Forestry carbon sink calculation model, forestry carbon credit calculation method and storage medium

Abstract

The invention relates to the technical field of carbon sink calculation, in particular to a forestry carbon sink calculation model, a forestry carbon credit calculation method and a storage medium. The model comprises an overground biological carbon sink amount estimation module used for calculating overground biological carbon sink data of a target area, an underground soil carbon sink amount estimation module used for calculating soil carbon sink data of the target area, and a litter carbon reserve estimation module used for calculating litter carbon sink data of the target area. The combined loss calculation module is used for calculating carbon loss data; and the final carbon sink total quantity fusion module is used for calculating the carbon sink quantity according to the aboveground biological carbon sink data, the soil carbon sink data, the litter carbon sink data and the carbon loss data. A forestry carbon sink calculation model based on multiple dimensions is established through multiple types of carbon sink estimation modules, and the problem that the requirement for rapid authentication of the global carbon market cannot be met due to high manual dependency is solved; the problems of low calculation precision, high cost and the like caused by insufficient model dynamics, single model calculation dimension and the like are solved.

Description

Technical Field

[0001] The present invention relates to the field of carbon sink calculation technology, in particular to a forestry carbon sink calculation model, a forestry carbon credit calculation method, and a storage medium. Background Art

[0002] As global climate change becomes increasingly severe, carbon emissions and carbon sequestration have become a focus of attention. Carbon sinks refer to the total amount of carbon dioxide removed from the atmosphere by ecosystems through biological uptake, chemical adsorption, and physical retention. Accurately calculating carbon sinks is crucial for assessing ecosystem services and developing carbon reduction policy tools.

[0003] Current forestry carbon credit calculations primarily rely on manual field surveys combined with static models, such as those recommended by the IPCC, or estimates based on single remote sensing data. These methods, due to their need for field surveys, use of static models, single considerations for carbon change, and inability to account for dynamic forest dynamics, are time-consuming, inefficient, costly, difficult to cover large forest areas, and suffer from low prediction accuracy. Summary of the Invention

[0004] The purpose of this application is to provide a forestry carbon sink calculation model, a forestry carbon credit calculation method, and a storage medium. By establishing a multi-dimensional forestry carbon sink calculation model through various types of carbon sink estimation modules, it solves the high dependence on manual labor and the inability to meet the rapid certification needs of the global carbon market; it solves the problems of low calculation accuracy and high cost caused by insufficient model dynamics and a single model calculation dimension.

[0005] This application provides a forestry carbon sequestration calculation model, including:

[0006] A module for estimating aboveground biomass carbon sinks used to calculate aboveground biomass carbon sink data in target areas;

[0007] A module for estimating underground soil carbon sequestration for calculating soil carbon sequestration data in target areas;

[0008] Litter carbon storage estimation module for calculating litter carbon sink data for the target area;

[0009] Joint loss calculation module for calculating carbon loss data;

[0010] The final carbon sink fusion module calculates the carbon sink amount based on aboveground biological carbon sink data, soil carbon sink data, litter carbon sink data and carbon loss data.

[0011] Furthermore, it also includes an uncertainty quantification module for providing confidence intervals for aboveground biological carbon sink data, confidence intervals for soil carbon sink data, and confidence intervals for litter carbon sink data.

[0012] Furthermore, it also includes a dynamic carbon sink correction module for adjusting the carbon sink amount according to changes in forest ecology; events that cause changes in forest ecology include fires, pests and diseases.

[0013] Furthermore, the training samples of the aboveground biological carbon sink estimation module include remote sensing images, phenological data and sampling data of aboveground biological sampling points;

[0014] The training samples of the underground soil carbon sequestration estimation module include soil information, climate factors, DEM, and remote sensing images;

[0015] The training samples of the litter carbon storage estimation module include forest type classification information, climate zone information and sampling data of litter sampling points.

[0016] Furthermore, it also includes a feature extraction module for extracting remote sensing time series feature information and environmental variable feature information, and fusing and splicing the extracted feature information to obtain a feature tensor.

[0017] On the other hand, the present application also provides a forestry carbon credit calculation method based on a forestry carbon sink calculation model, which uses any of the above-mentioned forestry carbon sink calculation models to calculate the carbon sink amount of the target area.

[0018] Further, the following steps are included:

[0019] Use the multimodal data acquisition module to collect basic data of the target area;

[0020] Performing spatiotemporal alignment and / or noise filtering processing on the basic data using a data fusion and quality control module;

[0021] The basic data that has been processed with spatiotemporal alignment and / or noise filtering is input into the forestry carbon sink calculation model to calculate the carbon sink amount.

[0022] Furthermore, the collecting of basic data of the target area using the multimodal data collection module includes:

[0023] Use high-resolution satellite remote sensing units to obtain information on canopy height and biomass density;

[0024] Use the multispectral scanning unit of the UAV to obtain the distribution information of understory vegetation and litter;

[0025] Use ground IoT units to obtain temperature, humidity, and CO2 flux information.

[0026] Furthermore, the collecting of basic data of the target area using the multimodal data collection module includes:

[0027] Access the ecological database interface to obtain tree species growth model information and historical fire data.

[0028] Furthermore, the data fusion and quality control module is used to perform spatiotemporal alignment and / or noise filtering on the basic data, including:

[0029] Use the spatiotemporal alignment engine to unify the temporal and spatial resolutions of various basic data;

[0030] Establish a mapping relationship between satellite imagery and ground sensor data;

[0031] Estimate data values for uncovered areas;

[0032] Use generative adversarial networks to repair abnormal information and / or missing data.

[0033] The abnormal data cleaning module training set includes at least one of good-quality remote sensing images, phenological data, normal sensor data, and climate factors in the target area.

[0034] Furthermore, the step of inputting the basic data processed by spatiotemporal alignment and / or noise filtering into the forestry carbon sink calculation model to calculate the carbon sink amount includes:

[0035] Use the feature extraction module to extract the time series feature information and environmental variable feature information of the processed basic data, and fuse and splice the extracted feature information to obtain the feature tensor;

[0036] Use the aboveground biomass carbon sink estimation module to calculate the aboveground biomass carbon sink data of the feature tensor;

[0037] Use the underground soil carbon sequestration estimation module to calculate the soil carbon sequestration data of the characteristic tensor;

[0038] Use the litter carbon storage estimation module to calculate litter carbon sink data of the characteristic tensor;

[0039] Use the joint loss calculation module to calculate carbon loss data for aboveground biomass carbon sink data, soil carbon sink data, and litter carbon sink data;

[0040] The final carbon sink fusion module is used to calculate the carbon sink based on aboveground biological carbon sink data, soil carbon sink data, litter carbon sink data and carbon loss data.

[0041] Furthermore, when forest ecological changes occur, it also includes: adjusting the carbon sink amount using a dynamic carbon sink correction module; events that cause forest ecological changes include fires, diseases and insect pests.

[0042] Furthermore, it also includes using historical fire data to predict future carbon emission risks.

[0043] In another aspect of the present application, a computer-readable storage medium is provided, which stores a computer program, and when the program is executed by a processor, the method as described in any one of the above items is implemented.

[0044] The forestry carbon sink calculation model, forestry carbon credit calculation method, and storage medium provided in this application adopt the design of an above-ground biomass carbon sink estimation module for calculating above-ground biomass carbon sink data in the target area; an underground soil carbon sink estimation module for calculating soil carbon sink data in the target area; a litter carbon storage estimation module for calculating litter carbon sink data in the target area; a joint loss calculation module for calculating carbon loss data; and a final carbon sink total fusion module for calculating carbon sinks based on above-ground biomass carbon sink data, soil carbon sink data, litter carbon sink data, and carbon loss data. In the forestry carbon sink calculation model, multi-dimensional modules corresponding to forest carbon production and carbon loss are adopted. Each module can estimate the carbon sink data of the corresponding dimension respectively. Combined with the joint loss calculation module and the final carbon sink total fusion module, a comprehensive and accurate calculation of forest carbon sinks is achieved.

[0045] This application also uses an uncertainty quantification module to provide the design of confidence intervals for aboveground biological carbon sink data, confidence intervals for soil carbon sink data, and confidence intervals for litter carbon sink data. The large model can be used for various types of forestry carbon sink calculations, and the large model can be simply adjusted according to the characteristics of forests and target areas in different target areas. At the same time, it will also output confidence intervals for carbon sink data in each dimension, ensuring the generalization ability and adaptability of the forestry carbon sink calculation model while improving the credibility of the calculation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0047] Figure 1 This is a flowchart of an embodiment of a forestry carbon credit calculation method based on a forestry carbon sink calculation model provided in this application. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0049] Carbon sinks refer to changes in carbon produced by activities, processes, or mechanisms that absorb atmospheric carbon dioxide and store it in organisms and soils through natural processes such as plant photosynthesis. For example, forest carbon sinks can be represented by the sum of changes in forest carbon storage over a given period. Carbon credits are greenhouse gas emission reduction or sequestration enhancement projects developed in accordance with the rules, procedures, and methodologies established by relevant standards or mechanisms, and are issued after verification, verification, and certification by an independent third-party organization. Each carbon credit represents the reduction or removal of one ton of carbon dioxide equivalent from the air.

[0050] Existing forestry carbon sink calculation methods have the following problems: 1. Due to the long artificial plot survey cycle (several months) and high cost, it is difficult to cover large areas of forests and is inefficient; 2. Because the static model does not take into account the dynamic changes of forests (such as natural disturbances and differences in growth stages), the error rate of carbon storage estimation is as high as 30% and the accuracy is insufficient; 3. The data dimension is single due to the use of only remote sensing spectral data, ignoring key carbon sink components such as soil carbon pools, understory vegetation, and litter; 4. Traditional algorithms have poor adaptability to forest types (such as tropical rainforests and boreal coniferous forests) and require repeated parameter adjustments, resulting in poor model generalization; 5. They are even unable to respond in real time to the impact of emergencies such as forest fires, pests and diseases on carbon sinks and lack timeliness. To solve the above problems, this application adopts the establishment of a multi-dimensional forestry carbon sink calculation model, which covers organisms such as aboveground trees, aboveground dead branches and leaves, underground soil, etc., so that the model covers all aspects of forestry carbon sources and carbon consumption. Moreover, the model-based forestry carbon credit calculation method provided in this application collects multi-modal basic data, and the basic data is aligned, filtered, integrated, and then analyzed by the model, ensuring the comprehensiveness and effectiveness of the input data, and thus achieving the accuracy and comprehensiveness of the output results.

[0051] An embodiment of the present application provides a forestry carbon sink calculation model, including:

[0052] A module for estimating aboveground biomass carbon sinks used to calculate aboveground biomass carbon sink data in target areas;

[0053] A module for estimating underground soil carbon sequestration for calculating soil carbon sequestration data in target areas;

[0054] Litter carbon storage estimation module for calculating litter carbon sink data for the target area;

[0055] Joint loss calculation module for calculating carbon loss data;

[0056] The final carbon sink fusion module calculates the carbon sink amount based on aboveground biological carbon sink data, soil carbon sink data, litter carbon sink data and carbon loss data.

[0057] The forestry carbon sink calculation model provided in this embodiment also includes an uncertainty quantification module for providing confidence intervals for aboveground biomass carbon sink data, soil carbon sink data, and litter carbon sink data. The confidence intervals for each data set enhance the credibility of the calculation results.

[0058] Each module in the forestry carbon sink calculation model provided in this embodiment is trained using corresponding data:

[0059] The training samples of the aboveground biological carbon sink estimation module include remote sensing images, phenological data and sampling data of aboveground biological sampling points;

[0060] The training samples of the underground soil carbon sequestration estimation module include soil information, climate factors, DEM, and remote sensing images;

[0061] The training samples of the litter carbon storage estimation module include forest type classification information, climate zone information and sampling data of litter sampling points.

[0062] In this embodiment, the aboveground biomass carbon sink estimation module is a regression model based on CNN+LSTM; the subsurface soil carbon sink estimation module is a regression model based on MLP or XGBoost; and the litter carbon storage estimation module is a random forest + ML (machine learning) model. The uncertainty quantification module extracts model parameters for the aboveground biomass carbon sink estimation module and calculates confidence intervals for aboveground biomass carbon sink data using a BNN (Bayesian Neural Network) model and Monte Carlo Dropout; extracts model parameters for the subsurface soil carbon sink estimation module and calculates confidence intervals for soil carbon sink data using a BNN (Bayesian Neural Network) model and Monte Carlo Dropout; and extracts model parameters for the litter carbon storage estimation module and calculates confidence intervals for litter carbon sink data using a BNN (Bayesian Neural Network) model and Monte Carlo Dropout.

[0063] The forestry carbon sink calculation model provided in this embodiment integrates prior knowledge of tree species physiology and ecology, adopts deep learning methods to estimate carbon reserves, and simultaneously calculates aboveground biomass, underground soil carbon, and litter carbon reserves to ensure the comprehensiveness of the calculation.

[0064] This embodiment also provides a confidence interval calculation method, which uses the Bootstrap+Ensemble method to calculate confidence intervals for aboveground biomass carbon sink data, soil carbon sink data, and / or litter carbon sink data. The uncertainty represented by the standard deviation and confidence interval is obtained.

[0065] Of course, heterogeneous noise modeling (Heteroscedastic Regression) can also be used, and the model directly outputs the error variance value. This heterogeneous noise model is more suitable when the data quality between plots is inconsistent.

[0066] The above training data of each module in this application are fused separately and then used to train the corresponding model. For example, fusion is performed at the input feature level. This can include using a unified encoder to share parameters: for example, remote sensing image input uses a common CNN or Transformer encoder to extract common features; then feature splicing and fusion: remote sensing image + meteorological data + soil data, multi-source splicing to form an input tensor. For another example, fusion is performed at the output prediction level. This can include training the joint loss calculation module, and the objective function is:

[0067]

[0068] Where: L AGB : aboveground biomass regression loss; L SOC : soil organic carbon loss; L Litter : litter carbon loss; λi: adjustable weight parameter. The adjustable weight parameter can be dynamically adjusted based on the aforementioned uncertainties, or based on experience or other methods. Those skilled in the art can select the uncertainty calculation method as needed, and this will not be elaborated here.

[0069] For example, the final carbon pool total fusion module is integrated: various carbon storage estimation results can be integrated into the total carbon storage in the following ways:

[0070] Total C=C AGB +C SOC +C Litter

[0071] Among them, C AGB : Aboveground biological carbon sink data; C SOC : Soil organic carbon sink data; C Litter : Litter carbon sink data.

[0072] The multi-task model employed in this application utilizes multiple branch modules to simultaneously model multiple carbon components, improving efficiency and consistency. Feature extraction is fused using a feature layer sharing method combined with a joint loss at the output layer, leveraging information from multiple sources to optimize overall performance. Uncertainty quantification utilizes a Bayesian NN / MC Dropout / Bootstrap approach to output credible intervals, enhancing model interpretability and practical application value.

[0073] The forestry carbon sink calculation model provided in this embodiment further includes a dynamic carbon sink correction module for adjusting the carbon sink amount according to changes in forest ecology; events that cause changes in forest ecology include fires, pests and diseases.

[0074] This application can establish independent fire carbon loss models and pest carbon loss assessment models to adapt to large-scale or serious emergencies. The input of the fire carbon loss model can include burned area, fire intensity, and aboveground biomass density, and the output can be fire carbon loss including direct gaseous carbon emissions and litter burn loss. Of course, these data sources can be MODIS fire point data, Sentinel-2 Burn Index (BAI), and ground records. The input of the pest carbon loss model can include pest area, biomass change, and vegetation index degradation rate, and the output can include carbon storage loss corresponding to biomass reduction. Of course, these data sources can be remote sensing identification of pest plates and / or forestry department pest reports.

[0075] The dynamic carbon sink module in this application can accurately capture the "additional" impact of emergencies on carbon sinks. The calculation is clear, easy to monitor, and meets the requirements of IPCC Tier 2 / 3.

[0076] To adapt to changing characteristics of fire and insect pests, this embodiment also employs online learning methods. For example, post-event observation data is used to rapidly adapt the model, updating parameters or structure. Small-scale models are retrained using the latest remote sensing / phenological observation data. Dynamic incremental training is performed on existing models. This embodiment optimizes the model in real time using the latest data to ensure continued accuracy and timeliness.

[0077] The forestry carbon sink calculation model described in this embodiment further includes a feature extraction module for extracting remote sensing time series feature information and environmental variable feature information, and fusing and splicing the extracted feature information to obtain a feature tensor.

[0078] like Figure 1 As shown, another embodiment of the present application provides a forestry carbon credit calculation method based on a forestry carbon sink calculation model, and the above-mentioned forestry carbon sink calculation model is applied to calculate the carbon sink amount of the target area.

[0079] The forestry carbon credit calculation method based on the forestry carbon sink calculation model provided in this embodiment specifically includes the following steps:

[0080] S1: Use the multimodal data acquisition module to collect basic data of the target area;

[0081] S2: Using a data fusion and quality control module to perform spatiotemporal alignment and / or noise filtering on the basic data;

[0082] S3: inputting the basic data processed by spatiotemporal alignment and / or noise filtering into the forestry carbon sink calculation model to calculate the carbon sink amount;

[0083] S4: The output and evidence storage module outputs the carbon sink amount to the visualization module and the blockchain evidence storage module.

[0084] The forestry carbon credit calculation method based on the forestry carbon sink calculation model provided in this embodiment uses a multimodal data acquisition module to collect basic data of the target area, including:

[0085] Use high-resolution satellite remote sensing units to obtain information on canopy height and biomass density;

[0086] Use the multispectral scanning unit of the UAV to obtain the distribution information of understory vegetation and litter;

[0087] Use ground IoT units to obtain temperature, humidity, and CO2 flux information;

[0088] And access the ecological database interface to obtain tree species growth model information and historical fire data.

[0089] In this embodiment, the high-resolution satellite remote sensing unit can use satellites such as Sentinel-1, Sentinel-2, and Landsat 8 to obtain canopy height, biomass density, and vegetation index. Canopy height, biomass density, and vegetation index are important ecological parameters that reflect the carbon sequestration capacity of plants and are closely related to carbon sequestration: higher canopy height indicates higher forest maturity, greater accumulated biomass, and greater long-term carbon sequestration potential; biomass density directly corresponds to the amount of carbon sequestered per unit area; greater biomass means more carbon sequestration; and higher vegetation indices (such as NDVI) indicate vigorous vegetation growth and strong photosynthesis, which means more CO2 is absorbed per unit time.

[0090] In this example, the NDVI (Near Infrared Reflectance Index) can be calculated using the formula NDVI = (Near Infrared Reflectance - Red Reflectance) / (Near Infrared Reflectance + Red Reflectance). Values range from 0.6-0.9 to indicate dense forest (strong carbon absorption); 0.2-0.5 typically indicates shrubland or grassland; and ≤0 typically indicates areas without vegetation, such as water bodies, sandy land, and urban areas. NDVI can be used to monitor long-term trends in carbon sequestration capacity.

[0091] In this embodiment, the multispectral scanning unit of the drone uses a multispectral or hyperspectral camera carried by the drone to obtain high-precision understory vegetation information, litter distribution and crown health status. The multispectral spectrum here includes multiple bands, each of which is used to obtain different information: blue light is used to obtain information related to water and soil characteristics, green light is used to obtain information related to vegetation health and chlorophyll concentration, red light (620-750nm) is used to obtain vegetation type and NDVI related information, red edge (700-740nm) is used to monitor plant stress and growth status information, and near infrared is used to obtain vegetation activity and NDVI related information. Thermal infrared and other methods can also be used to obtain soil moisture and temperature change information.

[0092] In this embodiment, each spectrum can be captured using a multi-lens multispectral camera, with each band using a separate lens to capture images, ensuring high spectral resolution. Naturally, the multiple lenses require time synchronization of the images captured, with the spatial alignment between the lenses controlled by a registration system. Spatial registration of images from multiple lenses is well known in the art and will not be discussed further here. Alternatively, an integrated multispectral camera can be used, with a single lens capturing images by switching between wavelengths using a beam splitter prism or filter wheel. While this type of camera may have slightly lower wavelength capture accuracy, it offers advantages such as cost savings and ease of operation. Those skilled in the art can select the camera type as needed, and this will not be discussed further here.

[0093] In order to reduce manual intervention in this embodiment, a route can be set for the drone, and each time the drone patrols, it patrols according to a highly overlapping route.

[0094] In this embodiment, the images acquired by the drone's multispectral unit can be directly input into the data fusion and quality control module without any processing. Of course, some simple processing can also be performed to eliminate related effects. For example, radiometric correction can be performed to eliminate brightness differences, geometric correction can be performed to align the various bands, or atmospheric correction can be performed to eliminate atmospheric effects.

[0095] The above basic data obtained in this embodiment can be transmitted to the data fusion and quality control module using 5G or satellite communication technology to ensure real-time transmission of data.

[0096] The forestry carbon credit calculation method based on the forestry carbon sink calculation model provided in this embodiment includes performing spatiotemporal alignment and / or noise filtering on the basic data using the data fusion and quality control module, including:

[0097] Use the spatiotemporal alignment engine to unify the temporal and spatial resolutions of various basic data;

[0098] Establish a mapping relationship between satellite imagery and ground sensor data;

[0099] Estimate data values for uncovered areas;

[0100] Use generative adversarial networks to repair abnormal information and / or missing data.

[0101] The training set of the abnormal data cleaning module includes at least one of good-quality remote sensing images, phenological data, normal sensor data, and climate factors in the target area.

[0102] In this embodiment, the spatiotemporal alignment engine is used to align the drone data with the remote sensing imagery, downsample the drone camera data to the same resolution as the remote sensing imagery, and unify the temporal and spatial resolutions of different data sources, making them comparable and fusible at the same scale. Establishing a mapping relationship between satellite imagery and ground sensor data involves modeling and matching the remote sensing imagery with the real-time ground sensor data in terms of space and attributes. Possible mapping models include linear / nonlinear regression models, or machine learning models such as decision trees (CART, XGBoost), random forests, support vector machines (SVM), and neural networks (MLP, CNN). Those skilled in the art can select these models as needed, and this will not be elaborated on here.

[0103] In this embodiment, the data values of the estimated uncovered areas can be used to predict the variable values of unmeasured areas using a model. The method that can be adopted is to use CNN or U-Net to directly extract spatial patterns from multispectral images and perform end-to-end modeling. The data used for model training is the data of the measured area. Of course, a supervised learning method based on labeled sample training can also be adopted to perform spatial interpolation prediction through a machine learning model (such as RF, SVM, MLP). Of course, it is also possible to use the values of known areas for spatial interpolation (such as Kriging, IDW), and combine remote sensing image guidance weighting.

[0104] This application first uses a spatiotemporal alignment engine to unify data of varying resolutions and temporal dimensions. Machine learning and modeling algorithms then establish a mapping relationship between remote sensing data and storefront data. Based on this comprehensive spatial and temporal mapping, the application estimates uninhabited areas, improving the accuracy of data on uninhabited areas. This estimation of uninhabited areas ensures the integrity of the entire forest data, which helps improve the accuracy of carbon sink calculations.

[0105] In this embodiment, the adversarial generative network is used to repair abnormal information and / or missing data, specifically to extract high-quality data collected in the target area to establish a subset to train the adversarial generative network. The generative adversarial network includes a generator and a discriminator. The training goal is to make the data generated by the generator "fake" and "fool" the discriminator. The training data can give priority to historical data with good quality in the target area, such as clear image fragments of the same plot, drone images at different times of the same day, and normal sensor data. Complete data from adjacent times or similar areas can also be used as supplementary training sets, such as data from nearby plots or plots in the same climate zone. GANs trained directly with high-quality historical data or with supplementary training sets offer several advantages: 1) strong spatial consistency, with ecological characteristics of the same or adjacent plots more similar; 2) strong temporal continuity, with the ability to capture growth trends and temporal patterns using time series information; 3) improved restoration accuracy, reducing noise and improving completion quality compared to external data; and 4) facilitating model convergence. Remote sensing data has high dimensions (spectral + spatial + temporal), and local consistency is conducive to GAN network training. Of course, if none of the above data are available, a general model pre-trained with remote sensing and ecological data can also be used. Those skilled in the art can make their own choices based on their needs, and this will not be elaborated on here.

[0106] In this embodiment, methods for detecting missing or abnormal monitoring data include detecting cloud obstruction data and determining sensor failure. Cloud shadows / obstruction are determined using methods such as Fmask, QA bands, and abnormal NDVI jumps. Sensor failure is determined when sensor values suddenly change, fluctuate abnormally, remain constant at zero, or drift.

[0107] In this embodiment, data restoration for abnormal or missing data involves: When a slightly missing or occluded area is detected, a GAN model and adjacent spatial / temporal data are used to "generate" images or sensor values to complete the details. When a severely missing or large area is occluded, spatial interpolation (such as Kriging or IDW) is first used for rough infill, followed by GAN optimization of texture details. When a sensor point is abnormal, LSTM, Prophet, or moving average is used to restore its trend.

[0108] In this embodiment, a "repair label" is retained after repair so that subsequent processing can identify the filled areas and improve the credibility of the carbon sink calculation results. In this embodiment, to prevent excessive data repair from affecting subsequent carbon sink calculations, when there is severe data loss or a large number of consecutive sensor data anomalies, the missing or abnormal data is marked as empty with a mask.

[0109] The forestry carbon credit calculation method based on the forestry carbon sink calculation model provided in this embodiment, wherein the basic data that has undergone spatiotemporal alignment and / or noise filtering is input into the forestry carbon sink calculation model to calculate the carbon sink amount, includes:

[0110] Use the feature extraction module to extract the time series feature information and environmental variable feature information of the processed basic data, and fuse and splice the extracted feature information to obtain the feature tensor;

[0111] Use the aboveground biomass carbon sink estimation module to calculate the aboveground biomass carbon sink data of the feature tensor;

[0112] Use the underground soil carbon sequestration estimation module to calculate the soil carbon sequestration data of the characteristic tensor;

[0113] Use the litter carbon storage estimation module to calculate litter carbon sink data of the characteristic tensor;

[0114] Use the joint loss calculation module to calculate carbon loss data for aboveground biomass carbon sink data, soil carbon sink data, and litter carbon sink data;

[0115] The final carbon sink fusion module is used to calculate the carbon sink based on aboveground biological carbon sink data, soil carbon sink data, litter carbon sink data and carbon loss data.

[0116] The forestry carbon credit calculation method based on the forestry carbon sink calculation model provided in this embodiment further includes: adjusting the carbon sink amount using a dynamic carbon sink correction module when forest ecological changes occur; events that cause forest ecological changes include fires, pests and diseases.

[0117] The dynamic carbon sink correction module includes a fire carbon loss model and / or an insect pest carbon impact model. The dynamic carbon correction module can adjust some parameters of the aboveground biomass carbon sink estimation module and some parameters of the litter carbon storage estimation module based on the identified event impact area. Furthermore, the model parameter adjustments can be combined with a disturbance coefficient.

[0118] The forestry carbon credit calculation method based on the forestry carbon sink calculation model provided in this embodiment also includes using historical fire data to predict future carbon emission risks.

[0119] Forest fires burn vegetation and soil organic matter, releasing large amounts of greenhouse gases such as carbon dioxide (CO2), methane (CH4), and carbon monoxide (CO). Historical fire data itself records past carbon emissions. Therefore, historical fire data can be used to estimate "past carbon emissions," that is, the contribution of carbon emissions caused by natural disturbance events. In terms of predicting future carbon emission risks, when establishing the above-mentioned model, disturbance factors based on fire frequency, fire intensity, seasonal factors, etc. can be established, and then the disturbance factors can be trained with historical fire data to predict future fire probabilities and scenario simulations. In terms of predicting future carbon emission risks, historical fire data + climate, topography, vegetation and other variables can also be used to train models such as MaxEnt and logistic regression to predict which areas are more likely to have fires in the future. If a fire occurs, how much will the loss be?

[0120] This embodiment provides a forestry carbon credit calculation method based on a forestry carbon sink calculation model. The output and evidence storage module outputs carbon sink amounts to a visualization module and a blockchain evidence storage module, enabling visual display and trusted evidence storage of carbon credit calculation results. The blockchain evidence storage module generates a unique hash value corresponding to the carbon sink amount in the current target area; this hash value is bound to the carbon sink data, ensuring that the data cannot be tampered with. The visualization module provides a three-dimensional visualization platform that displays the spatial gradient and temporal evolution of forest carbon sinks.

[0121] A computer-readable storage medium stores a computer program, which implements the method described above when executed by a processor.

[0122] The forestry carbon sink calculation model, forestry carbon credit calculation method, and storage medium provided in this embodiment achieve full automation of the entire process from data acquisition to carbon credit certification through multimodal data collection, data fusion and quality control, large-scale forestry model calculation, dynamic correction, and result storage. This achieves:

[0123] 1. Full process automation: No manual intervention is required from data collection to authentication, improving efficiency by over 85%.

[0124] 2. High-precision calculation: Multimodal data fusion reduces the carbon sink estimation error rate to less than 8% (30% for traditional methods).

[0125] 3. Dynamic response capability: Carbon loss caused by fire events can be updated to calculation results within 24 hours.

[0126] 4. Multi-dimensional coverage: Simultaneously calculate aboveground biomass, underground soil carbon, and litter carbon pools to cover the complete carbon sink components.

[0127] 5. Trusted Evidence: Blockchain technology ensures that data cannot be tampered with and supports international carbon market certification standards.

[0128] As used below, the term “unit,” “sub-unit,” or “module” may be a combination of software and / or hardware that implements a predetermined function.

[0129] Based on the same inventive concept, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method as described in any one of embodiments one to three.

[0130] Through the description of the above implementation methods, it can be known that those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary hardware. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, or can be embodied through the implementation process of data migration. All or part of the present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, mobile communication terminals, multi-processor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc.

[0131] Although the present application has been described with reference to the embodiments, those skilled in the art will appreciate that there are many modifications and variations to the present application without departing from the spirit of the present application. It is intended that the appended claims include these modifications and variations without departing from the spirit of the present application.

Claims

1. A forestry carbon sink calculation model, characterized in that ,include: A module for estimating aboveground biomass carbon sinks used to calculate aboveground biomass carbon sink data in target areas; A module for estimating underground soil carbon sequestration for calculating soil carbon sequestration data in target areas; Litter carbon storage estimation module for calculating litter carbon sink data for the target area; Joint loss calculation module for calculating carbon loss data; The final carbon sink fusion module calculates the carbon sink amount based on aboveground biological carbon sink data, soil carbon sink data, litter carbon sink data and carbon loss data.

2. The forestry carbon sink calculation model according to claim 1, characterized in that: It also includes an uncertainty quantification module for providing confidence intervals for aboveground biomass carbon sequestration data, soil carbon sequestration data, and litter carbon sequestration data.

3. The forestry carbon sink calculation model according to claim 1, characterized in that: It also includes a dynamic carbon sink correction module, which is used to adjust the carbon sink amount according to changes in forest ecology; events that cause forest ecological changes include fires, pests and diseases.

4. The forestry carbon sink calculation model according to claim 1, characterized in that: The training samples of the aboveground biological carbon sink estimation module include remote sensing images, phenological data and sampling data of aboveground biological sampling points; The training samples of the underground soil carbon sequestration estimation module include soil information, climate factors, DEM, and remote sensing images; The training samples of the litter carbon storage estimation module include forest type classification information, climate zone information and sampling data of litter sampling points.

5. A forestry carbon credit calculation method based on a forestry carbon sink calculation model, characterized in that: The forestry carbon sink calculation model as described in any one of claims 1 to 4 is used to calculate the carbon sink amount of the target area.

6. The method according to claim 5, characterized in that The following steps are involved: Use the multimodal data acquisition module to collect basic data of the target area; Performing spatiotemporal alignment and / or noise filtering processing on the basic data using a data fusion and quality control module; The basic data that has undergone spatiotemporal alignment and / or noise filtering processing is input into the forestry carbon sink calculation model to calculate the carbon sink amount.

7. The method according to claim 6, characterized in that The method of collecting basic data of the target area using the multimodal data collection module includes: Use high-resolution satellite remote sensing units to obtain information on canopy height and biomass density; Use the multispectral scanning unit of the UAV to obtain the distribution information of understory vegetation and litter; Use ground IoT units to obtain temperature, humidity, and CO2 flux information; The data fusion and quality control module is used to perform spatiotemporal alignment and / or noise filtering on the basic data, including: Use the spatiotemporal alignment engine to unify the temporal and spatial resolutions of various basic data; Establish a mapping relationship between satellite imagery and ground sensor data; Estimate data values for uncovered areas; Use generative adversarial networks to repair abnormal information and / or missing data; The step of inputting the basic data processed by spatiotemporal alignment and / or noise filtering into the forestry carbon sink calculation model to calculate the carbon sink amount includes: Use the feature extraction module to extract the time series feature information and environmental variable feature information of the processed basic data, and fuse and splice the extracted feature information to obtain the feature tensor; Use the aboveground biomass carbon sink estimation module to calculate the aboveground biomass carbon sink data of the feature tensor; Use the underground soil carbon sequestration estimation module to calculate the soil carbon sequestration data of the characteristic tensor; Use the litter carbon storage estimation module to calculate litter carbon sink data of the characteristic tensor; Use the joint loss calculation module to calculate carbon loss data for aboveground biomass carbon sink data, soil carbon sink data, and litter carbon sink data; The final carbon sink fusion module is used to calculate the carbon sink based on aboveground biological carbon sink data, soil carbon sink data, litter carbon sink data and carbon loss data.

8. The method according to claim 7, characterized in that When forest ecological changes occur, it also includes: adjusting the carbon sink amount using a dynamic carbon sink correction module; events that cause forest ecological changes include fires, diseases and insect pests.

9. The method according to claim 8, characterized in that It also includes using historical fire data to predict future carbon emission risks.

10. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed by a processor, the method according to any one of claims 5 to 9 is implemented.

Images