Southern grassland soil carbon reserve dynamic evaluation method and system fused with remote sensing monitoring

Abstract

The invention relates to a southern grassland soil carbon reserve dynamic evaluation method and system fused with remote sensing monitoring, and relates to the technical field of soil carbon reserve evaluation, and the method comprises the steps: drawing a target region topographic map based on multi-source satellite remote sensing data, inputting an initial soil carbon reserve inversion model, and outputting an initial soil carbon reserve distribution diagram and a deterministic thermodynamic diagram; extracting an uncertain area according to a certainty degree threshold value, and configuring and labeling supplementary sampling points according to a division rule; acquiring parameter information of the detection robot, generating a digital twinborn model in combination with the topographic map, configuring a supplementary sampling scheme, and sampling supplementary sampling points to obtain supplementary soil carbon reserve data; and returning the supplemented soil carbon reserve data to the initial soil carbon reserve inversion model, generating a corrected soil carbon reserve distribution diagram and a corrected determinacy thermodynamic diagram, judging that the task is completed when determinacy reaches the upper limit of resource consumption, otherwise, continuing to supplement sampling correction. The problems that traditional soil carbon reserve assessment is inaccurate and low in efficiency are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of soil carbon storage evaluation, in particular to a southern grassland soil carbon storage dynamic evaluation method and system fusing remote sensing monitoring. BACKGROUND

[0002] With the improvement of the data precision requirement of southern grassland soil carbon storage in ecological protection and carbon cycle research, the reliability of dynamic evaluation becomes a key demand.

[0003] At present, the traditional method for correcting soil carbon storage data obtained by remote sensing often uses unmanned aerial vehicle soil sampling, but the environmental adaptability of the unmanned aerial vehicle is poor, and it is difficult to land and sample in harsh weather or complex ground environment; at the same time, the existing technology lacks targeted sampling setting for the "uncertain area" with the highest model calibration value, and there are only two ways of high-precision low-efficiency ground grid sampling and low-precision high-efficiency pure remote sensing model, resulting in low efficiency of model correction of soil carbon storage data, which is difficult to support accurate evaluation. SUMMARY

[0004] The present application provides a southern grassland soil carbon storage dynamic evaluation method and system fusing remote sensing monitoring, which improves the current situation of poor environmental adaptability of traditional unmanned aerial vehicle sampling, lack of targeted sampling of "uncertain area" and low correction efficiency.

[0005] The embodiments of the present application disclose the following technical solutions: In a first aspect, the embodiments of the present application provide a southern grassland soil carbon storage dynamic evaluation method fusing remote sensing monitoring, which comprises: Based on the multi-source satellite remote sensing data of the target area, a topographic map of the target area is drawn, and the multi-source satellite remote sensing data of the target area is input into an initial soil carbon storage inversion model to output an initial soil carbon storage distribution map and a deterministic thermal map of the target area, wherein the deterministic thermal map is composed of a plurality of pixels labeled with a degree of certainty; Based on the initial soil carbon storage distribution map and the deterministic thermal map, an uncertain area is extracted according to a preset degree of certainty threshold, and a plurality of supplementary sampling points are configured for the uncertain area according to a preset division rule, and the plurality of supplementary sampling points are labeled on the topographic map; Obtain the parameter information of all detection robots in the target area, generate a digital twin model based on the parameter information of each detection robot and the topographic map of the target area, configure a supplementary sampling scheme based on the digital twin model, and non-destructively sample the plurality of supplementary sampling points to obtain a plurality of supplementary soil carbon storage data; The supplementary soil carbon storage data are fed back to the initial soil carbon storage inversion model to generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is deemed to be completed. Otherwise, supplementary sampling continues to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map.

[0006] Secondly, embodiments of this application provide a dynamic assessment system for soil carbon storage in southern grasslands that integrates remote sensing monitoring. The system includes: The initial inversion mapping module is used to draw a topographic map of the target area based on multi-source satellite remote sensing data of the target area, and input the multi-source satellite remote sensing data of the target area into the initial soil carbon storage inversion model, and output the initial soil carbon storage distribution map and deterministic heat map of the target area, wherein the deterministic heat map is composed of multiple pixels labeled with determinism; The uncertain area sampling configuration module is used to extract uncertain areas based on the initial soil carbon storage distribution map and the deterministic heat map according to a preset certainty threshold, and to configure several supplementary sampling points for the uncertain areas according to preset division rules, and to mark several supplementary sampling points on the topographic map; The digital twin sampling module is used to acquire parameter information of all exploration robots in the target area, generate a digital twin model based on the parameter information of each exploration robot and the topographic map of the target area, configure a supplementary sampling scheme based on the digital twin model, perform non-destructive sampling on several supplementary sampling points, and obtain several supplementary soil carbon storage data. The model correction and determination module is used to transmit the supplementary soil carbon storage data back to the initial soil carbon storage inversion model, generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is determined to be completed. Otherwise, supplementary sampling continues to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application proposes a method and system for dynamic assessment of soil carbon storage in southern grasslands that integrates remote sensing monitoring. Through collaborative operations such as inverting the initial carbon storage distribution and deterministic heat map from remote sensing data, configuring supplementary sampling points for uncertain areas, optimizing the robot sampling scheme using a digital twin model, acquiring measured data through non-destructive sampling, incrementally training and correcting the model, and iteratively evaluating, the system achieves accurate dynamic assessment of soil carbon storage in southern grasslands. First, based on multi-source satellite remote sensing data of the target area, a topographic map including terrain features is drawn. The remote sensing data is input into a pre-trained initial soil carbon storage inversion model, outputting an initial soil carbon storage distribution map and a deterministic heat map. Then, uncertain areas are extracted according to a preset certainty threshold, and initial sampling points are generated using a spatial random sampling algorithm. Points inaccessible to the robot are filtered out using the topographic map to determine supplementary sampling points. Next, the parameter information of the exploration robot is obtained, and a digital twin model is constructed using the topographic map to simulate the robot's reachable range and sampling time and energy consumption. The optimal supplementary sampling scheme is configured, and the exploration robot is controlled to carry a contact near-infrared spectral probe for non-destructive sampling to obtain supplementary soil carbon storage data. Finally, the supplementary data is transmitted back, and the measured organic carbon data and corresponding remote sensing data are used to form new training samples. The initial model is incrementally trained to obtain a corrected model. The remote sensing data is reprocessed to generate a corrected distribution map and heat map. The overall determinism or resource consumption is evaluated. If the criteria are met, the results are output; otherwise, supplementary sampling and model correction are repeated.

[0008] The technical solution proposed in this application solves the problems of poor environmental adaptability of UAVs, lack of targeted sampling in uncertain areas, and low model correction efficiency in traditional soil carbon storage assessment. It realizes the accuracy, efficiency, and ecological nature of soil carbon storage assessment in southern grasslands, and provides technical support for carbon sink research, ecological protection, and dynamic monitoring of carbon storage in southern grasslands. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart illustrating the dynamic assessment method for soil carbon storage in southern grasslands based on integrated remote sensing monitoring, provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of the dynamic assessment system for soil carbon storage in southern grasslands that integrates remote sensing monitoring, provided in an embodiment of this application.

[0011] The components represented by each number in the attached diagram are explained below: Initial inversion plotting module 01, uncertain region sampling configuration module 02, digital twin sampling module 03, model correction and judgment module 04. Detailed Implementation

[0012] This application provides a method and system for dynamic assessment of soil carbon storage in southern grasslands that integrates remote sensing monitoring. It addresses the technical problems in existing technologies, such as poor environmental adaptability of UAV sampling, lack of targeted sampling in "uncertain areas," and inefficient grid-based or low-precision pure remote sensing sampling, which leads to low efficiency in soil carbon storage data correction.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0015] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0016] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring. The method includes the following steps: S110: Based on multi-source satellite remote sensing data of the target area, draw a topographic map of the target area, and input the multi-source satellite remote sensing data of the target area into the initial soil carbon storage inversion model, and output the initial soil carbon storage distribution map and deterministic heat map of the target area, wherein the deterministic heat map is composed of multiple pixels labeled with determinism; In this embodiment of the application, in the scenario of dynamically assessing the carbon storage of grassland soil in the south, in order to obtain the basic topographic information and preliminary carbon storage distribution of the target area, and at the same time to clarify the credibility of the model's carbon storage prediction results, it is necessary to carry out topographic mapping and model calling based on multi-source satellite remote sensing data of the target area, so as to provide an initial basis for subsequent targeted supplementary sampling and model correction.

[0017] Specifically, firstly, a topographic map containing terrain features is generated based on multi-source satellite remote sensing data of the target area. The generation process can be achieved through existing remote sensing data direct mapping technology to ensure that the topographic map accurately reflects the terrain conditions of the target area.

[0018] Furthermore, by retrieving the pre-trained initial soil carbon storage inversion model and inputting multi-source satellite remote sensing data of the target area into the model, the model can output the initial soil carbon storage distribution map and deterministic heat map of the target area, so as to quickly obtain the preliminary distribution of soil carbon storage in the target area and the reliability of the model prediction results.

[0019] This step, by relying on a pre-trained model to efficiently process multi-source satellite remote sensing data, achieved a preliminary assessment and predictive labeling of soil carbon storage in southern grasslands, providing data support for subsequent accurate identification of uncertain areas requiring supplementary sampling and the development of scientific sampling plans.

[0020] Step S110 in the method provided in this application embodiment includes: Based on multi-source satellite remote sensing data of the target area, a topographic map of the target area including terrain features is generated; The pre-trained initial soil carbon storage inversion model is retrieved, multi-source satellite remote sensing data of the target area are input, and the initial soil carbon storage distribution map and deterministic heat map of the target area are output.

[0021] In this embodiment of the application, in order to establish a dynamic assessment basis for soil carbon storage in southern grasslands and avoid deviations in inversion results caused by single remote sensing data or insufficient model generalization ability, it is necessary to construct a reliable topographic map and a prediction model with self-assessment capabilities in the early stage of assessment, so as to improve the pertinence of subsequent sampling and correction steps and the overall assessment efficiency.

[0022] Specifically, a topographic map containing terrain features is first generated based on multi-source satellite remote sensing data of the target area. The topographic map is generated using existing remote sensing data terrain inversion methods. By extracting terrain-related feature information from multi-source satellite remote sensing data, the topographic details such as grassland undulations, valley distribution, steep slopes, and gentle slopes within the target area are accurately presented, meeting the needs of subsequent planning of sampling paths for exploration robots and selection of accessible sampling points.

[0023] Among them, multi-source satellite remote sensing data includes, but is not limited to, optical images, radar data, multispectral and hyperspectral data. By integrating the advantages of different types of data, comprehensive information on land cover, topographic structure and soil-related characteristics of the target area can be obtained.

[0024] While generating the topographic map, the pre-trained initial soil carbon storage inversion model is retrieved, and multi-source satellite remote sensing data of the target area are input to output the initial soil carbon storage distribution map and deterministic heat map.

[0025] The method provided in this application embodiment for obtaining the "initial soil carbon storage inversion model" includes: Several southern grasslands with similar hydrothermal conditions and vegetation types to the target area were selected as multiple sample grasslands; Historical multi-source satellite remote sensing data of multiple sample grasslands were acquired, as well as historical measured soil organic carbon data corresponding to the historical multi-source satellite remote sensing data. The historical measured soil organic carbon data included at least the sampling depth and soil organic carbon content. The initial soil carbon storage inversion model was built based on machine learning. The initial soil carbon storage inversion model was trained using the historical multi-source satellite remote sensing data and historical measured soil organic carbon data until convergence.

[0026] In this embodiment of the application, in order to enable the initial soil carbon storage inversion model to adapt to the characteristics of southern grasslands and invert soil carbon storage, it is necessary to go through the process of screening similar samples, collecting matching data, building a model framework and completing training, so as to ensure that the model can output reliable initial carbon storage distribution and deterministic assessment results based on multi-source satellite remote sensing data.

[0027] Specifically, several southern grasslands with similar hydrothermal conditions and vegetation types to the target area were first selected as sample grasslands. Among them, the hydrothermal conditions need to refer to key indicators such as the target area's average annual precipitation and average annual temperature, and the vegetation type needs to match the dominant herbaceous plant species in the target area.

[0028] For example, if the target area is a southern grassland with bermudagrass as the main vegetation in a subtropical monsoon climate, the sample grassland should be selected from areas with an average annual precipitation of 800-1500 mm, an average annual temperature of 15-22℃, and also with bermudagrass as the dominant vegetation, to ensure that the ecological environment of the sample grassland is highly similar to that of the target area.

[0029] Furthermore, historical multi-source satellite remote sensing data of the sample grassland, as well as historical measured soil organic carbon data for the corresponding historical periods, were obtained.

[0030] Among them, historical multi-source satellite remote sensing data should cover image information of the sample grassland at different time dimensions to comprehensively capture the dynamic characteristics of the sample grassland such as surface vegetation cover and soil moisture; historical measured soil organic carbon data should be obtained through field sampling and should at least include sampling depth and soil organic carbon content. For example, multiple sampling points should be set up in each sample grassland to collect soil samples from the 0-10cm, 10-20cm, and 20-30cm soil layers, and detect and record the organic carbon content of each soil layer to ensure that the measured data can provide accurate output labels for the model.

[0031] Furthermore, a basic framework for an initial soil carbon storage inversion model is built based on machine learning to establish a precise mapping relationship between "remote sensing data features and soil organic carbon content," providing reliable model support for subsequent carbon storage assessment of target areas.

[0032] Specifically, a random forest algorithm can be used as the basic framework to construct an initial soil carbon storage inversion model. Historical multi-source satellite remote sensing data are used as input features of the model; at the same time, historical measured soil organic carbon content is used as the output target of the model. By leveraging the multi-decision tree ensemble learning characteristics of the random forest algorithm, the advantages of different types of satellite remote sensing data can be effectively integrated.

[0033] Furthermore, the initial soil carbon storage inversion model was trained using collected historical multi-source satellite remote sensing data and historical measured soil organic carbon data until convergence.

[0034] Before training, the data needs to be preprocessed, namely, radiometric and geometric correction of multi-source satellite remote sensing data to eliminate atmospheric interference and image distortion; and outlier removal of measured data to avoid individual extreme data affecting the model training effect.

[0035] During training, the preprocessed dataset needs to be divided into a training set and a validation set, for example, in a 7:3 ratio, with 70% of the data used for model parameter learning and 30% used for real-time evaluation of the model's generalization ability.

[0036] During training, the training set is first input into the model, and multiple decision trees are constructed by Bootstrap sampling of the random forest algorithm. Each decision tree is split based on a different feature subset to reduce the risk of model overfitting. At the same time, the model fitting effect is optimized by adjusting parameters such as the number of decision trees, the maximum tree depth, and the minimum number of samples for node splitting.

[0037] After each training round, validation set data is input into the model to calculate the model's prediction error. Root mean square error (RMSE) is commonly used as an evaluation metric to intuitively reflect the degree of deviation between the predicted and measured values.

[0038] Furthermore, when the root mean square error of the model on the validation set stabilizes within a preset threshold and the error does not decrease significantly after 5-10 consecutive training rounds, the model training is considered to have converged. At this point, the model has a stable soil carbon storage inversion capability and can be used as the initial soil carbon storage inversion model for subsequent evaluation of the target area.

[0039] Furthermore, multi-source satellite remote sensing data of the target area are input into the pre-trained initial soil carbon storage inversion model. The model integrates and analyzes information such as vegetation cover reflected in optical images, soil surface structure captured by radar data, and spectral characteristics of soil organic matter contained in multispectral and hyperspectral data. It outputs an initial soil carbon storage distribution map and a deterministic heat map of the target area to present the preliminary spatial distribution pattern of carbon storage and the reliability of the model's prediction results for each area. This provides a core basis for subsequent extraction of uncertain areas and planning of supplementary sampling.

[0040] For example, if the target area is a southern grassland with bermudagrass as the main vegetation under a subtropical monsoon climate, the input of the initial soil carbon storage inversion model is multi-source satellite remote sensing data, including optical images of the area in summer, radar data in the rainy season, and hyperspectral data in the growing season.

[0041] After analysis, the model outputs an initial soil carbon storage distribution map that marks the approximate range of soil organic carbon content in different regions. The deterministic heat map uses different colors to indicate the degree of certainty, so as to clearly distinguish between credible and doubtful areas predicted by the model.

[0042] S120: Based on the initial soil carbon storage distribution map and the deterministic heat map, extract uncertain areas according to a preset certainty threshold, configure several supplementary sampling points for the uncertain areas according to preset division rules, and mark several supplementary sampling points on the topographic map; In this embodiment of the application, in order to accurately locate areas with low model prediction reliability, it is necessary to screen out uncertain areas that need to be verified in a key manner based on the initial carbon storage distribution and the model deterministic assessment results, and configure reasonable sampling points so as to obtain reliable measured data through targeted supplementary sampling, laying the foundation for subsequent model correction and improvement of carbon storage assessment accuracy.

[0043] Specifically, a certainty threshold is first preset. The setting of this threshold needs to be determined comprehensively based on the accuracy requirements of the evaluation task and the complexity of the terrain and vegetation in the target area, so as to ensure that the model prediction results can be accurately distinguished between credible and questionable areas.

[0044] Furthermore, pixel regions in the deterministic heatmap that do not meet the preset certainty threshold are classified as uncertain regions. This step allows for the identification of areas in the overall region where the model has low prediction confidence due to factors such as insufficient training samples, complex terrain, or blurred remote sensing signals, thus clarifying the key areas for subsequent sampling.

[0045] Furthermore, a spatial random sampling algorithm is employed to generate initial sampling points with a preset density in the uncertain region. The preset density needs to be set based on the size of the uncertain region, the complexity of the terrain, and the accuracy requirements of the evaluation task. Spatial random sampling ensures that the initial sampling points are evenly distributed within the uncertain region, thus comprehensively covering different sub-regions of the area.

[0046] Finally, the generated topographic map is retrieved, and the sampling points located in terrain inaccessible to the exploration robot are filtered out from the initial sampling points. The remaining points are several supplementary sampling points, which are then marked on the topographic map to provide guidance for the exploration robot to plan its sampling path.

[0047] Step S120 in the method provided in this application embodiment includes: Preset certainty threshold; Pixel regions in the deterministic heatmap that do not meet the determinism threshold are classified as uncertain regions. A spatial random sampling algorithm is used to generate initial sampling points with a preset density in uncertain regions; The topographic map is retrieved, and the sampling points in the initial sampling points that are located in terrain inaccessible to the exploration robot are filtered out. The remaining points are the supplementary sampling points.

[0048] In this embodiment of the application, in order to accurately identify areas with low model prediction reliability that require key verification, it is necessary to obtain supplementary sampling points that can be used to specifically calibrate the model by clarifying the reliability threshold, dividing the uncertain area, generating initial sampling points and then screening them, so as to improve the effectiveness of subsequent sampling data and the accuracy of model correction.

[0049] Specifically, a certainty threshold is first preset. The setting of this certainty threshold needs to be determined comprehensively based on the assessment accuracy requirements of the target area, the complexity of the terrain, and the characteristics of vegetation cover. For example, if the assessment task requires high accuracy of carbon storage data, the certainty threshold can be set to 80%, while if only a preliminary assessment of trends is required, it can be appropriately reduced to 70%.

[0050] At the same time, it is necessary to avoid excessively high thresholds that lead to large uncertain areas and a surge in sampling costs, or excessively low thresholds that result in low-confidence areas not being identified, so as to ensure that the credible areas and suspicious areas of the model prediction results can be accurately distinguished.

[0051] Furthermore, pixel regions in the obtained deterministic heatmap that do not meet the certainty threshold are classified as uncertain regions. These regions that do not meet the threshold are often areas with insufficient model training samples, complex terrain, mixed vegetation, or blurred remote sensing signals. Classifying these regions as uncertain regions allows subsequent sampling work to focus on the areas with the highest value for model calibration, avoiding invalid sampling of areas that the model can already reliably predict, and improving the targeting of sampling.

[0052] Furthermore, a spatial random sampling algorithm is used to generate initial sampling points with a preset density in uncertain regions.

[0053] The preset density needs to be set according to the size of the uncertain area, the heterogeneity of the terrain, and the accuracy requirements of the assessment task. For example, for an uncertain area with an area of ​​100 square kilometers and a relatively complex terrain, a density of 5 sampling points per square kilometer can be set to ensure that the initial sampling points can evenly cover different sub-regions within the area.

[0054] In addition, spatial random sampling algorithms can avoid excessive concentration or dispersion of sampling points, ensuring that the subsequent sampling data are representative and can comprehensively reflect the true situation of carbon storage in uncertain areas.

[0055] Finally, the topographic map containing terrain features generated from multi-source satellite remote sensing data is retrieved, and the sampling points in the initial sampling points that are located in terrain inaccessible to the exploration robot are filtered out. The remaining points are several supplementary sampling points.

[0056] Among them, the detection robot is limited by the chassis structure (such as a lightweight tracked or six-legged wheeled chassis), making it difficult to travel on terrains such as steep slopes (slopes exceeding 30°), deep ditches (depths exceeding 0.5 meters), or swamps. By screening the terrain map, these inaccessible points can be eliminated in advance, avoiding the robot being unable to complete the task due to terrain obstacles during subsequent sampling, and ensuring that the supplementary sampling points are actually operable.

[0057] Meanwhile, the selected supplementary sampling points not only focus on uncertain areas but also adapt to robot accessibility, laying the foundation for efficient non-destructive sampling and obtaining high-quality supplementary carbon storage data.

[0058] S130: Obtain parameter information of all detection robots in the target area, generate a digital twin model based on the parameter information of each detection robot and the topographic map of the target area, configure a supplementary sampling scheme based on the digital twin model, perform non-destructive sampling on several supplementary sampling points, and obtain several supplementary soil carbon storage data. In this embodiment of the application, in order to avoid low sampling efficiency or failure to complete the sampling task due to unclear parameters or unreasonable path planning of the detection robot, and to ensure accurate acquisition of soil carbon storage data of supplementary sampling points, it is necessary to configure a scientific sampling scheme and perform non-destructive sampling by acquiring robot parameters and constructing a digital twin model, so as to efficiently obtain reliable supplementary data to support the correction of the initial soil carbon storage inversion model and ensure the accuracy of the assessment.

[0059] Specifically, the first step is to acquire parameter information for all detection robots in the target area. This information includes at least initial position, battery level, movement speed, obstacle-crossing performance, and the status of core sensors. By comprehensively understanding the robots' basic parameters, the operational capabilities and limitations of each robot can be clearly defined, providing a foundation for subsequent solution configuration.

[0060] Furthermore, based on the parameter information of all the exploration robots and the previously generated terrain map containing terrain features, a digital twin model is created. This digital twin model can simulate the reachable range of each exploration robot, clearly present the terrain areas that the robots cannot pass through, and also calculate the sampling time and energy consumption of the robot from the initial position to each supplementary sampling point.

[0061] Furthermore, the locations of several supplementary sampling points marked on the topographic map, along with the initial positions of all exploration robots, are imported into a digital twin model, and sampling is simulated based on the model. The simulation provides a clear view of how different robots perform sampling tasks, allowing for the selection of the optimal supplementary sampling scheme that offers the shortest total time, lowest total energy consumption, and coverage of all supplementary sampling points.

[0062] Finally, following the optimal supplementary sampling plan, the probe robot was controlled to conduct non-destructive sampling at several supplementary sampling points. The supplementary soil carbon storage data obtained in this way ensures the authenticity and validity of the data while avoiding soil damage, and also meets the needs of grassland ecological protection in southern China.

[0063] Step S130 in the method provided in this application embodiment includes: Acquire parameter information of all detection robots in the target area, wherein the parameter information includes at least initial position, battery level, movement speed, obstacle crossing performance, and core sensor status; Based on the parameter information of all exploration robots and the terrain map, a digital twin model is created. The digital twin model can be used to simulate the reach of each exploration robot and the sampling time and energy consumption from the initial position to each supplementary sampling point. The digital twin model is used to import the locations of several supplementary sampling points marked on the topographic map and the initial positions of all exploration robots. Based on the digital twin model, simulated sampling is performed to obtain the optimal supplementary sampling scheme. Non-destructive sampling is then performed on several supplementary sampling points to obtain several supplementary soil carbon storage data.

[0064] In this embodiment of the application, in order to avoid low sampling efficiency of the detection robot due to unclear operating parameters and unreasonable sampling path planning, it is necessary to go through a process of fully understanding the robot parameters, building a digital twin model to optimize the sampling scheme, and adopting non-destructive sampling technology, so as to balance sampling efficiency, ecological protection and data validity and promote the orderly development of the assessment work.

[0065] Specifically, the first step is to acquire parameter information for all detection robots in the target area. This parameter information includes at least the initial position, battery level, movement speed, obstacle crossing performance, and the status of core sensors.

[0066] Among these factors, the initial position clearly defines the starting coordinates of the robot's operation, providing a starting point for subsequent path planning; the battery level directly relates to the robot's endurance, determining the sampling range it can cover in a single run; movement speed is a key indicator for calculating sampling time, affecting overall sampling efficiency; obstacle-crossing performance needs to be adapted to the rugged terrain of southern grasslands to determine whether the robot can reach specific sampling points; and the status of core sensors (i.e., whether they are available or unavailable) directly determines whether the robot can normally acquire soil-related data, avoiding invalid sampling data due to sensor malfunctions.

[0067] Furthermore, a digital twin model is created based on the parameter information of all the exploration robots and the previously generated topographic map containing terrain features.

[0068] Specifically, this digital twin model can fuse the parameters of the exploration robot with terrain information. On the one hand, it can simulate the reach of each exploration robot and clearly mark the areas that the robot cannot pass due to terrain limitations, thus avoiding the occurrence of unexecutable points in subsequent sampling schemes.

[0069] On the other hand, it can accurately calculate the sampling time and energy consumption of each robot from its initial position to each supplementary sampling point. For example, it can calculate the time based on the robot's movement speed and the terrain distance between the sampling points, and calculate the total energy consumption by combining the energy loss when crossing obstacles, providing a quantitative basis for subsequent optimization of the sampling scheme.

[0070] Furthermore, the locations of several supplementary sampling points marked on the topographic map and the initial positions of all exploration robots are imported into the constructed digital twin model, and simulation sampling is performed based on the model to obtain the optimal supplementary sampling scheme.

[0071] The method provided in this application embodiment, which involves "importing the marked locations of several supplementary sampling points on the topographic map and the initial positions of all exploration robots into the digital twin model, performing simulated sampling based on the digital twin model to obtain the optimal supplementary sampling scheme, performing non-destructive sampling on several supplementary sampling points, and obtaining several supplementary soil carbon storage data," includes: For each supplementary sampling point, the movement cost from the current position of any probe robot to the supplementary sampling point is calculated by combining the terrain features in the terrain map and the parameter information of the probe robot. The movement cost is a weighted function of the combined sampling time and sampling energy consumption. With the total sampling time and total sampling energy consumption as optimization objectives, a dedicated supplementary sampling point sequence is collaboratively allocated to all detection robots; For each probe robot, the optimal movement path is planned by combining the supplementary sampling point sequence with the topographic map; Each detection robot is controlled to sample along the optimal movement path. When an obstacle is encountered, the sequence of supplementary sampling points and the movement path are dynamically adjusted to obtain the optimal supplementary sampling scheme.

[0072] Specifically, for each supplementary sampling point, the movement cost from the current position of any probe robot to that supplementary sampling point is calculated by combining the terrain features in the topographic map and the parameter information of the probe robot.

[0073] Among the topographic features in the topographic map, special attention should be paid to the rugged terrain of the southern grasslands, such as the slope, the presence of valleys or protrusions, etc., as these features directly affect the robot's traversal difficulty and energy consumption. The movement cost, as a weighted function of the combined sampling time and sampling energy consumption, can have its weight adjusted according to the needs of the evaluation task. If rapid sampling is required, the weight of sampling time can be increased; if energy consumption needs to be controlled, the weight of sampling energy consumption can be increased. This calculation quantifies the cost for each probe robot to reach a sampling point, providing a basis for subsequent task allocation.

[0074] Furthermore, with the total sampling time and total sampling energy consumption as optimization objectives, a dedicated supplementary sampling point sequence is collaboratively allocated to all detection robots.

[0075] During the allocation process, it is necessary to simulate the total time and energy consumption under different allocation schemes based on the digital twin model to avoid an imbalance where some detection robots are overly concentrated in certain tasks and some robots are idle, while ensuring that all supplementary sampling points can be covered.

[0076] For example, a detection robot with an initial position close to a certain cluster of sampling points and sufficient battery power can be assigned to the sampling task in that area. A robot with high speed but average obstacle-crossing ability can be assigned to the sampling of flat areas, while a robot with strong obstacle-crossing ability but slow speed can be assigned to the sampling of areas with complex terrain. The total sampling cost can be minimized through collaborative allocation.

[0077] Furthermore, for each probe robot, an optimal movement path is planned based on the supplementary sampling point sequence assigned to it, combined with the topographic map.

[0078] Specifically, the planning should prioritize routes with gentle terrain and no obvious obstacles, avoiding steep slopes in southern grasslands that exceed the obstacle-crossing capabilities of the exploration robot, as well as deep valleys, to reduce energy consumption and malfunction risks during the robot's journey. At the same time, the round-trip distance between sampling points should be minimized to avoid increased time consumption due to detours.

[0079] For example, if the sampling point sequence of a certain detection robot is concentrated in the eastern part of the target area, and there is a gentle natural passage in the area, the robot can be prioritized to travel along the passage and arrive at each sampling point in sequence during path planning, so as to improve the passage efficiency.

[0080] Finally, each probe robot is controlled to sample along the optimal movement path. When encountering obstacles, the sequence of supplementary sampling points and the movement path are dynamically adjusted to obtain the best supplementary sampling scheme.

[0081] In the southern grasslands, there may be temporary obstacles not marked on the topographic map, such as fallen vegetation or temporary water puddles. When the robot encounters such obstacles and cannot proceed along the original path, it needs to use a digital twin model to provide real-time feedback on the obstacle information and quickly recalculate the movement cost based on the model. This allows for adjustments to the robot's supplementary sampling point sequence (e.g., prioritizing sampling points that are accessible around the obstacle, and assisting in sampling points behind the obstacle after other robots have completed their tasks) or replanning the movement path (e.g., going around to the other side of the obstacle).

[0082] In the method provided in this application embodiment, the detection robot is equipped with a contact near-infrared spectral probe, which can obtain the soil's supplementary soil carbon storage data without drilling.

[0083] Specifically, all the exploration robots use non-destructive sampling methods, are equipped with high-precision contact near-infrared spectral probes, and use lightweight tracked or multi-legged chassis to adapt to rugged terrain. After traveling to the supplementary sampling point, they closely attach the spectral probes to the ground surface, and can quickly obtain soil spectral data and calculate soil organic carbon content without drilling. This avoids the damage to grassland ecology caused by traditional drilling sampling and solves the problem that drones cannot sample in bad weather or complex terrain.

[0084] After the detection robot travels to the supplementary sampling point, the robotic arm adjusts the angle and height according to the preset program to smoothly lower the high-precision contact near-infrared spectral probe to the ground surface, ensuring that the probe is in close contact with the soil surface without gaps, thus avoiding air interference that could cause spectral data distortion.

[0085] Meanwhile, after the probe is attached, it will quickly emit near-infrared light and receive the spectral signal reflected by the soil. Simultaneously, the spectral data will be transmitted to the processor built into the detection robot. The processor will call the pre-calibrated soil organic carbon inversion model and calculate the supplementary soil organic carbon content at the supplementary sampling point by analyzing the characteristic absorption peaks in the spectral data.

[0086] In addition, after completing data acquisition at each supplementary sampling point, the probe automatically records the coordinates of that point, the sampling time, and the corresponding supplementary soil organic carbon content, forming a complete sampling record. Simultaneously, the probe transmits this data back to the cloud in real time via a wireless communication module, facilitating subsequent integration and model correction.

[0087] S140: The supplementary soil carbon storage data are fed back to the initial soil carbon storage inversion model to generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is determined to be completed. Otherwise, supplementary sampling is continued to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map.

[0088] In this embodiment of the application, in order to optimize the model inversion capability and improve the accuracy of carbon storage assessment results using measured sampling data, and at the same time avoid resource waste caused by unlimited sampling, it is necessary to send the supplementary soil carbon storage data back to the initial soil carbon storage inversion model to complete the correction and iterative assessment until the accuracy requirements are met or the resource consumption limit is reached, so as to output reliable dynamic assessment results of soil carbon storage.

[0089] Specifically, the obtained supplementary soil carbon storage data are first fed back into the initial soil carbon storage inversion model. That is, measured soil organic carbon data are extracted from the supplementary soil carbon storage data and combined with the corresponding multi-source satellite remote sensing data to form a new training sample. These training samples are used to incrementally train the initial soil carbon storage inversion model, resulting in a corrected soil carbon storage inversion model. This allows the model to optimize parameters based on the actual conditions of the target area, reducing prediction bias.

[0090] Furthermore, the multi-source satellite remote sensing data of the target area are reprocessed using the corrected model to generate a corrected soil carbon storage distribution map and a corrected deterministic heat map. The corrected soil carbon storage distribution map is closer to the actual carbon storage distribution, and the corrected deterministic heat map can update the certainty of each region to reflect the reliability of the prediction after the model is corrected.

[0091] Furthermore, the deterministic heatmap is evaluated and corrected. If the determinism of the entire domain reaches the determinism threshold, it means that the model prediction accuracy meets the standard. If it does not meet the standard but the task resource consumption (including total sampling time and total sampling energy consumption) has reached the preset upper limit, the cost needs to be weighed to determine whether the task is completed.

[0092] Conversely, if the certainty is not met and the resources are not exceeded, it is necessary to return to the supplementary sampling stage to obtain new supplementary data to revise the model and update the carbon storage distribution map again, and repeat the steps of "supplementary sampling - model revision - thermal map evaluation".

[0093] When the task is completed, the final corrected soil carbon storage distribution map will be used as the dynamic evaluation result. This result combines the advantages of the wide coverage of remote sensing data with the calibration value of measured data, and can serve as a reliable basis for the study of soil carbon storage in southern grasslands.

[0094] Step S140 in the method provided in this application embodiment includes: The measured soil organic carbon data in the supplemented soil carbon storage data and the corresponding multi-source satellite remote sensing data are used to form a new training sample. The initial soil carbon storage inversion model is incrementally trained to obtain the corrected soil carbon storage inversion model. The multi-source satellite remote sensing data of the target area are reprocessed using the modified soil carbon storage inversion model to generate a modified soil carbon storage distribution map and a modified deterministic heat map. The modified deterministic heatmap is evaluated. If its global determinism reaches the determinism threshold, or the task resource consumption has reached the preset upper limit, the task is judged to be completed, and the final modified soil carbon storage distribution map is output. The task resource consumption includes the total sampling time and the total sampling energy consumption. Otherwise, repeat the steps of supplementary sampling – revising the soil carbon storage inversion model – obtaining the revised deterministic heat map – evaluating the revised deterministic heat map until the task is completed; The final output of the corrected soil carbon storage distribution map will be used as the result of the dynamic assessment of soil carbon storage.

[0095] In this embodiment of the application, in order to continuously optimize the inversion accuracy of the initial soil carbon storage inversion model through measured data, and at the same time ensure that the assessment task is completed efficiently within a reasonable resource consumption range, it is necessary to combine the measured data obtained by supplementary sampling with remote sensing data to correct the model and update the assessment results. The task is closed through iterative iteration and threshold judgment to output accurate and practical dynamic assessment results of soil carbon storage in southern grasslands.

[0096] Specifically, the measured soil organic carbon data from the supplementary soil carbon storage data are first combined with the corresponding multi-source satellite remote sensing data to form a new training sample, and the initial soil carbon storage inversion model is incrementally trained.

[0097] Among them, the measured soil organic carbon data in the supplementary soil carbon storage data directly reflects the actual carbon storage situation in the target area, while the corresponding multi-source satellite remote sensing data contains the surface feature information of the measured point. The new training sample formed by the combination of the two can specifically make up for the limitation of the initial soil carbon storage inversion model relying only on historical sample grassland data.

[0098] Furthermore, incremental training does not require reconstructing the model framework; it only requires adjusting the model's internal parameters based on new samples. This not only preserves the learning results of the initial soil carbon storage inversion model but also allows the model to quickly adapt to the carbon storage distribution characteristics of the target area, ultimately resulting in a corrected soil carbon storage inversion model, which further enhances the model's inversion capability for the target area.

[0099] Furthermore, after obtaining the corrected soil carbon storage inversion model, the multi-source satellite remote sensing data of the target area are reprocessed using the model to generate a corrected soil carbon storage distribution map and a corrected deterministic heat map.

[0100] Specifically, the revised soil carbon storage distribution map, compared to the initial distribution map, can more accurately present the spatial differences in soil carbon storage in the target area. For example, an area that was originally predicted as "10-12 g / kg" by the initial soil carbon storage inversion model may be more in line with the actual "11-13 g / kg" after revision. The revised deterministic heat map will update the certainty of each pixel. After the model is calibrated with measured data, the prediction accuracy of some areas with low certainty will be improved, and the certainty value will be increased accordingly. Areas with prediction bias will remain with low certainty, providing an accuracy reference for subsequent assessment.

[0101] Furthermore, the deterministic heatmap is evaluated and revised to determine whether the dynamic assessment task has been completed. The evaluation focuses on two core conditions: first, whether the determinism of the entire target area has reached the preset determinism threshold. If the preset determinism threshold is reached, it indicates that the model's prediction of carbon storage in each region of the target area meets the accuracy requirements, and no further sampling is needed. Second, whether the task's resource consumption has reached the preset upper limit.

[0102] The task resource consumption includes total sampling time and total sampling energy consumption. If the resource consumption has reached the upper limit, even if the certainty of some areas is still not met, the task must be completed based on a trade-off between cost and efficiency in order to avoid resource waste caused by unlimited sampling.

[0103] Conversely, if the corrected deterministic heatmap shows that the overall determinism does not reach the preset determinism threshold, and the task resource consumption has not exceeded the upper limit, then the steps of "supplementary sampling - correcting the soil carbon storage inversion model - obtaining the corrected deterministic heatmap - evaluating the corrected deterministic heatmap" need to be executed repeatedly. That is, returning to the supplementary sampling stage, for the uncertain areas in the corrected deterministic heatmap that still do not reach the threshold, supplementary sampling points are reconfigured, and new supplementary soil carbon storage data is obtained through the detection robot. Then, the model correction and result update process is repeated with the new data until the task completion conditions are met.

[0104] Finally, when the task is deemed complete, the final output of the corrected soil carbon storage distribution map will be used as the dynamic assessment result of soil carbon storage. This result integrates the advantages of the wide coverage of multi-source satellite remote sensing data with the precise calibration value of the data measured by the detection robot. It can clearly reflect the overall distribution trend of soil carbon storage in the target area, and accurately present the details of carbon storage in local areas. It can serve as reliable data support for research on carbon sinks in southern grasslands and ecological protection planning.

[0105] The embodiments of this application, through the specific implementation methods described above, achieve the following technical effects: This application proposes a dynamic assessment method for soil carbon storage in southern grasslands based on integrated remote sensing monitoring. First, a topographic map incorporating terrain features is drawn based on multi-source satellite remote sensing data of the target area. Simultaneously, a pre-trained initial soil carbon storage inversion model is retrieved, and the remote sensing data is input into the model to output an initial soil carbon storage distribution map and a deterministic heat map. Next, uncertain regions are extracted from the deterministic heat map according to a preset certainty threshold. A spatial random sampling algorithm is used to generate initial sampling points in these uncertain regions. Points inaccessible to the exploration robots are filtered using the topographic map, and supplementary sampling points are identified and marked. Finally, the parameter information of all exploration robots is acquired. A digital twin model is constructed by combining topographic maps to simulate the reach of the exploration robot and the sampling time and energy consumption. Supplementary sampling points and the robot's initial position are imported to configure the optimal sampling scheme. The robot is controlled to carry a contact near-infrared spectral probe for non-destructive sampling to obtain supplementary soil carbon storage data. Finally, the supplementary data is transmitted back, and the measured organic carbon data and corresponding remote sensing data are used to form new training samples. The initial model is incrementally trained to obtain a corrected model. The remote sensing data is reprocessed to generate a corrected soil carbon storage distribution map and a deterministic heat map. The deterministic nature or resource consumption of the whole area is evaluated. If the criteria are met, the results are output. Otherwise, supplementary sampling and model correction are carried out in a loop.

[0106] The method provided in this application, through the technical solution of "initial inversion and deterministic labeling - precise sampling of uncertain areas - digital twin optimized sampling - non-destructive data acquisition - model iterative correction", solves the problems of poor environmental adaptability of UAVs, lack of targeted sampling, and low model correction efficiency in traditional soil carbon storage assessment. It realizes the accuracy, efficiency and ecological nature of soil carbon storage assessment in southern grasslands, reduces invalid sampling and resource waste, and provides technical support for carbon sink research, ecological protection and dynamic monitoring of carbon storage in southern grasslands.

[0107] Example 2, as shown in the appendix Figure 2 As shown, based on the inventive concept of the method for dynamic assessment of soil carbon storage in southern grasslands using integrated remote sensing monitoring provided in Embodiment 1, this application also provides a system for dynamic assessment of soil carbon storage in southern grasslands using integrated remote sensing monitoring, specifically including: The initial inversion mapping module 01 is used to draw a topographic map of the target area based on multi-source satellite remote sensing data of the target area, and input the multi-source satellite remote sensing data of the target area into the initial soil carbon storage inversion model, and output the initial soil carbon storage distribution map and deterministic heat map of the target area, wherein the deterministic heat map is composed of multiple pixels labeled with determinism; The uncertain area sampling configuration module 02 is used to extract uncertain areas based on the initial soil carbon storage distribution map and the deterministic heat map according to a preset certainty threshold, and to configure several supplementary sampling points for the uncertain areas according to preset division rules, and to mark several supplementary sampling points on the topographic map; The digital twin sampling module 03 is used to acquire parameter information of all detection robots in the target area, generate a digital twin model based on the parameter information of each detection robot and the topographic map of the target area, configure a supplementary sampling scheme based on the digital twin model, perform non-destructive sampling on several supplementary sampling points, and obtain several supplementary soil carbon storage data. The model correction and determination module 04 is used to transmit the supplementary soil carbon storage data back to the initial soil carbon storage inversion model, generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is determined to be completed. Otherwise, supplementary sampling continues to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map.

[0108] In one embodiment, the initial inversion plotting module 01 is further configured to: Based on multi-source satellite remote sensing data of the target area, a topographic map of the target area including terrain features is generated; a pre-trained initial soil carbon storage inversion model is retrieved, multi-source satellite remote sensing data of the target area is input, and the initial soil carbon storage distribution map and deterministic heat map of the target area are output.

[0109] Furthermore, the initial inversion plotting module 01 also includes: Several southern grasslands with similar hydrothermal conditions and vegetation types to the target area were selected as multiple sample grasslands. Historical multi-source satellite remote sensing data of the multiple sample grasslands and historical measured soil organic carbon data corresponding to the historical multi-source satellite remote sensing data were acquired. The historical measured soil organic carbon data included at least sampling depth and soil organic carbon content. The initial soil carbon storage inversion model was built based on machine learning. The initial soil carbon storage inversion model was trained using the historical multi-source satellite remote sensing data and the historical measured soil organic carbon data until convergence.

[0110] In one embodiment, the uncertain region sampling configuration module 02 is further configured to: A predetermined certainty threshold is set; pixel regions in the deterministic heatmap that do not meet the certainty threshold are divided into uncertain regions; a spatial random sampling algorithm is used to generate initial sampling points with a predetermined density in the uncertain regions; the terrain map is retrieved, and sampling points located in terrain inaccessible to the exploration robot are filtered out from the initial sampling points, and the remaining points are the supplementary sampling points.

[0111] In one embodiment, the digital twin sampling module 03 is also used for: Acquire parameter information for all exploration robots in the target area, including at least initial position, battery level, movement speed, obstacle crossing performance, and core sensor status. Based on the parameter information of all exploration robots and the topographic map, create a digital twin model. The digital twin model can be used to simulate the reachable range of each exploration robot and the sampling time and energy consumption from the initial position to each supplementary sampling point. Import the marked locations of several supplementary sampling points on the topographic map and the initial positions of all exploration robots into the digital twin model. Perform simulated sampling based on the digital twin model to obtain the optimal supplementary sampling scheme. Perform non-destructive sampling on several supplementary sampling points to obtain several supplementary soil carbon storage data.

[0112] Furthermore, the digital twin sampling module 03 also includes: For each supplementary sampling point, the movement cost from the current position of any probe robot to that supplementary sampling point is calculated by combining the terrain features in the topographic map and the parameter information of the probe robot. The movement cost is a weighted function of the comprehensive sampling time and sampling energy consumption. Taking the total sampling time and total sampling energy consumption as the optimization objective, a dedicated supplementary sampling point sequence is collaboratively allocated to all probe robots. For the supplementary sampling point sequence allocated to each probe robot, its optimal movement path is planned in combination with the topographic map. Each probe robot is controlled to sample according to the optimal movement path. When encountering obstacles, the supplementary sampling point sequence and movement path are dynamically adjusted to obtain the optimal supplementary sampling scheme.

[0113] Furthermore, the digital twin sampling module 03 also includes: The detection robot is equipped with a contact-type near-infrared spectral probe, which can obtain data on the replenishment of soil carbon storage without drilling.

[0114] In one embodiment, the model correction determination module 04 is further configured to: The measured soil organic carbon data from the supplemented soil carbon storage data, along with the corresponding multi-source satellite remote sensing data, form a new training sample. This sample is used to incrementally train the initial soil carbon storage inversion model, resulting in a corrected soil carbon storage inversion model. The corrected model is then used to reprocess the multi-source satellite remote sensing data of the target area, generating a corrected soil carbon storage distribution map and a corrected deterministic heatmap. The corrected deterministic heatmap is evaluated. If its global determinism reaches a certainty threshold, or if the task resource consumption reaches a preset upper limit, the task is considered complete, and the final corrected soil carbon storage distribution map is output. The task resource consumption includes total sampling time and total sampling energy consumption. Otherwise, the steps of supplemented sampling – correcting the soil carbon storage inversion model – obtaining the corrected deterministic heatmap – evaluating the corrected deterministic heatmap are repeated until the task is completed. The final output corrected soil carbon storage distribution map is used as the dynamic assessment result of soil carbon storage.

[0115] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0116] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0117] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for dynamic assessment of soil carbon storage in southern grasslands integrating remote sensing monitoring, characterized in that, include: Based on multi-source satellite remote sensing data of the target area, a topographic map of the target area is drawn, and the multi-source satellite remote sensing data of the target area is input into the initial soil carbon storage inversion model to output the initial soil carbon storage distribution map and deterministic heat map of the target area. The deterministic heat map consists of multiple pixels labeled with determinism. Based on the initial soil carbon storage distribution map and the deterministic heat map, uncertain regions are extracted according to a preset certainty threshold, and several supplementary sampling points are configured for the uncertain regions according to preset division rules, and several supplementary sampling points are marked on the topographic map. Obtain parameter information of all detection robots in the target area, generate a digital twin model based on the parameter information of each detection robot and the topographic map of the target area, configure a supplementary sampling scheme based on the digital twin model, perform non-destructive sampling on several supplementary sampling points, and obtain several supplementary soil carbon storage data. The supplementary soil carbon storage data are fed back to the initial soil carbon storage inversion model to generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is deemed to be completed. Otherwise, supplementary sampling continues to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map.

2. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 1, characterized in that, Based on multi-source satellite remote sensing data of the target area, a topographic map of the target area is drawn. The multi-source satellite remote sensing data of the target area is then input into the initial soil carbon storage inversion model, outputting an initial soil carbon storage distribution map and a deterministic heat map of the target area, including: Based on multi-source satellite remote sensing data of the target area, a topographic map of the target area including terrain features is generated; The pre-trained initial soil carbon storage inversion model is retrieved, multi-source satellite remote sensing data of the target area are input, and the initial soil carbon storage distribution map and deterministic heat map of the target area are output.

3. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 2, characterized in that, The method for obtaining the initial soil carbon storage inversion model includes: Several southern grasslands with similar hydrothermal conditions and vegetation types to the target area were selected as multiple sample grasslands; Historical multi-source satellite remote sensing data of multiple sample grasslands were acquired, as well as historical measured soil organic carbon data corresponding to the historical multi-source satellite remote sensing data. The historical measured soil organic carbon data included at least the sampling depth and soil organic carbon content. The initial soil carbon storage inversion model was built based on machine learning. The initial soil carbon storage inversion model was trained using the historical multi-source satellite remote sensing data and historical measured soil organic carbon data until convergence.

4. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 1, characterized in that, Based on the initial soil carbon storage distribution map and deterministic heat map, uncertain regions are extracted according to a preset certainty threshold. Then, according to preset division rules, several supplementary sampling points are configured for these uncertain regions, and these supplementary sampling points are marked on the topographic map, including: Preset certainty threshold; Pixel regions in the deterministic heatmap that do not meet the determinism threshold are classified as uncertain regions. A spatial random sampling algorithm is used to generate initial sampling points with a preset density in uncertain regions; The topographic map is retrieved, and the sampling points in the initial sampling points that are located in terrain inaccessible to the exploration robot are filtered out. The remaining points are the supplementary sampling points.

5. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 1, characterized in that, Obtain parameter information for all exploration robots in the target area. Generate a digital twin model based on the parameter information of each robot and the topographic map of the target area. Configure a supplementary sampling scheme based on the digital twin model, and perform non-destructive sampling at several supplementary sampling points to obtain several supplementary soil carbon storage data, including: Acquire parameter information of all detection robots in the target area, wherein the parameter information includes at least initial position, battery level, movement speed, obstacle crossing performance, and core sensor status; Based on the parameter information of all exploration robots and the terrain map, a digital twin model is created. The digital twin model can be used to simulate the reach of each exploration robot and the sampling time and energy consumption from the initial position to each supplementary sampling point. The digital twin model is used to import the locations of several supplementary sampling points marked on the topographic map and the initial positions of all exploration robots. Based on the digital twin model, simulated sampling is performed to obtain the optimal supplementary sampling scheme. Non-destructive sampling is then performed on several supplementary sampling points to obtain several supplementary soil carbon storage data.

6. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 5, characterized in that, The locations of several supplementary sampling points marked on the topographic map and the initial positions of all exploration robots are imported into the digital twin model. Simulated sampling is performed based on the digital twin model to obtain the optimal supplementary sampling scheme. Non-destructive sampling is then conducted at several supplementary sampling points to obtain supplementary soil carbon storage data, including: For each supplementary sampling point, the movement cost from the current position of any probe robot to the supplementary sampling point is calculated by combining the terrain features in the terrain map and the parameter information of the probe robot. The movement cost is a weighted function of the combined sampling time and sampling energy consumption. With the total sampling time and total sampling energy consumption as optimization objectives, a dedicated supplementary sampling point sequence is collaboratively allocated to all detection robots; For each probe robot, the optimal movement path is planned by combining the supplementary sampling point sequence with the topographic map; Each detection robot is controlled to sample along the optimal movement path. When an obstacle is encountered, the sequence of supplementary sampling points and the movement path are dynamically adjusted to obtain the optimal supplementary sampling scheme.

7. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 5, characterized in that, The detection robot is equipped with a contact-type near-infrared spectral probe, which can obtain data on the replenishment of soil carbon storage without drilling.

8. The method for dynamic assessment of soil carbon storage in southern grasslands based on integrated remote sensing monitoring as described in claim 1, characterized in that, The supplementary soil carbon storage data are fed back to the initial soil carbon storage inversion model to generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is deemed complete; otherwise, supplementary sampling continues to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map, including: The measured soil organic carbon data in the supplemented soil carbon storage data and the corresponding multi-source satellite remote sensing data are used to form a new training sample. The initial soil carbon storage inversion model is incrementally trained to obtain the corrected soil carbon storage inversion model. The multi-source satellite remote sensing data of the target area are reprocessed using the modified soil carbon storage inversion model to generate a modified soil carbon storage distribution map and a modified deterministic heat map. The modified deterministic heatmap is evaluated. If its global determinism reaches the determinism threshold, or the task resource consumption has reached the preset upper limit, the task is judged to be completed, and the final modified soil carbon storage distribution map is output. The task resource consumption includes the total sampling time and the total sampling energy consumption. Otherwise, repeat the steps of supplementary sampling – revising the soil carbon storage inversion model – obtaining the revised deterministic heat map – evaluating the revised deterministic heat map until the task is completed; The final output of the corrected soil carbon storage distribution map will be used as the result of the dynamic assessment of soil carbon storage.

9. A dynamic assessment system for soil carbon storage in southern grasslands integrating remote sensing monitoring, characterized in that, The system is used to execute the dynamic assessment method for soil carbon storage in southern grasslands based on fusion remote sensing monitoring as described in any one of claims 1-8, and the system comprises: The initial inversion mapping module is used to draw a topographic map of the target area based on multi-source satellite remote sensing data of the target area, and input the multi-source satellite remote sensing data of the target area into the initial soil carbon storage inversion model, and output the initial soil carbon storage distribution map and deterministic heat map of the target area, wherein the deterministic heat map is composed of multiple pixels labeled with determinism; The uncertain area sampling configuration module is used to extract uncertain areas based on the initial soil carbon storage distribution map and the deterministic heat map according to a preset certainty threshold, and to configure several supplementary sampling points for the uncertain areas according to preset division rules, and to mark several supplementary sampling points on the topographic map; The digital twin sampling module is used to acquire parameter information of all exploration robots in the target area, generate a digital twin model based on the parameter information of each exploration robot and the topographic map of the target area, configure a supplementary sampling scheme based on the digital twin model, perform non-destructive sampling on several supplementary sampling points, and obtain several supplementary soil carbon storage data. The model correction and determination module is used to transmit the supplementary soil carbon storage data back to the initial soil carbon storage inversion model, generate a corrected soil carbon storage distribution map and a corrected deterministic heat map of the target area. When the determinism of the entire target area reaches the determinism threshold or the predetermined resource consumption limit is reached, the dynamic assessment task is determined to be completed. Otherwise, supplementary sampling continues to correct the initial soil carbon storage inversion model and the initial soil carbon storage distribution map.

Images