by the above methods is implemented, and the method includes: S1. Obtain high-definition remote sensing images of the study area and, in combination with urban green space planning documents, determine the spatial distribution range of various types of green space within the study area. Calculate the normalized vegetation index based on the high-definition remote sensing images, extract grids with normalized vegetation index values ​​within a preset threshold range as urban green space, and generate corresponding sampling point layout plans based on the area proportions of various types of green space. S2. Conduct a field survey of the sampling points according to the sampling point layout plan, record the growth parameters of the trees and shrubs in the sample plot, including the diameter at breast height, plant height, and age of the trees, and the coverage area and type of the shrubs. Based on the growth parameters, calculate the biomass of the trees and shrubs in the sample plot, convert it into carbon storage in combination with the carbon content parameter, and summarize it to form the total carbon storage of the sample plot; S3. extracting the normalized vegetation index value and the corresponding carbon storage data of the sample point, establishing a relationship model between the normalized vegetation index and carbon density, and applying the relationship model to perform spatial calculations on the study area to generate a carbon density spatial distribution map; S4. Based on the relationship between the tree age and carbon storage of the trees in the sample plot, a carbon sequestration calculation model is constructed, and the annual carbon sequestration of the sample plot is calculated in combination with the carbon sequestration calculation model; S5. Extract the normalized vegetation index value and the corresponding carbon sink data of the sample point, establish a relationship model between the normalized vegetation index and the carbon sink, and apply the model to perform spatial calculations on the study area to generate a carbon sink spatial distribution map. Integrate the carbon density spatial distribution map and the carbon sink spatial distribution map to output the carbon storage capacity assessment results of different green space types in the study area.

[0095] Example 2 S1) Determine the area and geographical scope of the urban area to be studied and obtain high-definition remote sensing images; The research scope mainly includes the areas within the urban development boundary lines of various districts in a city and the green spaces in the surrounding areas. The regional green spaces mainly include mountain parks, suburban parks, scenic spots and strip riverside green spaces around the urban development boundaries.

[0096] S2) Obtain the spatial scope of various types of green space based on urban green space planning; Based on the vector file of a city's green space system planning, the regional green space around the urban development boundary was supplemented and delineated according to high-definition imagery. The spatial scope of the park green space (G1), protective green space (G2), square land (G3), ancillary green space (XG) within the urban development boundary and the regional green space (EG) around the urban development boundary were clarified.

[0097] S3) Extract the green space grid of the study area and identify the sampling point locations The high-definition image of the estimated area was interpreted and the NDVI map of the study area was obtained using the band operation method. The grid with NDVI in the range of [0.2, 1] was extracted and defined as the urban green space. The green space area of ​​each green space type was then counted, and G1 was 32.05 km 2 , G2 is 20.37km 2 , G3 is 0.76km 2 , XG is 266.89km 2 EG is 190.24km 2 The total green space grid area is 510.30km 2 Among them, residential green space is the main green space of the ancillary green space. Based on the area and minimum number of survey plots of different green space types, the relatively simple plant species in regional green spaces and the rich plant species in park green spaces, the final survey plot points for G1, G2, G3, XG and EG were set to 150, 50, 50, 650 and 50 respectively.

[0098] Using the Create Fishnet tool in ArcGIS, a certain number of sampling points were generated. Based on the spatial uniformity and distribution of the sampling points, the number or location of the quadrats was optimized. The locations of the quadrats were further adjusted based on the current problems encountered in the actual survey, and finally a spatial distribution map of the quadrats was obtained.

[0099] S4) Conduct sample surveys of various urban green spaces to obtain information on the growth of trees and shrubs in the sample plots Green space vegetation plot surveys were conducted according to longitude and latitude. The survey plots were 20m×20m. A ruler was used to measure the diameter at breast height and crown width, and a height meter was used to measure the tree height. The height, area, and type of shrubs were recorded. The vegetation coverage rate was obtained based on the vegetation coverage area of ​​the plots.

[0100] The relationship between tree age and biomass is affected by the tree's growth environment, soil fertility, differences between local and foreign tree species, and human activities. This study comprehensively considered the following four aspects to determine tree age: (1) Based on the latitude and longitude coordinates, high-definition remote sensing images of the study area in different years were compared to estimate the approximate range of tree age; (2) For branched trees, the main branches of the tree were identified, and the number of branches N of the tree was determined. N+1 is approximately the minimum age of the tree; (3) The stumps of felled trees in the sample area were found, and the tree age was determined based on the number of annual rings and the number of plant branches; (4) The morphology of each tree was photographed, and tree data was compiled based on the tree's height, diameter at breast height, branches, configuration structure, and surrounding environment. Three senior forestry experts were asked to judge the tree age. In summary, the tree age was finally determined. It is worth noting that there is a certain degree of subjectivity in the estimation of tree age.

[0101] S5) Calculation of carbon storage in vegetation plots Based on the tree species, diameter at breast height (D), plant height (H), and tree age (A) characteristics of the sample plot, the tree allometric growth model was used to calculate the tree biomass. Taking into account the maintenance and pruning of urban green spaces, a plant growth model that covers both diameter at breast height and plant height was selected. The plant growth mathematical model is as follows: W 乔 = a×(D 2 ×H) b ; Where W is the biomass of a single tree, and a and b are constants.

[0102] As shown in “Table 1 Plant Allometric Growth Model Standard Table”, this embodiment specifically adopts the plant allometric growth model standard of the Beijing local standard “Technical Specifications for Measurement and Monitoring of Forestry Carbon Sequestration (DB11 / T 953—2013)”.

[0103] Table 1 Standard table of plant allometric growth models

[0104] The product of the tree biomass in each quadrat and the carbon content conversion coefficient is the tree carbon storage, specifically: C 乔 =W 乔 ×T Where Ctree represents tree carbon storage, and T is the tree carbon conversion coefficient, which ranges from 0.43 to 0.52. The T value is based on DB11 / T 953-2013. Shrub carbon storage is the product of shrub area (S), biomass conversion factor (E), and carbon content (Tshrub), where Tshrub is set to 0.45. The specific formula is as follows: C 灌 =S ×E×T 灌 The sum of the carbon storage of trees in a plot is used to determine the total carbon storage of trees, while the sum of the carbon storage of trees and shrubs is used to determine the carbon storage of the plot. Green space carbon density is the ratio of carbon storage to area in a plot.

[0105] S6) Construct a relationship model between NDVI and green space carbon density to obtain the spatial pattern of urban green space carbon density; The NDVI values ​​of the sample points were extracted and the R language software was used to measure the carbon density and NDVI values ​​of various types of green space sample data. The linear model, exponential model, power function model, logarithmic model and other relationship fitting results were compared. value, get The model fitting of the highest carbon density and NDVI values. During the actual survey, some sample plots had lower carbon density levels, corresponding to NDVI values ​​between 0.15 and 0.2, and these data were also used in the model fitting.

[0106] The results show that the exponential function and power function are the most significant models for fitting carbon density and NDVI. Among them, the power function fitting models of G1, G3 and XG are Exponential function fitting model of highest, G2 and EG Therefore, using the ArcGIS raster calculator, the model formula was spatially calculated to obtain the spatial pattern of carbon density that distinguishes different green space types, which enables further analysis of the total carbon storage, mean, and carbon density levels in the study area.

[0107] S7) Calculation of vegetation carbon sequestration in green space quadrat The trees were classified according to their species. For tree species with more than 100 trees, the relationship between tree age and carbon storage was fitted separately. For tree species with relatively small numbers, tree species with similar growth characteristics were grouped and the relationship between tree age and carbon storage was fitted. Specifically, logarithmic model, power function model, exponential model, linear model, logistic model and Gompertz model fitting were carried out. According to the differences in plant species and growth characteristics of various green spaces, the relationship between tree age and carbon storage was calculated. The relationship fitting model with the highest value is shown in "Table 2 Fitting model of the relationship between tree age and carbon storage".

[0108] Table 2 Fitting model of the relationship between tree age and carbon storage

[0109] Based on the above relationship model and the carbon content conversion coefficient, the tree carbon storage of the current year and the previous year was calculated. The difference between the two values ​​was used to obtain the total tree carbon sink for the current year. The total tree carbon sink in the sample was then accumulated to obtain the sample carbon sink value.

[0110] S8) Construct a relationship model between sample carbon sequestration and NDVI to obtain the spatial pattern of urban green space carbon sequestration Different types of green spaces were distinguished, and the relationship between carbon sequestration and NDVI was fitted respectively. Linear or nonlinear fitting curves such as linear function, power function, exponential function, etc. were tried, and finally the results were obtained. The relationship model with the highest value was constructed using the Raster Calculator. Spatialization was performed on different green space types to generate a carbon sequestration spatial pattern map for the study area. This allowed for further analysis of carbon sequestration levels across administrative regions, green space types, and spatial scales, as well as the spatial patterns of carbon sequestration.

[0111] S9) Analyze the spatial characteristics of landscape pattern indicators in the study area and obtain the landscape pattern indicator values ​​of the sample points; The NDVI raster was reclassified into 7 categories. Using Fragstats 4.3 software, with a moving window of 30 m, 10 typical landscape indicators at the landscape pattern scale were analyzed: (1) Aggregation Index (AI) (2) Maximum Patch Index (LPI) (3) Landscape Division Index (DIVISION) (4) Shannon Diversity Index (SHDI) (5) Shannon Evenness Index (SHEI) (6) Patch Connectivity Index (COHESION) (7) Average Patch Area (AREA_MN) (8) Average Nearest Neighbor Distance (ENN_MN) (9) Patch Shape Index (SHAPE) (10) Fractal Dimension Index (FRACT). The spatial distribution map of landscape pattern indicators was obtained by running the software.

[0112] All indicators were opened in ArcGIS Pro, and the landscape pattern indicator values ​​at the sample points were extracted based on the latitude and longitude of the sample points. Due to the missing values ​​of non-green space in the NDVI raster map, data extraction was performed to finally extract the values ​​of 10 landscape pattern indicators for 290 sample points.

[0113] S10) Using XGBoost modeling and SHAP value interpretation, we analyzed the relationship between landscape pattern indicators and carbon density or carbon sequestration, and evaluated the relationship between landscape characteristics of various green spaces and carbon storage functions.

[0114] Using R language software packages such as "shapviz", "xgboost", and "Boruta", with 10 landscape indicators as independent variables and measured carbon density or carbon sequestration as dependent variables, the XGBoost model was applied to evaluate the marginal effects of landscape pattern indicators on carbon density or carbon sequestration and their relative importance. During the model analysis process, 70% of the samples were used for XGBoost model training and 30% of the data were used for model testing. The root mean square error (RMSE) and the fitting rate were used to fit the predicted values ​​and the test set data. The model fitting performance was quantitatively evaluated using the SHAP method. The SHAP value of each landscape indicator was calculated to quantify its relative importance to the carbon density or carbon sink prediction results. Further analysis was conducted to obtain the overall impact of the characteristics of a single independent variable on the results (SHAP value), and the image characteristics of the main independent variables (SHAP main value) and the interactive independent variables (SHAP interactive value) on the predicted dependent variables. Based on the above analysis results, the influence relationship of landscape pattern indicators in the interpretation of carbon density or carbon sink was obtained, thus characterizing the current status of the relationship between the construction of urban landscape beauty and richness and the carbon storage function. This application uses carbon density as an example to graphically present the analysis data of the study area.

[0115] Anything not described in this application can be achieved by adopting or drawing on existing technologies.

[0116] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0117] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for quantifying carbon storage in urban green spaces based on remote sensing and field surveys, characterized in that: include: S1. Obtain high-definition remote sensing images of the study area and, in combination with urban green space planning documents, determine the spatial distribution range of various types of green space within the study area. Calculate the normalized vegetation index based on the high-definition remote sensing images, extract grids with normalized vegetation index values ​​within a preset threshold range as urban green space, and generate corresponding sampling point layout plans based on the area proportions of various types of green space. S2. Conduct a field survey of the sampling points according to the sampling point layout plan, record the growth parameters of the trees and shrubs in the sample plot, including the diameter at breast height, plant height, and age of the trees, and the coverage area and type of the shrubs. Based on the growth parameters, calculate the biomass of the trees and shrubs in the sample plot, convert it into carbon storage in combination with the carbon content parameter, and summarize it to form the total carbon storage of the sample plot; S3. extracting the normalized vegetation index value and the corresponding carbon storage data of the sample point, establishing a relationship model between the normalized vegetation index and carbon density, and applying the relationship model to perform spatial calculations on the study area to generate a carbon density spatial distribution map; S4. Based on the relationship between the tree age and carbon storage of the trees in the sample plot, a carbon sequestration calculation model is constructed, and the annual carbon sequestration of the sample plot is calculated in combination with the carbon sequestration calculation model; S5. Extract the normalized vegetation index value and the corresponding carbon sink data of the sample point, establish a relationship model between the normalized vegetation index and the carbon sink, and apply the model to perform spatial calculations on the study area to generate a carbon sink spatial distribution map. Integrate the carbon density spatial distribution map and the carbon sink spatial distribution map to output the carbon storage capacity assessment results of different green space types in the study area.

2. The method according to claim 1, characterized in that The S1 step is specifically as follows: The resolution of the high-definition remote sensing image is 10 meters, and the boundaries of the study area are delineated by mask processing technology; The urban green space planning document includes vector data of park green space, protective green space, square land, ancillary green space and regional green space, which is used to assist in confirming the spatial distribution range of various types of green space; The calculation of the normalized vegetation index is achieved through a band operation method, and the preset threshold range is [0.2, 1]. The sampling point layout plan is generated through a GIS tool, and the number of sampling points is adjusted according to the area proportion of each type of green space. At the same time, the distribution of sampling points is optimized through spatial uniformity analysis to ensure that key areas of different green space types are covered.

3. The method according to claim 1, characterized in that The S2 step is specifically as follows: The field sample survey used a standard sample area of ​​20m×20m; The tree age is determined by comparing remote sensing images of different years, observing the number of branches, measuring the number of annual rings of felled tree stumps, or by expert experience and judgment; The coverage area of ​​the shrubs is measured by visual estimation or grid method, and the species are confirmed by field identification or specimen comparison; The tree biomass is calculated using an allometric growth model, and the shrub biomass is calculated based on the biomass per unit area parameters corresponding to the coverage area and type; The carbon content parameter is determined according to plant species or local standards, and the total carbon storage of the sample is generated through the conversion relationship between biomass and carbon storage.

4. The method according to claim 1, wherein The S3 step is specifically as follows: The relationship model between the normalized vegetation index and carbon density is fitted to the sample point data using a linear, exponential, power function or logarithmic model using R language software, and the model with the highest goodness of fit is selected as the final model; The spatial calculation is achieved through the raster calculator of GIS software to generate a carbon density spatial pattern map that distinguishes different green space types and annotates the regional total carbon storage and mean value.

5. The method according to claim 1, wherein The S4 step is specifically as follows: The carbon sequestration calculation model is grouped and fitted based on tree species or growth characteristics, and the carbon sequestration calculation model type includes a logistic model, a Gompertz model or other nonlinear models; The annual carbon sequestration is determined by calculating the difference between the carbon storage in the current year and the previous year, and is corrected based on the tree growth rate parameters.

6. The method according to claim 1, wherein The step S5 is specifically as follows: The relationship model between the normalized difference vegetation index and carbon sequestration was fitted to the sample point data using R language software, and the model with the highest goodness of fit was selected as the final model; The spatial calculation is achieved through a raster calculator in GIS software to generate a carbon sink spatial pattern map and mark the spatial distribution characteristics of high carbon sink areas and low carbon sink areas.

7. The method according to claim 1, characterized in that Also includes: The carbon storage capacity assessment results include the total regional carbon storage, average carbon density, spatial agglomeration characteristics and the contribution of different green space types; The decision support content includes recommendations for priority protection of high-carbon sink areas, recommendations for optimization and transformation of low-carbon sink areas, and scientific guidance on urban greening layout.

8. The method according to claim 5, characterized in that The carbon sequestration calculation model also includes: In view of the fact that tree growth rate is affected by climate change and soil fertility environmental factors, a correction coefficient K is introduced, K=1±Δ, where Δ is the comprehensive weight value of the influence of environmental factors; The carbon sequestration calculation model was retrospectively verified using the sample carbon storage data from the past five years to ensure that the deviation rate between the predicted values ​​and the measured values ​​of the carbon sequestration calculation model was less than 10%.

9. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps in the method according to any one of claims 1 to 8 are implemented.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps in the method according to any one of claims 1 to 8 are implemented.

Images