Forest carbon reserve unmanned aerial vehicle monitoring method and system
By combining UAV remote sensing and ground sensor data, an ecological stress correction model was constructed, which solved the problem of insufficient consideration of environmental factors in forest carbon storage monitoring and achieved high-precision, intelligent and adaptive carbon storage monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-23
- Publication Date
- 2026-03-31
AI Technical Summary
Existing forest carbon storage monitoring methods lack consideration for microhabitat factors, resulting in insufficient accuracy and reliability under complex environmental conditions. Furthermore, traditional monitoring processes lack intelligent and dynamic optimization capabilities.
By combining geometric structural features obtained from UAV remote sensing with microhabitat parameters perceived by ground sensors, an ecological stress correction model is constructed. Potential biomass is estimated through ecological stress index and allometric growth parameters, and the model parameters are self-learned and adjusted using a closed-loop optimization mechanism.
It has improved the accuracy and reliability of forest carbon storage monitoring, realized the intelligence and efficiency of ground surveys, and can adaptively optimize under different environments, thereby enhancing the interpretability and robustness of monitoring mechanisms.
Smart Images
Figure CN121762783A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of forest carbon storage monitoring and UAV remote sensing technology, specifically a UAV method and system for monitoring forest carbon storage. Background Technology
[0002] In the field of forest carbon sink monitoring and assessment, UAV remote sensing is widely used as an efficient technical means. Current mainstream methods mostly rely on geometric structural features such as canopy height and coverage extracted from remote sensing images to estimate forest biomass and carbon storage by establishing statistical regression models. However, these methods generally simplify the forest ecosystem into a physical structure model, rarely considering the actual stress effects of microhabitat factors such as soil nutrients and water conditions on vegetation growth. This approach results in insufficient interpretability of the estimation results, and the accuracy and reliability of the model are significantly affected when environmental conditions are complex and variable. In addition, traditional monitoring processes lack the ability to spatially assess the uncertainty of the model itself, and ground sampling verification work is often inefficient and costly due to the lack of intelligent guidance. Existing models are also generally static models, lacking a dynamic closed-loop feedback mechanism for self-learning and parameter optimization using newly added measured data.
[0003] Therefore, how to construct a forest carbon storage monitoring method that can deeply integrate macroscopic remote sensing observation with microscopic ecological process mechanisms and has the ability to self-assess uncertainties and intelligent closed-loop optimization is a problem that urgently needs to be solved by those skilled in the art.
[0004] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention
[0005] To address the aforementioned technical problems, this invention discloses a method and system for monitoring forest carbon storage using unmanned aerial vehicles (UAVs). Specifically, the technical solution of this invention is as follows:
[0006] A method for monitoring forest carbon storage using unmanned aerial vehicles (UAVs) includes:
[0007] Acquire geometric structural features collected by UAVs and microhabitat parameters collected by ground sensors;
[0008] S1, based on microhabitat parameters, calculates the ecological stress index;
[0009] S2, based on geometric structural features and preset allometric growth parameters, estimates potential biomass;
[0010] S3 combines potential biomass and ecological stress index, and uses a preset stress sensitivity coefficient to calculate the corrected biomass;
[0011] S4. Based on the corrected biomass and the preset biomass-carbon conversion coefficient, determine the forest carbon storage.
[0012] Optionally, the calculation of the ecological stress index includes:
[0013] Based on microhabitat parameters, preset weighting factors, preset optimal values of factors, and preset reference ranges of factors, the ecological stress index is obtained by weighted summation through the constraint factor law model.
[0014] Optionally, geometric features include canopy height and canopy coverage;
[0015] Estimate potential biomass includes:
[0016] Potential biomass was obtained by combining canopy height and canopy coverage and processing it using an allometric growth equation.
[0017] Optionally, the method further includes:
[0018] The model-observation conflict index is calculated based on potential biomass and corrected biomass.
[0019] Optionally, the method further includes:
[0020] If the model-observation conflict index is greater than a preset uncertainty threshold, a priority sampling map is generated.
[0021] If the model-observation conflict index is less than or equal to a preset uncertainty threshold, a priority sampling map will not be generated.
[0022] Optionally, the method further includes:
[0023] Obtain ground-based measured carbon storage based on priority sampling maps;
[0024] With the goal of minimizing the error between forest carbon storage and measured ground carbon storage, the preset undetermined parameters are optimized.
[0025] Among them, the parameters to be determined include the weighting factor and the stress sensitivity coefficient.
[0026] A forest carbon storage drone monitoring system includes:
[0027] The data acquisition module is used to acquire geometric structural features collected by the UAV and microhabitat parameters collected by ground sensors;
[0028] The stress index calculation module is used to calculate the ecological stress index based on microhabitat parameters;
[0029] The potential biomass estimation module is used to estimate potential biomass based on geometric structural features and preset allometric growth parameters.
[0030] The biomass correction module is used to combine potential biomass with the ecological stress index and calculate the corrected biomass using a preset stress sensitivity coefficient.
[0031] The carbon storage determination module is used to determine forest carbon storage based on the corrected biomass and the preset biomass-carbon conversion coefficient.
[0032] Optionally, the system also includes:
[0033] Uncertainty assessment module is used to calculate the model-observation conflict index based on potential biomass and corrected biomass;
[0034] The intelligent sampling navigation module generates a priority sampling map when the model-observation conflict index is greater than a preset uncertainty threshold, and does not generate a priority sampling map when the model-observation conflict index is less than or equal to the preset uncertainty threshold.
[0035] The closed-loop optimization module is used to obtain the measured ground carbon storage based on the priority sampling map, and optimize the preset undetermined parameters with the goal of minimizing the error between forest carbon storage and measured ground carbon storage. The undetermined parameters include weighting factors and stress sensitivity coefficients.
[0036] Compared with the prior art, the present invention has the following beneficial effects:
[0037] 1. This invention improves monitoring accuracy and mechanistic interpretability. This method innovatively integrates macroscopic geometric structures acquired through UAV remote sensing with microscopic habitat parameters perceived by ground sensors to construct a two-step estimation model for potential biomass and ecological stress correction. This model not only leverages the advantages of remote sensing technology in its wide coverage and high efficiency but also deeply integrates ecological process mechanisms. This makes the carbon storage estimation results no longer simply data fitting but possesses stronger mechanistic interpretability, enabling a more realistic response to environmental changes and significantly improving the accuracy and reliability of forest carbon storage monitoring.
[0038] 2. This invention achieves intelligent and efficient ground surveys. The method establishes a feedback loop from uncertainty self-assessment to intelligent sampling navigation. Through a unique model-observation conflict index, it can accurately identify the spatial areas with the highest uncertainty in model estimation and automatically generate a visualized priority sampling map. This mechanism transforms traditional, broad-based ground surveys into efficient, intelligent, and precise targeted sampling, greatly improving the efficiency and value of ground validation. It obtains the most critical data for model optimization with minimal sampling cost, effectively solving the pain points of difficult and costly acquisition of ground-based true data in large-scale carbon storage monitoring.
[0039] 3. This invention achieves adaptive optimization and continuous learning of the model. This method constructs a complete closed-loop optimization and adaptive learning mechanism. It can utilize high-value ground-based measured data acquired through intelligent sampling to automatically optimize key undetermined parameters within the model, aiming to minimize estimation errors. This design transforms the monitoring method from a static estimation model into a dynamic intelligent system capable of continuous self-learning and iterative optimization, enabling it to continuously absorb new knowledge and automatically calibrate, thereby significantly improving the estimation accuracy, universality, and robustness under different regions and environmental conditions.
[0040] 4. This invention achieves the scientific quantification and comprehensive assessment of environmental stress. This method proposes an ecological stress index calculation model based on the law of limiting factors, which can scientifically integrate multi-source, heterogeneous microhabitat parameters into a unified, dimensionless comprehensive stress index. This model not only considers the deviation of each environmental factor from optimal growth conditions but also innovatively introduces weighting factors that can be optimized through data learning to reflect the relative importance of different factors in a specific ecosystem. This approach makes the quantification of environmental stress closer to real ecological processes, providing a more reliable scientific basis for subsequent precise adjustments to biomass. Attached Figure Description
[0041] The present invention will be further explained below with reference to the accompanying drawings and embodiments:
[0042] Figure 1 This is a flowchart of the method of the present invention.
[0043] Figure 2 This is a system structure diagram of the present invention. Detailed Implementation
[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.
[0045] Example 1:
[0046] Please see Figure 1 A method for monitoring forest carbon storage using unmanned aerial vehicles (UAVs), comprising:
[0047] Acquire geometric structural features collected by UAVs and microhabitat parameters collected by ground sensors;
[0048] Calculate the ecological stress index based on microhabitat parameters;
[0049] Based on geometric structural features and preset allometric growth parameters, the potential biomass is estimated.
[0050] By combining potential biomass and ecological stress index, and using a preset stress sensitivity coefficient, the corrected biomass is calculated.
[0051] Forest carbon storage is determined based on the corrected biomass and the preset biomass-carbon conversion coefficient.
[0052] This embodiment provides a method for monitoring forest carbon storage using unmanned aerial vehicles (UAVs). The aim is to achieve high-precision and timely monitoring of forest carbon storage by dynamically fusing high-resolution remote sensing data with forest ecological process models. This method constitutes a complete technical closed loop. The acquisition of geometric structural features collected by the UAV and microhabitat parameters collected by ground sensors forms the data input foundation for the entire method. The purpose of geometric structural features is to macroscopically and physically characterize the three-dimensional morphology of the forest. In this embodiment, point cloud data is acquired using a lidar mounted on the UAV. From this data, a digital surface model and a digital elevation model are constructed. The difference between the two models generates a canopy height model, and the average height within the sample plot is calculated. This allows the extraction of the canopy height, which characterizes the stand growth status. ;
[0053] Simultaneously, high-resolution images are acquired using a hyperspectral sensor mounted on a drone, from which the Normalized Difference Vegetation Index (NDVI) is calculated. By setting an appropriate NDVI threshold, vegetation pixels are segmented, and the proportion of vegetation pixels to the total number of pixels is calculated, thereby interpreting the canopy cover. The purpose of microhabitat parameters is to precisely and ecologically characterize the local environmental conditions for forest growth. In this embodiment, a ground-based sensor network deployed in the forest area is used to collect real-time measured values of various stress factors, including soil available nitrogen concentration, water content, and pH value. ;
[0054] Ecological stress index is calculated based on microhabitat parameters; ecological stress index The aim is to integrate multidimensional and heterogeneous microenvironment parameters into a single indicator that can uniformly quantify the degree of environmental stress.
[0055] Based on geometric structural features and preset allometric growth parameters, potential biomass is estimated; potential biomass The aim is to establish a biomass benchmark under ideal conditions, that is, the maximum biomass that the current forest structure can support without any environmental stress.
[0056] By combining potential biomass and an ecological stress index, and using a preset stress sensitivity coefficient, the corrected biomass is calculated; this step is the core of the method of this invention, and its inherent logic lies in using an ecological stress index that represents the actual environmental impact. For ideal potential biomass Make adjustments to obtain a revised biomass that better reflects reality. ;
[0057] In this embodiment, a negative exponential function is used to express this correction relationship. This function can reasonably simulate the nonlinear saturation effect commonly found in ecological responses. The specific formula is as follows: ;in, The stress sensitivity coefficient is a dimensionless parameter that characterizes the sensitivity of the forest ecosystem to integrated environmental stresses. Its function is to regulate the intensity of the stress index's effect on potential biomass reduction. Its value is obtained through iterative learning via a subsequent closed-loop optimization feedback mechanism.
[0058] Forest carbon storage is determined based on the corrected biomass and a preset biomass-carbon conversion coefficient; this step aims to convert the estimated biomass into the final carbon storage result; in this embodiment, the corrected biomass... Multiply by a biomass-carbon conversion factor To obtain the final forest carbon storage. ,Right now Biomass-carbon conversion coefficient The ratio is used to convert a unit of biomass into a unit of carbon content. Its function is to complete the final conversion from biomass to carbon storage. It is set according to the guidelines issued by authoritative organizations such as the Intergovernmental Panel on Climate Change (IPCC), and is usually set to 0.5.
[0059] The method in this embodiment combines macroscopic geometric features acquired by UAV remote sensing with microscopic habitat parameters perceived by ground sensors to construct a two-step estimation model of potential biomass-ecological stress correction. This method not only utilizes the advantages of remote sensing technology in terms of its wide range and high efficiency, but also innovatively incorporates ecological process mechanisms. This makes the carbon storage estimation results no longer a simple data fitting, but rather has stronger mechanistic interpretability and responsiveness to environmental changes, thereby improving the accuracy and reliability of forest carbon storage monitoring.
[0060] Example 2:
[0061] The calculation of the ecological stress index includes:
[0062] Based on microhabitat parameters, preset weight factors, preset optimal values of factors, and preset reference ranges of factors, the ecological stress index is obtained by weighted summation through the constraint factor law model.
[0063] This embodiment, based on Embodiment 1, provides a specific implementation of the calculation of the ecological stress index. To more accurately quantify the comprehensive limiting effect of multiple environmental factors on vegetation growth, the calculation of the ecological stress index includes: based on microhabitat parameters, preset weight factors, preset optimal values of factors, and preset reference ranges of factors, a weighted summation process is performed using a limiting factor law model to obtain the ecological stress index.
[0064] The purpose of this calculation method is to scientifically integrate microhabitat parameters with different dimensions and influencing mechanisms into a unified, dimensionless stress index. Its implementation follows the limiting factor law in ecology, which states that vegetation growth is simultaneously influenced by multiple environmental factors. This embodiment extends it to a weighted model, and the specific calculation formula is as follows:
[0065] in, As an ecological stress index; Indicates the first Microhabitat parameters, The total number of factors; The weighting factor is a dimensionless parameter whose physical meaning lies in reflecting the weighting factor. The relative importance of stress-like factors to the biomass accumulation of this ecosystem is derived from data-driven learning and optimization through subsequent closed-loop optimization steps; The measured value of the factor refers to the first factor. Real-time on-site measurements of microhabitat parameters are provided by ground sensor networks or geographic information databases;
[0066] The optimal factor value refers to the ecologically most suitable value for the growth of this forest vegetation. The values of each environmental factor are determined based on authoritative ecological literature or prior knowledge of a specific tree species. The reference range is a benchmark used to normalize the deviation of each factor. Its function is to eliminate the inequivalence caused by different dimensions of different factors, ensuring that they can be comprehensively compared and weighted summed. To ensure the stability and clarity of the physical meaning of the normalization result, specifically, the range of the factor within the study area can be taken, i.e. ,in and The first The maximum and minimum values of each factor within the study area;
[0067] Calculate the measured value of each stress factor Its optimal value The degree of deviation, and then use Normalization is performed, and finally, the factor is determined based on its relative importance to the ecosystem. We perform a weighted summation; in this way, the invisible, multidimensional environmental stress field is transformed into a computable and comparable spatialized index. ;
[0068] This embodiment, by introducing a weighted summation model based on the law of limiting factors, can more comprehensively and scientifically quantify the combined stress effects of microenvironments. It not only considers the absolute influence of each factor but also uses optimizable weighting factors. This reflects the relative importance of different factors in a specific ecosystem, thus making the ecological stress index... The calculation results are closer to the real ecological process, providing a more reliable basis for the subsequent accurate correction of biomass and further improving the accuracy of the final carbon storage estimate.
[0069] It should be noted that the stress factor deviation calculation method used in this embodiment ( This is a simplified linear model designed to achieve unified quantification of multidimensional heterogeneous parameters. In certain specific application scenarios, if the dose-response relationship of some stress factors has obvious nonlinear characteristics, such as Gaussian function or piecewise function form, this linear deviation model can be replaced with a more mechanistic nonlinear response function to further improve the physical fidelity of the ecological stress index. This is a further optimization and improvement of the method of this invention.
[0070] Example 3:
[0071] Geometric features include canopy height and canopy coverage;
[0072] Estimate potential biomass includes:
[0073] Potential biomass was obtained by combining canopy height and canopy coverage and processing it using the allometric growth equation.
[0074] This embodiment, based on Embodiment 1, provides a specific implementation method for estimating potential biomass. To more accurately deduce the theoretical potential biomass from the geometric structural features obtained by the UAV, this embodiment clarifies the technical features used and provides a specific estimation model. The geometric structural features include canopy height and canopy coverage.
[0075] Estimating potential biomass involves: based on the combination of canopy height and canopy coverage, and processed using the allometric growth equation to obtain potential biomass;
[0076] The purpose of this estimation method is to select the combination of remote sensing features that has the strongest indicative significance for forest stand biomass density, and to calculate it using a recognized mechanistic model in the forestry field, so as to ensure the potential biomass. The scientific validity and accuracy of the estimation are ensured by employing a modified allometric growth equation, which estimates the average canopy height at the plot level. With canopy coverage The input variables are combined; the specific calculation formula is as follows:
[0077] in, Potential biomass, in tons per hectare ; Canopy height, in meters. This refers to point cloud data collected by the lidar mounted on the drone platform; , where is the canopy coverage, is a dimensionless value ranging from 0 to 1, and is high-resolution image data acquired by the UAV platform; and Allometric growth parameters are empirical coefficients that reflect the nonlinear relationship between biomass and geometric dimensions. Their function is to calibrate the growth pattern of a specific region or tree species. It should be noted that these two parameters are predetermined for a specific forest type or region by fitting a calibration dataset, which is independent of the real-time input during model runtime.
[0078] For example, a series of values can be collected, including a combination of stand average height and cover, and their corresponding measured biomass. Data points, including ground-measured biomass Biomass is obtained through standard forestry plot survey methods, such as measuring parameters like diameter at breast height (DBH) of each tree within the plot using the plot inventory method. This is then combined with the allometric growth equation for the specific tree species to calculate the biomass of each tree. Finally, the total biomass of the plot is summed, and the results are fitted using least squares regression analysis. and The value of ; to ensure dimensional consistency, the parameter The dimensions are ;
[0079] This formula uses canopy height, which characterizes the vertical growth of a forest, to represent the forest's vertical growth. Canopy coverage, which characterizes horizontal expansion density By multiplying these factors, a composite factor was constructed that can comprehensively reflect the three-dimensional structure and density of the forest stand. This composite factor serves as the independent variable in the allometric growth equation, through parameters... and Through regulation, the potential biomass under ideal conditions can be estimated. ;
[0080] Compared to traditional remote sensing methods that use only a single geometric feature, this embodiment employs a combination of canopy height and canopy cover. This combination more comprehensively reflects the spatial distribution density of forest stand biomass and is an excellent indicator of stand biomass density. Therefore, the potential biomass estimated in this way... As a benchmark value for subsequent corrections, it has higher initial accuracy, thus providing a solid foundation for the final accuracy of the entire model chain.
[0081] Example 4:
[0082] The method includes:
[0083] The model-observation conflict index is calculated based on potential biomass and corrected biomass.
[0084] Also includes:
[0085] If the model-observation conflict index is greater than a preset uncertainty threshold, a priority sampling map is generated.
[0086] If the model-observation conflict index is less than or equal to a preset uncertainty threshold, a priority sampling map will not be generated.
[0087] The model uncertainty assessment and intelligent sampling navigation mechanism are introduced, forming an organic whole;
[0088] This method also includes: calculating the model-observation conflict index based on potential biomass and corrected biomass; the model-observation conflict index... The aim is to quantitatively identify the region with the largest model correction magnitude, i.e., the highest model uncertainty; this is achieved by defining the relative correction rate of biomass; in Example 1... expression Substituting, we can obtain:
[0089] in, The model-observation conflict index is a dimensionless value between 0 and 1, and its calculation relies entirely on the existing stress sensitivity coefficient. and ecological stress index ; The higher the value, the more the biomass estimation results in this area depend on ecological stress correction, which means that the traditional model based solely on geometric features is at greater risk of failure in this area;
[0090] The method further includes: generating a priority sampling map in response to a model-observation conflict index greater than a preset uncertainty threshold; and not generating a priority sampling map in response to a model-observation conflict index less than or equal to a preset uncertainty threshold.
[0091] use The evaluation results of the index enable intelligent and targeted navigation of ground survey work, accurately allocating limited sampling resources to key areas where the model most needs validation and optimization. This is achieved as follows: a preset uncertainty threshold is set, for example, 0.4. This threshold can be determined based on historical data statistical analysis. Specifically, a historical validation dataset containing model estimates and ground truth values can be used. By plotting the receiver operating characteristic curve, the optimal balance point in distinguishing between high-error and low-error samples can be found. For example, the model-observation conflict index corresponding to the maximum Youden index. The value is set as a preset uncertainty threshold; for example, when the relative correction rate exceeds 40%, the probability of the model estimation error increasing significantly increases, so this is used as the boundary.
[0092] In a geographic information system, the entire study area is rasterized, and the value of each raster cell is calculated. Value; all The raster cells are marked as high-uncertainty areas, and a visual priority sampling map is generated. This map is then distributed to the ground survey team to guide them in prioritizing on-site investigation and precise sampling of these marked key areas to obtain high-value ground truth data. In areas where the model estimation results are deemed reliable, immediate ground validation is not required.
[0093] A feedback loop from uncertainty assessment to intelligent sampling navigation was established. By calculating the model-observation conflict index, this method can self-diagnose its spatial weaknesses. Based on this diagnosis, a priority sampling map is generated, transforming traditional ground surveys into an efficient, intelligent, and on-demand targeted sampling process. This mechanism greatly improves the efficiency and value of ground validation, obtaining the most critical data for model optimization with minimal sampling cost, and solving the pain points of difficult and costly acquisition of ground-based true data in large-scale carbon storage monitoring.
[0094] Example 5:
[0095] The method also includes:
[0096] Obtain ground-based measured carbon storage based on priority sampling maps;
[0097] With the goal of minimizing the error between forest carbon storage and measured ground carbon storage, the preset undetermined parameters are optimized.
[0098] This embodiment, based on embodiment 4, further describes how to use data acquired through intelligent sampling to optimize the model itself, thereby forming a complete closed-loop feedback system; the method also includes: acquiring ground-measured carbon storage based on priority sampling maps; optimizing preset undetermined parameters with the goal of minimizing the error between forest carbon storage and ground-measured carbon storage; wherein, the undetermined parameters include weighting factors and stress sensitivity coefficients;
[0099] The purpose of this closed-loop optimization mechanism is to enable the model to continuously learn from new measured data, thereby constantly improving its estimation accuracy and generalizability in complex environments; its implementation includes:
[0100] The generated priority sampling map allows the ground survey team to establish sample plots within designated high-uncertainty areas. Biomass within these plots is then precisely measured using traditional forestry survey methods and converted into measured ground carbon storage. ,in This represents the newly added sample point number;
[0101] The forest carbon storage estimated by the model at the corresponding sample locations. Compared with ground-measured carbon storage By comparing the two values, an objective function is constructed that aims to minimize the error between them. In this embodiment, the objective function is to minimize the sum of squared errors between the estimated and measured values, and its form is as follows:
[0102] in, and These are parameters to be determined, meaning parameters that need to be learned and optimized through data in the model. In this invention, they specifically refer to the weighting factors in the ecological stress index calculation formula. and the stress sensitivity coefficient in the biomass correction formula ; For the index of newly added sample points; This represents the total number of newly added samples. For the model to the first The estimated carbon storage value of each newly added sampling point is obtained from the previous steps; For the first The true values of ground-measured carbon storage at the newly added sampling points were obtained through ground surveys.
[0103] An optimization algorithm, such as gradient descent or particle swarm optimization, is used to automatically adjust the parameters to be determined. and The goal is to find the optimal value of the parameter so that the objective function is minimized. This process is iterated repeatedly until the parameters converge or the preset number of iterations is reached, ultimately yielding a set of optimal parameters. and Parameter combinations;
[0104] This closed-loop process effectively uses the high-value ground truth data acquired through intelligent sampling to drive the model's self-evolution; when there is a large deviation between the model's estimated values and the measured values, the objective function value will be large, and the optimization algorithm will adjust accordingly. and For example, if the model systematically overestimates the carbon reserves in a region, the algorithm might increase the weighting of the dominant stressor in that region. Or overall stress sensitivity coefficient This will increase the correction of potential biomass, making the next estimate closer to the true value.
[0105] This embodiment transforms the entire monitoring method from a static estimation model into a dynamic, self-learning, and iteratively optimizing intelligent system by establishing a closed-loop optimization mechanism. This design enables the model to continuously absorb new knowledge and automatically calibrate key parameters, thereby continuously improving its estimation accuracy and robustness under different regions and environmental conditions. The fully calibrated model can adjust the input stress factors... The value is used to simulate and predict changes in forest carbon storage under different future environmental scenarios, thus expanding the application value of this invention.
[0106] Example 6:
[0107] Please see Figure 2 The data acquisition module is used to acquire geometric structural features collected by the UAV and microhabitat parameters collected by ground sensors.
[0108] The stress index calculation module is used to calculate the ecological stress index based on microhabitat parameters;
[0109] The potential biomass estimation module is used to estimate potential biomass based on geometric structural features and preset allometric growth parameters.
[0110] The biomass correction module is used to combine potential biomass with the ecological stress index and calculate the corrected biomass using a preset stress sensitivity coefficient.
[0111] The carbon storage determination module is used to determine forest carbon storage based on the corrected biomass and the preset biomass-carbon conversion coefficient.
[0112] This embodiment provides a forest carbon storage drone monitoring system, which is implemented in a modular hardware and functional manner, and includes:
[0113] The data acquisition module is designed to provide the initial data input for the execution method, and is used to acquire the geometric features collected by the UAV and the micro-habitat parameters collected by the ground sensors.
[0114] The stress index calculation module is designed to perform the relevant steps in Example 1. This module is a software functional unit that receives microhabitat parameters from the data acquisition module and calculates the ecological stress index based on the microhabitat parameters.
[0115] The potential biomass estimation module is designed to perform the relevant steps in Example 1. This module is a software functional unit that receives geometric structural features from the data acquisition module and uses them to estimate potential biomass based on the geometric structural features and preset allometric growth parameters.
[0116] The biomass correction module is designed to perform the relevant steps in Example 1 and is the core calculation unit of the system. This module receives the ecological stress index and potential biomass, combines the potential biomass with the ecological stress index, and uses a preset stress sensitivity coefficient to calculate the corrected biomass.
[0117] The carbon storage determination module is designed to perform the relevant steps in Example 1 and output the final result. This module receives the corrected biomass and uses it to determine the forest carbon storage based on the corrected biomass and the preset biomass-carbon conversion coefficient.
[0118] The system in this embodiment clearly implements each functional step of the method through modular design; each module performs its own function and the data flow is clear, forming a complete processing chain from multi-source data acquisition to final carbon storage output, providing a complete system-level solution for efficient and accurate monitoring of forest carbon storage.
[0119] Example 7:
[0120] The system also includes:
[0121] Uncertainty assessment module is used to calculate the model-observation conflict index based on potential biomass and corrected biomass;
[0122] The intelligent sampling navigation module generates a priority sampling map when the model-observation conflict index is greater than a preset uncertainty threshold, and does not generate a priority sampling map when the model-observation conflict index is less than or equal to the preset uncertainty threshold.
[0123] The closed-loop optimization module is used to obtain the measured ground carbon storage based on the priority sampling map, and optimize the preset undetermined parameters with the goal of minimizing the error between forest carbon storage and measured ground carbon storage. The undetermined parameters include weighting factors and stress sensitivity coefficients.
[0124] This embodiment, based on the system of Embodiment 6, adds a series of functional modules to achieve model self-evaluation and closed-loop optimization, and also includes:
[0125] The uncertainty assessment module is designed to perform the functions of Example 4 and enable the model to self-diagnose. This module receives potential biomass and corrected biomass and uses them to calculate the model-observation conflict index based on the potential biomass and corrected biomass.
[0126] The intelligent sampling navigation module is designed to perform the functions of Embodiment 4 and guide efficient ground verification. This module receives the model-observation conflict index map and generates a priority sampling map in response to the model-observation conflict index being greater than a preset uncertainty threshold, and does not generate a priority sampling map in response to the model-observation conflict index being less than or equal to the preset uncertainty threshold.
[0127] The closed-loop optimization module aims to perform the functions of Example 5, enabling the model to learn itself and iteratively optimize. The functional chain of this module includes: acquiring measured ground carbon storage based on a priority sampling map; using this measured data as the true value, optimizing preset parameters to minimize the error between the model-estimated forest carbon storage and the measured ground carbon storage, where the parameters include weighting factors and stress sensitivity coefficients; and updating the optimized parameters in the stress index calculation module and the biomass correction module for use in the next round of carbon storage calculation.
[0128] This embodiment upgrades the monitoring system of Embodiment 6 from an open linear processing system to an adaptive intelligent system with self-assessment, intelligent navigation, and closed-loop learning capabilities by adding an uncertainty assessment module, an intelligent sampling and navigation module, and a closed-loop optimization module. These three modules work together to realize a complete intelligent closed loop of model calculation, problem discovery, guided verification, and learning optimization. This enables the system to continuously acquire high-value information from the external environment to improve itself, thereby continuously improving its monitoring accuracy and adaptability to different environments during long-term operation.
[0129] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims
1. A forest carbon stock unmanned aerial vehicle monitoring method, characterized in that, The method comprises the following steps: Obtaining geometric structure features collected by a UAV and microhabitat parameters collected by a ground sensor; S1. Calculating an ecological stress index based on the microhabitat parameters; S2. Estimating potential biomass based on the geometric structure features and preset allometric parameters; S3. Combining the potential biomass and the ecological stress index, and using a preset stress sensitivity coefficient to calculate a corrected biomass; S4. Determining the forest carbon storage based on the corrected biomass and a preset biomass-carbon conversion coefficient. 2.The forest carbon storage unmanned aerial vehicle monitoring method of claim 1, wherein, The calculation of the ecological stress index comprises: Based on the microhabitat parameters, a preset weight factor, a preset factor optimal value, and a preset factor reference range, the ecological stress index is obtained by weighted summation processing through a limiting factor law model. 3.The forest carbon storage unmanned aerial vehicle monitoring method of claim 1, wherein, The geometric structure features include canopy height and canopy coverage. The estimation of the potential biomass comprises: Based on the combination of the canopy height and the canopy coverage, the potential biomass is obtained by processing through an allometric equation. 4.The forest carbon storage unmanned aerial vehicle monitoring method of claim 1, wherein, The method further comprises: Calculating a model-observation conflict index based on the potential biomass and the corrected biomass.
5. The method of claim 4, wherein, The method further comprises: In response to the model-observation conflict index being greater than a preset uncertainty threshold, a priority sampling map is generated; In response to the model-observation conflict index being less than or equal to the preset uncertainty threshold, the priority sampling map is not generated.
6. The method of claim 5, wherein, The method further comprises: Obtaining a ground measured carbon storage collected based on the priority sampling map; Optimizing a preset to-be-determined parameter by minimizing the error between the forest carbon storage and the ground measured carbon storage, wherein the to-be-determined parameter includes the weight factor and the stress sensitivity coefficient. The method comprises:
7. A forest carbon storage UAV monitoring system based on the forest carbon storage UAV monitoring method of any one of claims 1-6, characterized in that, A data collection module for obtaining geometric structure features collected by a UAV and microhabitat parameters collected by a ground sensor; A stress index calculation module for calculating an ecological stress index based on the microhabitat parameters; A potential biomass estimation module for estimating potential biomass based on the geometric structure features and preset allometric parameters; A biomass correction module for combining the potential biomass and the ecological stress index, and using a preset stress sensitivity coefficient to calculate a corrected biomass; A carbon storage determination module for determining the forest carbon storage based on the corrected biomass and a preset biomass-carbon conversion coefficient. The method further comprises:
8. The forest carbon stock UAV monitoring system of claim 7, wherein, An uncertainty evaluation module for calculating a model-observation conflict index based on the potential biomass and the corrected biomass; An intelligent sampling navigation module for generating a priority sampling map in response to the model-observation conflict index being greater than a preset uncertainty threshold, and not generating the priority sampling map in response to the model-observation conflict index being less than or equal to the preset uncertainty threshold; A closed-loop optimization module for obtaining a ground measured carbon storage collected based on the priority sampling map, and optimizing a preset to-be-determined parameter by minimizing the error between the forest carbon storage and the ground measured carbon storage, wherein the to-be-determined parameter includes the weight factor and the stress sensitivity coefficient.