Crop irrigation regulation method and system based on effective leaf area index of photosynthesis

Abstract

The present application relates to the field of intelligent irrigation control, in particular to a crop irrigation regulation method and system based on effective leaf area index. The method comprises: collecting crop canopy image to calculate effective leaf area index and change rate, and switching prediction mode according to the change rate; obtaining transpiration driving factor sequence, combining leaf area index sequence and prediction time window length to obtain transpiration demand; calculating irrigation demand and water margin based on substrate weight, current weight and the like, and judging whether to irrigate based on water margin and threshold value; stopping irrigation according to the double conditions of cumulative irrigation amount and liquid inlet and outlet rate ratio, calculating actual transpiration amount based on water balance, comparing the predicted value and correcting the model when the conditions are met. The present application solves the problems of fixed transpiration prediction time scale, mismatch with crop canopy growth and lack of closed-loop correction, eliminates the cumulative deviation of prediction, improves the accuracy, real-time and stability of irrigation regulation, and realizes the dynamic adaptation of crop water supply and demand.

Description

Technical Field

[0001] This invention relates to the field of intelligent irrigation control, and in particular to a crop irrigation regulation method and system based on the effective photosynthetic leaf area index. Background Technology

[0002] Current crop irrigation control technologies mainly employ trigger-based control using substrate weight or moisture content thresholds, or estimate transpiration and determine irrigation amounts based on environmental factors such as photosynthetically active radiation, air temperature, and air saturation vapor pressure difference. These methods typically use fixed time steps or fixed prediction time windows for transpiration calculations, failing to dynamically adjust for the impact of crop canopy structure changes on transpiration capacity. This leads to a time-scale mismatch between transpiration predictions and actual water consumption during rapid crop growth or physiological decline stages, thus affecting the accuracy of irrigation control.

[0003] In addition, although some technologies introduce leaf area index into transpiration estimation, they are mostly used as static input parameters. They do not use the rate of change of leaf area index to reflect crop growth dynamics, nor do they establish a closed-loop correction mechanism for prediction errors based on water balance. This makes it difficult to adjust the parameters of the transpiration demand model online according to the actual water consumption during long-term operation, which easily leads to cumulative prediction errors and a systematic deviation between irrigation amount and actual transpiration demand. Summary of the Invention

[0004] The purpose of this invention is to provide a crop irrigation regulation method and system based on the effective photosynthetic leaf area index, in order to solve the technical problems in the prior art where the transpiration demand prediction time scale is fixed, it is difficult to reflect the physiological dynamic changes of the crop canopy, and there is a lack of a closed-loop correction mechanism for prediction errors based on water balance, which leads to a mismatch between transpiration prediction results and actual water consumption, and the irrigation control accuracy is prone to cumulative deviation.

[0005] The technical solution of this invention is: a crop irrigation regulation method based on the effective photosynthetic leaf area index, comprising:

[0006] Crop canopy images were collected at a preset sampling frequency and the effective photosynthetic leaf area index was calculated. The rate of change of the effective photosynthetic leaf area index within a preset time window was also calculated.

[0007] The prediction mode is switched according to the rate of change; the prediction mode includes a first prediction mode and a second prediction mode.

[0008] Obtain the transpiration driving factor sequence within a preset time window, and input the transpiration driving factor sequence, effective photosynthetic leaf area index sequence, and the predicted time window length into the transpiration demand model to obtain the transpiration demand within the predicted time window.

[0009] Calculate irrigation demand and water margin based on the current moisture status of the crop, and determine whether to start irrigation based on the water margin and the preset margin safety threshold.

[0010] After irrigation is started, the inflow rate, outflow rate and ratio of the crop are collected at a preset frequency. When the cumulative irrigation amount reaches the irrigation demand or the ratio drops to a preset saturation threshold, irrigation is stopped. The transpiration consumption is estimated based on the water balance, and the deviation between the transpiration consumption and the predicted transpiration demand is calculated. When the deviation meets the correction conditions, the parameters of the transpiration demand model are adaptively corrected.

[0011] Preferably, the method for switching the prediction mode according to the rate of change is as follows:

[0012] When the absolute value of the rate of change is greater than the preset rate of change threshold for N consecutive time windows, switch to the second prediction mode;

[0013] When the absolute value of the rate of change is within a preset stable range for N consecutive time windows, switch to the first prediction mode;

[0014] The preset rate of change threshold is greater than the upper limit of the preset stable range.

[0015] Preferably, the first prediction mode uses a preset fixed prediction time window, and the second prediction mode uses a dynamic prediction time window.

[0016] Preferably, the method for determining the length of the dynamic prediction time window is as follows:

[0017] Growth stages are identified based on the rate of change over N consecutive time windows; the growth stages include a rapid growth phase and a decline phase.

[0018] When the rate of change is positive for N consecutive time windows, it is determined to be a rapid growth period. The length of the predicted time window is determined based on the fixed predicted time window, the first adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index.

[0019] When the rate of change is negative for N consecutive time windows, it is determined to be a period of decline. The length of the predicted time window is determined based on the fixed predicted time window, the second adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index.

[0020] Preferably, the dynamic prediction time window length is set with a preset upper limit and a lower limit. When the calculated dynamic prediction time window length is greater than the upper limit, the upper limit is taken as the prediction time window length; when it is less than the lower limit, the lower limit is taken as the prediction time window length.

[0021] Preferably, the transpiration driving factor includes at least one of photosynthetically active radiation, air temperature, and air saturated vapor pressure difference.

[0022] Preferably, the method for calculating irrigation demand and water margin based on the current crop moisture status information is as follows: the moisture status information includes substrate weight, current weight and irrigation saturation weight; the current available water amount is obtained based on the difference between the current weight and the substrate weight; and the irrigation demand is obtained by comparing the transpiration demand within the prediction time window with the current available water amount.

[0023] The water margin is the difference between the current available water supply and the transpiration demand within the predicted time window.

[0024] The preferred method for estimating transpiration loss based on water balance is as follows:

[0025] Within the actual time period corresponding to the predicted time window, the cumulative irrigation volume and cumulative drainage volume are statistically analyzed, and the change in crop weight corresponding to the actual time period is recorded. Based on the cumulative irrigation volume, cumulative drainage volume, and change in crop weight, the transpiration consumption during the irrigation cycle is calculated.

[0026] On the other hand, this application also discloses a crop irrigation regulation system based on the effective photosynthetic leaf area index, comprising:

[0027] The leaf area index module is used to collect crop canopy images within a preset time window and calculate the effective photosynthetic leaf area index and its rate of change.

[0028] The transpiration demand prediction module is used to determine the prediction mode based on the rate of change. Under the prediction mode, the transpiration driving factor sequence within the preset time window is obtained. The transpiration driving factor sequence, the effective photosynthetic leaf area index sequence, and the prediction time window length are input into the transpiration demand model to obtain the transpiration demand within the prediction time window.

[0029] The moisture status monitoring module is used to monitor the moisture status information of crops, including substrate weight, crop weight, irrigation saturation weight, inflow rate and outflow rate.

[0030] The decision module calculates the irrigation demand and water margin based on the current moisture status information of the crop. Based on the water margin and the preset margin safety threshold, it determines whether to start irrigation. When irrigation is started, the inflow rate, outflow rate and the ratio of the two of the crop are collected at a preset frequency. When the cumulative irrigation amount reaches the irrigation demand or the ratio decreases to the preset saturation threshold, irrigation is stopped.

[0031] The model correction module is used to calculate the deviation between transpiration consumption and predicted transpiration demand based on water balance. When the deviation meets the correction conditions, the transpiration demand model parameters are adaptively corrected.

[0032] Compared with the prior art, the advantages of the present invention are:

[0033] (1) The present invention adaptively determines the transpiration prediction time window based on the effective photosynthetic leaf area index and its rate of change, so that the time scale of transpiration demand prediction matches the physiological dynamic changes of crop canopy, improves the prediction response sensitivity during the rapid growth period of crops, maintains prediction stability during the stable growth period, eliminates the problem of lag or over-prediction caused by fixed prediction time window from the time dimension, and improves the matching degree between transpiration demand prediction and actual water consumption.

[0034] (2) This invention establishes a water balance relationship based on the cumulative irrigation amount, drainage amount and weight change within the time interval corresponding to the prediction time window, calculates the actual transpiration consumption, and compares it with the predicted transpiration demand, thereby realizing the closed-loop error correction of the transpiration demand model. This allows the model parameters to be adjusted online according to the changes in crop physiological state and environmental conditions, avoiding the cumulative deviation generated by the long-term operation of the model.

[0035] (3) By combining the physiological dynamic-driven feedforward prediction mechanism with the water balance feedback correction mechanism, a continuously operating adaptive irrigation control structure is constructed, so that the irrigation decision is transformed from a single threshold response to a closed-loop control mode of prediction, verification and correction, thereby improving the matching degree between irrigation amount and actual evapotranspiration demand. Attached Figure Description

[0036] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0037] Figure 1 This is a flowchart of a crop irrigation regulation method based on the effective photosynthetic leaf area index as described in this invention;

[0038] Figure 2 This is a structural diagram of a crop irrigation regulation system based on the effective photosynthetic leaf area index as described in this invention. Detailed Implementation

[0039] The present invention will be further described in detail below with reference to specific embodiments:

[0040] This invention is applicable to intelligent irrigation scenarios in agriculture, such as greenhouses, intelligent plant factories, and large-scale field cultivation. In these scenarios, the dynamic changes in crop canopy structure are significant. Traditional fixed-time-window transpiration prediction methods are prone to deviating from actual water consumption and are prone to cumulative prediction errors over long-term operation. To address these issues, this invention achieves accurate prediction of transpiration demand and dynamic regulation of irrigation by dynamically monitoring the effective photosynthetic leaf area index and intelligently switching prediction modes, combined with closed-loop correction of the water balance model.

[0041] like Figure 1 As shown, a crop irrigation regulation method based on the effective photosynthetic leaf area index includes:

[0042] S1. Collect crop canopy images according to the preset sampling frequency and calculate the effective photosynthetic leaf area index, and calculate the rate of change of the effective photosynthetic leaf area index within the preset time window.

[0043] Specifically, the sampling frequency is the canopy monitoring time frequency set according to crop growth characteristics and planting scenarios. The effective photosynthetic leaf area index is the leaf area index calculated only for effective leaves with photosynthetic activity in the crop canopy. Its rate of change is the dynamic rate of change of the effective photosynthetic leaf area index within a preset time period, which is the core indicator reflecting the growth dynamics and transpiration capacity changes of the crop canopy.

[0044] In one implementation, monocular cameras are deployed at fixed points in the crop planting area, with a sampling frequency set to once a day, i.e., collecting top-down images of the crop canopy daily. The acquired images are then subjected to denoising, background segmentation, canopy region extraction, and invalid leaf removal. Based on the effective leaf image regions, a digital image analysis algorithm combined with a leaf area index calculation method commonly used in agriculture is employed to obtain the effective photosynthetic leaf area index value for each sampling time point. An effective photosynthetic leaf area index sequence is then constructed in chronological order.

[0045] A time window is set according to the crop growth cycle. The effective photosynthetic leaf area index is extracted from the end time point and the start time point of the time window. The difference between the two is taken and divided by the length of the time window to obtain the rate of change of the effective photosynthetic leaf area index.

[0046] In some other embodiments, to avoid potential errors from a single acquisition, such as image acquisition errors or slight changes in the canopy caused by short-term environmental changes, the sampling frequency is adjusted to a frequency of one hour or three hours, and the daily average, maximum or minimum value is taken as the effective photosynthetic leaf area index.

[0047] S2. Switch the prediction mode according to the rate of change; wherein the prediction mode includes the first prediction mode and the second prediction mode.

[0048] Specifically, the rate of change of effective photosynthetic leaf area index is a key basis for determining the trend of crop canopy transpiration capacity. The switching of prediction models takes the dynamic characteristics of this rate of change as the core criterion. Different prediction models are adapted to different changes in crop canopy, thereby matching the prediction time dimension of transpiration demand and improving the degree of matching between the prediction of transpiration demand and the actual physiological state of crops.

[0049] In one implementation, the calculated effective photosynthetic leaf area index change rate is used as the core judgment indicator, and its change characteristics within a continuous time window are continuously monitored. When the effective photosynthetic leaf area index shows significant dynamic fluctuation characteristics, that is, when the absolute value of the change rate in multiple windows is continuously greater than a preset threshold, the system switches to a second prediction mode adapted to rapid changes in the canopy. When the absolute value of the change rate tends to be stable and there are no obvious fluctuations, the system switches to a first prediction mode adapted to a stable canopy state, thereby realizing intelligent and precise switching of the prediction mode according to the growth dynamics of the crop canopy.

[0050] S3. Obtain the transpiration driving factor sequence within the preset time window, input the transpiration driving factor sequence, the effective photosynthetic leaf area index sequence, and the predicted time window length into the transpiration demand model to obtain the transpiration demand within the predicted time window.

[0051] Specifically, the transpiration driving factor sequence is a time-series dataset formed by continuously collecting various transpiration driving factor data in chronological order, and the transpiration driving factor sequence is extracted within a preset time window.

[0052] The transpiration demand model is a specialized model constructed by coupling the effective photosynthetic leaf area index with transpiration driving factors. It is used to calculate the water transpiration demand of crops within a specified period. By inputting relevant time series and the length of the prediction time window, it can output the cumulative transpiration demand for the corresponding period.

[0053] In one implementation, transpiration driving factor data is collected by environmental monitoring sensors deployed in the planting area. After removing outliers, a complete transpiration driving factor sequence is constructed, and then the transpiration driving factor sequence within a preset time window is extracted.

[0054] Extract the effective photosynthetic leaf area index sequence within the same time window as the transpiration driving factor sequence. Based on the prediction time window length determined by the prediction mode in step S2, input the above-mentioned transpiration driving factor sequence, effective photosynthetic leaf area index sequence, and prediction time window length into the trained transpiration demand model. After model calculation and processing, output the cumulative transpiration demand of the crop within the prediction time window.

[0055] In this embodiment, the transpiration demand model is a time-series cumulative prediction model constructed based on the coupling of effective photosynthetic leaf area index and transpiration driving factors. After being trained and verified with crop experimental data, it is a dedicated model specifically adapted for dynamic prediction of crop transpiration demand. The model takes the effective photosynthetic leaf area index sequence that reflects the physiological state of the crop and the transpiration driving factor sequence that reflects the environmental influence as core input features, and the prediction time window length as the time dimension parameter. Through the cumulative calculation of time-series data, it can accurately quantify the water transpiration consumption pattern of the crop within a specified prediction time window, and finally output the cumulative transpiration demand of the crop within the prediction time window set by the user.

[0056] S4. Calculate irrigation demand and water margin based on the current crop moisture status information, and determine whether to start irrigation based on the water margin and the preset margin safety threshold.

[0057] Specifically, the current moisture status information of crops is the core data reflecting the availability of water for crops. It includes the substrate weight, the current weight of the crop and the cultivation substrate, and the total weight under saturated irrigation conditions. Before crop transplanting, the dry substrate used for cultivation is weighed to obtain the dry weight of the substrate for a single cultivation unit. This value is fixed and is only recalibrated when the substrate is changed or the cultivation unit is adjusted. The saturated irrigation weight is a baseline value pre-calibrated at the beginning of planting and can be periodically checked and corrected as needed later.

[0058] Irrigation demand is the amount of effective water that needs to be added to meet the transpiration demand of crops within the predicted time window; water margin is the difference between the actual available water in the current cultivation system and the transpiration demand within the predicted time window, which directly reflects the water supply and demand balance; the margin safety threshold is the critical value for determining whether irrigation needs to be started, serving as a preset reference standard to ensure crop water supply.

[0059] In one implementation, the substrate weight and irrigation saturation weight calibrated at the initial stage of planting are retrieved first. At the same time, the current weight of the crop is obtained through monitoring equipment, and the current available water volume is calculated, which is the difference between the current total weight and the substrate weight. Then, the current available water volume is compared with the transpiration demand within the predicted time window. If the current available water volume is less than the transpiration demand, the irrigation demand is the difference between the two. If the current available water volume is greater than or equal to the transpiration demand, the irrigation demand is 0.

[0060] Calculate the water margin, which is the current available water amount minus the transpiration demand within the predicted time window. Finally, compare the calculated water margin with the preset margin safety threshold. If the water margin is lower than the threshold, it is determined that the current water supply of the crop cannot meet the subsequent transpiration demand, and irrigation needs to be started. If the water margin is higher than or equal to the threshold, it is determined that the water supply is sufficient, and irrigation does not need to be started.

[0061] S5. After irrigation is started, the inflow rate, outflow rate and ratio of the crop are collected at a preset frequency. When the cumulative irrigation amount reaches the irrigation demand or the ratio drops to the preset saturation threshold, irrigation is stopped. The transpiration consumption is estimated based on the water balance, and the deviation between the transpiration consumption and the predicted transpiration demand is calculated. When the deviation meets the correction conditions, the parameters of the transpiration demand model are adaptively corrected.

[0062] Specifically, the preset frequency is a fixed time interval for collecting water transmission data during irrigation, the inflow rate is the amount of irrigation water input into the cultivation system per unit time, and the outflow rate is the amount of excess water discharged from the cultivation system per unit time. The ratio of the two is the core indicator reflecting the water saturation level of the cultivation system.

[0063] The preset saturation threshold is the critical value of the ratio of liquid inflow to liquid outflow when the cultivation system reaches water saturation; the water balance is the comprehensive data of water input, output and storage changes in the cultivation system during the irrigation cycle, which is the basis for estimating the actual transpiration consumption; the correction condition is the preset deviation standard for determining whether the model parameters need to be adjusted, and the adaptive correction is the process of dynamically optimizing the core calculation parameters of the transpiration demand model according to the degree of deviation.

[0064] In one implementation, after the irrigation program is started, the real-time inflow rate and outflow rate of the cultivation system are collected by a flow sensor, the ratio of the inflow rate to the outflow rate of a single collection is calculated simultaneously, and the total irrigation volume during the irrigation process is accumulated.

[0065] The system compares the cumulative irrigation volume with the preset irrigation demand and the rate ratio with the saturation threshold in real time. When any condition is met, an irrigation stop command is immediately triggered to terminate the irrigation operation. Within the actual time period corresponding to the predicted time window, the system calculates the cumulative irrigation volume and cumulative drainage volume for that period. At the same time, the system records the change in the total weight of the crop and cultivation substrate through a weight sensor. Based on the water balance relationship, the system calculates the actual transpiration consumption within that period.

[0066] Calculate the deviation value and deviation rate between the actual transpiration consumption and the predicted transpiration demand, and compare them with the correction conditions. If the deviation meets the correction conditions, the core calculation parameters of the transpiration demand model are adaptively tuned and corrected using a parameter iteration algorithm. If the deviation does not meet the conditions, the original parameters of the model are retained, and only the deviation data is included in the model training sample library.

[0067] In summary, this invention provides a crop irrigation regulation method based on the effective photosynthetic leaf area index (APA). It constructs a comprehensive regulation system encompassing dynamic monitoring of APA, intelligent switching of prediction modes, precise prediction of transpiration demand, on-demand irrigation decision-making, dual-condition process control, and adaptive model correction. This system incorporates the dynamic physiological growth of the crop canopy into the core basis of transpiration prediction, couples crop physiological and environmental driving factors to scientifically quantify transpiration demand, accurately calculates irrigation demand based on water status, and intelligently determines irrigation start and stop. The irrigation process is precisely controlled by dual conditions: cumulative irrigation volume and the ratio of inflow to outflow rates. Furthermore, it constructs a prediction deviation mechanism based on water balance. The closed-loop correction mechanism specifically addresses the core problems of traditional irrigation technology, such as fixed transpiration prediction timescales that do not match actual water consumption and the lack of error correction that easily leads to cumulative biases. It not only ensures that transpiration demand predictions closely align with the actual growth and physiological state of crops, significantly improving prediction accuracy, but also enables precise on-demand irrigation supply, avoiding water waste and insufficient irrigation. Furthermore, it eliminates cumulative errors from long-term operation through adaptive model correction, ensuring the continuity, stability, and accuracy of irrigation regulation. This comprehensively enhances the intelligence and scientific level of crop irrigation regulation, adapting to the irrigation needs of various large-scale and refined agricultural planting scenarios.

[0068] To further illustrate the present invention, the key technical details of the above method will be described in detail below.

[0069] Step S2 switches the prediction mode based on the rate of change of the effective photosynthetic leaf area index. The specific method is as follows:

[0070] When the absolute value of the rate of change is greater than the preset rate of change threshold for N consecutive time windows, switch to the second prediction mode;

[0071] When the absolute value of the rate of change is within the preset stable range for N consecutive time windows, switch to the first prediction mode.

[0072] The first prediction mode uses a preset fixed prediction time window, while the second prediction mode uses a dynamic prediction time window.

[0073] The method for determining the length of the dynamic prediction time window in the second prediction mode is as follows:

[0074] Growth stages are identified based on the rate of change over N consecutive time windows; the growth stages include a rapid growth phase and a decline phase.

[0075] When the rate of change is positive for N consecutive time windows, it is determined to be a rapid growth period. The length of the predicted time window is determined based on the fixed predicted time window, the first adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index.

[0076] When the rate of change is negative for N consecutive time windows, it is determined to be a period of decline. The length of the predicted time window is determined based on the fixed predicted time window, the second adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index.

[0077] The dynamic prediction time window length has a preset upper limit and lower limit. When the calculated dynamic prediction time window length is greater than the upper limit, the upper limit is taken as the prediction time window length; when it is less than the lower limit, the lower limit is taken as the prediction time window length.

[0078] Specifically, N is a pre-set continuous window threshold, the preset rate of change threshold is the critical value for distinguishing whether the crop canopy is in a state of rapid change, and the preset stable interval is the numerical range of the rate of change when the crop canopy growth is stable. By determining the consistency of N consecutive time windows, the prediction mode is effectively avoided from being switched erroneously due to random factors, ensuring the accuracy and stability of the mode switching, and enabling the adjustment of the prediction mode to accurately match the real and continuous growth dynamics of the crop canopy.

[0079] The fixed forecast time window is a pre-set forecast time period suitable for when the crop canopy is stable; the dynamic forecast time window is a forecast time period that is adaptively adjusted according to the current growth stage of the crop and the change in the effective photosynthetic leaf area index.

[0080] The first and second adjustment coefficients are preset adjustment parameters adapted to the canopy change characteristics during the rapid growth and decline phases, respectively. Growth stage identification is based on the temporal change characteristics of the effective photosynthetic leaf area index to distinguish whether the crop is currently in a rapid growth phase, a stable phase, or a decline phase, so as to match the corresponding strategy and determine the length of the prediction time window.

[0081] In one implementation, the rate of change of the effective photosynthetic leaf area index between the current time window and the previous N-1 time windows is extracted. If the absolute value of the rate of change of N consecutive time windows is greater than a preset rate of change threshold, the second prediction mode is switched, and it is further determined whether the value of the above rate of change is positive or negative. If it is positive, it indicates that the crop is in a rapid growth period; if it is negative, it indicates that the crop is in a decline period.

[0082] When the plant is in a rapid growth phase, the length of the dynamic prediction time window adapted to the rapid growth phase is calculated by combining the fixed prediction time window, the first adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index. The specific calculation method is as follows:

[0083] ,

[0084] in, To determine the length of the dynamic prediction time window for the rapid growth period, To fix the prediction time window length, This is the first adjustment coefficient; in this embodiment, This is a weighted average of the rate of change in the current time window and the rate of change in the previous time window.

[0085] .

[0086] By smoothing the trend change rate of effective photosynthetic leaf area index within the current time window with the change rate of the previous time window, the obtained change rate reflects both the current physiological evolution rate and avoids sudden jumps caused by short-term environmental fluctuations or image acquisition errors. This suppresses frequent oscillations in the length of the prediction time window, improves the stability of stage identification and the continuity of prediction time scale adjustment, and makes irrigation decisions more consistent with actual growth trends.

[0087] In some other implementations, It is the weighted result of the change rate of the current time window and the change rates of multiple previous time windows, or the average of the change rates of multiple time windows.

[0088] When the crop is in the decline phase, the dynamic prediction time window length is calculated by combining the fixed prediction time window, the second adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index. This ensures that the prediction time window length is adapted to the dynamic changes of the canopy at different growth stages of the crop. The specific calculation method is as follows:

[0089] ,

[0090] in, The length of the dynamic forecasting time window for the recession period. This is the second adjustment coefficient.

[0091] The crop growth stage also includes a stationary period, during which the effective photosynthetic leaf area index changes little and remains within a stable range. When the crop is in a stationary period and the effective photosynthetic leaf area index changes within a stable range, a fixed prediction time window is used to predict transpiration demand. This avoids unnecessary dynamic adjustments on the time scale when the system is in a steady state, thereby improving the statistical stability of transpiration demand estimation, reducing the risk of over-response in the control process, and facilitating the gradual convergence of model parameters under steady-state conditions, thus enhancing the controllability and long-term operational stability of overall irrigation control.

[0092] The aforementioned forecasting model is used to predict the transpiration demand of crops within a defined forecasting time window. To predict transpiration demand, transpiration driving factors are needed as the basis for data-driven forecasting. Transpiration driving factors include at least one of photosynthetically active radiation, air temperature, and air saturated vapor pressure difference.

[0093] Specifically, transpiration driving factors are environmental input parameters upon which the model relies for data-driven predictions. They are key indicators reflecting the external driving conditions of crop transpiration and include at least one or more combinations of photosynthetically active radiation, air temperature, and air saturated vapor pressure difference. These factors can comprehensively characterize the external driving intensity of crop transpiration from dimensions such as light intensity, environmental heat, and air humidity deficit, providing a reliable data input basis for transpiration demand models and ensuring the accuracy and rationality of transpiration demand predictions.

[0094] The transpiration demand model uses the effective photosynthetic leaf area index sequence and the transpiration driving factor sequence as core inputs and the prediction time window as the time dimension constraint. By modeling the temporal characteristics of crop transpiration, it can quantify the total water consumption of crops within the prediction time window. It can couple the crop's own physiological state with the external environmental driving conditions, thereby improving the accuracy and reliability of the calculation of cumulative transpiration demand.

[0095] The effective photosynthetic leaf area index time series data, photosynthetically active radiation, air temperature, air saturated vapor pressure difference and other transpiration driving factors time series data collected within the preset time window, as well as the determined prediction time window length, are input into the trained transpiration demand model. The model performs feature fusion and cumulative calculation on the time series data and outputs the cumulative transpiration demand within the corresponding prediction time window.

[0096] In this embodiment, the time intervals of the transpiration driving factor time series data and the effective photosynthetic leaf area index time series data are consistent, but the time point intervals, i.e., the sampling frequencies, are different. The effective photosynthetic leaf area index uses low-frequency time series data collected daily, while the transpiration driving factor uses high-frequency time series data collected hourly. To achieve matching and coupling input of the two types of time series data with different frequencies, a fusion module is constructed in the input layer of the transpiration demand model. This module is used to extract features from the high-frequency time series data to achieve data dimensionality reduction, thereby performing channel splicing with the low-frequency time series data. This aligns the processed transpiration driving factor time series data with the effective photosynthetic leaf area index time series data in terms of data dimension, thus ensuring the temporal matching and computational rationality of the transpiration demand model input data.

[0097] In addition to the input layer, the transpiration demand model includes a backbone module and a cumulative output module. The backbone module adopts a lightweight gradient boosting regression model, which is trained using historical effective photosynthetic leaf area index sequences, transpiration driving factor sequences, and measured transpiration samples to achieve accurate prediction of daily transpiration. The cumulative output module accumulates the daily transpiration output from the backbone module based on the number of days included in the prediction time window to obtain the cumulative transpiration demand within the prediction time window.

[0098] In some other implementations, empirical or regression models are used to estimate the evapotranspiration demand for the forecast time window.

[0099] The method for calculating irrigation demand and water margin based on the current crop water status information is as follows: the water status information includes substrate weight, current weight and irrigation saturation weight. The current available water amount is obtained based on the difference between the current weight and the substrate weight. The irrigation demand is obtained by comparing the transpiration demand within the predicted time window with the current available water amount.

[0100] The water margin is the difference between the current available water supply and the evapotranspiration demand within the predicted time window.

[0101] Specifically, the substrate weight and irrigation saturation weight are pre-calibrated baseline values, while the current weight is a real-time monitoring value. The difference between the current weight and the substrate weight yields the current available water volume that can be absorbed and utilized by the crop within the cultivation system. This current available water volume is then compared with the transpiration demand within the predicted time window, and the difference determines the required additional irrigation. Finally, the water margin is obtained by subtracting the transpiration demand within the predicted time window from the current available water volume. This water margin visually reflects the extent to which the crop's water reserves are sufficient relative to subsequent transpiration consumption.

[0102] The method for estimating transpiration loss based on water balance is as follows:

[0103] Within the actual time period corresponding to the predicted time window, the cumulative irrigation volume and cumulative drainage volume are statistically analyzed, and the change in crop weight corresponding to the actual time period is recorded. Based on the cumulative irrigation volume, cumulative drainage volume, and change in crop weight, the transpiration consumption during the irrigation cycle is calculated.

[0104] Specifically, the water balance is calculated based on the actual time period corresponding to the predicted time window as a complete irrigation cycle. The core of the calculation is to statistically analyze and calculate the water input, water output, and crop weight changes of the cultivation system within this cycle.

[0105] Among them, the cumulative irrigation volume is the total amount of irrigation water actually input into the cultivation system during the irrigation cycle, and the cumulative drainage volume is the total amount of excess water discharged from the cultivation system during the cycle. Both are obtained by real-time statistics and accumulation through flow monitoring sensors. The change in crop weight is the difference between the total weight of the crop and the substrate at the beginning and end of the cycle. After deducting the pre-calibrated substrate weight, it is the net change in the weight of the crop itself.

[0106] Based on the principle of water balance, the total water input during the irrigation cycle is equal to the sum of the total water output and the change in crop weight. The total water output mainly includes crop transpiration consumption and cumulative drainage. Further calculation of transpiration consumption not only conforms to the water movement patterns in actual cultivation scenarios, but also ensures the accuracy and reliability of the calculation results through monitorable and quantifiable parameters, providing real and effective measured data support for the subsequent parameter correction of the transpiration demand model.

[0107] The actual transpiration consumption is compared with the predicted transpiration demand output by the transpiration demand model within the same period, and the deviation between the two is calculated. When the deviation exceeds the preset allowable deviation range, the model correction condition is met. Based on the deviation between the actual transpiration consumption and the predicted value, the relevant parameters in the transpiration demand model are adaptively corrected using gradient updates or weighted iterations, so that the predicted transpiration demand output by the model gradually approaches the actual transpiration consumption. When the deviation is within the preset allowable range, the existing parameters of the transpiration demand model are kept unchanged, thereby achieving continuous optimization of the model's prediction accuracy.

[0108] like Figure 2 As shown, the present invention also provides a crop irrigation regulation system based on the effective photosynthetic leaf area index, characterized in that it includes:

[0109] The leaf area index module is used to collect crop canopy images within a preset time window and calculate the effective photosynthetic leaf area index and its rate of change.

[0110] The transpiration demand prediction module is used to determine the prediction mode based on the rate of change. Under the determined prediction mode, the transpiration driving factor sequence within the preset time window is obtained. The transpiration driving factor sequence, the effective photosynthetic leaf area index sequence, and the prediction time window length are input into the transpiration demand model to obtain the transpiration demand within the prediction time window.

[0111] The moisture status monitoring module is used to monitor the moisture status information of crops, including substrate weight, crop weight, irrigation saturation weight, inflow rate and outflow rate.

[0112] The decision module calculates the irrigation demand and water margin based on the current moisture status information of the crop. Based on the water margin and the preset margin safety threshold, it determines whether to start irrigation. When irrigation is started, the inflow rate, outflow rate and the ratio of the two of the crop are collected at a preset frequency. When the cumulative irrigation amount reaches the irrigation demand or the ratio drops to the preset saturation threshold, irrigation is stopped.

[0113] The model correction module is used to calculate the deviation between transpiration consumption and predicted transpiration demand based on water balance. When the deviation meets the correction conditions, the transpiration demand model parameters are adaptively corrected.

[0114] The above embodiments are merely illustrative of the technical concept and features of the present invention, intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly, and should not be construed as limiting the scope of protection of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of the present invention is defined by the appended claims rather than the foregoing description, and thus all changes falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention.

Claims

1. A crop irrigation regulation method based on effective photosynthetic leaf area index, characterized in that, include: Crop canopy images were collected at a preset sampling frequency and the effective photosynthetic leaf area index was calculated. The rate of change of the effective photosynthetic leaf area index within a preset time window was also calculated. The prediction mode is switched according to the rate of change; the prediction mode includes a first prediction mode and a second prediction mode. Obtain the transpiration driving factor sequence within a preset time window, and input the transpiration driving factor sequence, effective photosynthetic leaf area index sequence, and the predicted time window length into the transpiration demand model to obtain the transpiration demand within the predicted time window. Calculate irrigation demand and water margin based on the current moisture status of the crop, and determine whether to start irrigation based on the water margin and the preset margin safety threshold. After irrigation is started, the inflow rate, outflow rate and ratio of the crop are collected at a preset frequency. When the cumulative irrigation amount reaches the irrigation demand or the ratio drops to a preset saturation threshold, irrigation is stopped. The transpiration consumption is estimated based on the water balance, and the deviation between the transpiration consumption and the predicted transpiration demand is calculated. When the deviation meets the correction conditions, the parameters of the transpiration demand model are adaptively corrected. The method for switching the prediction mode based on the rate of change is as follows: When the absolute value of the rate of change is greater than the preset rate of change threshold for N consecutive time windows, switch to the second prediction mode; When the absolute value of the rate of change is within a preset stable range for N consecutive time windows, switch to the first prediction mode; The preset rate of change threshold is greater than the upper limit of the preset stable interval; The first prediction mode uses a preset fixed prediction time window, while the second prediction mode uses a dynamic prediction time window.

2. The crop irrigation regulation method based on effective photosynthetic leaf area index according to claim 1, characterized in that, The method for determining the length of the dynamic prediction time window is as follows: Growth stages are identified based on the rate of change over N consecutive time windows; the growth stages include a rapid growth phase and a decline phase. When the rate of change is positive for N consecutive time windows, it is determined to be a rapid growth period. The length of the predicted time window is determined based on the fixed predicted time window, the first adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index. When the rate of change is negative for N consecutive time windows, it is determined to be a period of decline. The length of the predicted time window is determined based on the fixed predicted time window, the second adjustment coefficient, and the rate of change of the effective photosynthetic leaf area index.

3. The crop irrigation regulation method based on effective photosynthetic leaf area index according to claim 2, characterized in that, The dynamic prediction time window length is set with a preset upper limit and a lower limit. When the calculated dynamic prediction time window length is greater than the upper limit, the upper limit is taken as the prediction time window length; when it is less than the lower limit, the lower limit is taken as the prediction time window length.

4. The crop irrigation regulation method based on effective photosynthetic leaf area index according to claim 1, characterized in that, The transpiration driving factors include at least one of photosynthetically active radiation, air temperature, and air saturated vapor pressure difference.

5. The crop irrigation regulation method based on effective photosynthetic leaf area index according to claim 1, characterized in that, The method for calculating irrigation demand and water margin based on the current crop moisture status information is as follows: the moisture status information includes substrate weight, current weight and irrigation saturation weight. The current available water amount is obtained based on the difference between the current weight and the substrate weight. The irrigation demand is obtained by comparing the transpiration demand within the prediction time window with the current available water amount. The water margin is the difference between the current available water supply and the transpiration demand within the predicted time window.

6. The crop irrigation regulation method based on effective photosynthetic leaf area index according to claim 1, characterized in that, The method for estimating transpiration loss based on water balance is as follows: Within the actual time period corresponding to the predicted time window, the cumulative irrigation volume and cumulative drainage volume are statistically analyzed, and the change in crop weight corresponding to the actual time period is recorded. Based on the cumulative irrigation volume, cumulative drainage volume, and change in crop weight, the transpiration consumption during the irrigation cycle is calculated.

7. A crop irrigation regulation system based on effective photosynthetic leaf area index, used to implement the crop irrigation regulation method based on effective photosynthetic leaf area index as described in any one of claims 1-6, characterized in that, include: The leaf area index module is used to collect crop canopy images within a preset time window and calculate the effective photosynthetic leaf area index and its rate of change. The transpiration demand prediction module is used to determine the prediction mode based on the rate of change. Under the prediction mode, the transpiration driving factor sequence within the preset time window is obtained. The transpiration driving factor sequence, the effective photosynthetic leaf area index sequence, and the prediction time window length are input into the transpiration demand model to obtain the transpiration demand within the prediction time window. The moisture status monitoring module is used to monitor the moisture status information of crops, including substrate weight, crop weight, irrigation saturation weight, inflow rate and outflow rate. The decision module calculates the irrigation demand and water margin based on the current moisture status information of the crop. Based on the water margin and the preset margin safety threshold, it determines whether to start irrigation. When irrigation is started, the inflow rate, outflow rate and the ratio of the two of the crop are collected at a preset frequency. When the cumulative irrigation amount reaches the irrigation demand or the ratio decreases to the preset saturation threshold, irrigation is stopped. The model correction module is used to calculate the deviation between transpiration consumption and predicted transpiration demand based on water balance. When the deviation meets the correction conditions, the transpiration demand model parameters are adaptively corrected.

Images