A tobacco carbon accounting parameter analysis and prediction system

Abstract

This invention discloses a tobacco carbon accounting parameter analysis and prediction system, relating to the field of carbon accounting analysis, prediction, and management technology. The system includes: an initial accounting analysis module for continuously acquiring carbon accounting parameter sets for each period and analyzing the initial carbon emission index; a soil environment correction analysis module for extracting environmental correction factors based on soil environment data; a growth correction analysis module for extracting growth correction factors based on tobacco growth image sequences and a pre-trained growth status recognition model; and a prediction analysis module for calculating the comprehensive tobacco carbon emission index based on the analysis results of the above modules and predicting the comprehensive tobacco carbon emission index for the next period. This invention improves the accuracy of carbon accounting prediction by using the comprehensive tobacco carbon emission index for the next period in the prediction feedback management module to implement management measures for tobacco planting areas, thereby enabling timely detection of abnormal carbon emission fluctuations and the implementation of corresponding management measures.

Description

Technical Field

[0001] This invention relates to the field of carbon accounting analysis, prediction and management technology, specifically a tobacco carbon accounting parameter analysis and prediction system. Background Technology

[0002] As an important part of the national economy, the tobacco industry involves multiple stages in its production process, which contains complex carbon emission factors. Tobacco carbon accounting refers to the process of quantitatively assessing the carbon emissions and carbon sink behavior generated at each stage of the tobacco life cycle. It aims to comprehensively measure the contribution of tobacco production activities to regional carbon balance and climate impact. This accounting process usually covers stages such as planting, harvesting, curing, primary processing and distribution, and involves the identification and quantification of multiple carbon source and carbon sink elements such as energy consumption, mechanized operations and biological carbon sinks.

[0003] In the entire process, the planting stage is the most concentrated and variable link in terms of carbon emissions and carbon absorption activities, making carbon accounting work particularly crucial. Carbon accounting at this stage mainly focuses on the carbon emissions and carbon absorption caused by behaviors such as soil disturbance, fertilizer application, and plant growth during field operations. The aim is to dynamically capture the specific impact of agricultural operations on carbon flux, providing basic support for optimizing planting management strategies and implementing low-carbon agriculture, so as to reduce carbon emissions and achieve green and sustainable development.

[0004] However, traditional technologies generally rely on rough estimates based on annual average or periodic manual measurements of carbon emissions, making it difficult to achieve precise modeling and dynamic control of carbon emission behavior in specific planting areas and time periods. Furthermore, especially in the tobacco planting stage, traditional carbon emission accounting methods often lack a systematic consideration of the comprehensive impact of multiple factors such as changes in soil microenvironment, differences in growth stages, and agronomic practices, which can easily lead to insufficient accuracy in the accounting results.

[0005] The limitations of existing technologies include at least the following problems: Existing technologies cannot reflect the dynamic changes in carbon emissions in tobacco growing areas in real time. Carbon emissions in tobacco growing areas are affected by complex and interrelated factors such as soil and plant growth. These factors change dynamically over a period of time. For example, soil moisture fluctuates rapidly due to rainfall and evaporation, and the respiration rate of tobacco plants changes accordingly with the intensity of sunlight, thus affecting the overall level of carbon absorption and emission. As a result, carbon emissions in tobacco growing areas exhibit complex fluctuation characteristics throughout the entire growth cycle. Existing technologies lack the ability to integrate multi-dimensional data, making it difficult to collect and accurately integrate data from multiple influencing factors simultaneously. Consequently, it is difficult to conduct accurate carbon accounting by comprehensively considering multiple factors, which in turn makes it difficult to effectively capture and analyze the fluctuation characteristics of carbon emissions, thereby affecting the accuracy of management decisions. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a tobacco carbon accounting parameter analysis and prediction system, which solves the problem that existing technologies lack multi-dimensional data fusion, making it difficult to accurately capture the carbon emission characteristics of tobacco planting areas and affecting the accuracy of carbon accounting.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a tobacco carbon accounting parameter analysis and prediction system, comprising the following steps: an initial accounting analysis module, used to continuously acquire carbon accounting parameter sets for several cycles of a designated tobacco planting area and analyze the initial carbon emission index of the corresponding cycle; a soil environment correction analysis module, used to acquire soil environment data for each cycle of the designated tobacco planting area and analyze the soil environment correction factor of the corresponding cycle; a growth correction analysis module, used to acquire tobacco growth image sequence data for each cycle of the designated tobacco planting area, and analyze the growth correction factor of the corresponding cycle based on a pre-trained growth status recognition model; a prediction analysis module, used to analyze the comprehensive tobacco carbon emission index of the corresponding cycle based on the initial carbon emission index, soil environment correction factor, and growth correction factor of the designated tobacco planting area for each cycle, and predict the comprehensive tobacco carbon emission index of the next cycle; and a prediction feedback management module, used to take preset management measures for the designated tobacco planting area based on the comprehensive tobacco carbon emission index of the next cycle.

[0008] Furthermore, the carbon accounting parameter set includes planting area value, tobacco plant biomass value, tobacco plant respiration rate value, tobacco plant carbon sink value, soil carbon sink value, usage value of each agricultural activity, and corresponding carbon emission factors and application area value. The specific steps for analyzing and setting the initial carbon emission index for each cycle of the tobacco planting area are as follows: Based on the carbon accounting parameter set for each cycle of the tobacco planting area, analyze the accounting assessment set for the corresponding cycle, including the comprehensive carbon emission index and the comprehensive carbon sink index; obtain the carbon emission impact factor and carbon sink impact factor for each cycle of the tobacco planting area, and conduct a comprehensive analysis with the emission assessment set for the corresponding cycle to obtain the initial carbon emission index for the corresponding cycle.

[0009] Furthermore, the specific steps for analyzing the accounting and evaluation set of each cycle in the designated tobacco planting area are as follows: Based on the planting area value, tobacco plant biomass, tobacco plant respiration rate value, usage value of each agricultural activity, and corresponding carbon emission factors and application area value of each cycle in the designated tobacco planting area, analyze the comprehensive carbon emission index of the corresponding cycle; Based on the tobacco plant carbon sink value and soil carbon sink value of each cycle in the designated tobacco planting area, analyze the comprehensive carbon sink index of the corresponding cycle.

[0010] Furthermore, the specific steps for obtaining the carbon emission impact factors and carbon sink impact factors for each cycle in the designated tobacco planting area are as follows: obtain the tillage intensity value, tillage frequency value, and topsoil disturbance depth value for each cycle in the designated tobacco planting area, and analyze the carbon emission impact factors for the corresponding cycle; obtain the canopy uptake rate, chlorophyll relative content index, and soil structure stability index for each cycle in the designated tobacco planting area, and analyze the carbon sink impact factors for the corresponding cycle.

[0011] Furthermore, the soil environmental data includes soil thermal diffusivity, soil environmental regulation factors, root decomposition rate, soil electrical conductivity gradient, and soil dissolved oxygen concentration. The specific steps for analyzing the soil environmental correction factors for each cycle of the designated tobacco planting area are as follows: Obtain the climate factors for each cycle of the designated tobacco planting area, and combine them with the soil environmental data for the corresponding cycle to analyze the soil environmental assessment set for the corresponding cycle, including soil energy regulation factors and soil carbon transformation factors; Based on the soil environmental assessment set for each cycle of the designated tobacco planting area, analyze the soil environmental correction factors for the corresponding cycle.

[0012] Furthermore, the specific steps for analyzing the soil environmental assessment set for each cycle of the tobacco planting area are as follows: Based on the soil thermal diffusivity, soil environmental regulation factors, and soil electrical conductivity gradient values ​​of each cycle of the tobacco planting area, analyze the soil energy regulation factors for the corresponding cycle; based on the root decomposition rate and soil dissolved oxygen concentration values ​​of each cycle of the tobacco planting area, analyze the soil carbon transformation factors for the corresponding cycle.

[0013] Furthermore, the tobacco growth image sequence data includes tobacco growth image data for each time window, each including multi-band pixel values ​​and two-dimensional coordinates for each pixel. The specific steps for analyzing the growth correction factor for each cycle of the designated tobacco planting area are as follows: input the tobacco growth image data for each time window of each cycle of the designated tobacco planting area into a pre-trained growth status recognition model for comprehensive analysis to obtain the corresponding cycle's growth correction evaluation set, including carbon absorption index, tobacco vigor index, and tobacco light adaptation index; based on the growth evaluation set for each cycle of the designated tobacco planting area, analyze the corresponding cycle's growth correction factor.

[0014] Furthermore, the specific formula for calculating the growth correction factor for a given period in a tobacco-growing region is as follows: ;in, To determine the growth correction factor for a specific period in a tobacco-growing region. To determine the carbon uptake index for a tobacco-growing region over a specific period. The absorption coefficient is stored in the database. To establish a tobacco viability index for a specific period in a tobacco-growing region. The vitality coefficient is stored in the database. To determine the tobacco light adaptation index for a specific period in a tobacco-growing region. The fitness coefficients are stored in the database. These are adjustment coefficients stored in the database.

[0015] Furthermore, the growth status recognition model includes a feature extraction subnetwork, a temporal state modeling subnetwork, and a decoding output subnetwork. The specific steps for analyzing the growth evaluation set of each cycle in the designated tobacco planting area are as follows: In the feature extraction subnetwork of the growth status recognition model, tobacco growth image data for each time window of each cycle in the designated tobacco planting area are received and feature encoding is performed to obtain the feature vector of the corresponding time window; In the temporal state modeling subnetwork of the growth status recognition model, temporal analysis is performed on the feature vector of each time window of each cycle in the designated tobacco planting area to obtain the temporal feature vector of the corresponding cycle; In the decoding output subnetwork of the growth status recognition model, prediction processing is performed on the temporal feature vector of each cycle in the designated tobacco planting area to obtain the growth evaluation set of the corresponding cycle.

[0016] Furthermore, the specific formula for calculating the comprehensive tobacco carbon emission index for a given tobacco-growing region over a specific period is as follows: ;in, To establish a comprehensive tobacco carbon emission index for a tobacco-growing region over a specific period. To set an initial carbon emission index for a tobacco-growing region for a specific period. To determine the soil environmental correction factor for a specific period in a tobacco-growing region, The soil environmental coefficients stored in the database. To determine the growth correction factor for a specific period in a tobacco-growing region. The growth coefficients are stored in the database. These are the collaboration coefficients stored in the database.

[0017] The present invention has the following beneficial effects:

[0018] (1) The tobacco carbon accounting parameter analysis and prediction system analyzes data from multiple dimensions and obtains the initial carbon emission index, soil environment correction factor and growth correction factor, thereby achieving accurate capture of carbon emission characteristics in tobacco planting areas. The system cross-integrates the analyzed characteristics to accurately reflect the complex impact of tobacco plant growth and soil environment changes on carbon emissions, and thus can comprehensively assess the carbon emission status of tobacco planting areas. It can not only accurately calculate the carbon emission index of the current period, but also predict the carbon emission trend of the next period in real time, thereby effectively improving the accuracy of carbon emission accounting and prediction, and timely detecting abnormal carbon emission fluctuations so as to take corresponding management measures to improve the sustainability of tobacco planting.

[0019] (2) The tobacco carbon accounting parameter analysis and prediction system extracts the features of tobacco plant growth status from tobacco growth image sequence data based on a pre-trained growth status recognition model. It can not only accurately capture the details of tobacco plant growth process, but also deeply analyze how these details affect the formation and change of carbon emissions. If the tobacco growth is in the high photosynthesis stage, the plant's carbon absorption capacity is strong, which can easily lead to lower carbon emissions; conversely, when the growth is stagnant, carbon emissions are likely to increase. Thus, the system can accurately identify the impact of tobacco plant growth stage on carbon emissions and form a growth correction factor to optimize carbon emission accounting and prediction, thereby enhancing the accuracy of carbon emission accounting and prediction.

[0020] (3) The tobacco carbon accounting parameter analysis and prediction system predicts carbon emissions by setting a prediction window and combining the comprehensive tobacco carbon emission index of each period within the prediction window. This allows for dynamic prediction of the carbon emission change trend in the next period. When the system detects that the carbon emission fluctuation exceeds the preset threshold, it can adjust the prediction value in a timely manner and give an early warning, thereby providing an accurate carbon emission risk assessment. This enables the implementation of corresponding management measures in advance and effectively controls the fluctuation of carbon emissions, thereby improving the predictability of carbon emission accounting and helping to accurately achieve carbon emission control in the agricultural production process.

[0021] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0022] Figure 1 This is a block diagram of a tobacco carbon accounting parameter analysis and prediction system according to the present invention;

[0023] Figure 2 This is a flowchart illustrating the specific steps involved in analyzing and setting the initial carbon emission index for each cycle in a tobacco planting area within a tobacco carbon accounting parameter analysis and prediction system of the present invention.

[0024] Figure 3This is a time-series diagram illustrating the setting of the carbon absorption index of a tobacco planting area in a tobacco carbon accounting parameter analysis and prediction system of the present invention;

[0025] Figure 4 This is a time-series diagram illustrating the setting of tobacco vigor index for tobacco planting areas in a tobacco carbon accounting parameter analysis and prediction system of the present invention;

[0026] Figure 5 This is a time-series diagram illustrating the setting of the tobacco light adaptation index for a tobacco planting area in a tobacco carbon accounting parameter analysis and prediction system of the present invention;

[0027] Figure 6 This is a flowchart illustrating the specific steps involved in analyzing and setting growth correction factors for each cycle in a tobacco planting area within a tobacco carbon accounting parameter analysis and prediction system according to the present invention. Detailed Implementation

[0028] Please see Figure 1 This invention provides a technical solution: a tobacco carbon accounting parameter analysis and prediction system, comprising: an initial accounting analysis module, used to continuously acquire carbon accounting parameter sets for several cycles (each cycle being 10 days) of a designated tobacco planting area, and analyze the initial carbon emission index of the corresponding cycle; a soil environment correction analysis module, used to acquire soil environment data for each cycle of the designated tobacco planting area, and analyze the soil environment correction factor of the corresponding cycle; and a growth correction analysis module, used to acquire tobacco (plant) growth image sequence data for each cycle of the designated tobacco planting area, and analyze the growth correction factor of the corresponding cycle based on a pre-trained growth status recognition model.

[0029] The predictive analysis module is used to analyze the comprehensive tobacco carbon emission index for each cycle of a set tobacco planting area based on the initial carbon emission index, soil environment correction factor, and growth correction factor, and to predict the comprehensive tobacco carbon emission index for the next cycle. Specifically, it sets a prediction window, such as three consecutive cycles, and extracts the mean, standard deviation, and range (i.e., the difference between the maximum and minimum values ​​of the comprehensive tobacco carbon emission index within the prediction window) of the comprehensive tobacco carbon emission index. The standard deviation and range of the comprehensive tobacco carbon emission index are weighted to obtain the volatility index of the prediction window. It is then determined whether the volatility index is higher than a preset volatility index threshold. If it is higher, the mean of the comprehensive tobacco carbon emission index and the volatility index are added together to obtain the predicted comprehensive tobacco carbon emission index (which is then marked as the comprehensive tobacco carbon emission index for the next cycle). If it is not higher, the comprehensive tobacco carbon emission index is subtracted from the volatility index to obtain the comprehensive tobacco carbon emission index for the next cycle.

[0030] The predictive feedback management module is used to implement preset management measures for designated tobacco planting areas based on the comprehensive tobacco carbon emission index for the next cycle. Specifically, it determines whether the comprehensive tobacco carbon emission index exceeds a preset threshold. If it does, the preset management measures are implemented for the designated tobacco planting areas, namely, increasing the input of organic fertilizer, improving soil microbial activity, promoting carbon fixation capacity, reducing the carbon release rate, intercropping low-carbon sink plants at the edge of the tobacco planting area or intercropping them to enhance the overall carbon absorption capacity of the area, reducing the frequency of agricultural machinery operations, and using low-carbon agricultural machinery equipment to reduce mechanical fuel consumption and related carbon emissions. If the index does not exceed the threshold, the preset management measures are not implemented for the designated tobacco planting areas.

[0031] The specific formula for calculating the comprehensive tobacco carbon emission index for a given tobacco-growing region over a specific period is as follows: ;in, To establish a comprehensive tobacco carbon emission index for a tobacco-growing region over a specific period. To set an initial carbon emission index for a tobacco-growing region for a specific period. To determine the soil environmental correction factor for a specific period in a tobacco-growing region, The soil environmental coefficients stored in the database. To determine the growth correction factor for a specific period in a tobacco-growing region. The growth coefficients are stored in the database. These are the collaboration coefficients stored in the database.

[0032] It needs to be explained that the specific form of the tanh function is as follows: ,in, It is a natural constant, and in this example it can be taken as 2.71, with a domain of (−∞, +∞) and a range of (−1, +1).

[0033] Soil environmental coefficients stored in the database The acquisition steps are as follows: Historical soil temperature, soil moisture, and soil pH values ​​for several historical periods are obtained, and their standard deviations are processed separately. The results are then weighted based on the standard deviation data, and the weighted result is used as the soil environmental coefficient. ;

[0034] Growth coefficients stored in the database The steps for obtaining the historical growth correction factor are as follows: Obtain the historical growth correction factor for several historical periods, and analyze the mean and standard deviation of the historical growth correction factor respectively. Then, perform a ratio processing, i.e., historical growth correction factor standard deviation / historical growth correction factor mean. The result of this ratio processing is used as the growth coefficient. ;

[0035] Collaboration coefficients stored in the database The acquisition steps are as follows: Obtain historical growth correction factors and historical soil environment correction factors for several historical periods, and perform interactive analysis (historical growth correction factor × historical soil environment correction factor) to obtain the interactive value for each historical period. Then, perform weighted balancing and use the weighted balancing result as the synergy coefficient. .

[0036] Specifically, such as Figure 2 As shown, the carbon accounting parameter set includes planting area values ​​(measured in-situ using RTK and the results uploaded to the database), tobacco plant biomass values, tobacco plant respiration rate values, tobacco plant carbon sink values, soil carbon sink values, usage values ​​for each agricultural activity, and corresponding carbon emission factors (i.e., the conversion ratio of usage values ​​for various agricultural activities to corresponding carbon dioxide emissions, obtained from the carbon accounting manual stored in the database; for example, during a fertilization operation, the type of fertilizer used for the tobacco crop is recorded, and the corresponding carbon emission factor is found in the carbon accounting manual), and application area values. The specific steps for analyzing the initial carbon emission index for each cycle of the tobacco planting area are as follows: Based on the carbon accounting parameter set for each cycle of the tobacco planting area, analyze the corresponding cycle's accounting assessment set, including the comprehensive carbon emission index and the comprehensive carbon sink index; obtain the carbon emission impact factor and carbon sink impact factor for each cycle of the tobacco planting area, and perform a comprehensive analysis with its corresponding cycle's emission assessment set to obtain the initial carbon emission index for the corresponding cycle.

[0037] The specific formula for calculating the initial carbon emission index of a designated tobacco-growing area for a given period is as follows: ;in, To set an initial carbon emission index for a tobacco-growing region for a specific period. To set a comprehensive carbon emission index for a tobacco-growing region over a specific period. To determine the carbon emission impact factor for a tobacco-growing region over a specific period, To establish a comprehensive carbon sink index for a tobacco-growing region over a specific period. To determine the carbon sequestration impact factor for a tobacco growing region over a specific period.

[0038] Tobacco plant biomass is the total fresh weight of tobacco plants per unit area within a given period, reflecting the plant's growth status and organic matter accumulation. It is obtained through field sampling. Representative sample plots are selected, and several tobacco plants are randomly collected. Their fresh weight is weighed to obtain the tobacco plant dry matter value, which is then divided by the sampling area of ​​the sample plot to obtain the tobacco plant biomass value.

[0039] The tobacco plant respiration rate is the rate at which tobacco plants release carbon dioxide through respiration per unit area, reflecting the intensity of the plant's metabolic activity. It is obtained in real time through field monitoring using a closed photosynthesis-respiration measurement instrument (such as LI-COR LI-6800) and uploaded to a database. The instrument calculates the amount of carbon dioxide released per unit area per unit time (based on the instrument's built-in predictive algorithm) by measuring the rate of change of carbon dioxide concentration, chamber volume, measurement time, and plant leaf area of ​​the sample plant in a closed space.

[0040] The carbon sink value of tobacco plants is the net amount of carbon fixed by tobacco plants through photosynthesis, reflecting the plant's ability to absorb atmospheric carbon. It is obtained by reading the dry matter value of tobacco plants, using the carbon content coefficient of tobacco plants stored in the data, and then multiplying the dry matter value of tobacco plants by the carbon content coefficient of tobacco plants. The result is the carbon sink value of tobacco plants.

[0041] Soil carbon sequestration value refers to the net accumulation of organic carbon in soil, reflecting the soil's ability to absorb and store atmospheric carbon dioxide. It is obtained through joint analysis of soil organic carbon content measurement and soil respiration rate monitoring data. Representative soil samples from designated tobacco-growing areas are collected, and the organic carbon content in the samples is quantitatively analyzed using laboratory dry combustion or chemical oxidation methods to obtain the organic carbon content per unit mass of soil. Simultaneously, an in-situ automatic soil respiration meter (e.g., LI-8100) is used to measure the rate of change of carbon dioxide concentration in the soil surface layer in real time within a closed chamber, recording the change in carbon dioxide concentration per unit time. This data is then combined with the measurement chamber data. The volume and measurement area are used to calculate the carbon dioxide flux per unit area, specifically by multiplying the rate of change in carbon dioxide concentration by the chamber volume and dividing by the measurement area. Then, the carbon dioxide flux is converted into carbon flux based on the molar mass ratio of carbon to carbon dioxide, and finally converted into a soil respiration rate value. By comparing the changes in soil organic carbon content at different times and adding a full average, the change value of soil organic carbon content is generated. Combined with the total amount of soil respiration release (the result of multiplying the soil respiration rate value by the duration of the period), the net accumulation of soil organic carbon (change value of soil organic carbon content - total amount of soil respiration release) is calculated, which is the soil carbon sink value.

[0042] The usage values ​​for each agricultural activity are the unit input amounts for various operations such as fertilization, irrigation, agricultural film mulching, and mechanized farming (e.g., the amount of fertilizer used in a fertilization operation, or the amount of irrigation used in an irrigation operation), which can be obtained from the operation record table stored in the database.

[0043] The application area value is the area involved in each agricultural activity operation. For example, in a fertilization operation, after fertilization is completed, the land area fertilized in this fertilization is recorded, which can be obtained from the operation record table stored in the database.

[0044] The specific steps for analyzing the accounting and evaluation set of each cycle in the designated tobacco planting area are as follows: Based on the planting area value, tobacco plant biomass, tobacco plant respiration rate value, application rate value of each agricultural activity, and corresponding carbon emission factors and application area value of each cycle in the designated tobacco planting area, analyze the comprehensive carbon emission index of the corresponding cycle. Specifically, read the duration value and soil respiration rate value of each cycle, multiply the duration value, soil respiration rate value, and planting area value of the corresponding cycle to obtain the soil carbon emission value of the corresponding cycle, and multiply the duration value, soil respiration rate value, and planting area value of the corresponding cycle. The carbon emission value of the plant is obtained by multiplying the absorption rate values. The carbon emission values ​​of each agricultural activity in the corresponding period, along with the corresponding carbon emission factors and application area values, are multiplied together and summed to obtain the activity carbon emission value. The carbon emission values ​​of the soil, plant, and activity in the corresponding period are then weighted and summed to obtain the comprehensive carbon emission index for the corresponding period. Based on the carbon sink values ​​of tobacco plants and soil in each period of a designated tobacco planting area, the comprehensive carbon sink index for the corresponding period is analyzed (i.e., the carbon sink values ​​of tobacco plants and soil are weighted and summed).

[0045] The specific steps for obtaining the carbon emission impact factors and carbon sink impact factors for each cycle in a designated tobacco planting area are as follows: Obtain the tillage intensity, tillage frequency, and topsoil disturbance depth values ​​for each cycle in the designated tobacco planting area, and analyze the corresponding carbon emission impact factors (standardize the tillage intensity, tillage frequency, and topsoil disturbance depth values ​​for each cycle, and perform weighted processing based on the standardization results); Obtain the canopy uptake rate, relative chlorophyll content index, and soil structure stability index for each cycle in the designated tobacco planting area, and analyze the corresponding carbon sink impact factors (standardize the canopy uptake rate, relative chlorophyll content index, and soil structure stability index for each cycle, and perform weighted processing based on the standardization results).

[0046] The tillage intensity value is the sum of energy consumption during each operation within the cycle, reflecting the intensity of disturbance to the soil carbon cycle caused by tillage. The energy consumption during each operation is obtained by acquiring the total fuel consumption during each operation through a thermal sensor and obtaining the calorific value constant of the fuel type used stored in the database. The total fuel consumption is multiplied by the corresponding calorific value constant to obtain the energy consumption for each operation, and then summed to obtain the tillage intensity value.

[0047] The tillage frequency value is the cumulative number of tillage operations that occur within this time period, obtained from the agricultural production operation logs stored in the database.

[0048] The depth of soil disturbance in the topsoil is the vertical depth of soil disturbance caused by mechanical or manual operations. It is measured directly in the plot after tillage using a soil profile ruler (and uploaded to the database). Specifically, five representative points are selected, and after each mechanical or manual operation, the depth of disturbance traces is observed by excavating the profile and the average value is taken as the depth of soil disturbance in the topsoil.

[0049] Canopy absorptivity is the proportion of light energy absorbed by the canopy of a tobacco plant per unit area within the photosynthetic wavelength range (400–700 nm). It reflects the plant's light energy utilization efficiency and carbon absorption potential. It is obtained through a comparative method using photosynthetically active radiation (PAR) sensors installed in the field. Specifically, PAR sensors are placed above the tobacco canopy (incident light) and below the canopy (transmitted light) to collect the PAR values ​​of the upper and lower layers in real time per unit time. The real-time canopy absorptivity at each time point within this period is calculated as (incident PAR - transmitted PAR) / incident PAR. The average value is then used to obtain the canopy absorptivity.

[0050] The relative chlorophyll content index is an indicator of the chlorophyll concentration in the functional leaves of tobacco plants. It is measured by using a handheld SPAD meter (such as SPAD-502Plus) to measure the SPAD value of the third or fourth main functional leaf at the top of each tobacco plant. Each plant is sampled three times and the average is taken. At least 10 plants are sampled in each area, and the final average value is taken as the relative chlorophyll content index.

[0051] The soil structure stability index is the ability of soil aggregates (water-stable aggregates) to maintain structural integrity under external disturbances (such as rainfall and tillage). It reflects the soil's ability to protect organic matter and its carbon storage capacity. It is obtained through an indoor wet sieving test of aggregates. Specifically, the test involves collecting representative topsoil samples (0–20 cm), air-drying them, and sieving them through a 2 mm sieve. The soil samples are then classified for water stability using a wet sieving method to identify large, medium, and small aggregates (>2 mm, 0.25–2 mm, and <0.25 mm). The proportion of water-stable aggregates (>0.25 mm) is calculated, i.e., soil structure stability index = mass of water-stable aggregates / total mass of soil sample.

[0052] This implementation plan integrates on-site measurements and real-time sensor data to accurately reflect the growth status, respiration rate, carbon sequestration capacity, and soil carbon absorption of tobacco plants. This allows the system to capture changes in carbon emissions from various aspects, improving the accuracy of carbon accounting. Secondly, by combining agricultural activity usage with carbon emission factor calculations, it can not only analyze various carbon emission sources in the planting process in more detail but also assess the impact of different operations on carbon emissions, helping farmers optimize agricultural management measures. Finally, by analyzing carbon emission and carbon sequestration influencing factors, it can deeply analyze factors related to carbon emissions and carbon sequestration, thereby achieving a comprehensive assessment of the carbon emission process in tobacco planting areas and further improving the accuracy of carbon accounting.

[0053] Specifically, soil environmental data includes soil thermal diffusivity, soil environmental regulation factors, root decomposition rate, soil electrical conductivity gradient, and soil dissolved oxygen concentration. The specific steps for analyzing soil environmental correction factors for each cycle in the designated tobacco planting area are as follows: Based on the soil environmental data for each cycle in the designated tobacco planting area, analyze the corresponding soil environmental assessment set, including soil energy regulation factors (the indirect regulatory capacity of soil temperature, moisture, and oxygen dynamics on microbial carbon metabolism within the tobacco planting area) and soil carbon transformation factors (the dynamic distribution trend of organic carbon in the soil from gaseous carbon emissions to stable carbon pools, belonging to internal driving factors of carbon processes); obtain the climate factors for each cycle in the designated tobacco planting area, and combine them with the corresponding soil environmental assessment set to analyze the corresponding soil environmental correction factors (i.e., weighting the soil energy regulation factors, soil carbon transformation factors, and climate factors for each cycle, and the result is the soil environmental correction factor for the corresponding cycle, and the weighting process corresponds to the soil energy regulation factors, soil carbon transformation factors, and climate factors). The weight coefficients can be obtained through the particle swarm optimization algorithm. Specifically, in the particle swarm optimization algorithm, each particle represents a possible solution. Each solution consists of the weight coefficients corresponding to the soil energy regulation factor, soil carbon conversion factor, and climate factor. Initially, the position and velocity of the particles are randomly generated, and these positions correspond to different weight combinations. The fitness of each particle is evaluated by the objective function, which is related to the optimization of tobacco emissions or carbon conversion efficiency. The particle swarm optimization algorithm measures the quality of each solution by calculating the objective function value. A smaller objective function value indicates a better solution. Based on the value of the fitness function, the particle swarm optimization algorithm updates the velocity and position of the particles. The particles will adjust their velocity according to their historical best position and global best position, thereby searching in the solution space and gradually approaching the optimal solution. After multiple iterations, they gradually converge until the termination condition is met, such as reaching the maximum number of iterations or the fitness change being less than a preset threshold. After multiple iterations and updates, the particles will eventually find a set of optimal solutions, namely the weight coefficients corresponding to the soil energy regulation factor, soil carbon conversion factor, and climate factor.

[0054] Among them, the soil thermal diffusivity value is the soil's ability to transfer heat within a given period, reflecting the impact of soil temperature changes on microbial metabolism and carbon cycling. It is obtained through a sampling method, specifically by selecting areas within a designated tobacco planting region and setting multiple depth layers (e.g., 10cm, 20cm, and 30cm below the surface) along the vertical soil profile. Using a high-precision soil temperature sensor (e.g., Pt100 or a thermistor sensor) and a soil heat flux sensor (e.g., an HFP01 heat flux plate), the soil temperature and vertical heat flux at each set time window within that period are obtained. The system obtains the soil density value, and simultaneously acquires the soil density value (measured by a portable soil density meter and uploaded to the database) and specific heat capacity value (obtained through the soil texture specific heat capacity table stored in the database) for the corresponding set sampling time window (e.g., one day). Based on the soil temperature of each depth layer within the time window, the system calculates the temperature gradient value (i.e., standard deviation) and performs comprehensive analysis to obtain the thermal diffusivity for each set time window, i.e., vertical heat flux density value / (soil density value × specific heat capacity value × temperature gradient value). The system then performs mean processing, and the result is the soil thermal diffusivity value.

[0055] The soil environment regulation factor is the comprehensive influence of soil environmental conditions on the carbon transformation process within a given period. It is obtained by using soil temperature sensors, soil pH sensors, and soil moisture sensors to acquire soil temperature, soil pH, and soil moisture values ​​at each time point within the period, and then normalizing these values. The results are then weighted based on the normalization results, and the final result is the soil environment regulation factor.

[0056] The root decomposition rate is the rate at which organic matter in plant roots decomposes within a given period. It characterizes the speed at which root residues (such as dead roots and rhizosphere exudates) are decomposed by microorganisms in the soil, affecting the storage and release of soil organic carbon. It can be obtained through sampling, which involves selecting plant roots along with the attached soil within a designated tobacco planting area, excavating them as a whole, and dividing them into two parts of equal mass. One part is placed in a sealed container in a soil microbial activity culture medium and placed in a temperature-controlled incubator to simulate natural soil environmental conditions. The amount of carbon dioxide released from the culture medium is collected based on a preset sampling window (e.g., 30 minutes), and the carbon dioxide concentration is quantitatively detected using an infrared gas analyzer. The other part is dried to determine the initial dry weight of the root residues. The root decomposition rate within this preset sampling window is calculated as the amount of carbon dioxide released divided by the initial dry weight of the root residues. Multiple samplings are performed, and the average value is taken to obtain the root decomposition rate value.

[0057] The soil electrical conductivity gradient value is the rate of change of electrical conductivity between different soil depths. It is obtained by deploying multiple electrical conductivity sensors in the soil at different depths to acquire the electrical conductivity value at each time point within the cycle, and then processing the standard deviation. Based on the standard deviation processing result, the mean is processed, and the result is the soil electrical conductivity gradient value.

[0058] Soil dissolved oxygen concentration is the content of dissolved oxygen in soil pore water during a given period, which affects the aerobic respiration activity of soil microorganisms and the decomposition rate of organic matter. The dissolved oxygen concentration is measured at each time point during the period by burying dissolved oxygen sensors in the soil and then averaging the values. The result is the soil dissolved oxygen concentration value.

[0059] The specific steps for analyzing the soil environmental assessment set for each cycle of the tobacco planting area are as follows: Based on the soil thermal diffusivity, soil environmental regulation factors, and soil electrical conductivity gradient values ​​for each cycle of the tobacco planting area, analyze the soil energy regulation factors for the corresponding cycle (standardize the soil thermal diffusivity, soil environmental regulation factors, and soil electrical conductivity gradient values ​​for the corresponding cycle, and perform weighted processing based on the standardization results); Based on the root decomposition rate and soil dissolved oxygen concentration values ​​for each cycle of the tobacco planting area, analyze the soil carbon transformation factors for the corresponding cycle (standardize the root decomposition rate and soil dissolved oxygen concentration values ​​for the corresponding cycle, and perform weighted processing based on the standardization results).

[0060] The specific steps for obtaining the climate factors for each cycle in a designated tobacco growing area are as follows:

[0061] The system acquires the following data for each cycle in the designated tobacco planting area: air temperature (obtained through a temperature sensor for each set time window, averaged, and the result is the air temperature value), air humidity (obtained through a humidity sensor for each set time window, averaged, and the result is the air humidity value), rainfall (acquired through a tipping bucket rain gauge installed in the planting area, counting the number of tipping buckets within the cycle and converting the result into the corresponding rainfall amount, which is then uploaded to the database), and light intensity (acquired through a light sensor for each set time window, averaged, and the result is the air light intensity value). These data are then standardized, and a weighted average is applied based on the standardized results to obtain the corresponding climate factors for each cycle.

[0062] This implementation plan analyzes multiple soil environmental data to comprehensively understand the regulatory role of soil in microbial activity, carbon metabolism, and carbon transformation processes. For example, soil thermal diffusivity can help reveal the impact of soil temperature changes on carbon metabolism, while root decomposition rate can accurately reflect the decomposition rate of plant residues in the soil, correcting for soil carbon storage and release. Through standardization and weighting, various parameters are integrated into soil energy regulation factors and soil carbon transformation factors, thus providing a more intuitive and clearer view of the specific impact of soil on carbon emissions and carbon sinks in each cycle, which in turn helps to optimize carbon management in tobacco planting areas.

[0063] Specifically, the tobacco growth image sequence data includes tobacco growth image data for each time window (e.g., 2 days) (the image is a multispectral image), each including multi-band pixel values ​​(e.g., red light, green light, blue light, near-infrared reflectance, etc.) and two-dimensional coordinates for each pixel. The specific steps for analyzing the growth correction factor for each cycle of the designated tobacco planting area are as follows: Input the tobacco growth image data for each time window of each cycle of the designated tobacco planting area into a pre-trained growth status recognition model for comprehensive analysis to obtain the corresponding cycle's growth correction evaluation set, including carbon absorption index, tobacco vigor index, and tobacco light adaptation index; Based on the growth evaluation set for each cycle of the designated tobacco planting area, analyze the corresponding cycle's growth correction factor.

[0064] The specific formula for calculating the growth correction factor for a given period in a tobacco-growing region is as follows: ;in, To determine the growth correction factor for a specific period in a tobacco-growing region. To determine the carbon uptake index for a tobacco-growing region over a specific period. The absorption coefficient is stored in the database. To establish a tobacco viability index for a specific period in a tobacco-growing region. The vitality coefficient is stored in the database. To determine the tobacco light adaptation index for a specific period in a tobacco-growing region. The fitness coefficients are stored in the database. The adjustment coefficient is stored in the database and is set to 3.000 in this implementation example.

[0065] It needs to be explained that the absorption coefficient is stored in the database. The acquisition steps are as follows: Chlorophyll index features for each time window of each period are extracted based on the feature extraction sub-network. The mean and standard deviation of the chlorophyll index features are analyzed separately, and a ratio is calculated: standard deviation of chlorophyll index features / mean of chlorophyll index features. This ratio is used as the absorption coefficient. ;

[0066] Vitality coefficients stored in the database The acquisition steps are as follows: Based on the feature extraction sub-network, extract the stem diameter and plant height features for each time window of each cycle. Then, calculate the change rate of stem diameter and plant height features for adjacent time windows, and perform mean processing. Finally, weight the results based on the mean processing and use this weighted average as the vitality coefficient. ;

[0067] Fitness coefficients stored in the database The acquisition steps are as follows: The spectral response intensity features of each time window in each period are extracted based on the feature extraction sub-network, and standard deviation processing is performed. The results of the standard deviation processing are then normalized, and the normalized result is used as the fitness coefficient. .

[0068] The following is a specific implementation example of calculating the growth correction factor for a specific period in a designated tobacco growing area. The available data includes the carbon uptake index, tobacco vigor index, and tobacco light adaptation index for three periods in the designated tobacco growing area. (Details are as follows...) Figure 3-5 As shown:

[0069] Table 1. Example of time series data for setting up growth correction assessment in tobacco planting areas.

[0070] Carbon Absorption Index Tobacco Vitality Index Tobacco light adaptation index Period 1 0.712 0.734 0.684 Period 2 0.764 0.823 0.791 Period 3 0.812 0.857 0.826

[0071] Absorption coefficients stored in the database Approximately 0.137;

[0072] Vitality coefficients stored in the database Approximately 0.189;

[0073] Fitness coefficients stored in the database Approximately 0.174;

[0074] The adjustment coefficient stored in the database is 3.000;

[0075] Substituting the data from Table 1 and the coefficients mentioned above into the specific formula for calculating the growth correction factor for a given period in the tobacco planting area, we obtain:

[0076] Set the growth correction factor for the first cycle of the tobacco growing area = ;

[0077] Set the growth correction factor for the second cycle in the tobacco growing area = ;

[0078] Set the growth correction factor for the third cycle in the tobacco growing region = .

[0079] The growth status recognition model includes a feature extraction subnetwork, a temporal state modeling subnetwork, and a decoding output subnetwork. The specific steps for analyzing the growth evaluation set of each cycle in the designated tobacco planting area are as follows: In the feature extraction subnetwork of the growth status recognition model, tobacco growth image data for each time window of each cycle in the designated tobacco planting area are received and feature encoding is performed to obtain the feature vector of the corresponding time window. Specifically, based on the chlorophyll index feature extraction layer, plant density feature extraction layer, leaf area index feature extraction layer, plant height feature extraction layer, stem diameter feature extraction layer, spectral response intensity feature extraction layer, and texture uniformity feature extraction layer, the tobacco growth image data are respectively subjected to corresponding feature extraction, and the extracted features are combined into a feature vector.

[0080] In the temporal state modeling sub-network of the growth status recognition model, temporal analysis is performed on the feature vectors of each time window of each cycle in the tobacco planting area to obtain the temporal feature vectors of the corresponding cycle. Specifically, temporal correlation analysis is performed on the feature vectors of each time window based on the temporal convolutional network (TCN). For example, for the plant height feature of each time window, the rate of change of its adjacent time windows is calculated, such as (plant height feature of the first time window - plant height feature of the second time window) / plant height feature of the second time window, and then processed to extract the growth rate feature. For the spectral response intensity feature of each time window, the standard deviation is processed to extract the spectral stability. For the chlorophyll index feature, plant density feature, leaf area index feature, stem diameter feature, and texture uniformity of each time window, the mean is processed to obtain their corresponding feature stability, and these are combined into a temporal feature vector.

[0081] In the decoding output sub-network of the growth status recognition model, the temporal feature vector of each cycle of the set tobacco planting area is predicted to obtain the growth evaluation set of the corresponding cycle. Based on the carbon absorption index output layer, tobacco vigor index output layer, and tobacco light adaptation index output layer, the corresponding features of the temporal feature vector are extracted and weighted to output the carbon absorption index, tobacco vigor index, and tobacco light adaptation index.

[0082] The feature extraction subnetwork includes a chlorophyll index feature extraction layer, a plant density feature extraction layer, a leaf area index feature extraction layer, a plant height feature extraction layer, a stem diameter feature extraction layer, a spectral response intensity feature extraction layer, and a texture uniformity feature extraction layer.

[0083] The chlorophyll index feature extraction layer extracts the reflectance values ​​of red and near-infrared bands for each pixel of the tobacco growth image within each time window and analyzes them, i.e., (reflectance value of near-infrared band - reflectance value of red band) / (reflectance value of near-infrared band + reflectance value of red band). After obtaining the results, the mean value is processed to extract the chlorophyll index feature.

[0084] The plant density feature extraction layer extracts pixel values ​​from the green band (G channel) and near-infrared band (NIR) of the multispectral image. The plant region is binarized by a set threshold to obtain a binarized image. Morphological closing operation is performed on the binarized image, which includes dilation followed by erosion. Connected component labeling is performed on the closing operation result using structuring elements (such as a 5x5 square kernel), that is, all interconnected plant clusters in the image are identified. Each plant cluster corresponds to an independent plant region. The Euclidean distance (distance between cluster centers) between each plant cluster is calculated and the average value is taken, thus extracting the plant density feature.

[0085] The leaf area index feature extraction layer uses edge detection algorithms (such as Canny edge detection) to extract the leaf contours in the image, calculates the projected area of ​​the leaf contours, and converts it into the actual area value by combining the physical proportion of the image. It also analyzes the ratio of the total leaf area to the actual plot area to extract the leaf area index feature.

[0086] The plant height feature extraction layer uses edge detection algorithms (such as the Canny operator) to detect several plant contours. For each plant contour, a skeleton extraction algorithm is used to locate the main trunk centerline and determine the pixel positions corresponding to the plant base (lowest point of the bottom contour) and top (highest point). Through camera intrinsic parameters (focal length, principal point position) and extrinsic parameters (camera position, pose), a mapping relationship between image pixel coordinates and actual three-dimensional spatial coordinates is established (converting the two-dimensional coordinates of pixels in the image into spatial coordinates in the real world). The actual vertical height between the base and top pixels of the plant is calculated and averaged to extract the plant height feature.

[0087] The stem diameter feature extraction layer extracts several main stem contour lines based on edge detection algorithms (such as Canny edge detection), determines the longitudinal axis of the main stem, which is the longest continuous main stem region in the image. At multiple equally spaced sampling points along the longitudinal axis, the corresponding cross-sectional contours are extracted, forming a series of cross-sectional slices. For each cross-sectional slice, the maximum width of the contour is calculated as the stem diameter pixel value at that location. Statistical methods (such as removing outliers exceeding the mean ± 2 standard deviations) are used to filter out measurement errors, resulting in an effective set of cross-sectional diameters. A mapping relationship between image pixel coordinates and actual three-dimensional spatial coordinates is established. The effective stem diameter pixel values ​​in the effective set of cross-sectional diameters are converted into physical dimensions in millimeters or centimeters and mean-valued to extract the stem diameter feature.

[0088] The spectral response intensity feature extraction layer extracts the reflectance values ​​of the red, green, blue, and near-infrared bands for each pixel, analyzes the mean and standard deviation of each band, and performs weighted processing based on the analysis results to extract the spectral response intensity features.

[0089] The texture uniformity feature extraction layer calculates the texture features of the leaf region in the image based on the gray-level co-occurrence matrix (GLCM), including contrast, energy, entropy, etc., and performs mean processing on the values ​​in the texture features to extract texture uniformity features.

[0090] The decoding output subnetwork includes a carbon absorption index output layer, a tobacco viability index output layer, and a tobacco light adaptation index output layer.

[0091] The carbon absorption index output layer extracts the stability of chlorophyll index characteristics, plant density characteristics, and growth rate characteristics, and performs weighted processing to output the carbon absorption index. This index characterizes the efficiency of photosynthesis and carbon fixation capacity of tobacco plants during their growth period, reflects the sustainability of chlorophyll content, the rationality of plant spatial distribution, and the overall growth rate, and thus evaluates the utilization of environmental carbon resources and growth vitality of tobacco.

[0092] The tobacco vigor index output layer extracts the stability of leaf area index, plant height, and stem diameter features, and performs weighted processing to output the tobacco vigor index, which characterizes the growth health status and physical strength of tobacco plants, reflects the integrity of leaf unfolding, the vertical growth potential of the plant, and the thickness of the stem, thereby comprehensively evaluating the growth vigor of tobacco plants.

[0093] The tobacco light adaptability index output layer extracts spectral stability and texture uniformity feature stability, and performs weighted processing to output the tobacco light adaptability index. This index characterizes the tobacco plant's adaptability to the light environment and the uniformity of its spectral reflectance features. It reflects the stable reflectance state of tobacco leaves in different spectral bands and the uniform distribution of surface texture, thereby evaluating the physiological regulation and light energy utilization efficiency of tobacco under varying light conditions.

[0094] Furthermore, the pre-training process of the growth status recognition model is as follows:

[0095] In the pre-training stage of the growth status recognition model, it is first necessary to construct a training dataset that covers multi-dimensional growth features. Historical multi-source monitoring image data of tobacco planting areas under different growth stages and environmental conditions are collected. Data categories include, but are not limited to: environmental status data (such as temperature, humidity, light intensity, soil moisture, etc.), plant growth status data (such as chlorophyll index, plant density, leaf area index, plant height, stem diameter, etc.), and time series data (such as daily image data sequences, dynamic growth rate, etc.). For each type of data, time-series segmentation, image enhancement (such as rotation, scaling, cropping, etc.) and feature labeling operations are performed.

[0096] After data preparation is completed, each sub-network is pre-trained independently. The pre-training process adopts a supervised learning approach, inputting standardized time-series data of the corresponding type (such as image data of time windows and their corresponding growth features), and outputting known growth trends and feature labels.

[0097] The pre-training process of each sub-network uses the mean squared error (MSE) loss function to evaluate the error of various feature outputs and uses the Adam optimizer to adjust the network weights to ensure that the network can effectively capture subtle feature changes during the growth process.

[0098] After the independent pre-training of each sub-network is completed, the weight parameters obtained from the training are loaded into the complete growth status recognition model as initial values, and the multi-dimensional feature fusion optimization stage is entered. In this stage, joint training samples containing multiple cycles and multiple growth status features are constructed. Growth features of different dimensions (such as chlorophyll index, plant density, leaf area index, etc.) are synchronously input into the temporal state modeling sub-network, the obtained temporal feature vector is extracted, and it is input into the decoding output sub-network. In the decoding output sub-network, a fully connected fusion mechanism is used to jointly model the multi-source temporal features. The focus is on optimizing the connection weight parameters between the fusion layer and each output layer (such as the carbon absorption index output layer, tobacco vitality index output layer, and tobacco light adaptation index output layer). The training objective is to minimize the overall growth status assessment error.

[0099] Through this stage of optimization training, the model is equipped with the ability to accurately extract, fuse, and classify growth features under different growth states and environmental changes, thereby improving its accuracy in recognizing complex growth patterns. Ultimately, the trained model can quickly and accurately identify and evaluate growth patterns in practical applications.

[0100] In this implementation scheme, multispectral images from multiple time windows within each cycle are analyzed to ensure the temporal continuity of the data, thereby helping to comprehensively reflect the actual growth status of tobacco plants. Secondly, a pre-trained growth status recognition model is used to extract carbon absorption index, tobacco vigor index, and light adaptation index, which enables multi-dimensional assessment of the plant's photosynthetic efficiency, growth capacity, and environmental adaptability. At the same time, the absorption coefficient, vigor coefficient, and adaptation coefficient are all derived from the statistical characteristics of historical cycle data, making the current cycle assessment results highly comparable and thus improving the accuracy of tobacco growth assessment.

[0101] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0102] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A tobacco carbon accounting parameter analysis and prediction system, characterized in that, include: The initial carbon accounting analysis module is used to continuously acquire carbon accounting parameter sets for several cycles in a set tobacco planting area and analyze the initial carbon emission index for the corresponding cycle. The soil environment correction analysis module is used to acquire soil environment data for each period in a designated tobacco planting area and analyze the corresponding soil environment correction factors. The soil environment data includes soil thermal diffusivity, soil environment regulation factors, root decomposition rate, soil electrical conductivity gradient, and soil dissolved oxygen concentration. The specific steps for analyzing the soil environment correction factors for each period in the designated tobacco planting area are as follows: Based on soil environmental data for each cycle in a designated tobacco planting area, the corresponding soil environmental assessment set was analyzed, including soil energy regulation factors and soil carbon transformation factors. The climate factors for each cycle of the designated tobacco planting area are obtained, and the corresponding soil environmental assessment set is combined with the soil environmental correction factors for the corresponding cycle to analyze the soil environmental correction factors for the corresponding cycle. The growth correction analysis module is used to acquire tobacco growth image sequence data for each cycle in a designated tobacco planting area. Based on a pre-trained growth status recognition model, it analyzes the growth correction factor for the corresponding cycle. The specific steps are as follows: The growth status recognition model includes a feature extraction subnetwork, a temporal state modeling subnetwork, and a decoding output subnetwork. In the feature extraction subnetwork, tobacco growth image data for each time window of each cycle in a defined tobacco planting area are received and feature encoding is performed to obtain the feature vector of the corresponding time window. In the temporal state modeling subnetwork, the feature vectors of each time window of each cycle in the defined tobacco planting area are analyzed in a temporal manner to obtain the temporal feature vectors of the corresponding cycle. In the decoding output sub-network, the temporal feature vector of each cycle of the set tobacco planting area is predicted to obtain the growth evaluation set of the corresponding cycle, including carbon absorption index, tobacco vigor index and tobacco light adaptation index. Based on the growth assessment set of each cycle in the designated tobacco planting area, the growth correction factor for the corresponding cycle is analyzed. The predictive analysis module is used to analyze the comprehensive tobacco carbon emission index for each cycle of a set tobacco planting area based on the initial carbon emission index, soil environment correction factor, and growth correction factor, and to predict the comprehensive tobacco carbon emission index for the next cycle. The prediction and feedback management module is used to implement pre-set management measures for designated tobacco planting areas based on the comprehensive tobacco carbon emission index for the next cycle.

2. The tobacco carbon accounting parameter analysis and prediction system according to claim 1, characterized in that, The carbon accounting parameter set includes planting area value, tobacco plant biomass value, tobacco plant respiration rate value, tobacco plant carbon sink value, soil carbon sink value, usage value of each agricultural activity, and corresponding carbon emission factors and application area value. The specific steps for analyzing and setting the initial carbon emission index for each cycle of the tobacco planting area are as follows: Based on the carbon accounting parameter set for each cycle of a tobacco planting area, the corresponding accounting assessment set is analyzed, including the comprehensive carbon emission index and the comprehensive carbon sink index. The carbon emission impact factor and carbon sink impact factor of the designated tobacco planting area for each cycle are obtained, and a comprehensive analysis is performed with the emission assessment set of the corresponding cycle to obtain the initial carbon emission index of the corresponding cycle.

3. The tobacco carbon accounting parameter analysis and prediction system according to claim 2, characterized in that, The specific steps for analyzing and setting the accounting and evaluation set for each cycle of the tobacco planting area are as follows: Based on the planting area, tobacco plant biomass, tobacco plant respiration rate, amount of each agricultural activity, and corresponding carbon emission factors and application area values ​​for each cycle in the designated tobacco planting area, the comprehensive carbon emission index for the corresponding cycle is analyzed. Based on the carbon sequestration values ​​of tobacco plants and soil in a designated tobacco planting area for each cycle, the comprehensive carbon sequestration index for the corresponding cycle is analyzed.

4. The tobacco carbon accounting parameter analysis and prediction system according to claim 2, characterized in that, The specific steps to obtain the carbon emission impact factor and carbon sink impact factor for each cycle in a designated tobacco planting area are as follows: The tillage intensity, tillage frequency, and tillage disturbance depth values ​​for each cycle in a designated tobacco planting area were obtained, and the carbon emission impact factors for the corresponding cycle were analyzed. The canopy uptake rate, relative chlorophyll content index, and soil structure stability index of the designated tobacco planting area were obtained for each cycle, and the carbon sink influencing factors of the corresponding cycle were analyzed.

5. The tobacco carbon accounting parameter analysis and prediction system according to claim 1, characterized in that, The specific steps for analyzing and setting the soil environmental assessment set for each period in the tobacco planting area are as follows: Based on the soil thermal diffusivity, soil environmental regulation factors, and soil electrical conductivity gradient values ​​for each cycle in the tobacco planting area, the soil energy regulation factors for the corresponding cycle are analyzed. Based on the root decomposition rate and soil dissolved oxygen concentration values ​​for each cycle in the designated tobacco planting area, the soil carbon transformation factor for the corresponding cycle was analyzed.

6. The tobacco carbon accounting parameter analysis and prediction system according to claim 1, characterized in that, The specific formula for calculating the growth correction factor for a given period in a tobacco-growing region is as follows: ； in, , , , The parameters are, in order, growth correction factor, carbon uptake index, tobacco vigor index, and tobacco light adaptation index for a specific period in a tobacco-growing region. , , , The coefficients stored in the database are, in order: absorption coefficient, vitality coefficient, fitness coefficient, and regulation coefficient. The steps for obtaining the absorption coefficient stored in the database are as follows: Based on the feature extraction subnetwork, extract the chlorophyll index features of each time window of each cycle, and analyze the mean and standard deviation of the chlorophyll index features respectively, and perform ratio processing, and use the ratio processing result as the absorption coefficient. The steps for obtaining the vitality coefficient stored in the database are as follows: Based on the feature extraction subnetwork, extract the stem diameter feature and plant height feature of each time window of each cycle, and extract the stem diameter feature change rate and plant height feature change rate of adjacent time windows respectively, and perform mean processing. Based on the mean processing result, perform weighted processing to obtain the vitality coefficient. The steps for obtaining the fitness coefficient stored in the database are as follows: Extract the spectral response intensity features of each time window of each period based on the feature extraction sub-network, and perform standard deviation processing. Normalize the results based on the standard deviation processing, and use the normalized results as the fitness coefficients.

7. The tobacco carbon accounting parameter analysis and prediction system according to claim 1, characterized in that, The specific formula for calculating the comprehensive tobacco carbon emission index for a given tobacco-growing region over a specific period is as follows: ； in, , , , The components, in order, are the comprehensive tobacco carbon emission index, initial carbon emission index, soil environmental correction factor, and growth correction factor for a specific period in a tobacco-growing region. , , The parameters are, in order, the soil environmental coefficient, growth coefficient, and synergy coefficient stored in the database; The steps for obtaining the synergy coefficient stored in the database are as follows: obtain historical growth correction factors and historical soil environment correction factors for several historical periods, perform interactive analysis to obtain the interactive value for each historical period, and perform weighted average processing, using the weighted average processing result as the synergy coefficient.

Images