Irrigation decision optimization system and method based on complementation of soil moisture content and meteorological data

Abstract

The invention relates to the technical field of agricultural intelligent irrigation, in particular to an irrigation decision optimization system and method based on complementation of soil moisture content and meteorological data. Comprising a soil moisture content monitoring unit used for collecting layered soil moisture content and temperature in real time; the meteorological data acquisition unit is used for acquiring local actually measured meteorological parameters and integrating third-party remote sensing rainfall and evapotranspiration prediction data; the data fusion and processing unit is used for modeling the multi-dimensional data by using a machine learning model and extracting a field evapotranspiration feature vector; the decision optimization unit is used for analyzing the deviation degree of the actually measured moisture reduction rate and a theoretical predicted value based on a machine learning algorithm, dynamically correcting the crop coefficient in the current growth period and generating a precise irrigation instruction in combination with a water balance model; and the execution control unit is used for realizing closed-loop irrigation management and checking the infiltration effect through the intelligent ecological gateway. According to the invention, through deep fusion of the actually measured soil data and the weather forecast data, the defect of crop coefficient staticization in a traditional irrigation decision is corrected.

Description

Technical Field

[0001] This invention relates to the field of intelligent agricultural irrigation technology, and in particular to an irrigation decision optimization system and method based on the complementarity of soil moisture and meteorological data. Background Technology

[0002] Currently, agricultural water use accounts for a significant portion of my country's total water consumption. With the over-exploitation of groundwater, water scarcity is becoming increasingly prominent. Traditional irrigation methods rely on experience-based judgment, leading to inaccurate irrigation timing and unscientific water control, resulting in water waste and low utilization efficiency. While existing technologies include irrigation control systems based on soil moisture sensors or meteorological data, most systems rely solely on one type of data, failing to achieve deep integration and complementarity between soil moisture and meteorological data, thus hindering precise, dynamic, and adaptive irrigation decisions. Summary of the Invention

[0003] To overcome the above deficiencies, this invention provides an irrigation decision optimization system and method based on complementary soil moisture and meteorological data. The aim is to construct an intelligent irrigation decision model by collecting multi-level data such as soil moisture and temperature in real time, and combining field meteorological data with future weather forecasts, so as to achieve precise control of irrigation timing, irrigation volume and irrigation cycle.

[0004] In a first aspect, the present invention provides the following technical solution: an irrigation decision optimization system based on the complementarity of soil moisture and meteorological data, comprising:

[0005] The soil moisture monitoring unit is used to collect real-time data on soil moisture content and temperature at different depths.

[0006] The meteorological data acquisition unit is used to acquire field microclimate data in real time and integrate third-party remote sensing meteorological data to obtain future rainfall forecast data and future reference crop evapotranspiration forecast data.

[0007] The data fusion and processing unit connects the soil moisture monitoring unit and the meteorological data acquisition unit through an intelligent ecological gateway, and is used to upload the collected data to the cloud platform for cleaning, fusion and analysis.

[0008] The decision optimization unit is used to determine the effective water storage capacity and water storage capacity of the soil based on the water balance model and machine learning algorithm, and dynamically correct the crop coefficient by combining the predicted meteorological data, and generate decision instructions including irrigation timing, irrigation water volume and irrigation cycle.

[0009] The execution control unit is connected to the intelligent ecological gateway and is used to automatically control the irrigation equipment according to the decision instructions to achieve closed-loop irrigation management.

[0010] Preferably, in the soil moisture monitoring unit, the soil moisture monitoring unit adopts a layered soil moisture sensor based on FDR frequency domain reflectance technology; the layered soil moisture sensor has a built-in triaxial accelerometer sensor for detecting vibrations around the equipment and stopping data acquisition during vibrations to eliminate interference.

[0011] Preferably, the steps of acquiring future rainfall forecast data and future reference crop evapotranspiration forecast data in the meteorological data acquisition unit include:

[0012] Meteorological parameters, including atmospheric pressure, wind direction, wind speed, rainfall, total radiation, air temperature, and air humidity, are collected in real time using small field weather stations.

[0013] By integrating third-party remote sensing meteorological data through a cloud platform, we can obtain rainfall forecast data for the next 5 days and reference crop evapotranspiration forecast data for the next 7 days.

[0014] Preferably, in the data fusion and processing unit, the steps of uploading the collected data to the cloud platform for cleaning, fusion, and analysis include:

[0015] The real-time collected soil moisture and meteorological data are cleaned to identify and remove outlier data.

[0016] Multi-layer soil moisture data are spatiotemporally correlated to calculate the moisture weight of each layer within the crop root activity range, and feature vectors reflecting the current field evapotranspiration environment are generated by combining meteorological parameters.

[0017] The cleaned local data is aligned with the acquired third-party weather forecast data to establish a time series database.

[0018] Preferably, in the decision optimization unit, the steps of determining the effective water storage capacity and water holding capacity of the soil using real-time soil moisture content include:

[0019] Based on machine learning algorithms, the crop coefficient for the current growth stage is dynamically adjusted according to the correlation between historical soil moisture decline rate and meteorological factors during the corresponding period.

[0020] By combining the corrected crop coefficient with the predicted value of reference crop evapotranspiration, the actual future water requirement of the crop can be calculated.

[0021] Based on the actual water demand and future rainfall forecast data, the irrigation start time and target irrigation amount are determined using a water balance model.

[0022] Preferably, the step of dynamically correcting the crop coefficient for the current growth stage includes:

[0023] Based on the crop types and their current growth stages in the area to be monitored, a corresponding preset crop coefficient benchmark value is matched;

[0024] Using historical water loss data collected by the soil moisture monitoring unit and combined with real-time meteorological parameters from the same period, the actual water consumption change rate of crops under the current growth environment is obtained by fitting.

[0025] By comparing the actual rate of change in water consumption with the theoretical evapotranspiration level calculated based on meteorological environment, the degree of deviation between the actual water demand and the theoretical prediction is analyzed.

[0026] The preset crop coefficient benchmark value is corrected online based on the degree of deviation to generate a crop coefficient that reflects the current growth stage.

[0027] Preferably, in the execution control unit, the step of automatically controlling the irrigation equipment according to the decision command includes:

[0028] The intelligent ecological gateway receives decision instructions issued by the cloud platform and parses them into executable signals for the corresponding irrigation area control terminal.

[0029] During irrigation, the operating status of the water supply network and the feedback parameters of the irrigation equipment are monitored in real time to ensure that the irrigation operation is carried out according to the preset instructions;

[0030] The irrigation infiltration effect is dynamically verified based on the real-time moisture rise rate transmitted back by the soil moisture monitoring unit.

[0031] When the monitored soil moisture reaches the preset upper limit of moisture content or the preset irrigation time is completed, the irrigation equipment is automatically turned off, and the operation completion data is uploaded to the cloud platform to update the water balance model.

[0032] Secondly, the present invention provides the following technical solution: an irrigation decision optimization method based on the complementarity of soil moisture and meteorological data, used in any of the above-mentioned irrigation decision optimization systems, the method comprising:

[0033] S1. Real-time collection of soil moisture content and temperature data at different depths;

[0034] S2. Real-time acquisition of field microclimate data, and integration of third-party remote sensing meteorological data to obtain future rainfall forecast data and future reference crop evapotranspiration forecast data;

[0035] S3. Upload the collected data to the cloud platform for cleaning, fusion and analysis;

[0036] S4. Based on the water balance model and machine learning algorithm, the effective water storage capacity and water storage capacity of the soil are determined by real-time soil moisture content, and the crop coefficient is dynamically corrected by combining the predicted meteorological data to generate decision instructions including irrigation timing, irrigation water volume and irrigation cycle.

[0037] S5. Automatically control the irrigation equipment according to the decision instructions to achieve closed-loop irrigation management.

[0038] Preferably, in step S4, the step of determining the effective water storage capacity and water holding capacity of the soil using real-time soil moisture content includes:

[0039] Based on machine learning algorithms, the crop coefficient for the current growth stage is dynamically adjusted according to the correlation between historical soil moisture decline rate and meteorological factors during the corresponding period.

[0040] By combining the corrected crop coefficient with the predicted value of reference crop evapotranspiration, the actual future water requirement of the crop can be calculated.

[0041] Based on the actual water demand and future rainfall forecast data, the irrigation start time and target irrigation amount are determined using a water balance model.

[0042] Preferably, the step of dynamically correcting the crop coefficient for the current growth stage includes:

[0043] Based on the crop types and their current growth stages in the area to be monitored, a corresponding preset crop coefficient benchmark value is matched;

[0044] Using historical water loss data collected by the soil moisture monitoring unit and combined with real-time meteorological parameters from the same period, the actual water consumption change rate of crops under the current growth environment is obtained by fitting.

[0045] By comparing the actual rate of change in water consumption with the theoretical evapotranspiration level calculated based on meteorological environment, the degree of deviation between the actual water demand and the theoretical prediction is analyzed.

[0046] The preset crop coefficient benchmark value is corrected online based on the degree of deviation to generate a crop coefficient that reflects the current growth stage.

[0047] The present invention has the following beneficial effects:

[0048] 1. This invention deeply integrates real-time sensed soil moisture content with multi-dimensional meteorological forecast data, thus making up for the shortcomings of a single data source in terms of temporal foresight or spatial representativeness.

[0049] 2. This invention utilizes machine learning algorithms to deeply mine historical monitoring data, enabling online dynamic correction of crop coefficients. Compared to the traditional static lookup method, this solution can adaptively adjust irrigation strategies based on the actual crop growth and local microclimate characteristics.

[0050] 3. By supporting real-time monitoring of soil moisture at multiple levels and weighting the data according to the root activity patterns at different growth stages, the system can accurately capture the core areas where crops absorb water.

[0051] 4. This invention enables remote control and automatic execution through an intelligent ecological gateway, and has a complete anomaly monitoring and infiltration feedback mechanism, which greatly reduces the intensity of manual inspection.

[0052] 5. Through precise calculations using a water balance model, this invention maximizes the offsetting of natural rainfall, effectively reducing ineffective irrigation and deep seepage. Attached Figure Description

[0053] Figure 1 This is an architecture diagram of an irrigation decision optimization system based on complementary soil moisture and meteorological data provided in an embodiment of the present invention.

[0054] Figure 2 A flowchart illustrating the irrigation decision optimization method based on the complementarity of soil moisture and meteorological data provided in this embodiment of the invention;

[0055] Figure 3 The rainfall forecast for the next 5 days and the next 7 days for the Tongzhou wheat experimental field on May 12, 2018, provided for this embodiment of the invention. predict;

[0056] Figure 4 The Tongzhou wheat experimental field provided for embodiments of the present invention in 2017 Annual variation curve. Detailed Implementation

[0057] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Example 1

[0059] In a first embodiment of the present invention, the present invention provides an irrigation decision optimization system based on the complementarity of soil moisture and meteorological data, such as... Figure 1 As shown, it includes:

[0060] The soil moisture monitoring unit is used to collect real-time data on soil moisture content and temperature at different depths.

[0061] Preferably, in the soil moisture monitoring unit, the soil moisture monitoring unit adopts a layered soil moisture sensor based on FDR frequency domain reflectance technology; the layered soil moisture sensor has a built-in triaxial accelerometer sensor for detecting vibrations around the equipment and stopping data acquisition during vibrations to eliminate interference.

[0062] In this embodiment, a layered soil moisture sensor is configured with multiple monitoring points vertically, preferably in 10cm layers. Each monitoring point independently integrates a high-frequency sensing circuit and is placed in the crop root activity profile through pre-drilled holes, enabling synchronous acquisition of soil moisture content and temperature at different depths. The sensor is based on frequency domain reflectance (FDR) technology, utilizing high-frequency electromagnetic waves to sense changes in the soil dielectric constant. Since the dielectric constant of water is much higher than that of dry soil and air, the system collects the frequency signal output by the oscillation circuit and converts it into a voltage signal to calculate the soil volumetric water content.

[0063] The sensor incorporates a triaxial accelerometer to detect vibrations or displacements of the equipment in the X, Y, and Z axes in real time. When the detected acceleration value exceeds a preset stability threshold, it is determined to be due to external mechanical operation or human disturbance, and the system automatically executes logic protection. During vibration triggering, the system suspends sending acquisition commands to the high-frequency sensing circuit or directly discards the frequency feedback value for that period, ensuring that the data ultimately output to the gateway is not affected by noise caused by electrode displacement interference.

[0064] Through the above implementation method, the system has the ability to "resist physical interference" at the underlying perception stage, ensuring that the water data subsequently integrated into the cloud platform has extremely high authenticity and avoiding the risk of misjudging the timing of irrigation.

[0065] The meteorological data acquisition unit is used to acquire field microclimate data in real time and integrate third-party remote sensing meteorological data to obtain future rainfall forecast data and future reference crop evapotranspiration forecast data.

[0066] Preferably, the steps of acquiring future rainfall forecast data and future reference crop evapotranspiration forecast data in the meteorological data acquisition unit include:

[0067] Meteorological parameters, including atmospheric pressure, wind direction, wind speed, rainfall, total radiation, air temperature, and air humidity, are collected in real time using small field weather stations.

[0068] By integrating third-party remote sensing meteorological data through a cloud platform, we can obtain rainfall forecast data for the next 5 days and reference crop evapotranspiration forecast data for the next 7 days.

[0069] In this embodiment, the system deploys an automated small weather station in the field, which integrates a high-precision atmospheric pressure sensor, wind vane, wind speed sensor, tipping bucket rain gauge, total radiation sensor, and air temperature and humidity transmitter. Through these sensors, the system can capture the most realistic microclimate characteristics of the field in real time. For example, measured data on total radiation and air temperature and humidity can be used to verify the local instantaneous moisture evaporation intensity, ensuring that even in complex terrain or local microclimate environments, the system can still obtain accurate real-time background parameters.

[0070] The cloud platform periodically retrieves third-party remote sensing meteorological service data corresponding to the geographical location via a standard API interface. In this embodiment, the system focuses on extracting the probability and expected rainfall for the next 5 days, as well as the predicted evapotranspiration for the next 7 days. The rainfall forecast is used to determine whether there will be sufficient natural precipitation to cover the crop's water needs, thereby allowing for the early shutdown or delay of irrigation plans. The evapotranspiration forecast is used to obtain the theoretical water consumption trend of a standard reference crop over the next week.

[0071] Through the above implementation methods, the system combines real-time environmental monitoring with medium- and long-term weather forecasting, solving the problems of poor timeliness or insufficient local representativeness of data from single meteorological sources.

[0072] The data fusion and processing unit connects the soil moisture monitoring unit and the meteorological data acquisition unit through an intelligent ecological gateway, and is used to upload the collected data to the cloud platform for cleaning, fusion and analysis.

[0073] Preferably, in the data fusion and processing unit, the steps of uploading the collected data to the cloud platform for cleaning, fusion, and analysis include:

[0074] The real-time collected soil moisture and meteorological data are cleaned to identify and remove outlier data.

[0075] Multi-layer soil moisture data are spatiotemporally correlated to calculate the moisture weight of each layer within the crop root activity range, and feature vectors reflecting the current field evapotranspiration environment are generated by combining meteorological parameters.

[0076] The cleaned local data is aligned with the acquired third-party weather forecast data to establish a time series database.

[0077] In this embodiment, after receiving real-time collected soil moisture and meteorological data, the cloud platform first performs a quality assessment. Statistical methods such as the three-standard-deviation method or median filtering are used to identify outliers in the data sequence. For example, pulse noise caused by transient communication interference from sensors is removed, or invalid sampled values ​​triggered by accelerometer protection logic during agricultural machinery operations are removed, ensuring that the data used in subsequent analysis are within a reasonable physical and logical range.

[0078] The system performs spatiotemporal modeling on moisture data acquired from stratified soil moisture sensors at different depths, with an LSTM model being the preferred model for this part. Based on the current crop type and its growth stage, the system dynamically calculates the intensity of root activity at each depth, assigning different weight coefficients to each layer of moisture data to obtain a comprehensive water content reflecting the entire root activity zone. Since the crop root system's ability to absorb soil water at different depths varies at different growth stages, the system dynamically generates a set of weight coefficient arrays. For example, high weights are assigned to shallow water (0-20cm) in the early growth stage, while the weight of water in the 40-60cm depth is increased in the later growth stage. The temporal features extracted by the LSTM are weighted and summed with this weight coefficient array to calculate a comprehensive effective water content index reflecting the entire crop root activity zone. This weighted soil moisture index is then nonlinearly fused with locally measured meteorological parameters to generate a set of feature vectors that can characterize the actual evapotranspiration environment in the field in real time.

[0079] The cloud platform aligns the cleaned high-frequency local measured data with low-frequency weather forecast data obtained from third parties in terms of time step. Through resampling or interpolation algorithms, heterogeneous data from different sources and in different formats are uniformly mapped onto the same time axis, establishing a structured time series database.

[0080] Through the above implementation methods, the system realizes the transformation from raw data to feature information, which not only eliminates monitoring noise caused by the physical environment, but also truly restores the water transport characteristics of the crop-soil-atmosphere continuum through root weight allocation and spatiotemporal alignment.

[0081] The decision optimization unit is used to determine the effective water storage capacity and water storage capacity of the soil based on the water balance model and machine learning algorithm, and dynamically correct the crop coefficient by combining the predicted meteorological data, and generate decision instructions including irrigation timing, irrigation water volume and irrigation cycle.

[0082] Preferably, in the decision optimization unit, the steps of determining the effective water storage capacity and water holding capacity of the soil using real-time soil moisture content include:

[0083] Based on machine learning algorithms, the crop coefficient for the current growth stage is dynamically adjusted according to the correlation between historical soil moisture decline rate and meteorological factors during the corresponding period.

[0084] By combining the corrected crop coefficient with the predicted value of reference crop evapotranspiration, the actual future water requirement of the crop can be calculated.

[0085] Based on the actual water demand and future rainfall forecast data, the irrigation start time and target irrigation amount are determined using a water balance model.

[0086] In this embodiment, the system first obtains the crop coefficient benchmark value, which reflects the crop's physiological characteristics and environmental adaptability. This coefficient represents the proportion of water consumption of a specific crop at a specific growth stage relative to a standard reference crop. Machine learning algorithms are used to analyze the water depletion rate of soil moisture sensors over historical periods and to perform correlation analysis with meteorological factors during the same period. If the monitored actual water consumption deviates from the theoretical expectation, the system will automatically adjust the aforementioned crop coefficient. Online corrections are performed to generate personalized parameters that accurately represent the current vegetation growth and soil water supply capacity.

[0087] The system utilizes the aforementioned corrected real-time crop coefficients. Combined with the 7-day reference crop evapotranspiration forecast obtained from the meteorological data acquisition unit Calculated through product association This method calculates the actual daily water requirements of crops over the future forecast period. It addresses the response lag problem caused by traditional irrigation decisions that rely solely on past data.

[0088] By utilizing real-time monitored soil moisture content, combined with soil property parameters such as field capacity and wilting coefficient, the system determines the current effective soil water storage capacity and the remaining water storage space before reaching the irrigation ceiling. The system constructs a water balance equation, subtracting the predicted water demand for the future cycle from the current water storage capacity and adding the effective rainfall from the predicted rainfall data for the next 5 days. When the predicted moisture content calculated by the equation is about to reach the preset lower limit of soil moisture, the system determines this as the start time for irrigation, and uses the difference required to replenish the soil moisture to the upper limit as the irrigation target amount. Finally, it generates a decision instruction that includes the irrigation timing, water volume, and cycle.

[0089] Through the above implementation methods, the system uses machine learning for dynamic correction. The system can adapt to the differences in crop water consumption under different plots and growth conditions. With the help of the water balance model to pre-emptively offset rainfall, it maximizes the use of natural precipitation and significantly reduces the frequency of human intervention and irrigation costs.

[0090] Preferably, the step of dynamically correcting the crop coefficient for the current growth stage includes:

[0091] Based on the crop types and their current growth stages in the area to be monitored, a corresponding preset crop coefficient benchmark value is matched;

[0092] Using historical water loss data collected by the soil moisture monitoring unit and combined with real-time meteorological parameters from the same period, the actual water consumption change rate of crops under the current growth environment is obtained by fitting.

[0093] By comparing the actual rate of change in water consumption with the theoretical evapotranspiration level calculated based on meteorological environment, the degree of deviation between the actual water demand and the theoretical prediction is analyzed.

[0094] The preset crop coefficient benchmark value is corrected online based on the degree of deviation to generate a crop coefficient that reflects the current growth stage.

[0095] In this embodiment, the system first matches and retrieves the corresponding preset crop coefficient benchmark values ​​from a preset database based on the types of crops planted in the area to be monitored and their current growth stages, such as winter wheat, summer corn, etc., and their corresponding seedling, jointing, or maturity stages. These benchmark values ​​are based on the theoretical water consumption ratio derived under standard agricultural climate conditions.

[0096] The system utilizes historical water loss data collected by soil moisture monitoring units—that is, the decline curve of soil moisture content over a period of time—combined with real-time meteorological parameters from the same period. Through a fitting algorithm, it calculates the actual rate of change in crop water consumption under the current specific growth environment. This rate of change excludes non-agricultural evaporation factors and accurately reflects the intensity of water absorption and transpiration by plants through their roots.

[0097] The system compares the actual rate of change in water consumption with the theoretical evapotranspiration level calculated based on meteorological conditions. During this process, the system analyzes the degree of deviation between the actual water demand and the theoretical prediction. For example, if the measured rate of water consumption is significantly higher than the theoretical prediction, it is determined that the crop has a larger leaf area index, is growing better than usual, or that poor soil water retention capacity is causing increased water consumption.

[0098] Based on the aforementioned deviation, the system uses a machine learning model to correct the preset crop coefficient baseline value online. Through gain or loss compensation, personalized crop coefficients reflecting the current true vegetation growth and soil water supply capacity are generated. The corrected coefficient will be immediately incorporated into the next stage of water balance calculations to ensure the accuracy of irrigation instructions.

[0099] The above implementation method solves the problem of obtaining crop coefficients using the traditional table lookup method. The problem of overly generalized approaches failing to adapt to the differences in crop growth in specific fields has been addressed by implementing adaptive learning of crop water consumption patterns, thereby ensuring that the generated irrigation decision instructions can meet the actual needs of farmland.

[0100] The execution control unit is connected to the intelligent ecological gateway and is used to automatically control the irrigation equipment according to the decision instructions to achieve closed-loop irrigation management.

[0101] Preferably, in the execution control unit, the step of automatically controlling the irrigation equipment according to the decision command includes:

[0102] The intelligent ecological gateway receives decision instructions issued by the cloud platform and parses them into executable signals for the corresponding irrigation area control terminal.

[0103] During irrigation, the operating status of the water supply network and the feedback parameters of the irrigation equipment are monitored in real time to ensure that the irrigation operation is carried out according to the preset instructions;

[0104] The irrigation infiltration effect is dynamically verified based on the real-time moisture rise rate transmitted back by the soil moisture monitoring unit.

[0105] When the monitored soil moisture reaches the preset upper limit of moisture content or the preset irrigation time is completed, the irrigation equipment is automatically turned off, and the operation completion data is uploaded to the cloud platform to update the water balance model.

[0106] In this embodiment, the smart ecological gateway receives decision commands from the cloud platform via wireless communication methods such as 4G / 5G or LoRa. The gateway's built-in parsing module breaks down the commands into electrical signals that can be recognized by the control terminal of the specific irrigation area, driving the water pump to start or controlling the action of the valves in each zone.

[0107] During irrigation operations, the system monitors the operating status of the water supply network and the feedback parameters of the irrigation equipment in real time through sensors. This ensures that the irrigation operation is strictly executed according to the preset instructions. If a pipe bursts or the water pump runs dry, the system can immediately identify abnormal pressure or flow and trigger a protective shutdown.

[0108] The system calls upon data transmitted from the soil moisture monitoring unit to calculate the real-time rate of increase in soil moisture during irrigation. If the rate of increase is lower than expected, the system determines that surface runoff is severe or infiltration is obstructed, and will automatically adjust the irrigation pulse duration; if the rate of increase is extremely fast, there may be preferential flow around the sensor.

[0109] The system employs a dual-safety shutdown logic. When the monitored soil moisture reaches the preset upper limit, or the irrigation duration reaches the preset irrigation time, the system automatically sends a valve-closing command. After the operation is completed, the gateway uploads operational data such as the actual irrigation start time, end time, and total irrigation volume to the cloud platform. This real-world execution data serves as feedback input to update the water balance model for the next round, enabling continuous algorithm evolution.

[0110] Through the above implementation methods, the system constructs a complete closed loop from command issuance to physical execution and effect feedback, which not only ensures the safety of irrigation operations, but also continuously improves the accuracy of the water balance prediction model through the feedback mechanism.

[0111] Example 2

[0112] In a wheat experimental field in Tongzhou, after entering the jointing stage, the actual vegetation growth was significantly better than the conventional growth model for this variety due to the influence of local microclimate. At this time, relying solely on traditional irrigation guidelines to obtain crop coefficients cannot accurately reflect the additional transpiration demand caused by the large leaf area of ​​this plot, which can easily lead to delayed irrigation and thus affect wheat yield.

[0113] To address the aforementioned problems, the irrigation decision optimization method based on the complementarity of soil moisture and meteorological data, as provided in this invention, is adopted. The process is as follows: Figure 2 As shown. This method is based on data simulation from a wheat experimental field in Tongzhou. The specific implementation process is as follows:

[0114] S1: The multi-layer soil moisture acquisition system collects real-time data on soil volumetric moisture content and temperature at a depth of 10cm-40cm using sensors deployed at the wheat root profile in the experimental field. This underlying data provides a physical basis for determining the current water supply status of the wheat.

[0115] S2: The meteorological data integration system acquires real-time microclimate data such as air temperature, humidity, and radiation collected from small field weather stations. Simultaneously, it integrates third-party remote sensing meteorological services to obtain rainfall forecasts for the next 5 days and reference crop evapotranspiration for the next 7 days for the plot as of May 12, 2018. Forecast values, specific forecast trends as follows Figure 3 As shown.

[0116] S3: Cloud Platform Data Preprocessing. The intelligent ecological gateway synchronously uploads the soil moisture and meteorological data collected in steps S1 and S2 to the cloud platform. The cloud platform performs data cleaning, removes outliers and noise, and uses an LSTM model to perform spatiotemporal correlation analysis on multi-layer moisture data to extract feature vectors reflecting the current field evapotranspiration environment.

[0117] S4: Decision Instruction Generation. The decision optimization unit first uses real-time soil moisture content to determine the current effective water storage and remaining water storage capacity of the wheat root zone. Subsequently, referencing... Figure 4 The 2017 figure shown , and The annual variation curve, based on machine learning algorithms, compares the measured rate of water loss with... Figure 3 The theoretical forecast values ​​are used. Based on the deviation between actual water demand and theoretical predictions, the system dynamically corrects the crop coefficient for the current winter wheat growth stage online. Finally, combined with the revised version Based on future meteorological data, precise instructions including irrigation timing, irrigation volume, and irrigation cycle are generated through a water balance model.

[0118] S5: The closed-loop irrigation control execution control unit, based on the decision instructions generated by S4, sends signals through the intelligent gateway to drive the irrigation equipment in the experimental field. During operation, the system verifies the infiltration effect based on the real-time rate of moisture rise. When the soil moisture reaches the preset upper limit, the valve is automatically closed, completing this closed-loop irrigation management, and the operation record is fed back to the model library for parameter updates.

[0119] To quantitatively verify the advancement of the method of this invention, four sets of comparative experiments were set up under the same environmental parameters and initial soil moisture content in the Tongzhou wheat experimental field. The simulation results are shown in Table 1:

[0120] As shown in Table 1, the proposed solution significantly outperforms other single-data-source solutions in all key indicators. Regarding irrigation timing, the proposed solution dynamically adjusts the crop coefficient to ensure the minimum moisture content is maintained at 15.2%, effectively avoiding the crop drought risk (12.5% ​​and 13.8%) caused by empirical lookup methods and weather forecasting. In terms of water resource utilization, the proposed solution uses a water balance model to precisely offset 11.5 mm of effective rainfall, avoiding the overlapping of irrigation and rainfall caused by the lack of foresight in soil sensor-only solutions, thus reducing the water consumption per irrigation to 28 m³ / acre. Ultimately, the proposed solution achieves an irrigation water use efficiency of 94%, fully demonstrating that by deeply complementing soil moisture and meteorological data, it is possible to maximize irrigation water conservation while ensuring the physiological needs of crops, demonstrating strong scientific merit and economic benefits.

[0121] Table 1. Comparative experiments under the same simulation environment

[0122] Evaluation indicators Experience lookup table method Soil sensor only Weather forecast only Invention Solution Minimum moisture content before irrigation 12.5% ​​(Severely affected by drought) 15.0% (Trigger Threshold) 13.8% (Moderately Drought Affected) 15.2% (Smooth transition) Effective rainfall utilization (mm) 0 (Not considered) 2.5 (Post-irrigation rainfall) 8.0 (Theoretical Hedging) 11.5 (Precision Hedging) Water consumption per irrigation session (m³ / acre) 46 40 35 28 Irrigation water use efficiency 68% 75% 82% 94%

[0123] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An irrigation decision optimization system based on complementary soil moisture and meteorological data, characterized in that, include: The soil moisture monitoring unit is used to collect real-time data on soil moisture content and temperature at different depths. The meteorological data acquisition unit is used to acquire field microclimate data in real time and integrate third-party remote sensing meteorological data to obtain future rainfall forecast data and future reference crop evapotranspiration forecast data. The data fusion and processing unit connects the soil moisture monitoring unit and the meteorological data acquisition unit through an intelligent ecological gateway, and is used to upload the collected data to the cloud platform for cleaning, fusion and analysis. The decision optimization unit is used to determine the effective water storage capacity and water storage capacity of the soil based on the water balance model and machine learning algorithm, and dynamically correct the crop coefficient by combining the predicted meteorological data, and generate decision instructions including irrigation timing, irrigation water volume and irrigation cycle. The execution control unit is connected to the intelligent ecological gateway and is used to automatically control the irrigation equipment according to the decision instructions to achieve closed-loop irrigation management.

2. The irrigation decision optimization system based on complementary soil moisture and meteorological data as described in claim 1, characterized in that, In the soil moisture monitoring unit, the soil moisture monitoring unit adopts a layered soil moisture sensor based on FDR frequency domain reflectance technology; the layered soil moisture sensor has a built-in triaxial accelerometer sensor to detect vibrations around the equipment and stop data acquisition during vibrations to eliminate interference.

3. The irrigation decision optimization system based on complementary soil moisture and meteorological data as described in claim 1, characterized in that, The steps for acquiring future rainfall forecast data and future reference crop evapotranspiration forecast data in the meteorological data acquisition unit include: Meteorological parameters, including atmospheric pressure, wind direction, wind speed, rainfall, total radiation, air temperature, and air humidity, are collected in real time using small field weather stations. By integrating third-party remote sensing meteorological data through a cloud platform, we can obtain rainfall forecast data for the next 5 days and reference crop evapotranspiration forecast data for the next 7 days.

4. The irrigation decision optimization system based on complementary soil moisture and meteorological data as described in claim 1, characterized in that, In the data fusion and processing unit, the steps of uploading the collected data to the cloud platform for cleaning, fusion, and analysis include: The real-time collected soil moisture and meteorological data are cleaned to identify and remove outlier data. Multi-layer soil moisture data are spatiotemporally correlated to calculate the moisture weight of each layer within the crop root activity range, and feature vectors reflecting the current field evapotranspiration environment are generated by combining meteorological parameters. The cleaned local data is aligned with the acquired third-party weather forecast data to establish a time series database.

5. The irrigation decision optimization system based on complementary soil moisture and meteorological data according to claim 1, characterized in that, In the decision optimization unit, the steps for determining the effective water storage capacity and water holding capacity of the soil using real-time soil moisture content include: Based on machine learning algorithms, the crop coefficient for the current growth stage is dynamically adjusted according to the correlation between historical soil moisture decline rate and meteorological factors during the corresponding period. By combining the corrected crop coefficient with the predicted value of reference crop evapotranspiration, the actual future water requirement of the crop can be calculated. Based on the actual water demand and future rainfall forecast data, the irrigation start time and target irrigation amount are determined using a water balance model.

6. The irrigation decision optimization system based on complementary soil moisture and meteorological data according to claim 5, characterized in that, The steps for dynamically correcting the crop coefficient for the current growth stage include: Based on the crop types and their current growth stages in the area to be monitored, a corresponding preset crop coefficient benchmark value is matched; Using historical water loss data collected by the soil moisture monitoring unit and combined with real-time meteorological parameters from the same period, the actual water consumption change rate of crops under the current growth environment is obtained by fitting. By comparing the actual rate of change in water consumption with the theoretical evapotranspiration level calculated based on meteorological environment, the degree of deviation between the actual water demand and the theoretical prediction is analyzed. The preset crop coefficient benchmark value is corrected online based on the degree of deviation to generate a crop coefficient that reflects the current growth stage.

7. The irrigation decision optimization system based on complementary soil moisture and meteorological data according to claim 1, characterized in that, In the execution control unit, the steps of automatically controlling the irrigation equipment according to the decision instructions include: The intelligent ecological gateway receives decision instructions issued by the cloud platform and parses them into executable signals for the corresponding irrigation area control terminal. During irrigation, the operating status of the water supply network and the feedback parameters of the irrigation equipment are monitored in real time to ensure that the irrigation operation is carried out according to the preset instructions; The irrigation infiltration effect is dynamically verified based on the real-time moisture rise rate transmitted back by the soil moisture monitoring unit. When the monitored soil moisture reaches the preset upper limit of moisture content or the preset irrigation time is completed, the irrigation equipment is automatically turned off, and the operation completion data is uploaded to the cloud platform to update the water balance model.

8. An irrigation decision optimization method based on the complementarity of soil moisture and meteorological data, characterized in that, The method for the irrigation decision optimization system according to any one of claims 1-7 comprises: S1. Real-time collection of soil moisture content and temperature data at different depths; S2. Real-time acquisition of field microclimate data, and integration of third-party remote sensing meteorological data to obtain future rainfall forecast data and future reference crop evapotranspiration forecast data; S3. Upload the collected data to the cloud platform for cleaning, fusion and analysis; S4. Based on the water balance model and machine learning algorithm, the effective water storage capacity and water storage capacity of the soil are determined by real-time soil moisture content, and the crop coefficient is dynamically corrected by combining the predicted meteorological data to generate decision instructions including irrigation timing, irrigation water volume and irrigation cycle. S5. Automatically control the irrigation equipment according to the decision instructions to achieve closed-loop irrigation management.

9. The irrigation decision optimization method based on the complementarity of soil moisture and meteorological data according to claim 8, characterized in that, In step S4, the steps of determining the effective water storage capacity and water holding capacity of the soil using real-time soil moisture content include: Based on machine learning algorithms, the crop coefficient for the current growth stage is dynamically adjusted according to the correlation between historical soil moisture decline rate and meteorological factors during the corresponding period. By combining the corrected crop coefficient with the predicted value of reference crop evapotranspiration, the actual future water requirement of the crop can be calculated. Based on the actual water demand and future rainfall forecast data, the irrigation start time and target irrigation amount are determined using a water balance model.

10. The irrigation decision optimization method based on the complementarity of soil moisture and meteorological data according to claim 9, characterized in that, The steps for dynamically correcting the crop coefficient for the current growth stage include: Based on the crop types and their current growth stages in the area to be monitored, a corresponding preset crop coefficient benchmark value is matched; Using historical water loss data collected by the soil moisture monitoring unit and combined with real-time meteorological parameters from the same period, the actual water consumption change rate of crops under the current growth environment is obtained by fitting. By comparing the actual rate of change in water consumption with the theoretical evapotranspiration level calculated based on meteorological environment, the degree of deviation between the actual water demand and the theoretical prediction is analyzed. The preset crop coefficient benchmark value is corrected online based on the degree of deviation to generate a crop coefficient that reflects the current growth stage.

Images