Overground vegetation carbon reserve inversion method, device, equipment and medium

Abstract

The invention discloses an overground vegetation carbon reserve inversion method, device and equipment and a medium, and relates to the field of ecological monitoring, and the method comprises the steps: obtaining a hyperspectral image of a research region; classifying each pixel in the hyperspectral image by using a pre-trained classification model, and determining the vegetation type of each pixel; performing feature extraction on the hyperspectral image, and determining remote sensing features of each pixel; according to the remote sensing feature of each pixel, performing inversion on the biomass in the research area by adopting an inversion model to obtain the biomass of each pixel; the inversion model is obtained by training a training sample set in advance, the training sample set comprises remote sensing features and real biomass of each pixel in a plurality of sample plots, and the real biomass is determined based on hyperspectral images and three-dimensional structure parameters of the sample plots; and according to the carbon conversion coefficient of each vegetation and the vegetation type and biomass of each pixel, determining the aboveground vegetation carbon reserve of the research area. According to the invention, accurate inversion of the aboveground vegetation carbon reserves is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of ecological monitoring, and in particular to a method and device for retrieving aboveground vegetation carbon storage, equipment and medium. BACKGROUND

[0002] Aboveground biomass and carbon storage are important indicators for forest ecosystem carbon sink assessment and ecological monitoring. However, traditional biomass / carbon storage retrieval methods based on single spectral data of remote sensing are limited by the lack of structural information and scale inconsistency, and it is difficult to meet the needs of refined ecological monitoring. With the development of remote sensing and oblique photography technology, high spatial resolution spectral and three-dimensional data can be obtained, but how to effectively fuse multi-source information and accurately map plot biomass to image pixel scale remains a research difficulty. SUMMARY

[0003] The purpose of the present application is to provide a method and device for retrieving aboveground vegetation carbon storage, equipment and medium, which can realize pixel-level biomass retrieval and improve the accuracy of aboveground vegetation carbon storage.

[0004] To achieve the above purpose, the present application provides the following solutions: In a first aspect, the present application provides a method for retrieving aboveground vegetation carbon storage, comprising: obtaining hyperspectral images of a study area; classifying each pixel in the hyperspectral images using a pre-trained classification model to determine the vegetation type of each pixel; extracting features from the hyperspectral images to determine the remote sensing features of each pixel; retrieving the biomass in the study area using a retrieval model according to the remote sensing features of each pixel, to obtain the biomass of each pixel; wherein the retrieval model is pre-trained using a training sample set, which includes the remote sensing features of each pixel and the true biomass of each pixel in multiple plots; the true biomass is determined based on the hyperspectral images and three-dimensional structure parameters of the plot; determining the aboveground vegetation carbon storage of the study area according to the carbon conversion coefficient of each vegetation type, the vegetation type of each pixel and the biomass of each pixel.

[0005] In a second aspect, the present application provides a device for retrieving aboveground vegetation carbon storage, comprising: an image acquisition module for acquiring hyperspectral images of a study area; a pixel classification module for classifying each pixel in the hyperspectral images using a pre-trained classification model to determine the vegetation type of each pixel; a feature extraction module for extracting features from the hyperspectral images to determine the remote sensing features of each pixel; a biomass inversion module, configured to obtain the biomass of each pixel in the study area according to the remote sensing features of each pixel by using an inversion model, wherein the inversion model is obtained by training a training sample set in advance, the training sample set comprising the remote sensing features of each pixel in a plurality of sample plots and the real biomass of each pixel; and the real biomass is determined based on the hyperspectral image and the three-dimensional structure parameters of the sample plot; a carbon storage determination module, configured to determine the aboveground vegetation carbon storage in the study area according to the carbon conversion coefficient of each vegetation, the vegetation type of each pixel and the biomass of each pixel.

[0006] In a third aspect, the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned aboveground vegetation carbon storage inversion method.

[0007] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the above-mentioned aboveground vegetation carbon storage inversion method.

[0008] According to the specific embodiments provided by the present application, the present application has the following technical effects: the present application realizes the accurate inversion of the aboveground vegetation carbon storage by combining the hyperspectral image with the classification model, the inversion model and the carbon conversion coefficient. Specifically, first, the pixels of the hyperspectral image are accurately classified by means of the pre-trained classification model, and the vegetation type is determined, thereby providing a basis for subsequent carbon storage calculation; meanwhile, the pixel-level remote sensing features are obtained by feature extraction, and the inversion model trained by the real biomass of the sample plot is combined to efficiently and accurately invert the biomass of each pixel, and the real biomass is determined based on the hyperspectral image and the three-dimensional structure parameters of the sample plot, thereby further improving the reliability of the biomass inversion. Finally, the carbon storage is calculated by combining the carbon conversion coefficient of each vegetation, the vegetation type and the biomass, thereby realizing the accurate quantification of the aboveground vegetation carbon storage in the study area. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative effort.

[0010] Figure 1 FIG. 1 is a diagram of the application environment of an aboveground vegetation carbon storage inversion method according to an embodiment of the present application.

[0011] Figure 2A flowchart of a ground vegetation carbon storage inversion method provided by an embodiment of the present application is shown.

[0012] Figure 3 A functional module diagram of a ground vegetation carbon storage inversion device provided by an embodiment of the present application is shown.

[0013] Figure 4 A structural diagram of a computer device provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0014] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0015] The present application performs pixelization allocation of plot biomass through multi-source data fusion, and combines different vegetation types to perform carbon storage conversion, thereby constructing a complete and more refined ground carbon storage estimation method.

[0016] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0017] The ground vegetation carbon storage inversion method provided by the embodiments of the present application can be applied in an application environment as shown in Figure 1 The terminal 101 communicates with the server 102 through a network. The data storage system can store data required to be processed by the server 102. The data storage system can be separately arranged, or integrated on the server 102, or placed on a cloud or other server. The terminal 101 can send the hyperspectral image of the research area to the server 102, and the server 102 inverses the ground vegetation carbon storage of the research area based on the received hyperspectral image of the research area. The server 102 can feed back the ground vegetation carbon storage of the research area to the terminal 101.

[0018] The terminal 101 can be, but is not limited to, various desktop computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things device can be a smart speaker, a smart television, a smart air conditioner, a smart vehicle device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The server 102 can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.

[0019] In an exemplary embodiment, asFigure 2 As shown, a terrestrial vegetation carbon storage inversion method is provided, which is executed by a computer device, specifically, can be executed by a terminal or a server alone or jointly, in the embodiments of the present application, the method is applied to Figure 1 The server 102 in the system 100 is taken as an example for illustration, and includes the following steps 201 to 205.

[0020] In step 201, a hyperspectral image of a study area is acquired.

[0021] In a specific application example, a hyperspectral imaging system carried by a UAV is used to acquire a high spatial resolution hyperspectral image covering the study area, which provides rich spectral information for subsequent feature extraction and classification.

[0022] The hyperspectral image is further subjected to Gaussian filtering and noise band elimination in sequence.

[0023] In step 202, a pre-trained classification model is used to classify each pixel in the hyperspectral image to determine the vegetation type of each pixel. The classification model is a random forest model.

[0024] Specifically, the random forest model is used to perform fine vegetation classification on the hyperspectral image of the study area, and output a distribution map of dominant tree species or vegetation types in the study area (including the vegetation type of each pixel). The distribution map is used for subsequent carbon storage coefficient determination, which can improve the ecological significance and accuracy of carbon storage estimation.

[0025] In step 203, features of the hyperspectral image are extracted to determine the remote sensing features of each pixel.

[0026] In step 204, an inversion model is used to invert the biomass in the study area according to the remote sensing features of each pixel to obtain the biomass of each pixel.

[0027] The inversion model is obtained by training a training sample set, which includes the remote sensing features of each pixel in multiple sample plots and the true biomass of each pixel. The true biomass is determined based on the hyperspectral image and three-dimensional structure parameters of the sample plot.

[0028] In a specific application example, the training process of the inversion model includes the following steps (1) to (6).

[0029] (1) Acquire the hyperspectral image, multi-view image and sample plot survey data of multiple sample plots.

[0030] Specifically, the hyperspectral imaging system carried by the unmanned aerial vehicle collects hyperspectral images of the sample plot, and the oblique photography system of the unmanned aerial vehicle collects multi-view images, thereby providing accurate monitoring data for subsequent three-dimensional reconstruction to obtain a digital surface model (DSM) and a canopy surface area and other three-dimensional structure parameters.

[0031] In addition, the data collected in combination with the ground sample plot investigation, including tree species, diameter at breast height and height. The single-plant biomass is calculated according to the diameter at breast height and height, and the total biomass in the sample plot is calculated according to the single-plant biomass in the sample plot, thereby providing necessary real sample support for subsequent inversion model construction.

[0032] (2) For any sample plot, the hyperspectral images, multi-view images and sample plot investigation data of the sample plot are spatially calibrated.

[0033] First, the hyperspectral images are sequentially subjected to Gaussian filtering and noise band elimination. Then, the hyperspectral images, multi-view images and sample plot investigation data are spatially calibrated, so that different source data can be uniformly projected into the same coordinate system, thereby laying a foundation for subsequent analysis.

[0034] (3) The vegetation index of each pixel in the sample plot is calculated according to the calibrated hyperspectral images, and the canopy surface area of each pixel in the sample plot is calculated according to the calibrated multi-view images.

[0035] Specifically, the digital surface model is generated according to the calibrated multi-view images. The digital surface model includes the elevation grid of each pixel and the size of each pixel, i.e. three-dimensional structure parameters. The slope of each elevation grid in the digital surface model is calculated, and the slope is converted into radian to obtain the radian of each pixel: The canopy surface area of each pixel is calculated according to the radian of each pixel and the size of each pixel: . Wherein, is the radian of the pixel corresponding to the elevation grid in the k th row and the j th column of the digital surface model, is the slope of the elevation grid in the k th row and the j th column of the digital surface model, is the canopy surface area of the pixel corresponding to the elevation grid in the k th row and the j th column of the digital surface model, is the size of the pixel.

[0036] (4) The real biomass of each pixel in the sample plot is calculated according to the vegetation index of each pixel, the canopy surface area of each pixel and the sample plot investigation data in the sample plot.

[0037] The real biomass of the i-th pixel is calculated by the following formula: i i i i

[0038] In view of the problem that the plot scale is inconsistent with the image pixel scale in traditional biomass inversion, the present application realizes the mapping of the measured biomass of the plot to the pixel scale by calculating the vegetation index and the crown surface area of each pixel in the plot, and distributing the total biomass of the plot according to the proportion of the vegetation index and the structural characteristics (crown surface area) in the pixel, to generate the real biomass at the pixel level, and to provide accurate labeled data for subsequent model training.

[0039] (5) Feature extraction is performed on the calibrated hyperspectral image to obtain the remote sensing features of each pixel in the plot.

[0040] Specifically, the spectral features and texture features of the pixels in the plot are extracted from the hyperspectral image. Then, all the features are screened according to the importance of the features to the biomass, and the top 50 features are retained as the remote sensing features for model construction, so as to improve the calculation efficiency and enhance the stability of the model.

[0041] Among them, the feature screening process is: using all the features to train a random forest regression model. The random forest regression model can automatically evaluate the contribution of each feature to the prediction target variable (such as biomass) in the training process through the ensemble learning mechanism of multiple decision trees. Then, the importance of the features is calculated. In the model training process, the random forest quantifies the contribution of each feature to the improvement of the prediction performance based on the reduction of the Gini index or the decrease of the mean square error when the node is split. The higher the index, the stronger the explanatory power of the feature to the target variable. Then, all the features are sorted in descending order of their contribution scores to form a complete feature importance list. This list reflects the influence of the features on the prediction of the biomass relative to the entire model.

[0042] (6) A training sample set is constructed according to the real biomass of each pixel in each plot and the remote sensing features of each pixel, and the inversion model is trained using the training sample set.

[0043] ​​​​​​​​​​In addition, the application can also construct various inversion models based on the screened remote sensing features and the true biomass, including multiple linear regression, random forest, XGBoost, LightGBM, support vector regression, and artificial neural network. The performance of each model is evaluated comprehensively through cross-validation and precision indicators, and the optimal model is selected for aboveground biomass inversion in the study area to output the biomass results. The model finally selected by the application is the random forest model.

[0044] To solve the problems that the sample plot biomass sample is difficult to correspond to the high-precision remote sensing pixel and the carbon conversion coefficients of different vegetation types are different, the application uses the hyperspectral image and three-dimensional structure data (such as crown surface area) obtained by the unmanned aerial vehicle to finely spatially distribute the biomass in the sample plot at the pixel scale, realize the accurate mapping of the biomass sample to the pixel level, and on this basis, combine the multi-source remote sensing features to construct an inversion model, and determine the carbon conversion coefficients of different vegetation types by using the vegetation classification patches, and finally realize the high-precision estimation of the aboveground vegetation carbon storage. Compared with the traditional uniform distribution method based on the sample plot mean, the application can retain the internal structure differences of the sample plot, improve the consistency and representativeness of the training sample, and thus significantly enhance the reliability and precision of the remote sensing inversion results.

[0045] In step 205, the aboveground vegetation carbon storage of the study area is determined according to the carbon conversion coefficient of each vegetation, the vegetation type of each pixel, and the biomass of each pixel.

[0046] In one specific application example, for any pixel, the carbon conversion coefficient of the pixel is determined according to the vegetation type of the pixel. The carbon storage of the pixel is calculated according to the carbon conversion coefficient of the pixel and the biomass of the pixel. The carbon storage is the product of the carbon conversion coefficient and the biomass. The aboveground vegetation carbon storage of the study area is determined according to the carbon storage of each pixel. The aboveground vegetation carbon storage of the study area is the sum of the carbon storage of each pixel.

[0047] Specifically, the biomass inversion results are processed by different vegetation types according to the vegetation type distribution map obtained in step 202. Then, the aboveground vegetation carbon storage of the study area is calculated according to the carbon conversion coefficients of various vegetation types to form the final carbon storage spatial data product, which provides a scientific basis for regional ecological monitoring and carbon sink evaluation.

[0048] The application significantly improves the precision and stability of aboveground biomass inversion by fusing hyperspectral images and three-dimensional structure parameters. The proposed pixelization method of sample plot biomass effectively solves the problem of inconsistency between sample plot scale and pixel scale in traditional methods, making the model training more accurate. In addition, the hierarchical inversion method based on vegetation classification makes the final biomass and carbon storage results more consistent with ecological characteristics, realizing high-precision and high-resolution biomass estimation.

[0049] Based on the same inventive concept, the embodiments of the present application also provide a device for implementing the above-mentioned method. The device provides a solution to the implementation similar to the implementation described in the above-mentioned method, and therefore the specific limitations in one or more device embodiments provided below can refer to the limitations of the method described above, which will not be described here again.

[0050] In one exemplary embodiment, as shown in Figure 3 a terrestrial vegetation carbon storage inversion is provided, which includes the following functional modules.

[0051] The image acquisition module 301 is configured to acquire hyperspectral images of the study area.

[0052] The pixel classification module 302 is configured to classify each pixel in the hyperspectral images by using a pre-trained classification model to determine the vegetation type of each pixel.

[0053] The feature extraction module 303 is configured to extract features from the hyperspectral images to determine the remote sensing features of each pixel.

[0054] The biomass inversion module 304 is configured to invert the biomass in the study area according to the remote sensing features of each pixel by using an inversion model to obtain the biomass of each pixel. The inversion model is obtained by training a training sample set in advance, and the training sample set includes the remote sensing features of each pixel and the real biomass of each pixel in multiple sample plots. The real biomass is determined based on the hyperspectral images and three-dimensional structure parameters of the sample plots.

[0055] The carbon storage determination module 305 is configured to determine the aboveground vegetation carbon storage of the study area according to the carbon conversion coefficient of each vegetation type, the vegetation type of each pixel, and the biomass of each pixel.

[0056] In an exemplary embodiment, a computer device is provided, which can be a server or a terminal, and its internal structure diagram can be as shown in Figure 4As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the hyperspectral image of the study area. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to realize a method for retrieving aboveground vegetation carbon storage.

[0057] Those skilled in the art can understand that, Figure 4 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0058] In an exemplary embodiment, a computer readable storage medium is provided, which stores a computer program. The computer program is executed by a processor to implement the steps in each of the method embodiments described above.

[0059] In an exemplary embodiment, a computer program product is provided, which includes a computer program. The computer program is executed by a processor to implement the steps in each of the method embodiments described above.

[0060] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0061] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. The non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. The volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, the RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0062] The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, etc., without being limited thereto.

[0063] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application.

[0064] The principles and implementation modes of the present application are described by using specific examples in the present application. The above-mentioned embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In conclusion, the content of the present application should not be understood as a limitation.

Claims

1. A method for inverting aboveground vegetation carbon storage, characterized in that, The method includes: Acquire hyperspectral images of the study area; A pre-trained classification model is used to classify each pixel in the hyperspectral image to determine the vegetation type of each pixel; Feature extraction is performed on the hyperspectral image to determine the remote sensing features of each pixel; Based on the remote sensing characteristics of each pixel, an inversion model is used to invert the biomass in the study area to obtain the biomass of each pixel; wherein, the inversion model is trained in advance using a training sample set, which includes the remote sensing characteristics of each pixel in multiple sample plots and the actual biomass of each pixel; the actual biomass is determined based on the hyperspectral image and three-dimensional structural parameters of the sample plot. The aboveground vegetation carbon storage in the study area was determined based on the carbon conversion coefficient of each vegetation type, the vegetation type of each pixel, and the biomass of each pixel.

2. The method for retrieving aboveground vegetation carbon storage according to claim 1, characterized in that, Before classifying each pixel in the hyperspectral image using a pre-trained classification model, the method further includes: The hyperspectral image is then subjected to Gaussian filtering and noise band removal processes in sequence.

3. The method for retrieving aboveground vegetation carbon storage according to claim 1, characterized in that, The training process of the inversion model includes: Acquire hyperspectral images, multi-view images, and sample plot survey data from multiple sample plots; For any given plot, the hyperspectral imagery, multi-view imagery, and plot survey data of the plot are spatially calibrated. The vegetation index of each pixel in the plot is calculated based on the calibrated hyperspectral image, and the canopy surface area of ​​each pixel in the plot is calculated based on the calibrated multi-view image. The actual biomass of each pixel in the sample plot is calculated based on the vegetation index, canopy surface area of ​​each pixel, and sample plot survey data. Feature extraction is performed on the calibrated hyperspectral image to obtain the remote sensing features of each pixel within the sample plot; A training sample set is constructed based on the actual biomass of each pixel in each sample plot and the remote sensing features of each pixel, and the inversion model is trained using the training sample set.

4. The method for retrieving aboveground vegetation carbon storage according to claim 3, characterized in that, The canopy surface area of ​​each pixel within the sample plot is calculated based on the calibrated multi-view imagery, including: A digital surface model is generated based on the calibrated multi-view images; the digital surface model includes the elevation grid of each pixel and the size of each pixel. Calculate the slope of each elevation grid in the digital surface model, and convert the slope into radians to obtain the radians of each cell; The canopy surface area of ​​each pixel is calculated based on the curvature and size of each pixel.

5. The method for retrieving aboveground vegetation carbon storage according to claim 3, characterized in that, The data from the sample plot survey refers to the total biomass within the sample plot; The following formula is used to calculate the first... i The actual biomass of a single pixel: ； in, For the first i The true biomass of a single pixel For the first i The canopy surface area of ​​each pixel For the first i The vegetation index of each pixel This represents the total biomass within the sample plot.

6. The method for retrieving aboveground vegetation carbon storage according to claim 1, characterized in that, Both the classification model and the inversion model are random forest models.

7. The method for retrieving aboveground vegetation carbon storage according to claim 1, characterized in that, Based on the carbon conversion coefficient of each vegetation type, the vegetation type of each pixel, and the biomass of each pixel, the aboveground vegetation carbon storage in the study area was determined, including: For any given pixel, the carbon conversion coefficient of that pixel is determined based on the vegetation type of that pixel; The carbon storage of a pixel is calculated based on the pixel's carbon conversion coefficient and the pixel's biomass. The carbon storage of aboveground vegetation in the study area was determined based on the carbon storage of each pixel.

8. A device for retrieving carbon storage in aboveground vegetation, characterized in that, The apparatus performs the aboveground vegetation carbon storage inversion method according to any one of claims 1-7, and the apparatus comprises: The image acquisition module is used to acquire hyperspectral images of the study area. The pixel classification module is used to classify each pixel in the hyperspectral image using a pre-trained classification model and determine the vegetation type of each pixel. The feature extraction module is used to extract features from the hyperspectral image and determine the remote sensing features of each pixel; The biomass inversion module is used to invert the biomass of the study area based on the remote sensing characteristics of each pixel using an inversion model, thereby obtaining the biomass of each pixel. The inversion model is pre-trained using a training sample set, which includes the remote sensing characteristics of each pixel in multiple sample plots and the actual biomass of each pixel. The actual biomass is determined based on the hyperspectral image and three-dimensional structural parameters of the sample plot. The carbon storage determination module is used to determine the aboveground vegetation carbon storage in the study area based on the carbon conversion coefficient of each vegetation, the vegetation type of each pixel, and the biomass of each pixel.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the aboveground vegetation carbon storage inversion method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the aboveground vegetation carbon storage inversion method as described in any one of claims 1-7.

Images