Carbon sink dynamic monitoring regulation and control system based on agricultural multiple scenes

Abstract

The invention discloses a carbon sink dynamic monitoring regulation and control system based on agricultural multiple scenes, and relates to the field of agricultural carbon sink monitoring, and the system comprises a data collection module which obtains carbon flux original data, crop physiological state parameter data and historical standard trend data of preset sampling points of a crop junction area; the feature construction module processes original data based on a time sliding window and generates a composite trend feature vector and a standard composite trend vector. The difference analysis module calculates trend deviation indexes, and the affiliation state judgment module divides clear affiliation, non-affiliation or cross overlapping states according to threshold values. The first verification module carries out repeated difference matching on the cross overlapping points, and the second verification module carries out fitting compensation processing on the non-attribution points through trend fluctuation stability calculation and adjacent points. The result output module is used for generating a structured judgment result for carbon sink monitoring and value estimation; according to the invention, carbon sink data fine attribution and high-credibility carbon flux dynamic monitoring in a complex boundary scene of an agricultural field can be realized.

Description

Technical Field

[0001] The present invention relates to the field of agricultural carbon sink monitoring, and specifically to a carbon sink dynamic monitoring and control system based on multiple agricultural scenarios. Background Art

[0002] As global climate change becomes increasingly prominent, agricultural carbon sinks, as an important carbon absorption pathway, have attracted widespread attention. Agricultural carbon sinks refer to the process of fixing atmospheric carbon dioxide through agricultural production activities and farmland ecosystems, thereby reducing greenhouse gas emissions.

[0003] Agricultural ecosystems are complex and changeable, especially in areas where multiple crops meet. Accurate monitoring and attribution of carbon fluxes face many challenges: First, it is impossible to effectively handle the attribution of carbon flux data in areas where crops meet, resulting in inaccurate carbon sink valuations; second, existing systems have difficulty handling abnormal sampling points that appear during data collection, which can easily lead to misjudgments or attribution issues. In addition, different crops exhibit large differences in carbon flux characteristics under the same environment and are sensitive to changes in physiological state, which can easily lead to misjudgments or attribution issues of carbon fluxes. The existing technology lacks a dynamic monitoring mechanism that can combine time series trend analysis and spatial consistency verification, making it difficult to achieve accurate judgments at the sampling point level.

[0004] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a carbon sink dynamic monitoring and control system based on multiple agricultural scenarios.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows: In a first aspect, the present invention discloses a carbon sink dynamic monitoring and control system based on multiple agricultural scenarios, comprising: A data acquisition module is used to obtain raw carbon flux data, crop physiological state parameter data, and historical standard trend data at preset sampling points in the crop boundary area of an agricultural field; A feature construction module is used to process the carbon flux raw data and crop physiological state parameter data based on a preset time sliding window to construct a composite trend feature vector; and to construct a standard composite trend vector of the target crop based on the historical standard trend data; a difference analysis module, configured to perform sliding window difference matching between the composite trend feature vector and the standard composite trend vector, and calculate a trend deviation index within a time window; An attribution status determination module is configured to determine whether the trend deviation index of the sampling point is lower than a first preset threshold value: if so, it is marked as a clear attribution state; otherwise, it is further determined whether it exceeds a second preset threshold value: if so, it is marked as an unattributable state; otherwise, it is marked as an overlapping state; A first verification module is used to repeatedly perform sliding window difference matching on the overlapping sampling points and re-mark them as a clear attribution state or an unattributable state; The second verification module is used to calculate the trend fluctuation stability of the sampling points that cannot be attributed and determine whether to discard them. If yes, they are discarded. Otherwise, the sampling points are further fitted with the consistency of the attribution of the neighboring points and determine whether to discard them. If yes, they are discarded. Otherwise, the sampling points are fitted with the neighboring points for compensation. The result output module is used to output a structured determination result including a state label and a crop type, and the structured determination result is used for carbon sink monitoring and valuation.

[0007] In a second aspect, the present invention discloses a method for dynamic monitoring and control of carbon sinks based on multiple agricultural scenarios, comprising the following steps: Obtaining raw carbon flux data, crop physiological parameter data, and historical standard trend data at preset sampling points in crop boundary areas of agricultural fields; Processing the carbon flux raw data and crop physiological state parameter data based on a preset time sliding window to construct a composite trend feature vector; constructing a standard composite trend vector for the target crop based on the historical standard trend data; Performing sliding window difference matching on the composite trend feature vector and the standard composite trend vector, and calculating a trend deviation index within a time window; Determine whether the trend deviation index of the sampling point is lower than a first preset threshold: if yes, mark it as a clear attribution state; otherwise, further determine whether it exceeds a second preset threshold: if yes, mark it as an unattributable state; otherwise, mark it as a cross-overlapping state; Repeat the sliding window difference matching for the overlapping sampling points and re-mark them as clearly attributed or unattributed; For the sampling points that cannot be assigned, the trend fluctuation stability calculation is performed to determine whether to discard them. If yes, they are discarded. Otherwise, the sampling point is further assigned to the neighboring points for consistency fitting and whether to discard them. If yes, they are discarded. Otherwise, the sampling point is fitted with the neighboring points for compensation. The output includes a structured determination result of the state label and the crop type, and carbon sink monitoring and valuation are performed based on the structured determination result.

[0008] Compared with the prior art, the present invention has the following beneficial effects: By fusing multidimensional physiological state parameters with carbon flux data to construct a composite trend feature vector, a comprehensive characterization of inter-crop differences is achieved, solving the problems of ambiguous sampling point attribution and easy overlap of trend features in traditional methods; by introducing trend deviation indicators and multi-level verification mechanisms, a clear attribution judgment mechanism is established, which can effectively divide the attribution status of sampling points; by introducing trend stability assessment and spatial consistency fitting for uncertain points, dynamic optimization and compensation are achieved, the robustness of the system is improved, and the problems of insufficient response of traditional models to temporal dynamic changes, inability to adapt to short-term fluctuations, and insufficient utilization of spatial neighborhood features are effectively solved; through structured result output, reliable data support is provided for subsequent carbon sink monitoring, modeling, policy making and other applications, ultimately achieving fine attribution of carbon sink data and high-reliability dynamic monitoring of carbon fluxes in complex boundary scenarios of agricultural fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0010] Figure 1 This is an overall block diagram of a system according to an embodiment of the present invention; Figure 2 This is a flowchart of the overall execution of the system according to the first embodiment of the present invention; Figure 3 This is an overall block diagram of the method according to the second embodiment of the present invention. DETAILED DESCRIPTION

[0011] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0012] Application Overview: In existing technologies, carbon sequestration monitoring of agricultural fields usually relies on average values or static model estimates in fixed areas, which makes it difficult to cope with dynamic changes in crop boundary areas. When different crop types are planted in an interlaced manner, local carbon flux characteristics are significantly affected by physiological states and environmental factors. Traditional methods cannot effectively distinguish the attribution relationship, resulting in deviations in monitoring data or inability to determine. For example, in areas where corn and soybeans are mixed, the carbon flux trends of the two crops show different fluctuation patterns under the same environmental conditions. Static models cannot capture the dynamic differences in the time dimension, resulting in misjudgment of attribution. To address these issues, it is necessary to design a monitoring system that can dynamically track carbon flux trends and integrate multi-source data. Analysis revealed that misidentification of crop boundary areas is primarily caused by trend feature mismatches and spatial fluctuations. Traditional methods only consider data comparisons at a single point in time, ignoring the patterns of trend changes within the time window. Based on this, the inventors proposed introducing a sliding window mechanism to dynamically fuse carbon flux and physiological parameters to construct a composite trend vector. At the same time, historical standard trends are combined for differential matching, and a multi-level verification mechanism is used to address overlap and inability to attribute.

[0013] Example 1: like Figure 1-2 As shown, the carbon sink dynamic monitoring and control system based on agricultural multi-scenario includes a data acquisition module, a feature construction module, a difference analysis module, an attribution status determination module, a first verification module, a second verification module and a result output module.

[0014] The data acquisition module is used to acquire raw carbon flux data, crop physiological parameter data, and historical standard trend data from pre-set sampling points within the crop boundary area of an agricultural field. Raw carbon flux data includes the CO2 exchange rate per unit area at a specific time-stamped sampling point; crop physiological parameter data includes leaf temperature, leaf surface wetness, and light intensity, acquired simultaneously with the carbon flux data; and historical standard trend data includes standard carbon flux trend vectors and standard physiological trend vectors for different crop types under pre-set environmental conditions. Raw carbon flux data is collected using an infrared gas analyzer, while leaf temperature in the crop physiological parameter data is collected using an infrared thermal imaging device, leaf surface wetness is acquired using a surface humidity sensor, and light intensity is collected using a photon sensor. Data collected by all sensors is aligned using a unified timestamp and a fixed sampling frequency.

[0015] In this embodiment, the infrared gas analysis device adopts the LI-6400XT portable photosynthesis measurement system, which has the ability to measure CO2 concentration with high precision, a measurement range of 0-3000μmol / mol, and an accuracy of ±1μmol / mol. The infrared thermal imaging device adopts the FLIR T640 thermal imager, with a temperature measurement range of -40°C to 500°C and a thermal sensitivity of 0.035°C. The surface humidity sensor adopts the SHT75 digital humidity sensor, with a measurement range of 0-100%RH and an accuracy of ±1.8%RH. The photon sensor adopts the LI-190R photon sensor, with a measurement range of 0-10000μmol / (m 2 s), with an accuracy of ±5%. The sampling frequency is set to once every 10 minutes to ensure the temporal continuity and integrity of the data.

[0016] Through the above technical solution, this application can accurately distinguish the dynamic attribution of carbon fluxes in crop boundary areas, reducing misjudgments due to mismatched trend characteristics. By matching trends within a time window and verifying spatial consistency, it addresses the issues of cross-overlap and fluctuation interference that traditional methods cannot handle. The resulting structured determination provides a reliable data foundation for carbon sink valuation and supports precise carbon sink monitoring in multiple agricultural scenarios.

[0017] The feature construction module processes raw carbon flux data and crop physiological parameter data within a preset time sliding window to construct a composite trend feature vector. A standard composite trend vector for the target crop is constructed based on historical standard trend data. When constructing the composite trend feature vector, a sliding average is applied to the carbon flux curve to smooth noise. Leaf temperature, humidity, and light parameters are each normalized to their minimum and maximum values, and then concatenated in a preset order to form a unified vector format. The standard composite trend vector is generated from data on similar crops in a historical database. This process involves time-normalizing the historical carbon flux data and physiological parameter data, then concatenating them into a historical composite trend vector.

[0018] In this embodiment, the preset time sliding window length is 24 hours, and the sliding step is 1 hour. The sliding average processing of the carbon flux change curve adopts a 5-point sliding average algorithm, that is, each data point takes the average value of itself and the two points before and after it, which effectively reduces the noise impact caused by short-term fluctuations. The minimum-maximum normalization processing maps each parameter value to the [0,1] interval, and the calculation formula is: X_norm=(X-X_min) / (X_max-X_min), where X is the original data, X_min and X_max are the minimum and maximum values of the data, respectively. The preset splicing order is carbon flux data, leaf temperature data, leaf surface humidity data, and light intensity data to form a unified vector format. The time normalization processing uniformly maps historical data of different lengths to the standard time axis to ensure the comparability of data of different time spans.

[0019] Traditional methods typically directly use raw data or a single parameter for trend analysis, failing to account for noise interference and dimensional differences between multi-source data. For example, a simple arithmetic average of carbon flux is performed without addressing the numerical distribution characteristics of parameters such as leaf temperature, resulting in an imbalance in the contributions of different parameters when matching trends. This approach, through a combination of sliding averaging and normalization, eliminates data acquisition noise and effectively integrates multiple parameters, enabling the composite vector to accurately reflect the coordinated changes in crop physiological state and carbon flux.

[0020] Through the above technical solution, this application effectively solves the problem of inaccurate trend characteristics caused by noise interference from multi-source data in crop boundary areas. Through standardization processing, parameters of different dimensions are made comparable, which significantly improves the discrimination accuracy of composite trend vectors in cross-crop type matching and provides a reliable data basis for subsequent attribution status determination.

[0021] The difference analysis module is used to perform sliding window difference matching between the composite trend feature vector and the standard composite trend vector, calculating the trend deviation index within the time window. The trend deviation index calculation process includes: setting a sliding window of fixed length; calculating three difference indicators between the two vectors within the sliding window: directional consistency: counting the proportion of time periods when the two vectors change in the same direction; inflection point position difference: calculating the mean time offset of the occurrence of the same inflection point; amplitude offset: calculating the mean absolute difference of the corresponding data points; and taking the weighted sum of these three difference indicators as the trend deviation index.

[0022] In this embodiment, the sliding window of fixed duration is set to 12 hours, and the sliding step is 1 hour. When calculating the directional consistency, the first-order difference is first calculated for each of the two vectors, and then the proportion of points with the same sign of the difference value to the total number of points is counted. The inflection point is defined as the point where the sign of the first-order difference changes, and the inflection point position difference is calculated as the average of the absolute difference between the corresponding inflection point timestamps in the two vectors. The amplitude offset is calculated as the average of the absolute value of the difference between the corresponding position data points of the two vectors. The weights of the three difference indicators are set as follows: directional consistency weight 0.5, inflection point position difference weight 0.3, amplitude offset weight 0.2. The calculation formula of the trend deviation index is: TDI=0.5×(1-DC)+0.3×IPD+0.2×AD, where TDI is the trend deviation indicator, DC is the directional consistency, IPD is the normalized inflection point position difference, and AD is the normalized amplitude offset.

[0023] Traditional methods typically only use single numerical differences or static trend matching, which cannot effectively distinguish between timing deviations caused by differences in crop growth stages and true abnormal fluctuations. Existing technologies often evaluate trend deviations based on fixed thresholds or single-dimensional indicators, such as only considering amplitude differences or directional consistency, which can easily lead to misjudgments. This solution, through a joint analysis of the three dimensions of direction, timing, and amplitude, can more accurately distinguish between natural growth fluctuations and abnormal deviations, significantly improving the accuracy of trend matching, especially in complex scenarios at the boundary of crops. Through the above technical solution, this application can effectively solve the problem of misjudgment of crop boundary areas caused by the superposition of time offset and amplitude fluctuation, and reduce the probability of misjudgment caused by the limitations of a single indicator through comprehensive evaluation of multi-dimensional trend characteristics.

[0024] The attribution status determination module is used to determine whether the trend deviation index of the sampling point is lower than the first preset threshold: if so, it is marked as a clear attribution state; otherwise, it is further determined whether it exceeds the second preset threshold: if so, it is marked as an unattributable state; otherwise, it is marked as a cross-overlapping state.

[0025] In this embodiment, the first preset threshold is set to 0.3, and the second preset threshold is set to 0.7. When the trend deviation index is lower than 0.3, it indicates that the carbon flux trend of the sampling point is highly consistent with the standard trend of a certain crop type and can be clearly attributed to that crop type; when the trend deviation index exceeds 0.7, it indicates that the carbon flux trend of the sampling point is significantly different from the standard trend of any known crop type and cannot be attributed to any known crop type; when the trend deviation index is between 0.3 and 0.7, it indicates that the carbon flux trend of the sampling point may be affected by multiple crop types and is in a cross-overlapping state.

[0026] Through the above technical solution, this application can effectively solve the problem of misidentification of sampling points in crop boundary areas due to environmental interference or mixed planting. Through a continuous multi-day trend feature verification mechanism, the accuracy of sampling point attribution in overlapping areas is improved, ensuring the temporal and spatial consistency of carbon flux monitoring results, and providing a reliable data foundation for subsequent carbon sink valuation.

[0027] The first verification module is configured to repeatedly perform sliding window difference matching on overlapping sampling points, re-marking them as either a clear attribution state or an unattributable state. The process of repeatedly performing sliding window difference matching on overlapping sampling points includes extracting a composite trend feature vector for the same time period between the previous and next day of the sampling point and recalculating a trend deviation index; if the trend deviation index remains below a first preset threshold for N consecutive days, the state is updated to a clear attribution state; otherwise, the state is updated to an unattributable state, where N is a positive integer.

[0028] In this embodiment, the N value is set to 3, meaning that the sampling point's attribution status is updated to clear attribution only if the trend deviation index remains below the first preset threshold of 0.3 for three consecutive days. This multi-day verification mechanism effectively eliminates the impact of short-term environmental fluctuations or measurement errors, improving the reliability of attribution determination. If a clear attribution status cannot be determined after three consecutive days of verification, the sampling point is updated to unattributable, thus avoiding carbon sink estimation errors caused by uncertainty.

[0029] Traditional methods rely solely on single-point-in-time or static spatial data for attribution, failing to effectively address data anomalies caused by dynamic changes in crop growth or local environmental disturbances. This approach, by integrating trend fluctuation stability analysis over time and verifying spatial neighborhood consistency, can distinguish between true overlapping areas and data anomalies, mitigating the impact of invalid data on carbon sink valuations. Through the above technical solution, this application can effectively filter out invalid sampling points caused by sensor noise and transient environmental interference, while retaining overlapping points that are actually located in crop boundary areas, thus avoiding carbon sink valuation bias caused by data misjudgment. For outliers with clear spatial attribution structure, a neighboring point data compensation mechanism can improve the spatial continuity of carbon flux attribution determination and ensure the reliability of monitoring results.

[0030] The second verification module is used to calculate the trend fluctuation stability of sampling points that cannot be assigned and determine whether to discard them. If so, they are discarded. Otherwise, the sampling point is further assigned a neighboring point consistency fit and determined whether to discard them. If so, they are discarded. Otherwise, the sampling point is fitted with neighboring points for compensation. The process of calculating the trend fluctuation stability of sampling points that cannot be assigned and determining whether to discard them includes: extracting the composite trend feature vector for a preset number of days before and after the current sampling point; counting the number of first-order difference direction changes within all sliding windows; and discarding the sampling point when the number of direction changes exceeds a preset ratio threshold of the total number of windows. The process of performing neighboring point consistency fit and determining whether to discard the sampling point includes: extracting the attribution determination results of the sampling point's neighbors. If the attribution ratio of a certain crop type exceeds a preset spatial consistency threshold, the spatial attribution structure is determined to be clear. Otherwise, the sampling point is discarded. The neighboring points are the eight neighborhood points centered on the current sampling point. The process of fitting and compensating the sampling point using neighboring points includes: if the attribution ratio of a certain crop type exceeds the preset spatial consistency threshold, the mean carbon flux of the neighboring points of the crop type in the current window is extracted, and the mean is used as the compensation value of the current sampling point and the attribution status is updated.

[0031] In this embodiment, the preset number of days is set to 7, meaning that the composite trend feature vectors for 3.5 days before and after the current sampling point are extracted to calculate trend fluctuation stability. The preset ratio threshold is set to 60%, meaning that when the number of direction changes exceeds 60% of the total number of windows, the data at that sampling point is considered to have excessive fluctuations and lack stability and is discarded. The spatial consistency threshold is set to 75%, meaning that when six or more of the eight neighboring points belong to the same crop type, the spatial attribution structure of the area where the sampling point is located is considered clear. The average carbon flux of the neighboring points is used as the compensation value for the current sampling point, and its attribution status is updated to clear attribution.

[0032] Traditional methods typically discard sampling points that cannot be attributed, resulting in reduced data utilization and increased bias in carbon sink estimates in border areas. This approach, however, introduces a spatial neighborhood compensation mechanism. While retaining valid data, it also uses the statistical characteristics of carbon flux trends at neighboring points to correct for outliers. This approach not only avoids the impact of isolated points on overall monitoring results, but also enhances the robustness of data assessment in border areas.

[0033] Through the above technical solution, this application can significantly improve the accuracy of carbon flux attribution determination in multiple agricultural scenarios, especially in areas with staggered crop planting. Through spatial consistency verification and data compensation, it can effectively reduce misjudgments or missed judgments caused by local data anomalies, ensuring the integrity of carbon sink monitoring results and the reliability of valuation calculations.

[0034] The result output module is used to output structured determination results, including the attribution status label and the crop type. These structured determination results are used for carbon sink monitoring and valuation. Specifically, the structured determination results include: an attribution status field that stores four types of labels: clear attribution, weak attribution, overlapping, and unattributable; and a crop type field that stores the crop code for the determined attribution. Carbon sink monitoring and valuation are performed for all sampling points with a clear attribution status field.

[0035] In this embodiment, the structured determination results are output in JSON format and include the following fields: sampling point ID, geographic coordinates (latitude and longitude), attribution status tag, crop type code, carbon flux value, timestamp, and credibility score. The crop type code uses the internationally recognized crop classification system, such as CR001 for rice, CW001 for wheat, and CM001 for corn. Carbon sink monitoring and valuation are based on data from sampling points with clearly defined attribution status. The carbon sink amount per unit area is calculated, and the carbon sink value is estimated based on the carbon sequestration coefficients of different crop types.

[0036] Traditional methods typically record only raw measurements or simple classification results, without establishing a standardized, structured output format for the determination results. Existing technologies lack a multi-level classification mechanism for attribution status, making it difficult to effectively filter high-confidence data for subsequent processing. Furthermore, the identification of crop types is confusing, making it difficult to support cross-regional data integration. This solution achieves standardized and scalable data output by defining structured fields and a coding system.

[0037] Through the above technical solution, this application effectively addresses the valuation error caused by ambiguous attribution of carbon flux data in crop boundary areas. A structured labeling system enables data tiering, reducing the impact of low-quality data on monitoring results. Furthermore, standardized crop coding fields provide a foundation for multi-scenario data fusion, supporting dynamic monitoring and accurate valuation of agricultural carbon sinks.

[0038] Example 2: like Figure 3 As shown, the carbon sink dynamic monitoring and control method based on agricultural multi-scenario includes the following steps: obtaining carbon flux raw data, crop physiological state parameter data and historical standard trend data of preset sampling points in the crop boundary area of the agricultural field; Processing the carbon flux raw data and crop physiological state parameter data based on a preset time sliding window to construct a composite trend feature vector; constructing a standard composite trend vector for the target crop based on the historical standard trend data; Performing sliding window difference matching on the composite trend feature vector and the standard composite trend vector, and calculating a trend deviation index within a time window; Determine whether the trend deviation index of the sampling point is lower than a first preset threshold: if yes, mark it as a clear attribution state; otherwise, further determine whether it exceeds a second preset threshold: if yes, mark it as an unattributable state; otherwise, mark it as a cross-overlapping state; Repeat the sliding window difference matching for the overlapping sampling points and re-mark them as clearly attributed or unattributed; For the sampling points that cannot be assigned, the trend fluctuation stability calculation is performed to determine whether to discard them. If yes, they are discarded. Otherwise, the sampling point is further assigned to the neighboring points for consistency fitting and whether to discard them. If yes, they are discarded. Otherwise, the sampling point is fitted with the neighboring points for compensation. The output includes a structured determination result of the state label and the crop type, and carbon sink monitoring and valuation are performed based on the structured determination result.

[0039] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.

[0040] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0041] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A carbon sink dynamic monitoring and control system based on multiple agricultural scenarios, characterized by: include: A data acquisition module is used to obtain raw carbon flux data, crop physiological state parameter data, and historical standard trend data at preset sampling points in the crop boundary area of an agricultural field; A feature construction module is used to process the carbon flux raw data and crop physiological state parameter data based on a preset time sliding window to construct a composite trend feature vector; constructing a standard composite trend vector of the target crop based on the historical standard trend data; a difference analysis module, configured to perform sliding window difference matching between the composite trend feature vector and the standard composite trend vector, and calculate a trend deviation index within a time window; An attribution status determination module is configured to determine whether the trend deviation index of the sampling point is lower than a first preset threshold value: if so, it is marked as a clear attribution state; otherwise, it is further determined whether it exceeds a second preset threshold value: if so, it is marked as an unattributable state; otherwise, it is marked as an overlapping state; A first verification module is used to repeatedly perform sliding window difference matching on the overlapping sampling points and re-mark them as a clear attribution state or an unattributable state; The second verification module is used to calculate the trend fluctuation stability of the sampling points that cannot be attributed and determine whether to discard them. If yes, they are discarded. Otherwise, the sampling points are further fitted with the consistency of the attribution of the neighboring points and determine whether to discard them. If yes, they are discarded. Otherwise, the sampling points are fitted with the neighboring points for compensation. The result output module is used to output a structured determination result including a state label and a crop type, and the structured determination result is used for carbon sink monitoring and valuation.

2. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 1 is characterized by: The carbon flux raw data includes: CO2 exchange rate at a specific time stamp sampling point per unit area; the crop physiological state parameter data includes leaf temperature, leaf surface humidity and light intensity acquired synchronously with the carbon flux data; the historical standard trend data includes the standard carbon flux trend vector and standard physiological state trend vector of different crop types under preset environmental conditions. The carbon flux raw data is collected by an infrared gas analysis device, the leaf temperature in the crop physiological state parameter data is collected by an infrared thermal imaging device, the leaf surface humidity is obtained by a surface humidity sensor, and the light intensity is collected by a photon sensor; The data collected by all sensors are aligned with a unified timestamp and the sampling frequency is fixed.

3. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 2 is characterized by: The specific process of constructing the composite trend vector includes: When constructing the composite trend feature vector, the carbon flux change curve was processed using sliding average to smooth the noise, and the leaf temperature, humidity and light parameters were normalized to the minimum and maximum values respectively, and then spliced in a preset order to form a unified vector format.

4. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 3 is characterized by: The standard composite trend vector is generated from the same type of crop data in the historical database. The generation process includes respectively performing time normalization processing on the historical carbon flux data and the physiological state parameter data and then splicing them into the historical composite trend vector.

5. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 4 is characterized by: The calculation process of the trend deviation indicator includes: Set a sliding window of fixed length; calculate three difference indicators of the two vectors within the sliding window: Directional consistency: counts the proportion of time periods when two vectors change in the same direction; Inflection point position difference: calculate the mean time offset of the same inflection point; Amplitude offset: Calculate the mean of the absolute differences of the corresponding data points; The weighted sum of the three difference indices is used as the trend deviation indicator.

6. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 5 is characterized by: The process of repeatedly performing sliding window difference matching on the cross-overlapping sampling points includes: Extract the composite trend feature vector of the same period of the day before and the day after the sampling point and recalculate the trend deviation index; If the trend deviation indicator continues to be lower than the first preset threshold for N consecutive days, it is updated to the clear attribution state; otherwise, it is updated to the unattributable state, where N is a positive integer.

7. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 6 is characterized by: The process of calculating the trend fluctuation stability of the sampling points that cannot be assigned and determining whether to discard them includes: extracting the composite trend feature vector of a preset number of days before and after the current sampling point; counting the number of first-order difference direction changes in all sliding windows; and discarding the sampling point when the number of direction changes exceeds a preset ratio threshold of the total number of windows; The process of fitting the attribution consistency of neighboring points for the sampling point and determining whether to discard the sampling point includes: extracting the attribution determination results of the neighboring points of the sampling point; if the attribution ratio of a certain crop type exceeds a preset spatial consistency threshold, then the spatial attribution structure is determined to be clear; otherwise, the sampling point is discarded; the neighboring points are eight neighborhood points centered on the current sampling point.

8. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 7 is characterized by: The process of fitting and compensating the sampling point using neighboring points includes: if the attribution ratio of a certain crop type exceeds a preset spatial consistency threshold, extracting the average carbon flux of the neighboring points of the crop type in the current window, using the average as the compensation value of the current sampling point and updating the attribution status.

9. The carbon sink dynamic monitoring and control system based on agricultural multi-scenario according to claim 8 is characterized by: The structured determination results specifically include: Attribution status field: stores four types of labels: clear attribution, weak attribution, overlap, and unattributable. Crop type field: stores the crop code to determine the crop type; Carbon sink monitoring and valuation are carried out for all sampling points with clear attribution in the attribution status field.

10. A method for dynamic monitoring and control of carbon sinks based on multiple agricultural scenarios based on the system according to any one of claims 1 to 9, comprising the following steps: Obtaining raw carbon flux data, crop physiological parameter data, and historical standard trend data at preset sampling points in crop boundary areas of agricultural fields; Processing the carbon flux raw data and crop physiological state parameter data based on a preset time sliding window to construct a composite trend feature vector; constructing a standard composite trend vector of the target crop based on the historical standard trend data; Performing sliding window difference matching on the composite trend feature vector and the standard composite trend vector, and calculating a trend deviation index within a time window; Determine whether the trend deviation index of the sampling point is lower than a first preset threshold: if yes, mark it as a clear attribution state; otherwise, further determine whether it exceeds a second preset threshold: if yes, mark it as an unattributable state; otherwise, mark it as a cross-overlapping state; Repeat the sliding window difference matching for the overlapping sampling points and re-mark them as clearly attributed or unattributed; For the sampling points that cannot be assigned, the trend fluctuation stability calculation is performed to determine whether to discard them. If yes, they are discarded. Otherwise, the sampling point is further assigned to the neighboring points for consistency fitting and whether to discard them. If yes, they are discarded. Otherwise, the sampling point is fitted with the neighboring points for compensation. The output includes a structured determination result of the state label and the crop type, and carbon sink monitoring and valuation are performed based on the structured determination result.

Images