Method and system for estimating organic carbon content of soil based on multi-modal data

Abstract

The invention provides a method and system for estimating the organic carbon content of soil based on multi-modal data, and particularly relates to the technical field of forestry carbon sink, and the scheme comprises the following steps: obtaining a multi-modal data set of a target region in a target time period, the multi-modal data set at least comprising remote sensing data, climate data and soil attribute data of the target region; and processing the multi-modal data set by using the trained soil organic carbon dynamic model to obtain a soil organic carbon content estimation result. According to the scheme, the remote sensing data, the climate data and the soil attribute data of the target region are processed by utilizing the trained soil organic carbon dynamic model, high-precision estimation of the SOC content of the large-scale target region can be realized, dynamic monitoring can be realized, and the method has relatively high applicability and generalizability.

Description

Technical Field

[0001] The present invention relates to the field of forestry carbon sink technology, and in particular to a method and system for estimating soil organic carbon content based on multimodal data. Background Art

[0002] Soil organic carbon (SOC) is one of the important indicators for assessing soil health. At present, the assessment of soil organic carbon mainly relies on manual sampling and laboratory analysis of soil-related data in a small area. However, the sampling and laboratory analysis processes are time-consuming and labor-intensive, cannot meet the needs of dynamic monitoring, and cannot efficiently collect and monitor a large range of soil-related data. It can be seen that the existing soil organic carbon assessment methods have the problem of low efficiency in assessing soil organic carbon content and inability to achieve dynamic monitoring. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the object of the present invention is to provide a method and system for estimating soil organic carbon content based on multimodal data, aiming to solve the problems in the prior art of low efficiency in assessing soil organic carbon content and inability to achieve dynamic monitoring.

[0004] In order to achieve the above object, the present invention provides a method for estimating soil organic carbon content based on multimodal data in a first aspect, comprising:

[0005] Acquire a multimodal data set of a target area within a target time period, wherein the multimodal data set includes at least remote sensing data, climate data, and soil property data of the target area;

[0006] The multimodal data set is processed using the trained soil organic carbon dynamic model to obtain an estimation result of the soil organic carbon content.

[0007] Optionally, the multimodal data set is processed using a trained soil organic carbon dynamic model to obtain a soil organic carbon content estimation result, including:

[0008] Based on the climate data, determine the average temperature and average rainfall;

[0009] Based on the remote sensing data, obtaining vegetation coverage, easily decomposable plant materials and resistant to decomposition plant materials;

[0010] determining plant fertilizer input based on the soil property data;

[0011] The average temperature, the average rainfall, the vegetation coverage, the easily decomposable plant materials, the resistant to decomposition plant materials and the plant fertilizer input are used as inputs of the trained soil organic carbon dynamic model to obtain an estimated result of the soil organic carbon content.

[0012] Optionally, the training process of the soil organic carbon dynamic model includes:

[0013] Initializing a soil organic carbon dynamic model of the target area;

[0014] Collect remote sensing data samples, climate data samples, soil property data samples and actual soil organic carbon content of the target area within the simulation period;

[0015] The soil organic carbon dynamic model is iteratively updated using the remote sensing data samples, the climate data samples, the soil property data samples and the actual soil organic carbon content corresponding to the preset time step until the simulation period or the preset estimation accuracy is reached, thereby obtaining a trained soil organic carbon dynamic model.

[0016] Optionally, the iterative updating of the soil organic carbon dynamic model using the remote sensing data samples, the climate data samples, the soil property data samples and the actual soil organic carbon content corresponding to a preset time step includes:

[0017] Calculating the decomposition rate of soil organic carbon during the simulation period using the remote sensing data samples, the climate data samples and the soil property data samples;

[0018] Determining easily decomposable plant material samples and resistant to decomposition plant material samples based on the remote sensing data samples;

[0019] Determining a plant fertilizer input amount sample based on the soil property data sample;

[0020] Obtaining an estimated value of soil organic carbon content based on the easily decomposable plant material sample, the resistant decomposition plant material sample and the plant fertilizer input amount sample corresponding to each time step as inputs to the soil organic carbon dynamic model;

[0021] The soil organic carbon dynamic model is iteratively updated based on the estimated value corresponding to each time step and the actual soil organic carbon content.

[0022] Optionally, the soil organic carbon dynamic model includes an active pool model, a slow pool model and a passive pool model, and the easily decomposable plant material sample, the resistant decomposable plant material sample and the plant fertilizer input amount sample corresponding to each time step are used as inputs of the soil organic carbon dynamic model to obtain an estimated value of the soil organic carbon content, including:

[0023] updating the active reservoir model using the easily decomposable plant material and the decomposition rate;

[0024] updating the slow pool model using the decomposition-resistant plant material, the active pool, and the decomposition rate;

[0025] updating a passive sink model based on the slow sink model, the decomposition rate, and the plant fertilizer input;

[0026] Based on the dynamically updated active pool model, the slow pool model and the passive pool model, an estimated value of soil organic carbon content is obtained.

[0027] Optionally, the calculating the decomposition rate of soil organic carbon by using the remote sensing data, the climate data and the soil property data includes:

[0028] Based on the climate data, determine the average temperature and average rainfall within the time step;

[0029] Determining soil vegetation coverage based on the remote sensing data;

[0030] Based on the average temperature, the average rainfall and the soil vegetation coverage, the decomposition rate of soil organic carbon within the time step is calculated.

[0031] Optionally, after acquiring the multimodal dataset of the target region within the target time period, the method further includes:

[0032] Different types of data in the multimodal data set are standardized according to a preset spatial resolution, and the multimodal data set is updated.

[0033] A second aspect of the present invention provides a system for estimating soil organic carbon content based on multimodal data, the system comprising:

[0034] A data acquisition module, used to obtain a multimodal data set of a target area within a target time period, wherein the multimodal data set at least includes remote sensing data, climate data, and soil property data of the target area;

[0035] The soil organic carbon content estimation module is used to process the multimodal data set using the trained soil organic carbon dynamic model to obtain a soil organic carbon content estimation result.

[0036] A third aspect of the present invention provides an intelligent terminal, which includes a memory, a processor, and a program for estimating soil organic carbon content based on multimodal data, which is stored in the memory and can be run on the processor. When the program for estimating soil organic carbon content based on multimodal data is executed by the processor, any step of the above-mentioned method for estimating soil organic carbon content based on multimodal data is implemented.

[0037] A fourth aspect of the present invention provides a computer-readable storage medium, on which is stored a program for estimating soil organic carbon content based on multimodal data. When the program for estimating soil organic carbon content based on multimodal data is executed by a processor, the program implements any step of the above-mentioned method for estimating soil organic carbon content based on multimodal data.

[0038] Compared with the prior art, the beneficial effects of this solution are as follows:

[0039] The present invention mainly utilizes satellite remote sensing data to obtain key parameters such as vegetation coverage and organic matter input, and combines climate data and soil property data to construct a soil organic carbon dynamic model that considers the input, decomposition and accumulation processes of organic carbon in the soil. The constructed soil organic carbon dynamic model can respectively use three sub-models to estimate the contents of easily decomposable, difficultly decomposable and very difficultly decomposable soil organic carbon components, and can achieve high-precision estimation of the organic carbon content in large-scale target areas, and can realize real-time dynamic monitoring. The method has strong applicability and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0041] Figure 1 A flow chart of a method for estimating soil organic carbon content based on a multimodal data set according to the present invention;

[0042] Figure 2 It is a schematic diagram of the data collection and preprocessing process of the present invention;

[0043] Figure 3 It is a flow chart of soil organic carbon dynamic model training of the present invention;

[0044] Figure 4 A schematic diagram of the target region selected for the present invention;

[0045] Figure 5The curve diagram of soil organic carbon estimation results for the selected target area of ​​the present invention;

[0046] Figure 6 A schematic diagram of a system module for estimating soil organic carbon content based on a multimodal data set according to the present invention;

[0047] Figure 7 It is a schematic diagram of the structure of the intelligent terminal of the present invention. DETAILED DESCRIPTION

[0048] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0049] It should be understood that when used in this specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0050] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0051] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0052] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] In view of the problem that the evaluation efficiency of soil organic carbon content is low and dynamic monitoring cannot be achieved in the prior art, the present invention proposes a method for estimating soil organic carbon content based on a multimodal data set, which mainly uses satellite remote sensing data to obtain key parameters such as vegetation coverage and organic matter input, and combines climate data and soil property data to construct a SOC dynamic model that considers SOC input, decomposition and accumulation processes. Through model calibration and verification, high-precision estimation of SOC content in large-scale target areas can be achieved, and dynamic monitoring can be achieved. This method has strong applicability and scalability.

[0055] The embodiment of the present invention provides a method for estimating soil organic carbon content based on a multimodal data set, which is deployed on electronic devices such as computers and servers and applied to soil carbon sink assessment or carbon trading scenarios. It is aimed at estimating soil organic carbon content in a large area using multimodal data, wherein the multimodal data includes but is not limited to remote sensing data, climate data and soil property data of the studied area. Specifically, Figure 1 As shown, the steps of the method in this embodiment include:

[0056] Step S100: acquiring a multimodal data set of a target region within a target time period, wherein the multimodal data set at least includes remote sensing data, climate data, and soil property data of the target region;

[0057] Specifically, Figure 2 As shown, the boundary of the target area is determined, and then multimodal data sets within the target time period are obtained respectively, including collecting satellite remote sensing data from Sentinel-2 and Landsat 8, so as to obtain the Normalized Difference Vegetation Index (NDVI) using light data reflected by different spectral bands, which is used to measure the growth and density of vegetation; collecting climate data provided by ERA5-Land, including rainfall, evaporation and average temperature; collecting soil property data provided by ISRIC, including soil clay content and other information. It is easy to understand that since this embodiment uses satellite-assisted measurement of remote sensing data, climate data and soil property data of the target area, it is possible to collect data from a large area, so the area of ​​the target area that can be studied by the present invention is not limited, and it can be a small area or a large area.

[0058] In one embodiment, after acquiring the multimodal dataset of the target region within the target time period, the method further includes:

[0059] Different types of data in the multimodal data set are standardized according to a preset spatial resolution, and the multimodal data set is updated.

[0060] Specifically, first, atmospheric correction, geometric correction and cloud detection are performed on remote sensing data to obtain updated remote sensing data, and the updated multimodal data set is used to generate changes in vegetation coverage and vegetation type differences within the target time period. Among them, atmospheric correction is used to eliminate the influence of the atmosphere on surface reflectivity and radiation, and obtain the true reflectivity or radiance of the surface; geometric correction is used to eliminate the geometric distortion caused by factors such as sensors, earth curvature, atmospheric refraction, etc. during the imaging process of remote sensing images, such as using ground control points for polynomial transformation or automatic correction based on geocoding information to ensure that the corrected image is consistent with the standard geographic coordinate system; cloud detection is used to identify and remove clouds and cloud shadows in the image to avoid interference with surface information analysis, such as using methods based on spectral features (such as red and blue band reflectivity differences), texture features, time series analysis, etc.; vegetation coverage is estimated by classification methods (such as supervised classification, unsupervised classification) or threshold methods based on vegetation indices (such as NDVI). NDVI is calculated by the reflectivity of the red band and near-infrared band, and is often used to monitor the health of vegetation growth, density, and changes in vegetation coverage.

[0061] Then, the climate data is spatially interpolated to have the same spatial resolution as the remote sensing data, so as to interpolate the observation data of the climate station into a continuous spatial distribution map so as to match it spatially with the remote sensing data. Interpolation methods such as inverse distance weighted interpolation, kriging interpolation, and spline interpolation can be selected by considering the spatial distribution characteristics of the data, the size of the data, and other requirements. Similarly, the soil data is standardized to ensure spatial consistency with the remote sensing data and climate data, and to ensure spatial consistency and comparability of the soil data with the remote sensing data and climate data. If the spatial resolution of the soil data is lower than that of the remote sensing data and climate data, the spatial resolution is matched by resampling or aggregation processing; by unifying the coordinate system, the coordinate system of all data is ensured to be consistent to avoid spatial position errors; by converting the soil data into the same format as the remote sensing data and climate data, it is convenient for subsequent data processing and analysis. Furthermore, these data can also be cleaned to remove outliers or missing values ​​to ensure the validity and smoothness of the data.

[0062] Finally, the processed remote sensing data, climate data, and soil data are integrated into the target geographic information system platform to prepare for subsequent use of the data.

[0063] Step S200: using the trained soil organic carbon dynamic model to process the multimodal data set to obtain a soil organic carbon content estimation result.

[0064] Specifically, based on climate data, the average temperature and average rainfall in the target time period are extracted to determine the average temperature and average rainfall; based on remote sensing data, vegetation coverage data are obtained through remote sensing image processing (such as NDVI analysis, classification, etc.), and vegetation coverage, easily decomposable plant materials, and resistant plant materials are obtained. Based on soil property data, the amount of plant fertilizer input is determined through soil testing, agricultural practice records, or model estimation. The amount of plant fertilizer input includes the input of organic fertilizers and inorganic fertilizers, as well as the natural return of plant residues (such as roots, fallen leaves, etc.).

[0065] Then, the average temperature, average rainfall, vegetation coverage, easily decomposable plant materials, resistant to decomposition plant materials and plant fertilizer input were normalized and used as inputs to the trained soil organic carbon dynamic model to obtain the estimated results of soil organic carbon content. Among them, the trained soil organic carbon dynamic model is based on the biological and physical principles of soil carbon cycle, taking into account carbon input (such as plant residues, fertilizers, etc.), output (such as respiration, leaching, etc.) and physical and chemical processes in the soil. The model is constructed and trained to accurately estimate the soil organic carbon content in the target area.

[0066] Furthermore, the estimated results of soil organic carbon content can be used to generate a spatial distribution map of SOC content in the target area, or to obtain the statistical characteristics of SOC content in any period of time, which is conducive to analyzing the changing trend of SOC content in the time dimension and the distribution pattern in the spatial dimension.

[0067] In this embodiment, satellite remote sensing data is used to obtain key parameters such as vegetation coverage and organic matter input, and combined with climate data and soil property data, a SOC dynamic model that takes into account the SOC input, decomposition and accumulation process is constructed. Through model calibration and verification, high-precision estimation of SOC content in large-scale target areas is achieved, and dynamic monitoring can be achieved. This method has strong applicability and scalability, significantly improves the efficiency of soil organic carbon content assessment, and realizes the function of dynamic monitoring of soil organic carbon content in any time period.

[0068] like Figure 3 As shown, in one embodiment, the training process of the soil organic carbon dynamic model includes:

[0069] Step M100: Initializing the soil organic carbon dynamic model of the target area;

[0070] Specifically, based on the biological and physical principles of soil carbon cycle, the soil organic carbon dynamic model of the target area is initialized by considering carbon input (such as plant residues, fertilizers, etc.), output (such as respiration, leaching, etc.) and physical and chemical processes in the soil.

[0071] Step M200: collecting remote sensing data samples, climate data samples, soil property data samples and actual soil organic carbon content of the target area within a simulation period;

[0072] Specifically, Figure 2 As shown in the figure, the boundary of the target area is determined, and then multimodal data sets are obtained within the target time period, including: collecting satellite remote sensing data from Sentinel-2 and Landsat 8, including information such as vegetation cover and land use change, to extract NDVI, which helps to understand the input and output process of soil organic carbon; collecting climate data provided by ERA5-Land, including temperature, rainfall, humidity, etc., which are important factors affecting the decomposition and accumulation of soil organic carbon; collecting soil property data provided by ISRIC, including soil clay content, soil texture, pH value, soil moisture and other soil physical and chemical property data, which directly affect the storage and transformation of soil organic carbon. In addition, the measured SOC data in the study area are collected, including sampling location, sampling depth, SOC content and other information, so as to use the measured SOC data as the benchmark for model training and verification. It is easy to understand that the historical sample data and the measured SOC data collected during the training model should be consistent in terms of data type, sampling location, sampling depth, etc., to ensure the accuracy of the trained model and lay a good foundation.

[0073] It is easy to understand that in order to use the data in the multimodal dataset within the target time period to perform multiple iterative optimization operations on the soil organic carbon dynamic model, the data within a preset time step within the target time period is used as training data in each iterative training. For example, if the preset time step is a month, then the remote sensing data samples, climate data samples, and soil attribute data samples of each month in the multimodal dataset are used as training data, where the climate data samples are the monthly rainfall, monthly evaporation, and monthly average temperature of a certain month.

[0074] Step M300: Iteratively update the soil organic carbon dynamic model using the remote sensing data samples, the climate data samples, the soil property data samples and the actual soil organic carbon content corresponding to the preset time step until the simulation period or the preset estimation accuracy is reached to obtain a trained soil organic carbon dynamic model.

[0075] Specifically, by comprehensively considering the above-mentioned various input data, a model suitable for simulating the dynamic changes of soil organic carbon is constructed, so that these data are used as the input of the model to predict the changes of soil organic carbon. Basic models such as empirical models, statistical models (such as multivariate linear regression, random forest, etc.) and process models (such as DNDC, RothC, etc.) can be selected according to actual application requirements to build. In the process of iterative updating of the model, according to research needs and data availability, the initial parameters are set for the model, and the time step (such as year, quarter, month) and simulation cycle of the model update are set. The length of the time step can be adjusted accordingly according to the length of the simulation cycle. The time step used in this embodiment is month, that is, the data in the simulation cycle is divided into multiple parts by month, so that the monthly data is used as the data of one iteration. In each time step, the processed remote sensing data, climate data and soil attribute data in the corresponding time are input into the model, and the model calculates the predicted value of the soil organic carbon content of the current time step according to the input data, and compares the predicted value with the real soil organic carbon content data in each iteration, calculates the error, and continuously adjusts the model parameters using optimization methods such as gradient descent and genetic algorithm. Repeat the above steps until the simulation cycle ends or the model prediction accuracy reaches a preset threshold.

[0076] In this embodiment, historical sample data and real SOC data are used to optimize model parameters and structure through multiple iterations, so as to improve the adaptability and accuracy of the model.

[0077] In one embodiment, the iterative updating of the soil organic carbon dynamic model using the remote sensing data samples, the climate data samples, the soil attribute data samples and the actual soil organic carbon content corresponding to the preset time step includes:

[0078] Step M310: Calculate the decomposition rate of soil organic carbon in the simulation period using the remote sensing data samples, the climate data samples and the soil property data samples;

[0079] Specifically, since the decomposition rate is the rate of soil organic carbon decomposition process, which is affected by temperature, humidity and soil cover, this embodiment determines the average temperature and average rainfall within the time step based on climate data; determines the soil vegetation coverage based on remote sensing data; and calculates the decomposition rate of soil organic carbon within the time step based on the average temperature, average rainfall and soil vegetation coverage. The decomposition rate calculation formula is as follows:

[0080] decom_rate=rate_temp×rate_moisture×rate_cover,

[0081] Among them, decom_rate represents the SOC decomposition rate, and the temperature decomposition rate factor rate_temp = eα×(temp-temp_0) , temp is the average temperature in the time step (in degrees Celsius), α is the temperature factor, which indicates the sensitivity of temperature to decomposition rate, and temp_0 is a constant, indicating the reference temperature. For example, temp-temp_0 indicates the difference between the actual temperature temp and the reference temperature temp_0, which reflects the degree of deviation of the current temperature conditions from the benchmark conditions. When the actual temperature is higher than 10, the difference is positive, indicating that microbial activity may accelerate; when the actual temperature is lower than 10, the difference is negative, indicating that microbial activity may slow down; when the actual temperature is exactly 10, the difference is zero, indicating that microbial activity is at the benchmark level. Using an exponential function to represent the sensitivity of temperature to decomposition rate can ensure that the temperature effect is nonlinear and more in line with the actual change process. Moisture decomposition rate factor rate_moisture = β × moisture, moisture is the amount of rainfall in the time step (in millimeters), β is the humidity factor, indicating the degree of influence of humidity on the decomposition rate. Soil cover influence factor rate_cover = e -γ×soil_cover , which is used to indicate the soil vegetation coverage rate in the target area, soil_cover is the vegetation index NDVI, and β is the humidity factor, which indicates the inhibitory effect of soil cover on the decomposition rate.

[0082] In this embodiment, by respectively calculating the effects of the temperature, humidity and vegetation coverage of the target area on the decomposition rate, an estimated value of the comprehensive decomposition rate of the current time step is obtained, which comprehensively considers the effects of various factors on the SOC decomposition rate, and is beneficial to improving the calculation accuracy.

[0083] Step M320: Determine the easily decomposable plant material sample and the resistant-to-decomposable plant material sample based on the remote sensing data sample;

[0084] Specifically, the remote sensing data samples are preprocessed by atmospheric correction, geometric correction, cloud and shadow removal to ensure the accuracy of the data, and the spectral characteristics of the remote sensing images (such as red edge, near-infrared band) and vegetation indices (such as NDVI, EVI) are used to classify vegetation and distinguish different vegetation types (such as forests, grasslands, farmlands, wetlands, etc.). Then, based on this information, the correspondence between the remote sensing classification results and the decomposition properties of plant materials is established. For example, some types of grasslands may be rich in easily decomposable herbaceous plants, while some forest types may contain more resistant woody materials. According to the correspondence between the remote sensing classification results and the decomposition properties, representative sample areas are selected on the remote sensing images. The accuracy of remote sensing classification and the inference of decomposition properties are verified by real SOC data, and the remote sensing classification method and the inference rules of decomposition properties are adjusted according to the verification results to improve the accuracy of sample selection.

[0085] Step M330: determining a plant fertilizer input amount sample based on the soil property data sample;

[0086] Specifically, based on the soil attribute data samples, the input of organic fertilizers and inorganic fertilizers is determined, including but not limited to, and the fertility of the soil is judged by the natural return of plant residues (such as roots, fallen leaves, etc.), thereby determining the plant fertilizer input amount sample.

[0087] Step M340: obtaining an estimated value of soil organic carbon content based on the easily decomposable plant material sample, the resistant plant material sample and the plant fertilizer input amount sample corresponding to each time step as inputs of the soil organic carbon dynamic model;

[0088] Specifically, in order to construct an impact of factors in multiple dimensions on the dynamic changes of SOC, this embodiment simulates the soil organic carbon dynamic model by constructing an active reservoir model, a slow reservoir model and a passive reservoir model. That is, the soil organic carbon dynamic model is constructed by these three sub-models. Among them, the active reservoir contains easily decomposable organic carbon components, which decompose quickly and are mainly composed of microorganisms and fresh organic matter. The slow reservoir contains difficult-to-decompose organic carbon components, which decompose at a medium rate and are mainly composed of humus and relatively stable organic matter. The passive reservoir contains very difficult-to-decompose organic carbon components, which decompose extremely slowly and are mainly composed of highly stable organic matter and mineral-bound organic carbon.

[0089] Then, the active pool model is updated using the easily decomposable plant material and the decomposition rate, namely:

[0090] active_pool t+1 =active_pool t +dpm-(decom_rate×active_pool t ),

[0091] Among them, active_pool represents the organic carbon content in the active pool, the amount of easily degradable plant material dpm = plant_residue×dpm_rpm_ratio, plant_residue represents the amount of plant residue within the time step, dpm_rpm_ratio represents the estimated value of the decomposability of plant materials, that is, the dpm / rpm ratio, and rpm represents the amount of resistant plant material.

[0092] Update the slow pool model using decomposition resistant plant material, active pool and decomposition rate, i.e.:

[0093] slow_pool t+1 =slow_pool t +rpm+(ε×decom_rate×active_poolt )-(decom_rate×slow_pool t ),

[0094] Among them, slow_pool represents the organic carbon content in the slow pool, the amount of decomposition-resistant plant material rpm=plant_residue-dpm, ε represents the proportion of decomposition products in the active pool entering the slow pool, for example, ε=0.5 means that 50% of the decomposition products in the active pool enter the slow pool.

[0095] The passive sink model is updated based on the slow sink model, decomposition rate and plant fertilizer input, namely:

[0096] passive_pool t+1 =passive_pool t +(ρ×decom_rate×slow_pool t )-(decom_rate×passive_pool t )+fym_input,

[0097] Among them, passive_pool represents the organic carbon content in the passive pool, fym_input represents the plant fertilizer input within the time step, and ρ represents the proportion of decomposition products of the slow pool entering the passive pool. For example, ρ = 0.1 means that 10% of the decomposition products of the slow pool enter the passive pool.

[0098] Finally, based on the dynamically updated active pool model, slow pool model and passive pool model, the estimated value of soil organic carbon content total_soc is obtained t , whose expression is:

[0099] total_soc t =active_pool t +slow_pool t +passive_pool t ,

[0100] It can be seen that total_soc t Represents the sum of SOC of each carbon pool within each time step.

[0101] Step M350: iteratively updating the soil organic carbon dynamic model based on the estimated value corresponding to each time step and the actual soil organic carbon content.

[0102] Specifically, the estimated SOC value obtained in each iteration corresponding to each time step is compared with the real SCO data in the corresponding time period, and the key parameters in the model, such as reference temperature, temperature factor, humidity factor, soil cover factor, etc., are identified to determine the degree of influence on the SOC estimation results, and the model parameters are adjusted through the optimization algorithm to minimize the error between the model estimate and the measured value. Furthermore, the model can also be evaluated using an independent validation data set to further improve the accuracy and reliability of the model. Obviously, after multiple iterations of the soil organic carbon dynamic model are optimized, the values ​​of each key parameter in the soil organic carbon dynamic model are determined, and the entire soil organic carbon dynamic model is trained. It is easy to understand that for different target regions, the values ​​of each key parameter in the trained soil organic carbon dynamic model are different, that is, the values ​​of the variable parameters in the model change with the changes in the region where the soil under study is located, and it is precisely the different values ​​of these variable parameters that reflect the differences in soil organic carbon content in different regions.

[0103] In summary, the present invention collects remote sensing data, climate data and soil property data of a large-scale target area as the basis for model training, and continuously optimizes the key parameters in the model through multiple iterative training, thereby training a soil organic carbon dynamic model that can accurately measure the SOC content of the target area. The soil organic carbon dynamic model constructed by the present invention can use three sub-models to estimate the contents of easily decomposable, difficult to decompose and very difficult to decompose soil organic carbon components, respectively, so as to obtain an accurate total soil organic carbon content.

[0104] The effectiveness of the method provided by the present invention is verified by experiments as follows: Figure 4 The area within the gray irregular frame shown is taken as the target area. The soil organic carbon content of the target area from 2015 to 2019 is estimated using the method proposed in the present invention. The estimated curve of soil organic carbon content is as follows: Figure 5 As shown, the horizontal axis represents the year, and the estimation is performed with a half-year time step (for example, 2015.0 represents January 2015, and 2015.5 represents June 2015), and the vertical axis represents the estimated value of soil organic carbon SOC (in tons / hectare). The results show that the estimated result is very close to the actual soil organic carbon content in the corresponding year in the region, thereby effectively verifying the effectiveness of the method of the present invention.

[0105] like Figure 6 As shown, corresponding to the above-mentioned method for estimating soil organic carbon content based on multimodal data, an embodiment of the present invention further provides a system for estimating soil organic carbon content based on multimodal data, and the above-mentioned system for estimating soil organic carbon content based on multimodal data includes:

[0106] A data acquisition module, used to obtain a multimodal data set of a target area within a target time period, wherein the multimodal data set at least includes remote sensing data, climate data, and soil property data of the target area;

[0107] The soil organic carbon content estimation module is used to process the multimodal data set using the trained soil organic carbon dynamic model to obtain a soil organic carbon content estimation result.

[0108] Specifically, in this embodiment, the specific functions of the system for estimating soil organic carbon content based on multimodal data can also refer to the corresponding description in the method for estimating soil organic carbon content based on multimodal data, which will not be repeated here.

[0109] Based on the above embodiments, the present invention further provides an intelligent terminal, whose principle block diagram can be shown as follows: Figure 7 As shown. The above-mentioned intelligent terminal includes a processor, a memory, a network interface and a display screen connected through a system bus. Among them, the processor of the intelligent terminal is used to provide computing and control capabilities. The memory of the intelligent terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a program for estimating the organic carbon content of soil based on multimodal data. The internal memory provides an environment for the operation of the operating system and the program for estimating the organic carbon content of soil based on multimodal data in the non-volatile storage medium. The network interface of the intelligent terminal is used to communicate with an external terminal through a network connection. When the program for estimating the organic carbon content of soil based on multimodal data is executed by the processor, the steps of any one of the above-mentioned methods for estimating the organic carbon content of soil based on multimodal data are implemented. The display screen of the intelligent terminal can be a liquid crystal display screen or an electronic ink display screen.

[0110] Those skilled in the art will understand that Figure 7 The principle block diagram shown in the figure is only a block diagram of a partial structure related to the scheme of the present invention, and does not constitute a limitation on the smart terminal to which the scheme of the present invention is applied. The specific smart terminal may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0111] In one embodiment, a smart terminal is provided, which includes a memory, a processor, and a program for estimating soil organic carbon content based on multimodal data stored in the memory and executable on the processor. When the program for estimating soil organic carbon content based on multimodal data is executed by the processor, the steps of any one of the methods for estimating soil organic carbon content based on multimodal data provided in an embodiment of the present invention are implemented.

[0112] An embodiment of the present invention also provides a computer-readable storage medium, on which is stored a program for estimating soil organic carbon content based on multimodal data. When the program for estimating soil organic carbon content based on multimodal data is executed by a processor, the steps of any one of the methods for estimating soil organic carbon content based on multimodal data provided in the embodiments of the present invention are implemented.

[0113] It should be understood that the serial numbers of the steps in the above embodiments do not imply a sequence of execution, and the execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0114] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the above-mentioned device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0115] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0116] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0117] In the embodiments provided by the present invention, it should be understood that the disclosed apparatus / terminal equipment and method can be implemented in other ways. For example, the apparatus / terminal equipment embodiments described above are only illustrative, for example, the division of the above modules or units is only a logical function division, and in actual implementation, other division methods can be used, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0118] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features therein may be replaced by equivalents. However, these modifications or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for estimating soil organic carbon content based on a multimodal data set, characterized in that: The following steps are involved: Acquire a multimodal data set of a target area within a target time period, wherein the multimodal data set includes at least remote sensing data, climate data, and soil property data of the target area; The multimodal data set is processed using the trained soil organic carbon dynamic model to obtain an estimation result of the soil organic carbon content.

2. The method for estimating soil organic carbon content based on a multimodal data set according to claim 1, characterized in that: The multimodal data set is processed using the trained soil organic carbon dynamic model to obtain a soil organic carbon content estimation result, including: Based on the climate data, determine the average temperature and average rainfall; Based on the remote sensing data, obtaining vegetation coverage, easily decomposable plant materials and resistant to decomposition plant materials; determining plant fertilizer input based on the soil property data; The average temperature, the average rainfall, the vegetation coverage, the easily decomposable plant materials, the resistant to decomposition plant materials and the plant fertilizer input are used as inputs of the trained soil organic carbon dynamic model to obtain an estimated result of the soil organic carbon content.

3. The method for estimating soil organic carbon content based on a multimodal data set according to claim 1, characterized in that: The training process of the soil organic carbon dynamic model includes: Initializing a soil organic carbon dynamic model of the target area; Collect remote sensing data samples, climate data samples, soil property data samples and actual soil organic carbon content of the target area within the simulation period; The soil organic carbon dynamic model is iteratively updated using the remote sensing data samples, the climate data samples, the soil property data samples and the actual soil organic carbon content corresponding to the preset time step until the simulation period or the preset estimation accuracy is reached, thereby obtaining a trained soil organic carbon dynamic model.

4. The method for estimating soil organic carbon content based on a multimodal data set according to claim 3, characterized in that: The iterative updating of the soil organic carbon dynamic model using the remote sensing data samples, the climate data samples, the soil attribute data samples and the actual soil organic carbon content corresponding to the preset time step includes: Calculating the decomposition rate of soil organic carbon during the simulation period using the remote sensing data samples, the climate data samples and the soil property data samples; Determining easily decomposable plant material samples and resistant to decomposition plant material samples based on the remote sensing data samples; Determining a plant fertilizer input amount sample based on the soil property data sample; Obtaining an estimated value of soil organic carbon content based on the easily decomposable plant material sample, the resistant decomposition plant material sample and the plant fertilizer input amount sample corresponding to each time step as inputs to the soil organic carbon dynamic model; The soil organic carbon dynamic model is iteratively updated based on the estimated value corresponding to each time step and the actual soil organic carbon content.

5. The method for estimating soil organic carbon content based on a multimodal data set according to claim 4, characterized in that: The soil organic carbon dynamic model includes an active pool model, a slow pool model and a passive pool model. The easily decomposable plant material sample, the resistant decomposable plant material sample and the plant fertilizer input amount sample corresponding to each time step are used as inputs of the soil organic carbon dynamic model to obtain an estimated value of the soil organic carbon content, including: updating the active reservoir model using the easily decomposable plant material and the decomposition rate; updating the slow pool model using the decomposition-resistant plant material, the active pool, and the decomposition rate; updating a passive sink model based on the slow sink model, the decomposition rate, and the plant fertilizer input; Based on the dynamically updated active pool model, the slow pool model and the passive pool model, an estimated value of soil organic carbon content is obtained.

6. The method for estimating soil organic carbon content based on a multimodal data set according to claim 4, characterized in that: The step of calculating the decomposition rate of soil organic carbon by using the remote sensing data, the climate data and the soil property data comprises: Based on the climate data, determine the average temperature and average rainfall within the time step; Determining soil vegetation coverage based on the remote sensing data; Based on the average temperature, the average rainfall and the soil vegetation coverage, the decomposition rate of soil organic carbon within the time step is calculated.

7. The method for estimating soil organic carbon content based on a multimodal data set according to claim 1, characterized in that: After obtaining the multimodal dataset of the target region within the target time period, the method further includes: Different types of data in the multimodal data set are standardized according to a preset spatial resolution, and the multimodal data set is updated.

8. A system for estimating soil organic carbon content based on multimodal data, characterized in that: The system comprises: A data acquisition module, used to obtain a multimodal data set of a target area within a target time period, wherein the multimodal data set at least includes remote sensing data, climate data, and soil property data of the target area; The soil organic carbon content estimation module is used to process the multimodal data set using the trained soil organic carbon dynamic model to obtain a soil organic carbon content estimation result.

9. An intelligent terminal, characterized in that: The intelligent terminal includes a memory, a processor, and a program for estimating soil organic carbon content based on multimodal data, which is stored in the memory and can be run on the processor. When the program for estimating soil organic carbon content based on multimodal data is executed by the processor, the steps of the method for estimating soil organic carbon content based on multimodal data as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program for estimating soil organic carbon content based on multimodal data. When the program for estimating soil organic carbon content based on multimodal data is executed by a processor, the steps of the method for estimating soil organic carbon content based on multimodal data as described in any one of claims 1 to 7 are implemented.

Images