Intelligent water-saving control method and system for agricultural irrigation device

Abstract

The invention provides an intelligent water-saving control method and system for an agricultural irrigation device, and the method comprises the following steps: generating a farmland state prediction index based on soil comprehensive information, weather information and crop information, and generating a farmland irrigation strategy; calculating an irrigation emergency degree based on the farmland state prediction index; if the irrigation emergency degree reaches the highest emergency degree, the filtering irrigation module is controlled to execute alternate irrigation operation; if the irrigation emergency degree does not reach the highest emergency degree, calculating to obtain the backwashing emergency degree of the filtering irrigation module; if the backwash emergency degree is greater than or equal to the irrigation emergency degree, generating a backwash strategy, and sequentially executing the backwash strategy and the farmland irrigation strategy; if the backwash emergency degree is smaller than the irrigation emergency degree, the filtering irrigation module is controlled to execute alternate irrigation operation on the target farmland. According to the invention, intelligent irrigation and backwashing control can be realized, and a better water-saving effect is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of irrigation water-saving control, and in particular relates to an intelligent water-saving control method and system for agricultural irrigation devices. Background Art

[0002] In the context of increasingly tight global water resources and huge challenges facing agricultural production, on the one hand, population growth and climate change have led to an intensification of the contradiction between water supply and demand, and agriculture, as a major water user, is facing increasing pressure to save water. On the other hand, in order to meet the growing demand for food, agricultural production must improve efficiency and output. In this context, how to achieve the dual goals of water conservation and increased production has become a core issue in the development of agricultural irrigation technology.

[0003] Traditional irrigation systems generally lack intelligent and refined management capabilities. Irrigation devices often rely on fixed irrigation plans or manual experience and are difficult to flexibly adjust according to actual conditions. This "one-size-fits-all" approach will result in over-irrigation in some areas and under-irrigation in others. Ultimately, this will not only affect crop yields, but also lead to a serious waste of water resources. Summary of the invention

[0004] The present invention provides an intelligent water-saving control method and system for an agricultural irrigation device, so as to solve the problem of serious waste of water resources that is prone to occur during the application of a traditional irrigation device.

[0005] In a first aspect, the present invention provides an intelligent water-saving control method for an agricultural irrigation device, which is applied to the agricultural irrigation device, wherein the agricultural irrigation device comprises a filtering irrigation module and a backwashing module, wherein the irrigation unit of the filtering irrigation module is connected to a target farmland to be irrigated, and the irrigation unit is located at one end of the target farmland and is provided with a multifunctional irrigation outlet head, and the backwashing module is used to backwash the filtering irrigation module, and the method comprises the following steps:

[0006] Collecting comprehensive soil information of the target farmland through a first sensor preset in the target farmland;

[0007] Acquiring weather information of the target farmland and crop information in the target farmland;

[0008] Based on the comprehensive soil information, the meteorological information and the crop information and using a preset farmland state prediction model, a farmland state prediction index of the target farmland is generated, and a farmland irrigation strategy is generated based on the farmland state prediction index, wherein the farmland state prediction model is a CNN-LSTM model;

[0009] Based on the farmland state prediction index, the irrigation urgency corresponding to the farmland irrigation strategy is calculated through a fuzzy logic system;

[0010] If the irrigation urgency reaches the highest level of urgency, controlling the filtering irrigation module to perform an alternate irrigation operation on the target farmland according to the farmland irrigation strategy;

[0011] If the irrigation urgency level does not reach the highest level of urgency, the filtering component data of the filtering irrigation module is collected through a second sensor preset inside the filtering irrigation module, and the historical backwashing data of the backwashing module is obtained;

[0012] Combining the filter component data, the historical backwash data and the farmland irrigation strategy, and calculating the backwash urgency of the filter irrigation module through a fuzzy logic system;

[0013] If the backwashing urgency is greater than or equal to the irrigation urgency, then generating a backwashing strategy for the backwashing module to perform a backwashing operation for the filter irrigation module based on the filter component data, and controlling the backwashing module and the filter irrigation module to sequentially perform the backwashing strategy and the farmland irrigation strategy;

[0014] If the backwash urgency is less than the irrigation urgency, the filtering irrigation module is controlled to perform an alternating irrigation operation on the target farmland according to the farmland irrigation strategy.

[0015] Optionally, the comprehensive soil information includes soil moisture information and soil spatial distribution information, and the crop information includes crop remote sensing information and crop type information.

[0016] Optionally, the generating the farmland state prediction index of the target farmland based on the soil comprehensive information, the meteorological information and the crop information and using a preset farmland state prediction model comprises the following steps:

[0017] preprocessing the soil moisture information, the soil spatial distribution information and the meteorological information;

[0018] fusing the preprocessed soil moisture information and the soil spatial distribution information into soil moisture distribution data in a spatial dimension;

[0019] fusing the crop remote sensing information and the crop type information into the soil spatial distribution information in a spatial dimension to obtain crop spatial distribution data;

[0020] The soil moisture distribution data and the crop spatial distribution data are input into a preset farmland state prediction model, and a soil state prediction index and a crop state prediction index are output through a convolutional neural network module in the farmland state prediction model;

[0021] Inputting the preprocessed meteorological information into the farmland state prediction model, and outputting meteorological prediction indicators through the long short-term memory network module in the farmland state prediction model;

[0022] The soil state prediction index, the crop state prediction index and the meteorological prediction index are integrated into a farmland state prediction index.

[0023] Optionally, fusing the preprocessed soil moisture information and the soil spatial distribution information into soil moisture distribution data in the spatial dimension comprises the following steps:

[0024] Optionally, spatial registration processing is performed on the soil moisture information and the soil spatial distribution information so that the soil moisture information and the soil spatial distribution information are in the same spatial reference system;

[0025] Resampling the soil moisture information and the soil spatial distribution information after spatial registration processing to a spatial grid of a preset size;

[0026] For each of the spatial grids, all the soil moisture information in the spatial grid is aggregated in time series to obtain soil moisture aggregated data;

[0027] Traversing all the spatial grids, for a null-valued spatial grid where the soil moisture aggregated data does not exist, estimating the soil moisture aggregated data of the null-valued spatial grid based on all the soil moisture aggregated data of the adjacent spatial grids of the null-valued spatial grid and using a spatial interpolation method;

[0028] The soil moisture aggregation data of all the spatial grids are fused with the soil spatial distribution information into soil moisture distribution data in a multi-layer grid form.

[0029] Optionally, fusing the crop remote sensing information and the crop type information into the soil spatial distribution information in a spatial dimension to obtain crop spatial distribution data comprises the following steps:

[0030] resampling the crop remote sensing information to the spatial grid;

[0031] Calculating the crop growth status in each of the spatial grids using vegetation index;

[0032] estimating crop root distribution data within each of the spatial grids in combination with the crop growth status and the crop type;

[0033] The crop growth status, the crop type and the crop root distribution data are fused with the soil spatial distribution information into crop spatial distribution data in a multi-layer grid form.

[0034] Optionally, the step of calculating the irrigation urgency corresponding to the farmland irrigation strategy based on the farmland state prediction index through a fuzzy logic system comprises the following steps:

[0035] Using the farmland state prediction index as the first fuzzy input variable of the fuzzy logic system;

[0036] defining a fuzzy set and a first membership function for each of the first fuzzy input variables;

[0037] combining the fuzzy set and the first membership function, and using a fuzzy inference method to map all the first fuzzy input variables to fuzzy output variables according to a preset irrigation fuzzy inference rule;

[0038] The fuzzy output variable is defuzzified by using the center of gravity method to obtain the irrigation urgency corresponding to the farmland irrigation strategy.

[0039] Optionally, the filter component data includes the inlet and outlet water pressure difference, filtered water flow rate, filtered water turbidity and cumulative filtered water volume.

[0040] Optionally, combining the filter component data, the historical backwash data and the farmland irrigation strategy, and calculating the backwash urgency of the filter irrigation module through a fuzzy logic system comprises the following steps:

[0041] The Kalman filter method is used to merge the inlet and outlet water pressure difference, the filtered water flow rate and the turbidity of the filtered water into real-time filtering data;

[0042] Based on the real-time filtering data, the predicted filtering data is obtained by predicting through a preset filtering state prediction model, wherein the filtering state prediction model is an LSTM model;

[0043] using the predicted filtering data and the historical backwashing data as second fuzzy input variables of a fuzzy logic system;

[0044] defining a second membership function for each of the second fuzzy input variables;

[0045] Based on the second membership function and according to the preset backwashing fuzzy inference rule, a backwashing association relationship matrix is ​​established;

[0046] Analyzing the second fuzzy input variable by the hierarchical analysis method to obtain a matrix weight vector of the backwashing association relationship matrix;

[0047] Combining the backwash association relationship matrix and the matrix weight vector and obtaining a backwash reasoning evaluation result through fuzzy synthesis operation;

[0048] The backwashing reasoning evaluation result is defuzzified by using the center of gravity method to obtain the initial backwashing urgency degree;

[0049] An irrigation influencing factor is constructed based on the irrigation duration and irrigation water volume in the farmland irrigation strategy, and the initial backwashing urgency is corrected using the irrigation influencing factor to obtain the backwashing urgency of the filtering irrigation module.

[0050] Optionally, the expression of the irrigation influencing factor is as follows:

[0051]

[0052] Where: δ represents the irrigation impact factor, α represents the irrigation water impact weight, q represents the irrigation water impact weight, n represents the irrigation water volume, Q represents the preset backwash interval water volume, Q M represents the cumulative filtered water volume of the filtering irrigation module, β represents the influence weight of irrigation time, t n represents the irrigation duration, and T represents the preset maximum irrigation duration.

[0053] In a second aspect, the present invention also provides an intelligent water-saving control system for an agricultural irrigation device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the intelligent water-saving control method for an agricultural irrigation device as described in the first aspect is implemented.

[0054] The beneficial effects of the present invention are:

[0055] Compared with the prior art, the present invention realizes more accurate and efficient water resource management, and greatly improves the overall operation efficiency of the irrigation system. First, the present invention provides a comprehensive and accurate basis for irrigation decision-making through multi-dimensional data collection and analysis, including comprehensive soil information, meteorological information and crop information, and effectively avoids the problem of improper irrigation caused by insufficient information in traditional methods. Secondly, the present invention adopts the advanced CNN-LSTM model as the farmland state prediction model, which can more accurately predict the farmland state, so as to formulate the optimal irrigation strategy. This method based on deep learning greatly improves the accuracy of irrigation and effectively reduces the waste of water resources. On the other hand, the present invention introduces a fuzzy logic system to evaluate the urgency of irrigation and the urgency of backwashing. This flexible decision-making mechanism can better balance the relationship between irrigation demand and equipment maintenance, and ensure that the system always operates in the best state. The present invention also takes into account the maintenance needs of the irrigation system itself, and realizes the coordinated optimization of irrigation and equipment maintenance by intelligently judging the necessity and urgency of backwashing. This not only prolongs the service life of the equipment, but also improves the reliability and stability of the entire irrigation system. In summary, compared with the prior art, the present invention has achieved significant improvements in water-saving effect, irrigation accuracy, system reliability and operating efficiency, providing strong technical support for modern agricultural production. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of an intelligent water-saving control method for an agricultural irrigation device in one embodiment of the present application.

[0057] Figure 2 This is a schematic diagram of the system structure of an intelligent water-saving control system for an agricultural irrigation device in one embodiment of the present application. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0059] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0060] The present invention discloses an intelligent water-saving control method for an agricultural irrigation device. The method is applied to the agricultural irrigation device. Figure 1 The agricultural irrigation device consists of a filtering irrigation module and a backwashing module. The filtering irrigation module is connected to the target farmland and is equipped with a multifunctional irrigation outlet. The backwashing module is used to clean the filtering irrigation module. The operating principle of the system is as follows:

[0061] First, the first sensor collects comprehensive soil information of the target farmland, including soil moisture and spatial distribution data; at the same time, the smart agriculture management system obtains local meteorological information and crop information (including remote sensing and type data). The central processor uses the CNN-LSTM model as a farmland state prediction model, generates farmland state prediction indicators based on the above collected information, and formulates farmland irrigation strategies accordingly. Then, the system calculates the irrigation urgency through fuzzy logic. If the highest level of urgency is reached, the system directly controls the filter irrigation module to perform alternating irrigation operations; if not, the second sensor collects the filter component data, and combines the historical backwash data to calculate the backwash urgency again through fuzzy logic. The system will compare the backwash urgency with the irrigation urgency. If the former is greater than or equal to the latter, the backwash operation will be performed first and then irrigation; otherwise, the alternating irrigation operation will be performed directly.

[0062] This intelligent control method fully considers the soil, weather, crop conditions and the working status of irrigation equipment. Through multi-sensor data collection, deep learning model prediction and fuzzy logic decision-making, it realizes the precision and intelligence of agricultural irrigation. The core of the system lies in its adaptability and predictability. It can make the best decision according to real-time environmental changes and equipment status, which not only ensures the water demand of crops, but also saves water resources to the maximum extent. The central processor, as the brain of the system, integrates the CNN-LSTM model for farmland status prediction and the LSTM model for processing time series data. This deep learning architecture can effectively capture the complex dynamic changes of the farmland environment. The introduction of the fuzzy logic system enhances the flexibility and robustness of decision-making, and can handle the uncertainty and ambiguity in agricultural production. The whole system forms a closed-loop intelligent control process: from environmental perception, data analysis, status prediction to decision execution, and then to feedback adjustment, it realizes the all-round intelligent management of agricultural irrigation.

[0063] Figure 2 FIG. 1 is a flow chart of an intelligent water-saving control method for an agricultural irrigation device in one embodiment. It should be understood that although Figure 2 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2 At least part of the steps in the above method may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps. Figure 2 As shown, the intelligent water-saving control method for an agricultural irrigation device disclosed in the present invention specifically includes the following steps:

[0064] S101. Collect comprehensive soil information of the target farmland through a first sensor preset in the target farmland.

[0065] Among them, the first sensor is pre-arranged in the target farmland to collect comprehensive soil information, which includes soil moisture information and soil spatial distribution information. These sensors are high-precision multifunctional soil monitoring equipment that can simultaneously measure soil moisture and spatial distribution characteristics. Soil moisture information is obtained through capacitive or time-domain reflective sensors, which can accurately measure soil moisture content at different depths. The sensor network covers the entire farmland and is distributed in a grid shape. A sensor is installed at each grid point, usually at an interval of 10-20 meters. This dense arrangement ensures the spatial representativeness of the data. Soil spatial distribution information is obtained through georesistance imaging technology, which uses an electrode array to send a weak current underground and infers the spatial distribution of soil type, structure and moisture content by measuring resistivity changes. The sensor collects data every 15 minutes to ensure real-time performance. The collected data is transmitted wirelessly to the central data processing unit, and after preliminary processing, a complete soil comprehensive information data set is formed. These data include soil volumetric water content at different depths (such as 0-10cm, 10-30cm, 30-60cm), soil temperature, electrical conductivity and other parameters, as well as a three-dimensional soil structure model that reflects soil spatial heterogeneity. The implementation of this step provides key basic data for subsequent precision irrigation decisions, which can reflect the dynamic changes and spatial differences in farmland soil moisture conditions, and effectively avoid the problem of irrigation imbalance caused by traditional single-point measurement methods.

[0066] S102. Obtaining weather information at the location of the target farmland and crop information in the target farmland.

[0067] Among them, Figure 1 As shown in the figure, meteorological information and crop information can be obtained through the smart agriculture management system, where crop information includes crop remote sensing information and crop type information. The acquisition of meteorological information mainly relies on automatic weather stations set up around farmland. These weather stations are equipped with a variety of sensors that can monitor parameters such as temperature, humidity, wind speed, rainfall, solar radiation, etc. in real time. The weather station records data every 5 minutes and transmits it to the smart agriculture management system in real time via GPRS or 4G network. At the same time, the mesoscale numerical weather forecast model data provided by the regional meteorological department will be connected to obtain weather forecast information for the next 3-7 days. After downscaling, these forecast data can provide more accurate predictions of farmland microclimate.

[0068] The acquisition of crop information combines ground observation and remote sensing technology. Crop remote sensing information is mainly obtained through multispectral satellite images and drone aerial photography. Satellite remote sensing data (such as Sentinel-2 satellite) is acquired every 5-10 days with a resolution of up to 10 meters. Drones are equipped with multispectral cameras and conduct low-altitude aerial photography once a week with a resolution of up to centimeters. After pre-processing such as atmospheric correction, geometric correction and radiation calibration, these remote sensing data are used to calculate vegetation indices such as normalized vegetation index (NDVI) and leaf area index (LAI) to reflect crop growth conditions. Crop type information is obtained through historical planting records.

[0069] By comprehensively utilizing these multi-source data, a comprehensive farmland environment and crop growth database has been formed. Meteorological information can reflect the water evaporation demand and precipitation replenishment of farmland, crop remote sensing information can timely capture the dynamic changes of crop growth, and crop type information provides a basic reference for crop water demand. The integration of this information provides comprehensive and accurate input data for subsequent farmland status prediction and irrigation decision-making, greatly improving the scientificity and accuracy of irrigation management.

[0070] S103. Generate farmland status prediction indicators for the target farmland based on the comprehensive soil information, meteorological information and crop information and using a preset farmland status prediction model, and generate a farmland irrigation strategy based on the farmland status prediction indicators.

[0071] Among them, the farmland state prediction model adopts an advanced CNN-LSTM hybrid architecture, which fully utilizes the advantages of convolutional neural networks (CNN) in spatial feature extraction and long short-term memory networks (LSTM) in time series data processing. The input data of the model is first preprocessed. Soil moisture information and soil spatial distribution information are converted into a unified spatial grid data format through spatial interpolation and time series aggregation methods. Crop remote sensing information and crop type information are also resampled to the same spatial grid. Meteorological data are organized into time series form.

[0072] The CNN part is used to process spatial data. It contains multiple convolutional layers and pooling layers to extract the characteristics of soil moisture distribution and crop spatial distribution. A typical CNN structure may include 3 convolutional layers, each using a 3x3 convolution kernel, followed by ReLU activation function and 2x2 maximum pooling layer. The output of this part of the network is a series of feature maps, representing the spatial characteristics of the farmland status. The LSTM part processes time series data, mainly meteorological information and historical farmland status data. The LSTM network usually contains 2-3 layers of LSTM units, each containing 128-256 neurons. This part of the network can capture long-term and short-term time dependencies and predict future meteorological changes and their impact on farmland status. The outputs of CNN and LSTM are fused through a fully connected layer to finally output farmland status prediction indicators. These indicators include key parameters such as predicted soil moisture content, crop transpiration, and root water absorption. The model is trained using historical data, including all monitoring data from the past few growing seasons and the corresponding actual farmland status. The training process uses the Adam optimizer, the learning rate is set to 0.001, the batch size is 64, and the number of training rounds is 200. To prevent overfitting, dropout (ratio 0.5) and L2 regularization techniques are used.

[0073] Based on the generated farmland status prediction indicators, the system further formulates farmland irrigation strategies. This process involves multi-objective optimization, taking into account factors such as water resource utilization efficiency, crop yield and economic benefits. Specifically, irrigation strategies include irrigation time, irrigation amount and irrigation method. For example, if the prediction indicators show that soil moisture will be lower than the critical value of crop water requirement in the next 3 days and there is not enough rainfall, the system will recommend irrigation within 24 hours. The calculation of irrigation amount is based on the predicted soil moisture deficit and crop water requirement, while taking into account the efficiency of the irrigation system. The choice of irrigation method (such as drip irrigation, sprinkler irrigation or flooding) is based on crop type, growth stage and soil characteristics.

[0074] This method of farmland status prediction and irrigation strategy generation based on the CNN-LSTM model can make full use of multi-source data to achieve accurate prediction of farmland status and precise assessment of irrigation needs, thereby greatly improving water resource utilization efficiency and reducing unnecessary irrigation. At the same time, it ensures that crops receive the best water supply, ultimately achieving the goal of increasing yield and quality.

[0075] S104. Based on the farmland status prediction index, the irrigation urgency corresponding to the farmland irrigation strategy is calculated through the fuzzy logic system.

[0076] Among them, first, the farmland state prediction indicators are used as the input variables of the fuzzy logic system. These indicators usually include the predicted soil moisture content, crop water requirement, expected rainfall, etc. Corresponding fuzzy sets and membership functions are defined for each input variable. For example, for the soil moisture content, three fuzzy sets of "drought", "moderate", and "wet" can be defined. The membership functions adopt common triangular or trapezoidal functions.

[0077] Taking the soil moisture content as an example, its membership function may be defined as follows:

[0078] "Drought": μ(x) = 1 when x ≤ 15%; μ(x) = (25% - x) / 10% when 15% < x < 25%; μ(x) = 0 when x ≥ 25%

[0079] "Moderate": μ(x) = (x - 15%) / 10% when 15% < x ≤ 25%; μ(x) = (35% - x) / 10% when 25% < x < 35%; μ(x) = 0 in other cases

[0080] "Wet": μ(x) = 0 when x ≤ 25%; μ(x) = (x - 25%) / 10% when 25% < x < 35%; μ(x) = 1 when x ≥ 35%. Among them, x represents the percentage of the soil moisture content, and μ(x) represents the membership degree.

[0081] Next, a series of fuzzy rules are formulated based on expert knowledge and historical experience. These rules describe the irrigation urgency under different combinations of input variables. For example: If the soil moisture content is "drought" and the crop water requirement is "high", then the irrigation urgency is "very high"; If the soil moisture content is "moderate" and the expected rainfall is "heavy", then the irrigation urgency is "low". Then, the input variables are mapped to the output variable (irrigation urgency) through these rules using a fuzzy inference method (such as the Mamdani method). This process includes four steps: fuzzification, rule evaluation, rule aggregation, and defuzzification. Finally, the centroid method is used for defuzzification to obtain a definite numerical value of the irrigation urgency.

[0082] S105. If the irrigation urgency reaches the highest level of urgency, then control the filtered irrigation module to perform an alternating irrigation operation on the target farmland according to the farmland irrigation strategy.

[0083] Among them, the current irrigation urgency level is judged according to the pre-set threshold value to determine whether it has reached the highest level. The highest level of urgency usually means that the crop is facing a serious risk of water shortage and irrigation measures need to be taken immediately. The setting of this threshold is based on the crop type, growth stage and local climate conditions, and is usually determined based on historical data and experience. Once it is confirmed that the irrigation urgency level has reached the highest level, the filtration irrigation module will be activated immediately. The filtration irrigation module is a comprehensive system that integrates water sources, filtration systems, water pipe networks and irrigation equipment. It first checks the water source conditions to ensure that there is enough water supply. Then, the water in the water source is processed by the filtration system to remove impurities and suspended matter that may clog the irrigation equipment. The filtration system usually includes multi-stage filtration devices such as sand and gravel filters, mesh filters and laminated filters, which can effectively remove impurities of different sizes and types.

[0084] In this embodiment, the farmland irrigation strategy is actually an irrigation strategy for the alternating irrigation process. The alternating irrigation operation is an innovative irrigation method. When the irrigation strategy is executed, the target farmland is first divided into multiple irrigation areas according to the soil spatial distribution information and the crop distribution. Then these areas are irrigated in sequence according to a predetermined order. For example, in a farmland divided into four areas, the first area will be irrigated for 15 minutes, then the second area will be irrigated for 15 minutes, and so on. After completing a round of irrigation for all areas, a new round of alternating irrigation is restarted until the total irrigation time or water volume specified in the farmland irrigation strategy is reached.

[0085] This method of alternating irrigation has several significant advantages: First, alternating irrigation allows water to be more evenly distributed throughout the soil. In traditional blanket irrigation, water may accumulate in some areas while other areas do not receive adequate water. Alternating irrigation allows water to better penetrate and distribute in the soil by giving each area intermittent "rest" time. Second, alternating irrigation can reduce surface runoff and deep infiltration, thereby improving water resource utilization efficiency. Because each area is irrigated for a shorter period of time, water is more likely to be absorbed by the soil rather than being lost or seeping into the groundwater. This method also improves soil aeration. During the intervals between irrigations, the soil has a chance to drain excess water, allowing air to enter the soil pores, providing a better growth environment for crop roots.

[0086] S106. If the irrigation urgency level does not reach the highest level of urgency, the filtering component data of the filtering irrigation module is collected through a second sensor preset inside the filtering irrigation module, and the historical backwashing data of the backwashing module is obtained.

[0087] If the irrigation emergency level does not reach the highest emergency level, the second sensor network preset inside the filter irrigation module is activated. These sensors are specifically used to monitor various parameters of the filter irrigation module, including but not limited to:

[0088] Inlet and outlet water pressure difference: measured by pressure sensors installed at the filter inlet and outlet. This parameter reflects the degree of filter blockage.

[0089] Filtered water flow rate: measured using an electromagnetic flow meter or ultrasonic flow meter. This can reflect the actual water supply capacity of the irrigation system.

[0090] Turbidity of filtered water: measured with an online turbidity meter. This directly reflects the filtering effect.

[0091] Cumulative filtered water volume: calculated by integrating the flow rate. This data is used to evaluate the service life of the filter media.

[0092] These sensors are usually high-precision, waterproof and dustproof industrial-grade devices that can work stably for a long time in harsh environments. The data collection frequency is usually set to once a minute to capture subtle changes in system status.

[0093] At the same time, the system will retrieve the historical backwash data of the backwash module from the database, including the time of the last backwash, the duration of the backwash, the amount of water used for backwashing, the change in pressure difference before and after backwashing, and the backwashing effect (such as the degree of improvement in filtration efficiency after backwashing). These historical data are usually kept for the last 3-6 months for long-term trend analysis.

[0094] The data collection process adopts a distributed architecture, and each sensor node is equipped with a local processing unit, which can perform preliminary data cleaning and anomaly detection. For example, if the reading of a sensor suddenly jumps abnormally, the local processing unit will immediately mark this data point as a suspicious value and increase the sampling frequency to confirm whether it is a device failure or a real system anomaly. All collected data are transmitted to the central control system in real time via industrial Ethernet or wireless transmission (such as LoRa, NB-IoT and other low-power wide area network technologies). Data transmission uses an encryption protocol to ensure the security and integrity of the data. After receiving the data, the central control system will immediately perform data fusion and preliminary analysis. For example, the current inlet and outlet water pressure difference is compared with historical data to evaluate the clogging trend of the filter; the turbidity of the filtered water is compared with the irrigation water quality standard to determine whether the current filtration effect meets the requirements. These real-time collected filter component data and historical backwashing data provide important basis for the next decision. They are not only used to evaluate the current operating status of the irrigation system, but also to predict possible problems in the future. For example, if the pressure differential between inlet and outlet water is found to be increasing but has not yet reached the backwash threshold, the system may schedule a preventive backwash in advance to avoid the problem of reduced filtration efficiency during a critical irrigation period.

[0095] S107. Combine the filter component data, historical backwash data and farmland irrigation strategy, and calculate the backwash urgency of the filter irrigation module through the fuzzy logic system.

[0096] The collected real-time filtered data is preprocessed and fused. Specifically, the following steps are included:

[0097] Data cleaning: Remove outliers and noise. For example, use median filtering to remove short-term spike interference.

[0098] Data standardization: Convert data of different dimensions (such as pressure, flow, turbidity) to a unified standard scale, usually using the minimum-maximum standardization method.

[0099] Data fusion: The Kalman filter algorithm is used to fuse the parameters such as the inlet and outlet pressure difference, filtered water flow rate and filtered water turbidity into a comprehensive real-time filtration status indicator. Kalman filtering can effectively handle measurement noise and system errors and provide optimal estimates.

[0100] Next, the long short-term memory network (LSTM) model is used to predict the future filtration status. The input of the LSTM model includes the real-time filtration status indicator time series of the past 24 hours, relevant features in the historical backwashing data (such as the running time after the last backwashing, the accumulated filtered water volume, etc.), and the expected irrigation demand in the next 24 hours (extracted from the farmland irrigation strategy). The output of the LSTM model is the predicted filtration status indicator in the next 24 hours. This prediction takes into account the decay trend of filter performance over time and the potential impact of future irrigation demand on the filter.

[0101] Then the predicted filtration status index and historical backwashing data are used as input variables of the fuzzy logic system. Fuzzy sets and membership functions are defined for each input variable. For example, for the predicted inlet and outlet pressure difference, four fuzzy sets of "normal", "mild blockage", "moderate blockage" and "severe blockage" can be defined. The membership function can use Gaussian function or trapezoidal function. Based on these fuzzy sets and membership functions, backwashing fuzzy inference rules are established. These rules are derived from expert knowledge and historical data analysis. For example: if the pressure difference is "severe blockage" and the time since the last backwash is "long", the backwashing urgency is "very high"; if the pressure difference is "mild blockage" and the predicted irrigation demand in the next 24 hours is "high", the backwashing urgency is "medium".

[0102] Next, the analytic hierarchy process (AHP) is used to determine the weight of each input variable in the decision. AHP constructs a judgment matrix and calculates the eigenvector to obtain the weight of each factor. This step takes into account the relative importance of different parameters to the backwashing decision. Then, the fuzzy reasoning rules and parameter weights are combined to obtain the backwashing reasoning evaluation results through fuzzy synthesis operations. Commonly used fuzzy synthesis algorithms include the maximum-minimum synthesis method and the weighted average method. The centroid method is used to defuzzify the fuzzy evaluation results to obtain a clear initial backwashing urgency value. After obtaining the initial backwashing urgency, the impact of the irrigation plan in the farmland irrigation strategy on the backwashing decision is also considered and an irrigation influence factor is constructed. The final backwashing urgency is obtained by multiplying the initial urgency by the irrigation influence factor.

[0103] S108. If the backwashing urgency is greater than or equal to the irrigation urgency, a backwashing strategy is generated based on the filtering component data for the backwashing module to perform backwashing operations on the filtering irrigation module, and the backwashing module and the filtering irrigation module are controlled to execute the backwashing strategy and the farmland irrigation strategy in turn.

[0104] First, a detailed backwashing strategy is generated based on the filter component data. This strategy includes the following key parameters:

[0105] Backwash time: Determined by the degree of blockage, usually between 30 seconds and 3 minutes.

[0106] Backwash flow rate: usually set to 2-3 times the normal filtration flow rate to ensure effective removal of accumulated impurities.

[0107] Backwash pressure: usually 20-30% higher than normal filtration pressure to provide sufficient impact.

[0108] Backwash sequence: For a multi-tank filtration system in parallel, determine the backwash sequence for each filter.

[0109] The generation of a backwash strategy takes into account several factors:

[0110] Current blockage degree: determined by the difference in inlet and outlet water pressure.

[0111] Water quality: Adjust according to the turbidity of raw water and the quality of water after filtration.

[0112] Historical backwashing effects: Analyze the relationship between the duration and effect of past backwashing.

[0113] Energy efficiency: Select the best combination of backwash parameters to strike a balance between cleaning effect and energy consumption.

[0114] After the strategy is generated, the system will control the backwash module to perform the backwash operation. The specific steps are as follows:

[0115] Close the normal filtration water path and open the backwash water path. This is usually achieved through electric valves or hydraulic valves to ensure fast and reliable switching.

[0116] Start the backwash pump to reverse the clean water through the filter at high pressure. The backwash pump usually adopts variable frequency control, which can accurately adjust the flow and pressure.

[0117] Monitor various parameters during the backwash process, including pressure, flow rate, and turbidity. Compare the differences between these parameters and expected values ​​in real time. If an abnormality is found (such as a sudden drop in pressure), the backwash strategy will be adjusted immediately or an alarm will be issued.

[0118] The backwash process is automatically terminated according to the preset backwash time or water quality index (such as the turbidity of the outlet water is lower than a certain threshold).

[0119] After the backwash is complete, a short forward flush is performed to stabilize the filter media and remove any remaining impurities.

[0120] After backwashing is completed, it will immediately switch to executing the previously formulated farmland irrigation strategy.

[0121] S109. If the backwashing urgency is less than the irrigation urgency, the filtering irrigation module is controlled to perform alternating irrigation operations on the target farmland according to the farmland irrigation strategy.

[0122] In one embodiment, generating a farmland state prediction index of a target farmland based on soil comprehensive information, meteorological information and crop information and using a preset farmland state prediction model comprises the following steps:

[0123] Preprocessing soil moisture information, soil spatial distribution information and meteorological information;

[0124] The preprocessed soil moisture information and soil spatial distribution information are fused into soil moisture distribution data in the spatial dimension;

[0125] The crop remote sensing information and crop type information are integrated into the soil spatial distribution information in the spatial dimension to obtain the crop spatial distribution data;

[0126] The soil moisture distribution data and the crop spatial distribution data are input into a preset farmland state prediction model, and the soil state prediction index and the crop state prediction index are output through the convolutional neural network module in the farmland state prediction model;

[0127] The preprocessed meteorological information is input into the farmland state prediction model, and the meteorological prediction index is output through the long short-term memory network module in the farmland state prediction model;

[0128] The soil state prediction indicators, crop state prediction indicators and meteorological prediction indicators are integrated into the farmland state prediction indicators.

[0129] In this embodiment, the preprocessing step is to clean, standardize and enhance the raw data to prepare for subsequent data fusion and model input. For soil moisture information, outlier detection and processing are first performed, and the median absolute deviation method (MAD) is used to identify and remove outliers. For example, if a sensor reading deviates from the median by more than 3 times the MAD, it is considered an outlier. Next, the data is interpolated in time series to fill in possible data gaps. The commonly used methods are linear interpolation or spline interpolation. Then, a sliding average filter is applied to smooth the data to reduce the impact of short-term fluctuations. For soil spatial distribution information, spatial registration is first performed to ensure that all data points are in the same coordinate system. Then, the Kriging interpolation method is used to spatially interpolate discrete sampling points to generate a continuous soil property distribution map. For meteorological information, unit unification and scale transformation are first performed, such as converting temperature to degrees Celsius and converting wind speed to meters per second. Then, the time series data is seasonally adjusted and trend decomposed, and the seasonal, trend and residual components are separated using the STL (Seasonal and Trend decomposition using Loess) method. Finally, all preprocessed data are standardized, usually using the Z-score standardization method, to make different types of data comparable. Through these preprocessing steps, the original heterogeneous data are transformed into clear, consistent and information-rich data sets, laying a solid foundation for subsequent analysis and modeling.

[0130] Next, the two data are resampled into a predefined unified spatial grid, and the grid size is usually set to 5 meters × 5 meters or 10 meters × 10 meters to balance spatial resolution and computational efficiency. For each grid cell, the time series aggregation value of all soil moisture measurement points inside it is calculated, such as the mean, median or weighted mean. The weight can be based on the distance from the measurement point to the center of the grid, using the inverse distance weighting method. For grid cells without direct measurement data (null value grids), geostatistical methods are used for spatial interpolation. Specifically, Ordinary Kriging is used for interpolation, which takes into account spatial autocorrelation and can provide the best linear unbiased estimate. The semivariogram model of Kriging interpolation selects a spherical model, an exponential model or a Gaussian model, which is determined according to the spatial structure characteristics of the actual data. Finally, the continuous soil moisture field obtained by interpolation is fused with the original soil spatial distribution information (such as soil type, texture, organic matter content, etc.) to generate multi-layer raster data. Each layer represents a different soil attribute or moisture condition at a time point. This fusion method not only provides a high-resolution spatial distribution map of soil moisture, but also retains the physical property information of the soil itself, providing rich and accurate input data for subsequent farmland status prediction.

[0131] For remote sensing information, atmospheric correction and geometric correction are required to eliminate the influence of atmosphere and terrain on remote sensing images. Then, the corrected remote sensing images are resampled to the same spatial grid as the soil spatial distribution information, usually using bilinear interpolation or cubic convolution interpolation methods. For each spatial grid, vegetation indices such as normalized vegetation index (NDVI) and enhanced vegetation index (EVI) are calculated, which can effectively reflect the growth status and biomass of crops. For example, the calculation formula of NDVI is: NDVI = (NIR-RED) / (NIR+RED), where NIR and RED represent the reflectance of the near-infrared band and the red light band, respectively. Next, combined with crop type information, a specific crop type label is assigned to each spatial grid. If a grid contains multiple crops, the dominant crop type is determined by area weighting. Based on crop type and growth stage, the crop root distribution of each grid is estimated. Root distribution can be described using an exponential decay model, such as R(z) = 1-βz, where R(z) is the root density at depth z and β is the attenuation coefficient, which is related to crop type and growth stage. Finally, the crop growth status (such as NDVI value), crop type and root distribution information are fused with soil spatial distribution information to generate multidimensional raster data. Each raster unit contains multiple attribute layers such as soil properties, crop type, growth status and root distribution. This fused crop spatial distribution data not only reflects the above-ground status of the crop, but also contains information about the underground part, providing comprehensive data support for precision irrigation decisions.

[0132] The soil moisture distribution data and crop spatial distribution data are organized into a four-dimensional tensor form (batch size, number of channels, height, width), and each channel represents an attribute layer (such as soil moisture, soil type, NDVI, crop type, etc.). The architecture of the CNN module usually contains multiple convolutional layers, pooling layers, and fully connected layers. For example, a structure similar to the VGG network can be used, which contains 5 convolutional blocks, each block contains 2-3 3x3 convolutional layers and a 2x2 maximum pooling layer. The activation function uses ReLU to introduce nonlinearity and alleviate the gradient vanishing problem. After the last convolutional block, a global average pooling layer is used to reduce the number of parameters, and then the fully connected layer is connected. The output of the fully connected layer is divided into two parts: soil state prediction indicators and crop state prediction indicators. Soil state prediction indicators may include soil moisture content and soil temperature in the next few days, and crop state prediction indicators may include leaf area index, biomass, etc. The model training adopts the mini-batch stochastic gradient descent method, and the loss function is the weighted sum of the mean square error (MSE) and the mean absolute error (MAE).

[0133] The preprocessed meteorological data is organized into a three-dimensional tensor form (batch size, time step, number of features). Each time step contains multiple meteorological features, such as temperature, humidity, rainfall, wind speed, etc. The architecture of the LSTM module usually contains multiple layers of LSTM units and an output layer. For example, two layers of LSTM can be used, each containing 128 hidden units, followed by a fully connected layer. The core of the LSTM unit is its gating mechanism, including the forget gate, input gate, and output gate. The forget gate determines which information is discarded, the input gate determines which information is updated, and the output gate determines which information is output. The model training adopts the back propagation through time (BPTT) algorithm, and the Adam optimizer is used for parameter update. The loss function selects the mean square error (MSE), and the L2 regularization term is introduced to prevent overfitting. The output of the model is the meteorological forecast indicators for the next few days, including key parameters such as average daily temperature, sunshine hours, and rainfall.

[0134] The integration of soil state prediction index, crop state prediction index and meteorological prediction index into farmland state prediction index is a comprehensive data fusion process. First, the three types of prediction indexes are standardized to ensure the comparability of indexes of different scales and units. Then, the principal component analysis (PCA) method is used to reduce the dimension of each type of index and extract the main features. The principal components with a cumulative explained variance of more than 85% are selected as the new feature set. Next, the weighted summation method is used to combine the principal components of different categories into a unified farmland state prediction index. The weights are determined using the analytic hierarchy process (AHP), taking into account expert opinions and historical data analysis results. For example, the possible weight distribution is: 40% for soil state, 35% for crop state, and 25% for meteorological state. The final farmland state prediction index calculation formula is: I = ∑ (wi*PCi), where wi is the weight of the i-th principal component and PCi is the score of the i-th principal component. For ease of interpretation and application, the comprehensive index is normalized to the range of 0-100, where 0 represents the worst state and 100 represents the best state. In addition, warning thresholds are set according to the indicator values, such as 0-20 is defined as a serious warning, 20-40 as a moderate warning, 40-60 as a slight warning, 60-80 as good, and 80-100 as excellent.

[0135] In one embodiment, fusing the preprocessed soil moisture information and soil spatial distribution information into soil moisture distribution data in the spatial dimension comprises the following steps:

[0136] Perform spatial registration processing on soil moisture information and soil spatial distribution information so that the soil moisture information and soil spatial distribution information are in the same spatial reference system;

[0137] Resample the soil moisture information and soil spatial distribution information after spatial registration processing to a spatial grid of a preset size;

[0138] For each spatial grid, all soil moisture information in the spatial grid is aggregated in time series to obtain soil moisture aggregate data;

[0139] Traverse all spatial grids, for null-valued spatial grids where there is no aggregated soil moisture data, estimate the aggregated soil moisture data of the null-valued spatial grid using a spatial interpolation method based on all the aggregated soil moisture data of the adjacent spatial grids of the null-valued spatial grid;

[0140] The aggregated soil moisture data of all spatial grids are fused with the soil spatial distribution information into soil moisture distribution data in the form of multi-layer grids.

[0141] In this embodiment, a unified target coordinate system is determined, and the Universal Transverse Mercator Projection (UTM) or a commonly used projection system in the local area is usually selected. For soil moisture information, these data usually come from sensors distributed throughout the farmland, and each sensor has its specific geographic coordinates. Convert these coordinates from their original system (such as longitude and latitude) to the target system. The conversion process uses a coordinate transformation formula. For soil spatial distribution information, these data may come from soil surveys or remote sensing images and may already be in a specific projection system. If this system is different from the target system, reprojection is required. The reprojection process involves coordinate conversion and resampling, using methods such as bilinear interpolation or cubic convolution to ensure the spatial continuity of the data. When performing coordinate conversion, it is also necessary to consider the effects of different ellipsoids and datums, and use appropriate conversion parameters (such as seven parameters or Helmert conversion) for adjustment. After the registration is completed, the two data sets are completely aligned in space, and each data point has the same coordinate system and projection.

[0142] The soil moisture information and soil spatial distribution information after spatial registration are resampled to a spatial grid of a preset size. Specifically, the appropriate grid size is first determined, which is usually determined based on the area of ​​the study area, the original resolution of the data, and the required accuracy. For example, for a 100-hectare farmland, a grid size of 10 meters × 10 meters may be selected, so that the entire area will be divided into 10,000 grid cells. The resampling process uses different methods for soil moisture information and soil spatial distribution information. For soil moisture information, since these data are usually discrete point data, they need to be converted into continuous raster data using interpolation methods. For soil spatial distribution information, if the original data is in vector format (such as polygons), it needs to be converted into raster format first. During the conversion process, each grid cell is assigned the attribute value of the soil type with the largest coverage area. If the original data is already in raster format but with different resolutions, a resampling algorithm such as the nearest neighbor method, bilinear interpolation, or cubic convolution is used to adjust the resolution. The nearest neighbor method is suitable for categorical data, while bilinear interpolation and cubic convolution are suitable for continuous data. After resampling, both datasets were converted to grids of the same size and resolution, with each grid cell containing soil moisture values ​​and soil property information.

[0143] For each grid, all measurements of all soil moisture sensors within the grid range in the past period (such as a month or a growing season) are collected to form a time series data set. Then, statistical analysis and feature extraction are performed on this time series. Common aggregation methods include calculating statistics such as mean, median, maximum, minimum, standard deviation, etc. In addition, time series features such as autocorrelation coefficient and partial autocorrelation coefficient can also be calculated to capture the time dependence of soil moisture. For grids with multiple sensors, the data of each sensor is first processed as above, and then the results are spatially averaged or weighted averaged, and the weight can be based on the distance from the sensor to the center of the grid. Finally, each grid obtains a set of aggregation indicators, such as average humidity, humidity coefficient of variation, humidity change trend, etc.

[0144] Next, all spatial grids are traversed to identify all the null grids, which may be missing data due to sensor failure, installation location restrictions, or data transmission problems. Then, for each null grid, its neighboring grids are determined. The neighboring grids usually include 8 directly adjacent grids (Moore neighborhood), but in some cases may be extended to a larger range, such as 16 or 24 nearest grids. Then a suitable spatial interpolation method is selected. Commonly used methods include inverse distance weighted (IDW), ordinary kriging, and radial basis function (RBF). For each null grid, the selected interpolation method is applied to estimate its value using the aggregated soil moisture data of the neighboring grids. During the interpolation process, considering auxiliary variables such as topography and soil type can improve the estimation accuracy. For example, co-kriging can be used to incorporate these auxiliary variables into the interpolation model. After the interpolation is completed, the estimation results are checked for rationality to ensure that the estimated values ​​are within a reasonable range and no obvious outliers appear. The interpolation accuracy can be evaluated using cross-validation methods, such as leave-one-out cross-validation (LOOCV). Through this approach, the data of all the empty grids were successfully filled, creating a complete and spatially continuous soil moisture distribution map.

[0145] During the last and most critical step of data integration, the spatial resolution and range of the final raster data need to be determined, which is usually consistent with the spatial grid used in the previous steps. Then, a multidimensional data structure is created for each grid cell, containing multiple data layers. The first layer is soil moisture aggregate data, including indicators such as average moisture, moisture coefficient of variation, and moisture trend. Subsequent layers include soil spatial distribution information, such as soil type, soil texture, organic matter content, pH value, etc. Each attribute is a separate data layer. During the data fusion process, these layers are superimposed together using geographic information system (GIS) software or custom algorithms. For example, raster algebra operations are used to combine data from different layers into a multidimensional array, and each grid cell contains a complete set of attribute values. During the fusion process, attention should be paid to the consistency of data types. Continuous variables (such as moisture values) and categorical variables (such as soil type) need to be processed separately. For continuous variables, standardization may be required. For categorical variables, encoding methods such as one-hot encoding are usually used to convert them into numerical form. The fused multi-layer raster data contains not only spatial information, but also information in the time dimension (through the time series characteristics of soil moisture). This data structure allows for complex spatial analysis and spatiotemporal pattern recognition. For example, it is easy to extract moisture trends under specific soil types or analyze the impact of different soil textures on moisture distribution.

[0146] In one embodiment, the crop remote sensing information and the crop type information are integrated into the soil spatial distribution information in the spatial dimension to obtain the crop spatial distribution data, which includes the following steps:

[0147] Resample crop remote sensing information to a spatial grid;

[0148] The vegetation index is used to calculate the crop growth status in each spatial grid;

[0149] Estimate the crop root distribution data in each spatial grid based on crop growth status and crop type;

[0150] The crop growth status, crop type and crop root distribution data are fused with the soil spatial distribution information into crop spatial distribution data in the form of multi-layer grids.

[0151] In this embodiment, resampling crop remote sensing information to a spatial grid is a key data preprocessing step, which first involves the acquisition and preliminary processing of remote sensing images. Remote sensing images usually come from satellites or UAV platforms, including multiple spectral bands, such as red light, near infrared, and short-wave infrared. These raw images first need to be geometrically corrected and atmospherically corrected. Geometric correction ensures that each pixel of the image corresponds to the correct geographic location, which is usually achieved using ground control points and accurate digital elevation models (DEM). Atmospheric correction eliminates the effects of atmospheric scattering and absorption on remote sensing signals, using atmospheric radiation transfer models such as MODTRAN or 6S. The corrected images usually have a very high spatial resolution, which may reach sub-meter level.

[0152] Next, a resampling algorithm is used to convert the corrected high-resolution remote sensing image to a spatial grid. Common resampling methods include the nearest neighbor method, bilinear interpolation, and cubic convolution. The nearest neighbor method simply assigns the source pixel value closest to the center of the target pixel to the target pixel, which is suitable for classified data. The bilinear interpolation method uses distance-weighted average to calculate the new pixel value, which is suitable for continuous data and can maintain a certain spatial smoothness. The cubic convolution method uses a weighted average of the 16 nearest neighbor pixels, which can provide smooth results while maintaining edge sharpness. After resampling, each spatial grid contains reflectance values ​​for multiple spectral bands. These data are organized into a multidimensional array, where each element corresponds to a grid and contains all the spectral information of the grid.

[0153] Using vegetation indices to calculate crop growth status within each spatial grid is a key step in converting remote sensing data into meaningful crop physiological information. Vegetation indices are mathematical combinations based on reflectances of different bands and can effectively reflect the biophysical characteristics of vegetation, such as leaf area index, biomass, and chlorophyll content. This process first involves selecting an appropriate vegetation index. The most commonly used is the Normalized Difference Vegetation Index (NDVI), which is calculated as follows: NDVI = (NIR-RED) / (NIR+RED), where NIR is the reflectance of the near-infrared band and RED is the reflectance of the red band. NDVI values ​​range from -1 to 1, where higher positive values ​​indicate healthier and more lush vegetation. For each spatial grid, the selected vegetation index is calculated using the resampled multispectral data. This process can be efficiently completed through raster algebra operations. For example, when using NDVI, the following operations are performed for each grid: extract the reflectance values ​​of the near-infrared and red bands of the grid; apply the NDVI formula to calculate the NDVI value of the grid; and store the calculation results as new attributes of the grid.

[0154] Next, the root distribution is estimated using a root growth model based on the crop type and growth state. A commonly used model is the exponential decay model, which has the basic form of: R(z) = 1-βz, where R(z) is the relative root density at depth z and β is the decay coefficient. The β value varies depending on the crop type and growth stage. For example, shallow-rooted crops (such as wheat) have larger β values, while deep-rooted crops (such as corn) have smaller β values. In order to make the model more accurate, the effect of crop growth state is introduced. In this way, crops with better growth status (higher NDVI values) will have deeper root distribution. For each spatial grid, the root density distribution at different depths is calculated according to the above model. Usually, the root density value is calculated every 10 cm in the depth range of 0-100 cm. In this way, a 10-dimensional root density vector is obtained for each grid. Finally, the estimated root distribution data is integrated with the original spatial grid data structure. In addition to the aboveground information (such as NDVI, crop type), each grid now also contains a vector describing the vertical distribution of the root system.

[0155] Finally, a multidimensional data structure is created, where each spatial grid corresponds to a data unit and contains multiple attribute layers.

[0156] These layers include:

[0157] 1. Crop growth status layer: contains multiple vegetation indices, such as NDVI, EVI, etc. Each index is an independent sublayer.

[0158] 2. Crop type layer: Use encoding to represent different crop types. For example, you can use integer encoding (1 = corn, 2 = soybeans, 3 = wheat, etc.) or one-hot encoding (create multiple binary layers, one for each crop).

[0159] 3. Crop root distribution layer: contains root density estimates at multiple depths, with each depth as a separate sublayer.

[0160] 4. Soil spatial distribution layer: contains multiple soil attributes, such as texture, organic matter content, pH value, etc., each attribute as an independent sublayer.

[0161] Data fusion is achieved using raster algebra and geographic information system (GIS) tools. During the fusion process, attention should be paid to the consistency of data types. Continuous variables (such as NDVI, root density) and categorical variables (such as crop type, soil type) need to be processed separately. For continuous variables, standardization is performed, and for categorical variables, encoding methods are used to convert them into numerical form.

[0162] For example, for crop types, the following methods can be used:

[0163] 1. Integer encoding: directly use integers to represent different types (1 = corn, 2 = soybeans, etc.)

[0164] 2. One-hot encoding: Create a binary layer for each crop type

[0165] The fused multi-layer raster data structure can be expressed as:

[0166] D(i,j)=[V1(i,j), V2(i,j),...,Vn(i,j)]

[0167] Where D(i, j) is the data vector at position (i, j), and Vk(i, j) is the value of the kth attribute layer.

[0168] In one embodiment, the irrigation urgency corresponding to the farmland irrigation strategy is calculated based on the farmland state prediction index and through a fuzzy logic system, including the following steps:

[0169] The farmland state prediction index is used as the first fuzzy input variable of the fuzzy logic system;

[0170] defining a fuzzy set and a first membership function for each first fuzzy input variable;

[0171] Combining the fuzzy set and the first membership function, and using the fuzzy reasoning method to map all the first fuzzy input variables to the fuzzy output variables according to the preset irrigation fuzzy reasoning rules;

[0172] The fuzzy output variables are defuzzified using the centroid method to obtain the irrigation urgency corresponding to the farmland irrigation strategy.

[0173] In this embodiment, the farmland state prediction index includes multiple dimensions such as soil moisture prediction value, crop growth state prediction value and meteorological prediction value. Each indicator is regarded as an independent fuzzy input variable. For example, the soil moisture prediction value may range from 0% to 100%, the crop growth state is represented by the normalized difference vegetation index (NDVI) ranging from -1 to 1, and the meteorological prediction value includes the probability of rainfall and evapotranspiration in the next few days. These precise numerical values ​​need to be converted into language variables such as "low", "medium", "high", etc., so that the fuzzy logic system can process them. This conversion enables the system to simulate the decision-making process of human experts and deal with the uncertainty and ambiguity inherent in agricultural production.

[0174] Next, we need to define the fuzzy sets and the first membership function. Taking the soil moisture prediction value as an example, three fuzzy sets can be defined: "dry", "moderate" and "wet". The membership function determines the degree to which a specific soil moisture value belongs to these fuzzy sets. Commonly used membership functions include triangular functions, trapezoidal functions and Gaussian functions. For example, the membership function of "dry" defined using the trapezoidal function can be expressed as:

[0175]

[0176] Where x is the soil moisture value. Similarly, corresponding fuzzy sets and membership functions are defined for crop growth status and meteorological forecast values. This definition allows the input variables to smoothly transition between different fuzzy sets, better reflecting the fuzziness in reality. Through carefully designed fuzzy sets and membership functions, the system can more accurately capture subtle changes in farmland status and provide richer input information for subsequent fuzzy reasoning. Combining fuzzy sets and first membership functions, and using fuzzy reasoning methods to map all first fuzzy input variables to fuzzy output variables according to preset irrigation fuzzy reasoning rules is the core reasoning process of the fuzzy logic system. For example, a rule may be: "If soil moisture is dry, crop growth status is good, and there is no rainfall forecast in the future, then the irrigation urgency is high." These rules are expressed using IF-THEN structures, covering a variety of possible input combinations.

[0177] Next, these rules are evaluated using fuzzy reasoning methods such as Mamdani reasoning. For each rule:

[0178] 1. Calculate the truth degree of the antecedent (IF part) using the membership function of the fuzzy set.

[0179] 2. Use fuzzy implication operators (such as the minimum method) to determine the activation strength of the rule.

[0180] 3. Truncate the fuzzy set of the consequent (THEN part) according to the activation strength.

[0181] The outputs of all rules are aggregated, usually using the maximum method. This process produces a comprehensive set of fuzzy outputs that represent the irrigation urgency. In this way, the system can comprehensively consider multiple factors, simulate the decision-making process of human experts, and come up with reasonable irrigation recommendations. Finally, the centroid method is used to defuzzify the fuzzy output variables to obtain the irrigation urgency corresponding to the farmland irrigation strategy. The centroid method is a commonly used and effective defuzzification method. It calculates the centroid of the fuzzy output set and uses it as the final crisp output value. In this way, the fuzzy irrigation urgency is converted into a specific value, such as a scale of 0 to 100, where 0 means no irrigation is required at all and 100 means an extremely urgent need for irrigation.

[0182] In one embodiment, the filter component data includes the inlet and outlet water pressure difference, the filtered water flow rate, the turbidity of the filtered water, and the accumulated filtered water volume. Combining the filter component data, the historical backwashing data, and the farmland irrigation strategy, and calculating the backwashing urgency of the filter irrigation module through the fuzzy logic system includes the following steps:

[0183] The Kalman filter method is used to integrate the inlet and outlet water pressure difference, filtered water flow rate and filtered water turbidity into real-time filtering data;

[0184] Based on the real-time filtering data, the predicted filtering data is obtained through the preset filtering state prediction model, and the filtering state prediction model is an LSTM model;

[0185] The predicted filtering data and the historical backwashing data are used as the second fuzzy input variables of the fuzzy logic system;

[0186] defining a second membership function for each second fuzzy input variable;

[0187] Based on the second membership function and according to the preset backwashing fuzzy inference rule, a backwashing association relationship matrix is ​​established;

[0188] The second fuzzy input variable is analyzed by the analytic hierarchy process to obtain the matrix weight vector of the backwashing correlation matrix;

[0189] Combining the backwashing association matrix and the matrix weight vector and using fuzzy synthesis operation to obtain the backwashing reasoning evaluation result;

[0190] The center of gravity method is used to defuzzify the backwashing reasoning evaluation results to obtain the initial backwashing urgency.

[0191] The irrigation influencing factor is constructed based on the irrigation duration and irrigation water volume in the farmland irrigation strategy, and the irrigation influencing factor is used to correct the initial backwashing urgency to obtain the backwashing urgency of the filter irrigation module.

[0192] In this embodiment, the Kalman filter method is used to merge the inlet and outlet water pressure difference, the filtered water flow rate and the turbidity of the filtered water into real-time filtering data. Kalman filtering is a recursive estimation algorithm that can effectively process multi-source and noisy data. In this embodiment, the system state equation and observation equation are first established. The state equation describes the dynamic characteristics of the filtration system, including the changing laws of pressure difference, flow rate and turbidity. The observation equation describes the measurement process of these variables. Kalman filtering is iteratively performed through two steps: prediction and update. In the prediction step, the current state is estimated based on the state of the previous moment; in the update step, the prediction is corrected in combination with the actual observation value. The specific calculation involves the update of the covariance matrix and the calculation of the Kalman gain.

[0193] The predicted filtering data is obtained based on the real-time filtering data and predicted by a preset filtering state prediction model, wherein the filtering state prediction model is an LSTM model. LSTM (Long Short-Term Memory) is a special recurrent neural network that is particularly suitable for processing and predicting time series data. In this embodiment, the LSTM model receives the real-time filtering data processed by the Kalman filter as input. The model structure includes an input layer, multiple LSTM layers and an output layer. Each LSTM unit includes an input gate, a forget gate and an output gate, which can effectively capture the long-term dependencies of the data. The model training uses historical filtering data and optimizes the network parameters through a back-propagation algorithm. When predicting, the model outputs a filtering state prediction for a period of time in the future based on the current and historical data sequences.

[0194] The predicted filtration data and historical backwashing data are used as the second fuzzy input variables of the fuzzy logic system. The predicted filtration data includes parameters such as pressure difference, flow rate and turbidity predicted by the LSTM model in the future. The historical backwashing data includes information such as the frequency, duration and effect of past backwashing. These precise numerical data need to be converted into language variables that can be processed by the fuzzy logic system. For example, the predicted pressure difference may be classified as "low", "medium", and "high", while the historical backwashing frequency may be described as "frequent", "moderate", and "scarce". This conversion enables the system to simulate the decision-making process of experts and deal with the uncertainty and ambiguity in the operation of the filtration system.

[0195] Next, the second membership function is defined for each second fuzzy input variable. For predicting the pressure difference in the filtered data, a triangular membership function can be defined:

[0196] μlow(x)=max(0,min((mx) / (ma),1))

[0197] Where a is the minimum pressure difference and m is the median pressure difference. Similarly, corresponding membership functions are defined for parameters such as flow rate and turbidity. For historical backwash data, such as backwash frequency, a trapezoidal membership function can be used:

[0198] μfrequent(x)=max(0,min((xa) / (ba),1,(dx) / (dc)))

[0199] Among them, [a, b, c, d] define the four characteristic points of the fuzzy set "frequent". These carefully designed membership functions allow the input variables to smoothly transition between different fuzzy sets, better reflecting the fuzziness and uncertainty in actual operation.

[0200] Based on the second membership function and according to the preset backwashing fuzzy inference rules, a backwashing association relationship matrix is ​​established. Specifically, a series of backwashing fuzzy inference rules are first formulated, such as "If the predicted pressure difference is high and the historical backwashing frequency is low, the backwashing urgency is high." Then, these rules are converted into a fuzzy relationship matrix R. The rows of the matrix R correspond to the fuzzy sets of the input variables, and the columns correspond to the fuzzy sets of the output variables (backwashing urgency). The matrix element rij represents the degree of association between the i-th input fuzzy set and the j-th output fuzzy set, which is usually determined using fuzzy implication operations. In this way, all the inference rules are encoded into a unified matrix, providing a mathematical basis for subsequent fuzzy reasoning.

[0201] Next, the second fuzzy input variable is analyzed by the analytic hierarchy process to obtain the matrix weight vector of the backwash association matrix. The analytic hierarchy process (AHP) first establishes a decision hierarchy structure and decomposes the backwash decision problem into three levels: goals, criteria, and plans. Then, the judgment matrix A is constructed by comparing the importance of the indicators one by one. For example, if the predicted pressure difference is considered to be twice as important as the historical backwash frequency, the corresponding matrix element aij=2. Next, the maximum eigenvalue λmax and the corresponding eigenvector W of the judgment matrix are calculated. The eigenvector W is the weight vector after normalization. At the same time, the consistency of the judgment is tested by calculating the consistency ratio, and then the backwash association matrix and the matrix weight vector are combined and the backwash reasoning evaluation result is obtained through fuzzy synthesis operation. Specifically, the weight vector W is first fuzzy synthesized with the association matrix R, using the maximum-minimum synthesis method: in represents the fuzzy composition operator. The result B is a fuzzy set, which represents the membership distribution of the backwash urgency. For example, B may be [0.2, 0.5, 0.3], corresponding to the membership of the three backwash urgency levels of low, medium, and high. The centroid method is used to defuzzify the backwash reasoning evaluation results to obtain the initial backwash urgency. The centroid method calculates the centroid of the fuzzy set B and uses it as the final output value. The calculated initial backwash urgency is a specific value between 0 and 100.

[0202] The irrigation impact factor is constructed based on the irrigation duration and irrigation water volume in the farmland irrigation strategy, and the irrigation impact factor is used to correct the initial backwash urgency to obtain the backwash urgency of the filter irrigation module. First, the irrigation impact factor is constructed, and the expression of the irrigation impact factor is as follows:

[0203]

[0204] In the formula: δ represents the irrigation impact factor, α represents the irrigation water volume impact weight, q represents the irrigation water volume impact weight, n represents the irrigation water volume, Q represents the preset backwash interval water volume, Q M represents the cumulative filtered water volume of the filtering irrigation module, β represents the influence weight of irrigation time, and t n represents the irrigation duration, and T represents the preset maximum irrigation duration.

[0205] The initial backwash urgency is then multiplied by the irrigation impact factor to correct the initial backwash urgency and obtain the final backwash urgency. This process takes into account the immediate impact of the current irrigation operation on the filtration system, making the backwash decision more flexible and real-time.

[0206] The present invention also discloses an intelligent water-saving control system for an agricultural irrigation device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the intelligent water-saving control method for an agricultural irrigation device described in any one of the above-mentioned embodiments is implemented.

[0207] Among them, the processor can adopt a central processing unit (CPU). Of course, according to actual usage, other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be adopted. The general-purpose processor can adopt a microprocessor or any conventional processor, etc., and this application does not impose any restrictions on this.

[0208] Among them, the memory can be an internal storage unit of a computer device, such as a hard disk or memory of a computer device, or an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (SMC), a secure digital card (SD) or a flash memory card (FC) equipped on the computer device, etc., and the memory can also be a combination of an internal storage unit and an external storage device of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or is to be output, and this application does not impose any restrictions on this.

[0209] A person skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of protection of the present application is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as above, which are not provided in detail for the sake of simplicity.

[0210] One or more embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the present application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present application should be included in the protection scope of the present application.

Claims

1. An intelligent water-saving control method for an agricultural irrigation device, characterized in that: The invention is applied to an agricultural irrigation device, the agricultural irrigation device comprises a filtering irrigation module and a backwashing module, the irrigation unit of the filtering irrigation module is connected to a target farmland to be irrigated, the irrigation unit is located at one end of the target farmland and is provided with a multifunctional irrigation outlet head, the backwashing module is used to backwash the filtering irrigation module, and the method comprises the following steps: Collecting comprehensive soil information of the target farmland through a first sensor preset in the target farmland; Acquiring weather information of the target farmland and crop information in the target farmland; Based on the comprehensive soil information, the meteorological information and the crop information and using a preset farmland state prediction model, a farmland state prediction index of the target farmland is generated, and a farmland irrigation strategy is generated based on the farmland state prediction index, wherein the farmland state prediction model is a CNN-LSTM model; Based on the farmland state prediction index, the irrigation urgency corresponding to the farmland irrigation strategy is calculated through a fuzzy logic system; If the irrigation urgency reaches the highest level of urgency, controlling the filtering irrigation module to perform an alternate irrigation operation on the target farmland according to the farmland irrigation strategy; If the irrigation urgency level does not reach the highest level of urgency, the filtering component data of the filtering irrigation module is collected through a second sensor preset inside the filtering irrigation module, and the historical backwashing data of the backwashing module is obtained; Combining the filter component data, the historical backwash data and the farmland irrigation strategy, and calculating the backwash urgency of the filter irrigation module through a fuzzy logic system; If the backwashing urgency is greater than or equal to the irrigation urgency, then generating a backwashing strategy for the backwashing module to perform a backwashing operation for the filter irrigation module based on the filter component data, and controlling the backwashing module and the filter irrigation module to sequentially perform the backwashing strategy and the farmland irrigation strategy; If the backwash urgency is less than the irrigation urgency, the filtering irrigation module is controlled to perform an alternating irrigation operation on the target farmland according to the farmland irrigation strategy.

2. The intelligent water-saving control method for an agricultural irrigation device according to claim 1, characterized in that: The comprehensive soil information includes soil moisture information and soil spatial distribution information, and the crop information includes crop remote sensing information and crop type information.

3. The intelligent water-saving control method for agricultural irrigation device according to claim 2, characterized in that: The method of generating the farmland state prediction index of the target farmland based on the soil comprehensive information, the meteorological information and the crop information and using a preset farmland state prediction model comprises the following steps: preprocessing the soil moisture information, the soil spatial distribution information and the meteorological information; fusing the preprocessed soil moisture information and the soil spatial distribution information into soil moisture distribution data in a spatial dimension; fusing the crop remote sensing information and the crop type information into the soil spatial distribution information in a spatial dimension to obtain crop spatial distribution data; The soil moisture distribution data and the crop spatial distribution data are input into a preset farmland state prediction model, and a soil state prediction index and a crop state prediction index are output through a convolutional neural network module in the farmland state prediction model; Inputting the preprocessed meteorological information into the farmland state prediction model, and outputting meteorological prediction indicators through the long short-term memory network module in the farmland state prediction model; The soil state prediction index, the crop state prediction index and the meteorological prediction index are integrated into a farmland state prediction index.

4. The intelligent water-saving control method for agricultural irrigation device according to claim 3, characterized in that: The step of fusing the pre-processed soil moisture information and the soil spatial distribution information into soil moisture distribution data in the spatial dimension comprises the following steps: Performing spatial registration processing on the soil moisture information and the soil spatial distribution information so that the soil moisture information and the soil spatial distribution information are in the same spatial reference system; Resampling the soil moisture information and the soil spatial distribution information after spatial registration processing to a spatial grid of a preset size; For each of the spatial grids, all the soil moisture information in the spatial grid is aggregated in time series to obtain soil moisture aggregated data; Traversing all the spatial grids, for a null-valued spatial grid where the soil moisture aggregated data does not exist, estimating the soil moisture aggregated data of the null-valued spatial grid based on all the soil moisture aggregated data of the adjacent spatial grids of the null-valued spatial grid and using a spatial interpolation method; The soil moisture aggregation data of all the spatial grids are fused with the soil spatial distribution information into soil moisture distribution data in a multi-layer grid form.

5. The intelligent water-saving control method for agricultural irrigation device according to claim 4, characterized in that: The step of fusing the crop remote sensing information and the crop type information into the soil spatial distribution information in a spatial dimension to obtain crop spatial distribution data comprises the following steps: resampling the crop remote sensing information to the spatial grid; Calculating the crop growth status in each of the spatial grids using vegetation index; estimating crop root distribution data within each of the spatial grids in combination with the crop growth status and the crop type; The crop growth status, the crop type and the crop root distribution data are fused with the soil spatial distribution information into crop spatial distribution data in a multi-layer grid form.

6. The intelligent water-saving control method for agricultural irrigation device according to claim 3, characterized in that: The method of calculating the irrigation urgency corresponding to the farmland irrigation strategy based on the farmland state prediction index and through a fuzzy logic system comprises the following steps: Using the farmland state prediction index as the first fuzzy input variable of the fuzzy logic system; defining a fuzzy set and a first membership function for each of the first fuzzy input variables; combining the fuzzy set and the first membership function, and using a fuzzy inference method to map all the first fuzzy input variables to fuzzy output variables according to a preset irrigation fuzzy inference rule; The fuzzy output variable is defuzzified by using the center of gravity method to obtain the irrigation urgency corresponding to the farmland irrigation strategy.

7. The intelligent water-saving control method for agricultural irrigation device according to claim 1, characterized in that: The filter component data includes the inlet and outlet water pressure difference, filtered water flow rate, filtered water turbidity and accumulated filtered water volume.

8. The intelligent water-saving control method for agricultural irrigation device according to claim 7, characterized in that: The step of combining the filter component data, the historical backwash data and the farmland irrigation strategy and calculating the backwash urgency of the filter irrigation module through a fuzzy logic system comprises the following steps: The Kalman filter method is used to merge the inlet and outlet water pressure difference, the filtered water flow rate and the turbidity of the filtered water into real-time filtering data; Based on the real-time filtering data, the predicted filtering data is obtained by predicting through a preset filtering state prediction model, wherein the filtering state prediction model is an LSTM model; using the predicted filtering data and the historical backwashing data as second fuzzy input variables of a fuzzy logic system; defining a second membership function for each of the second fuzzy input variables; Based on the second membership function and according to the preset backwashing fuzzy inference rule, a backwashing association relationship matrix is ​​established; Analyzing the second fuzzy input variable by the hierarchical analysis method to obtain a matrix weight vector of the backwashing association relationship matrix; Combining the backwash association relationship matrix and the matrix weight vector and obtaining a backwash reasoning evaluation result through fuzzy synthesis operation; The backwashing reasoning evaluation result is defuzzified by using the center of gravity method to obtain the initial backwashing urgency degree; An irrigation influencing factor is constructed based on the irrigation duration and irrigation water volume in the farmland irrigation strategy, and the initial backwashing urgency is corrected using the irrigation influencing factor to obtain the backwashing urgency of the filtering irrigation module.

9. The intelligent water-saving control method for agricultural irrigation device according to claim 8, characterized in that: The expression of the irrigation influencing factor is as follows: Where: δ represents the irrigation impact factor, α represents the irrigation water impact weight, q represents the irrigation water impact weight, n represents the irrigation water volume, Q represents the preset backwash interval water volume, Q M represents the cumulative filtered water volume of the filtering irrigation module, β represents the influence weight of irrigation time, t n represents the irrigation duration, and T represents the preset maximum irrigation duration.

10. An intelligent water-saving control system for an agricultural irrigation device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the intelligent water-saving control method for an agricultural irrigation device as described in any one of claims 1 to 9 is implemented.

Images