Meteorological data fused water-saving irrigation control method, device, equipment and medium

Abstract

The invention relates to the technical field of agricultural intelligent irrigation, and discloses a meteorological data fused water-saving irrigation control method, device and equipment and a medium. According to the method, a historical meteorological data set is constructed through multi-source meteorological data fusion, a dynamic water demand table is generated in combination with a crop water demand characteristic database, and a crop water demand model is established based on soil moisture content data. A grid irrigation unit division and growth period coupling soil moisture content response matrix construction technology is adopted, a cooperative constraint is established between a water demand threshold value and a water saving benefit through a multi-objective optimization learning algorithm, and a personalized irrigation scheme is generated. A soil moisture content dynamic evaluation matrix containing a dynamic time warping operator is designed, and dynamic matching of the soil layered soil moisture content and a standard template is achieved. The contradiction between meteorological response lag and low water resource utilization rate in traditional irrigation is effectively solved, accurate irrigation decision is realized through multi-dimensional data fusion and an intelligent optimization algorithm, and the water-saving benefit and the agricultural water resource utilization efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the field of agricultural irrigation, and in particular relates to a water-saving irrigation control method, device, equipment and medium integrating meteorological data. Background Art

[0002] With the development of agricultural water-saving irrigation technology, the concept of precision agriculture has promoted the widespread application of intelligent irrigation systems. In existing technologies, irrigation control methods have gradually evolved from single parameter monitoring (such as relying solely on soil moisture or rainfall) to multi-source data integration. By integrating meteorological factors such as temperature, humidity, and light, a dynamic assessment of crop water requirements is achieved. It has initially acquired the characteristics of environmental adaptability and can adjust irrigation strategies according to real-time climate conditions. Compared with traditional fixed-cycle irrigation modes, it has certain progress. In traditional technologies, irrigation control usually adopts empirical threshold methods or open-loop control methods based on single-point soil moisture conditions: by pre-setting fixed irrigation cycles and water volumes, or triggering irrigation based solely on surface soil moisture sensor data, it lacks comprehensive consideration of the water requirements of crops throughout their growth period, the hydraulic characteristics of the terrain, and the dynamic changes of water sources. Current irrigation control methods generally have the following problems: First, the fusion accuracy of multi-source meteorological data is insufficient, and traditional methods such as weighted averaging cannot handle the coupled effects of nonlinear meteorological factors on crop transpiration, resulting in deviations in water demand calculations; second, there is a lack of coupled modeling of the dynamic changes in soil moisture conditions and the water absorption characteristics of crop roots, and irrigation plans cannot accurately match the water demand of the root zone; third, there is a lack of a coordinated optimization mechanism for water source flow constraints and water-saving benefits, and there is often a contradiction where the irrigation volume exceeds the water source carrying capacity or excessive water saving affects crop yield; fourth, the closed-loop feedback regulation capability is weak, and it is difficult to respond in real time to sudden meteorological changes or equipment operation deviations during the irrigation process, resulting in a difficult balance between water-saving efficiency and crop growth assurance. Summary of the Invention

[0003] Based on this, it is necessary to provide a water-saving irrigation control method, device, equipment and medium that integrates meteorological data to solve the above problems.

[0004] In a first aspect, the present application provides a water-saving irrigation control method integrating meteorological data, comprising:

[0005] Obtain historical multi-source meteorological data and fuse them to obtain a historical meteorological dataset;

[0006] By matching historical meteorological data sets with a pre-established crop water requirement database, a table of crop water requirements under different meteorological conditions is generated.

[0007] Obtain historical soil moisture data and combine it with water demand tables to generate crop water demand models;

[0008] Obtain real-time multi-source meteorological data, soil moisture data, and water flow data, and combine them with crop water demand models to generate irrigation allocation plans.

[0009] In one embodiment, real-time multi-source meteorological data, soil moisture data, and water source flow data are obtained and combined with a crop water demand model to generate an irrigation allocation plan, including:

[0010] Obtain the grid coordinate information of crops and divide them into different irrigation units according to the terrain and hydraulic characteristics;

[0011] Obtain real-time soil moisture data for different irrigation units and construct a coupled soil moisture response matrix during the growing season in combination with the crop water demand model;

[0012] Acquire real-time multi-source meteorological data and water source flow data for different irrigation units, and combine them with the coupled soil moisture response matrix during the growth period to calculate the synergistic constraint conditions for plant growth water requirement thresholds and maximizing water-saving benefits;

[0013] Based on collaborative constraints, a multi-objective optimization learning algorithm is used to generate personalized irrigation plans.

[0014] In one embodiment, constructing a soil moisture dynamic assessment matrix includes:

[0015] The soil moisture dynamic assessment matrix is constructed using the following formula:

[0016]

[0017] Where d∈{1, 2, 3} is the index of soil stratification, where 1, 2, and 3 correspond to the soil layer, root zone layer, and subsoil layer, respectively. represents the deep dynamic weight, represents the sensitivity coefficient of the dth soil layer during the growth period, κ represents the adjustment parameter, represents the soil moisture response coefficient, τ p Indicates the number of days in the current reproductive period, T max and T min Respectively represent the preset maximum and minimum time ranges of the reproductive period p, S d，t represents the actual soil moisture value of the dth layer at time t, μ d represents the base soil moisture value of the dth layer, g represents the crop growth function, φ p represents the threshold soil moisture of the growing period p, DTW(·) represents the dynamic time warping operator, represents the actual moisture content sequence of the dth layer from the initial moment to time t, Represents the standard soil moisture process template for growing period p, represents the process stability factor, represents the change in soil moisture in the dth layer within the time interval Δ, Indicates the critical soil moisture change threshold during the growth period p.

[0018] In one embodiment, the multi-objective optimization learning algorithm uses the following mathematical formula:

[0019] Objective function:

[0020]

[0021] in:

[0022]

[0023] Constraints:

[0024]

[0025] D i，t ≤x i，

[0026]

[0027] In the above formula, x=(x i，t ) represents the irrigation decision matrix, x i，t represents the irrigation amount of the i-th irrigation unit in time window t, M i，t represents the soil moisture dynamic evaluation matrix, represents the dynamic water demand threshold, τ p represents the response coefficient of the reproductive period, represents the deep dynamic weight, represents the soil moisture sensitivity coefficient, Q max，t represents the upper limit of water source flow in time window t, DTW(·) represents the dynamic time warping operator, γ∈(0,1) represents the nonlinear adjustment factor of water-saving benefit, represents the process stability factor, f(S, x) represents the soil moisture dynamic evaluation model, Indicates the critical soil moisture change threshold during the growth period p.

[0028] In one embodiment, the method further comprises:

[0029] Generate irrigation control instructions based on personalized irrigation plans;

[0030] The dynamic data variational analysis algorithm is used to analyze the real-time changes in multi-source meteorological data, soil moisture data, and water source flow data within the irrigation unit when the irrigation control command is executed, and an irrigation status monitoring data set is generated;

[0031] Based on the irrigation status monitoring dataset, a closed-loop feedback mechanism is used to update the irrigation parameters corresponding to the personalized irrigation plan and generate optimized irrigation control instructions.

[0032] In one embodiment, the closed-loop feedback mechanism includes:

[0033] Build a corresponding system characteristic compensation model based on the irrigation system type, and calculate the deviation between the actual irrigation and the preset plan based on the irrigation status monitoring data set;

[0034] Based on the deviation, a multi-agent reinforcement learning algorithm is used to generate irrigation parameter compensation values. The compensation values include but are not limited to pipe network flow adjustment and irrigation duration correction values.

[0035] The compensation value is mapped to the execution parameter space of the personalized irrigation plan in real time to generate an anti-interference optimized irrigation plan.

[0036] In one embodiment, historical multi-source meteorological data is obtained and fused to obtain a historical meteorological dataset, including:

[0037] Acquire historical multi-source meteorological data and use a fusion algorithm based on Kalman filtering to perform fusion processing to obtain historical meteorological fusion data; meteorological factors of historical multi-source meteorological data include but are not limited to temperature, humidity, wind speed, wind direction, light intensity, and rainfall;

[0038] Based on the meteorological factors of meteorological data, a dynamic meteorological weight distribution model is constructed using the analytic hierarchy process;

[0039] The dynamic meteorological weight distribution model is used to optimize and adjust the historical meteorological fusion data to obtain the historical meteorological dataset.

[0040] In a second aspect, the present application also provides a water-saving irrigation control device integrating meteorological data, comprising:

[0041] Multi-source meteorological data fusion module, used to obtain historical multi-source meteorological data and fuse them to obtain historical meteorological data sets;

[0042] The crop water requirement characteristic matching module is used to match and calculate the historical meteorological data set with the pre-established crop water requirement characteristic database to generate a water requirement table for crops under different meteorological conditions;

[0043] The crop water demand modeling module is used to obtain historical soil moisture data and combine it with the water demand table to generate a crop water demand model;

[0044] The real-time irrigation allocation decision module is used to obtain real-time multi-source meteorological data, soil moisture data and water source flow data, and generate irrigation allocation plans in combination with the crop water demand model.

[0045] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned method steps for water-saving irrigation control by integrating meteorological data when executing the computer program;

[0046] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned water-saving irrigation control method integrating meteorological data;

[0047] The aforementioned meteorological data-integrated water-saving irrigation control method, device, equipment, and medium address the spatial and temporal heterogeneity of regional meteorological elements and soil moisture conditions, making it difficult for traditional irrigation systems to integrate these elements by acquiring multi-source meteorological data and fusing them together. By matching dynamic meteorological data sets with a database of crop water requirements, a water demand table is calculated and generated, establishing a dynamic correlation between meteorological fluctuations and crop water sensitivity during the growth period to avoid excessive irrigation or water stress. A dynamic water demand model is generated based on real-time soil moisture conditions, taking into account the coordinated optimization of water supply capacity and dynamic meteorological water demand, minimizing irrigation interruptions or water resource waste. Acquiring historical water source flow data and generating irrigation allocation plans enables flexible response to precipitation events, avoiding surface runoff and soil nutrient loss, thereby achieving the goal of water-saving irrigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 This is a step diagram of the water-saving irrigation control method integrating meteorological data according to the present invention;

[0050] Figure 2 This is a flow chart of the water-saving irrigation control method integrating meteorological data according to the present invention;

[0051] Figure 3 This is a structural diagram of the water-saving irrigation control device integrating meteorological data according to the present invention;

[0052] Figure 4 This is a structural diagram of an embodiment of a water-saving irrigation control device integrating meteorological data according to the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0054] The present invention provides a water-saving irrigation control method, device, equipment, and medium that integrates meteorological data. The hardware involved includes a data acquisition terminal, a server, and a user terminal, all of which work together via a network connection. In areas with water shortages or tight water allocation, such as arid regions or large farms, the data acquisition terminal acquires data and transmits it to a server or user terminal. The server or user terminal integrates multi-source meteorological data with information such as soil moisture to generate an irrigation allocation plan. The plan is dynamically adjusted based on real-time feedback, achieving precise irrigation and improving water resource utilization efficiency.

[0055] In one embodiment, Figure 1 As shown, a water-saving irrigation control method integrating meteorological data is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, or to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0056] S101, obtain historical multi-source meteorological data and fuse them to obtain a historical meteorological dataset.

[0057] Historical multi-source meteorological data can come from weather stations, satellite remote sensing, and radar detection, providing basic meteorological parameters such as temperature, humidity, air pressure, wind speed, wind direction, precipitation, and light intensity and duration. Data from different sensor sources are uniformly processed based on their formats and spatiotemporal characteristics to ensure consistent dimensions. For example, time stamp matching and interpolation algorithms can be used to bridge the time interval differences between data from multiple sources to create a meteorological dataset.

[0058] S102 , using historical meteorological data sets and a pre-established crop water requirement characteristic database for matching calculations to generate a crop water requirement table under different meteorological conditions.

[0059] Among them, the pre-established crop water requirement characteristics database contains the water requirement patterns of different crop varieties in each growth stage (seedling stage, flowering stage, maturity stage, etc.), such as the division of growth stages and the critical period of water requirement (peak water requirement during the filling stage); crop transpiration coefficient, soil water effective utilization coefficient; and the impact model of meteorological factors (such as temperature and light) on water requirement (such as the evapotranspiration calculation formula). The historical meteorological dataset can be split according to the time resolution (such as every 2 hours as a node) to extract the meteorological factor combination of each node (such as {temperature 25℃, light 800μmol / m 2 / s, 60% humidity}) and align the time intervals corresponding to meteorological factor combinations with the crop growth period to ensure that meteorological data within the same growth period matches the water demand characteristics of that stage. Crop evapotranspiration calculation methods (such as the Penman-Montes formula) can be used to calculate water requirements by growth period for all meteorological factor combinations in historical data (covering different temperature, humidity, light, rainfall, and other scenarios) (for example, the water requirements of the same crop on dry days and wet days during the seedling stage) to generate a water demand table.

[0060] S103, obtaining historical soil moisture data and combining it with the water demand table to generate a crop water demand model.

[0061] Soil moisture data includes time series of moisture content (e.g., volumetric moisture content of the 0-20cm, 20-50cm, and 50-100cm layers), soil texture (sand / clay / loam), and field capacity for different soil layers (e.g., surface layer, root zone layer, and subsoil layer). This data can be aligned over time (e.g., moisture content of the three layers at 8:00 AM daily) to form a structured dataset. Water requirement thresholds for different growth stages in the water requirement table (e.g., root zone moisture content must be ≥50% of field capacity during the seedling stage and ≥70% during the flowering stage) are used as samples to train crop water demand models. This model is then constructed through either mechanism-driven physical modeling or data-fitting-driven modeling.

[0062] S104: Acquire real-time multi-source meteorological data, soil moisture data, and water source flow data, and generate an irrigation allocation plan based on a crop water demand model.

[0063] Among them, real-time multi-source meteorological data, soil moisture data and water source flow data (which can come from the real-time water supply capacity of reservoirs, pumping stations, and channels) are obtained and aligned in time and space. The multi-source meteorological data and soil moisture data are input into the crop water demand model to obtain the irrigation volume required by the crops under the corresponding meteorological conditions. Then, constraints are imposed based on the water source flow data, such as water balance: the total irrigation volume ≤ the real-time water source flow upper limit (for example, the current flow of the pump station is 50m 3 / h, the total irrigation volume must be controlled within this range); soil moisture safety: ensure that the moisture content of each soil layer after irrigation is within the range required by crops, neither too high (for waterlogging prevention) nor too low (for drought prevention), and generate an irrigation allocation plan.

[0064] The above-mentioned water-saving irrigation control method integrating meteorological data obtains historical multi-source meteorological data and fuses them to obtain a historical meteorological data set, thereby solving the problem of distortion in meteorological element representation caused by traditional methods relying on single data; the data set is used to match and calculate the crop water requirement table under different meteorological conditions with a pre-established crop water requirement characteristic database, and a dynamic correlation is established between meteorological fluctuations and crop water requirement sensitivity during the growth period to avoid excessive irrigation or water stress; historical soil moisture data is obtained and combined with the water requirement table to generate a crop water demand model, considering the coupling of the dynamic changes in soil moisture in stratification and the water absorption characteristics of crop roots, so that the irrigation plan can accurately match the root zone water demand; real-time multi-source meteorological data, soil moisture data and water source flow data are obtained and combined with the crop water demand model to generate an irrigation distribution plan, considering the coordinated optimization of water source supply capacity and meteorological dynamic water demand, establishing the correlation between irrigation distribution and water source carrying capacity, reducing sudden irrigation interruptions or water resource waste, balancing water-saving benefits and crop growth guarantees, and improving the efficiency of agricultural water resource utilization.

[0065] In one embodiment, real-time multi-source meteorological data, soil moisture data, and water source flow data are obtained and combined with a crop water demand model to generate an irrigation allocation plan, including:

[0066] S201, obtaining grid coordinate information of crops and dividing them into different irrigation units according to topographical and hydraulic characteristics;

[0067] S202, obtaining real-time soil moisture data of different irrigation units, and constructing a coupled soil moisture response matrix during the growth period in combination with a crop water demand model;

[0068] S203, obtaining real-time multi-source meteorological data and water source flow data for different irrigation units, and combining them with the coupled soil moisture response matrix during the growth period to calculate the synergistic constraint conditions for plant growth water requirement threshold and maximizing water-saving benefits;

[0069] S204: Based on the collaborative constraints, a multi-objective optimization learning algorithm is used to generate a personalized irrigation plan.

[0070] Specifically, spatial coordinate data of crop planting areas can be obtained through a geographic information system (GIS) or satellite remote sensing to establish a geographic grid (e.g., 10 m × 10 m). Based on hydraulic parameters such as soil texture (sand / clay / loam), field water holding capacity, and permeability, combined with terrain slope and drainage direction, grids with similar hydrological response characteristics are clustered into irrigation units. Real-time soil moisture data is obtained from soil sensors within each irrigation unit. Combined with a crop water demand model (e.g., the water sensitivity of different crops at different growth stages), a growth-period-coupled soil moisture response matrix is constructed to quantify the dynamic response of crops to soil moisture at different growth stages. Taking into account real-time meteorological data (temperature, humidity, light, etc., which influence crop transpiration), the growth-period-coupled soil moisture response matrix is used to determine the water demand range for each unit under current meteorological conditions and growth stages. Water source flow data (e.g., reservoir water flow rate and pumping station head limit) are input as constraints. Through matrix operations, the water demand interval is coupled with the water source constraint to form a feasible solution space for the optimization problem. For example, when water resources are tight, priority is given to satisfying the irrigation of critical units in the water demand interval, while reducing the irrigation volume of units in non-critical periods. Avoid large-scale production reductions due to insufficient water resources or waste of water resources due to excessive irrigation under corresponding climatic conditions. Through a multi-objective optimization algorithm, the optimal irrigation strategy under the current collaborative constraints is solved, and a differentiated solution is generated for each irrigation unit, including irrigation time (such as avoiding high temperature periods to reduce evaporation), water distribution (such as different flow control for drip irrigation areas and sprinkler irrigation areas), and irrigation methods (such as triggering drip irrigation when the root zone is short of water, and triggering sprinkler irrigation when the surface is short of water), to achieve precise irrigation with one policy for each area.

[0071] In one embodiment, constructing a soil moisture dynamic assessment matrix includes:

[0072] S301, calculate the soil moisture dynamic evaluation matrix using the following formula:

[0073]

[0074] Where d∈{1, 2, 3} is the index of soil stratification, where 1, 2, and 3 correspond to the soil layer, root zone layer, and subsoil layer, respectively. represents the deep dynamic weight, represents the sensitivity coefficient of the dth soil layer during the growth period, κ represents the adjustment parameter, represents the soil moisture response coefficient, τ p Indicates the number of days in the current reproductive period, T max and T min Respectively represent the preset maximum and minimum time ranges of the reproductive period p, S d，t represents the actual soil moisture value of the dth layer at time t, μ d represents the base soil moisture value of the dth layer, g represents the crop growth function, φ prepresents the threshold soil moisture of the growing period p, DTW(·) represents the dynamic time warping operator, represents the actual moisture content sequence of the dth layer from the initial moment to time t, Represents the standard soil moisture process template for growing period p, represents the process stability factor, represents the change in soil moisture in the dth layer within the time interval Δ, Indicates the critical soil moisture change threshold during the growth period p.

[0075] For example, the distribution of soil moisture in the vertical direction (surface layer, root zone layer, subsoil layer) has significant effects on crop growth. Dynamic weights are assigned to different soil layers to highlight the role of key soil layers. The response of crops to soil moisture deviation is not linear, so the logistic function can be used, i.e. Quantify the nonlinearity of soil moisture deviation and fit the threshold response characteristics of biological systems; crops have different sensitivities to soil moisture changes at different growth stages, and a soil moisture response coefficient related to the number of days in the growth period can be introduced is the crop growth function g at the threshold soil moisture condition φ p The derivative at reflects the instantaneous impact rate of soil moisture changes on crop growth; the actual soil moisture changes are time series and need to be compared with historical or standard soil moisture templates, such as the standard soil moisture process template for the growth period p. Comparison is done to ensure that the current soil moisture trend is consistent with the crop growth law. The dynamic time warping operator DTW(·) is used to measure sequence similarity and solve the time series alignment problem (such as irrigation delays and time series offsets caused by weather fluctuations). Soil moisture may fluctuate drastically due to factors such as short-term rainfall and evaporation. The process stability factor The weight is adjusted according to the ratio of the soil moisture change to the critical threshold to suppress the impact of abnormal fluctuations. The soil moisture of the surface layer (evaporation-dominated), the root zone (water absorption core), and the subsoil (water reserve) are integrated in the spatial dimension, and the current soil moisture deviation (static value S d，t With the reference value μ d The soil moisture dynamic evaluation matrix M is obtained by matching the standard template with the DTW operator to evaluate whether the dynamic evolution is reasonable. t , reflecting the comprehensive suitability of soil moisture under the current state - historical trends - growth period requirements.

[0076] In one embodiment, S401, the multi-objective optimization learning algorithm uses the following mathematical formula:

[0077] Objective function:

[0078]

[0079] in:

[0080]

[0081] Constraints:

[0082]

[0083] D i，t ≤x i，

[0084]

[0085] In the above formula, x=(x i，t ) represents the irrigation decision matrix, x i，t represents the irrigation amount of the i-th irrigation unit in time window t, M i，t represents the soil moisture dynamic evaluation matrix, represents the dynamic water demand threshold, τ p represents the response coefficient of the reproductive period, represents the soil moisture sensitivity coefficient, Q max，t represents the upper limit of water source flow in time window t, DTW(·) represents the dynamic time warping operator, γ∈(0,1) represents the nonlinear adjustment factor of water-saving benefit, represents the process stability factor, f(S, x) represents the soil moisture dynamic evaluation model, Indicates the critical soil moisture change threshold during the growth period p.

[0086] Specifically, multi-objective optimization requires a balance between meeting crop water demand and saving water resources, which requires a multi-objective optimization method. 需水 (x) By measuring the actual irrigation volume x i，t and the dynamic evaluation matrix M of soil moisture i，t (obtained in step S301) Dynamic water demand threshold value under corresponding weight The difference reflects the degree of satisfaction of crop water demand. 节水 (x) By constraining the irrigation volume to account for the upper limit of water source flow Q max，t proportion The dynamic time warping (DTW) is used to evaluate the rationality of irrigation time distribution and quantify the water-saving benefits. The weight parameters λ1 and λ2 adjust the priority of water demand and water conservation, which can be dynamically adjusted according to the actual scenario requirements (such as giving priority to water demand during drought). t ≤Q max，t Ensure that the total irrigation volume does not exceed the water supply capacity; D i，t ≤x i， tSet upper and lower limits of irrigation based on the minimum water requirement of crops and the water holding capacity of soil; By dynamically evaluating the matrix M based on the moisture content i，t The evaluation model f(S,x) limits the amplitude of soil moisture changes after irrigation to prevent excessive wetting or drought.

[0087] In one embodiment, the method further comprises:

[0088] S501, generating irrigation control instructions according to the personalized irrigation plan;

[0089] S502, using a dynamic data variational analysis algorithm to analyze real-time changes in multi-source meteorological data, soil moisture data, and water source flow data within the irrigation unit when the irrigation control command is executed, and generating an irrigation status monitoring data set;

[0090] S503: Based on the irrigation status monitoring data set, a closed-loop feedback mechanism is used to update the irrigation parameters corresponding to the personalized irrigation plan, and an optimized irrigation control instruction is generated.

[0091] For example, the generated personalized irrigation plan is converted into an executable device instruction by

[0092] Instruction mapping mechanism such as based on traffic demand (such as 5m 3 / h) to control the pump frequency; set the solenoid valve on / off timing according to a time window (e.g., 10:00-12:00), generate control instructions, and send them to the irrigation unit's actuators (e.g., valves, pump stations, etc.) via communication protocols such as Modbus and MQTT. A dynamic data variational analysis algorithm uses multi-source meteorological data, soil moisture data, and water source flow data to establish the system state equation and generate a prediction model. By minimizing the difference between observed data and model predictions, it estimates hidden variables (e.g., unknown changes in soil water permeability) and performs anomaly detection (e.g., calculating the residual between actual and predicted values and marking residuals exceeding a threshold as abnormal events). This generates a structured irrigation status monitoring dataset, including but not limited to the following information: timestamp, irrigation unit ID, actual irrigation amount, soil moisture changes, meteorological fluctuations, water source flow consumption, and whether anomalies exist. Based on the irrigation status monitoring dataset, a closed-loop update mechanism can be used to compensate for execution deviations and environmental interference (such as calculating the deviation of meteorological, soil, and water source change data and generating corresponding compensation values, which are superimposed on the original irrigation plan), and generate optimized irrigation control instructions according to the instruction mapping mechanism.

[0093] In one embodiment, the closed-loop feedback mechanism includes:

[0094] S601, constructing a corresponding system characteristic compensation model based on the irrigation system type, and calculating the deviation between the actual irrigation and the preset plan based on the irrigation status monitoring data set;

[0095] S602: Based on the deviation, a multi-agent reinforcement learning algorithm is used to generate irrigation parameter compensation values, which include but are not limited to pipe network flow adjustment and irrigation duration correction values.

[0096] S603: Map the compensation value to the execution parameter space of the personalized irrigation plan in real time to generate an anti-interference optimized irrigation plan.

[0097] Specifically, a compensation model is established based on the differences in physical characteristics of different irrigation systems (such as drip irrigation, sprinkler irrigation, and underground irrigation) (such as pipeline water pressure decay and emitter flow-pressure relationship) to eliminate the impact of equipment nonlinearity on irrigation results. Taking the drip irrigation system as an example, the hydraulic characteristic equation can be established as follows: Among them, Q 实际 Indicates the actual flow rate, Q 预设 Indicates the preset flow rate. It represents the ratio of the actual pressure of the pipeline to the rated pressure, β represents the friction coefficient of the pipeline, and L represents the length of the pipeline; the valve opening delay equation is established as t 开启 =t 指令 +τ·(1-e -k·ΔP ), where t 开启 Indicates the time required for the valve to open, t 指令Represents the time required to generate the instruction, τ represents the time constant, k represents the gain coefficient of the influence of pressure difference on the valve opening speed, and ΔP represents the pressure difference; the compensation model is obtained according to the established equation, and the deviation between the actual irrigation and the preset plan is calculated according to the irrigation status monitoring data set, including but not limited to the irrigation amount deviation (the ratio of irrigation amount deviation to the preset irrigation amount), soil moisture deviation (the ratio of the soil moisture change of each irrigation unit to the critical soil moisture change threshold of the crop growth period p), and response deviation (the ratio of the soil moisture change rate to the preset change rate). According to the deviation, according to the multi-agent reinforcement learning algorithm (each irrigation unit is regarded as an agent, and each agent makes autonomous decisions based on local monitoring data (such as soil moisture, weather, water flow, equipment status, etc.). Its situation space can be the previous soil moisture dynamic evaluation matrix, deviation, water source flow upper limit, crop growth period stage, etc.; the output is the parameter to be adjusted (such as flow adjustment amount ΔQ, duration correction value ΔT). The reward function can be used to balance the water-saving benefits (such as the deviation between the actual water consumption and the planned amount) and the crop water demand satisfaction (such as whether the soil moisture is maintained within the threshold). Historical data or real-time data is used for training to generate compensation values for parameters such as pipe network flow and irrigation time. For example, when the actual flow is monitored to be 10% lower than the preset value, the algorithm automatically calculates and outputs a compensation value for increasing the pipe network flow by 5%. At the same time, information is shared through the communication network to collaboratively optimize the global irrigation strategy.) According to the generated compensation value such as flow compensation value ΔQ = +2m 3 / h, and a duration correction value of ΔT = -10 minutes, to perform parameter space mapping. For example, for a drip irrigation system, ΔQ corresponds to the solenoid valve opening adjustment, and for a sprinkler irrigation system, ΔT corresponds to the pump start and stop time correction. By dynamically adjusting parameters, irrigation plans can adapt to real-time changes. For example, when a sudden increase in wind speed causes sprinkler evaporation to increase, the irrigation duration is automatically extended to compensate for evaporation losses. When water flow suddenly decreases, irrigation for crops in critical growth periods is prioritized, while quotas for non-critical areas are reduced. These steps, through a closed-loop monitoring-analysis-compensation-execution system, address the poor environmental adaptability and weak anti-interference capabilities of traditional open-loop control.

[0098] In one embodiment, historical multi-source meteorological data is obtained and fused to obtain a historical meteorological dataset, including:

[0099] S701, acquiring historical multi-source meteorological data and fusing it using a fusion algorithm based on Kalman filtering to obtain historical meteorological fusion data; meteorological factors of the historical multi-source meteorological data include but are not limited to temperature, humidity, wind speed, wind direction, light intensity, and rainfall;

[0100] S702, based on the meteorological factors of meteorological data, a dynamic meteorological weight distribution model is constructed using the analytic hierarchy process;

[0101] S703: Optimize and adjust the historical meteorological fusion data using a dynamic meteorological weight distribution model to obtain a historical meteorological data set.

[0102] For example, multi-source meteorological data may include satellite remote sensing, ground-based meteorological stations, and numerical weather prediction models. The data obtained includes, but is not limited to, temperature, humidity, wind speed, wind direction, light intensity, and rainfall. Because data from these sources can differ significantly or conflict in temporal and spatial resolution, accuracy, and sampling frequency, a Kalman filter fusion algorithm can be employed. This algorithm treats meteorological factors as system state variables, constructs state equations and observation equations, and performs prediction-correction. This algorithm dynamically adjusts the weights of each data source by combining historical data with current observations. For example, when satellite data aligns with historical trends, it is given a higher weight; when anomalies or conflicts occur, its weight is reduced. After standardizing the temporal and spatial dimensions and accuracy of data acquisition, and filling or removing data gaps caused by different sampling frequencies, fused historical meteorological data is generated. The analytic hierarchy process can be used to rank the importance of meteorological factors to irrigation decisions. For example, temperature, which affects crop transpiration rate, has a higher weight in hot seasons; humidity, which affects soil evaporation, has a higher weight in arid regions; and rainfall, which directly replenishes soil moisture, has a significantly higher weight in the rainy season. The weights of fused historical meteorological data are updated in real time based on temporal and spatial factors. For example: in the time dimension, the weight of temperature at noon in summer is the highest, and the weight of humidity increases at night; in the spatial dimension, the weight of rainfall in arid areas is higher than that in humid areas; by dynamically adjusting the weight of data sources and the importance of data changes over time and space, the final data set is closer to actual irrigation needs.

[0103] To further illustrate the technical steps of the present invention, in an exemplary embodiment, Figure 2 The process shown in the figure uses the following water-saving irrigation control method that integrates meteorological data in corn agricultural production activities in a certain area:

[0104] S11 uses weather stations, soil sensors, and water source monitoring equipment to obtain historical multi-source meteorological data (temperature, humidity, wind speed, and rainfall), soil moisture data (surface layer, root zone layer, and subsoil moisture content), and real-time water flow data. A Kalman filter algorithm is used to perform spatiotemporal fusion of multi-source meteorological data, eliminating sensor noise and generating a high-precision historical meteorological dataset.

[0105] S12: Combined with the crop water requirement database (including corn growth period water requirement thresholds), the Penman-Montes formula is used to calculate the theoretical water requirements for each growth period under different meteorological conditions, creating a dynamic water requirement table. For example, during the tasseling period, corn requires 6.5 mm of water per day at an average daily temperature of 28°C and a humidity of 60%.

[0106] S13 analyzes the relationship between root zone moisture content and crop growth based on historical soil moisture data, and uses a regression model to fit a soil moisture-water requirement curve. Real-time meteorological data is input into the model to predict crop water requirements for the next three days and prioritize irrigation (for example, triggering irrigation when root zone moisture content falls below 60% during flowering).

[0107] S14, based on the terrain slope and soil permeability coefficient, the corn field is divided into 10 irrigation units. Using the multi-objective optimization algorithm, with water demand satisfaction and water saving benefits as the dual objectives, under the total water source flow limit (reservoir daily water supply 500m 3 ) to allocate irrigation duration and water volume to each unit. For example, prioritize irrigation for unit A, which is at its critical water demand stage, while reducing the quota for unit B, which is at its mature stage.

[0108] S15: After executing the irrigation plan, the system monitors soil moisture changes and weather fluctuations in real time. If the actual moisture content of unit C is 10% lower than the predicted value, the system dynamically adjusts the irrigation flow rate using a reinforcement learning algorithm, increasing the irrigation volume by 5% to compensate for evaporation losses and updating the irrigation plan for the next three hours.

[0109] The aforementioned water-saving irrigation control method, which integrates meteorological data, significantly improves the accuracy of crop water demand predictions based on nonlinear meteorological factors such as temperature, humidity, and light through the integration of multi-source meteorological data and a dynamic weight allocation model. Combining a dynamic soil moisture assessment matrix for stratified soil moisture with a deep dynamic weighting algorithm, it accurately quantifies the spatiotemporal coupling between crop root water absorption characteristics and root zone moisture. The innovative embedding of a dynamic time warping operator into the soil moisture response model during the growth period enables dynamic matching of actual soil moisture sequences with standard templates, effectively addressing the problems of delayed meteorological response and one-sided soil moisture monitoring in traditional irrigation. A multi-objective optimization algorithm establishes a synergistic mechanism between water demand thresholds and water-saving benefits under the constraints of water source flow, enabling irrigation allocation to improve water-saving efficiency while ensuring crop growth. The closed-loop feedback system, relying on a reinforcement learning algorithm, compensates for pipe network pressure fluctuations and evaporation losses in real time, improving anti-interference response speed. Through grid unit division and clustering of terrain hydraulic characteristics, hourly irrigation strategy optimization can be achieved in thousands of acres of farmland, computing efficiency is improved, and ultimately an intelligent irrigation system with multi-dimensional coordination of meteorology, soil, and water sources is constructed, which maximizes water-saving benefits while ensuring reasonable and stable irrigation of crops.

[0110] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0111] Based on the same inventive concept, embodiments of the present application also provide a water-saving irrigation control device for implementing the aforementioned meteorological data-integrated method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of a water-saving irrigation control device for meteorological data-integrated provided below can be found in the above-described limitations of the water-saving irrigation control method for meteorological data-integrated method, and are not further elaborated here.

[0112] In an exemplary embodiment, Figure 3 As shown, a water-saving irrigation control device integrating meteorological data is provided, comprising:

[0113] The multi-source meteorological data fusion module 11 is used to obtain historical multi-source meteorological data and fuse them to obtain a historical meteorological data set;

[0114] The crop water requirement characteristic matching module 12 is used to match and calculate the historical meteorological data set with the pre-established crop water requirement characteristic database to generate a water requirement table for crops under different meteorological conditions;

[0115] The crop water demand modeling module 13 is used to obtain historical soil moisture data and generate a crop water demand model based on the water demand table;

[0116] The real-time irrigation allocation decision module 14 is used to obtain real-time multi-source meteorological data, soil moisture data and water source flow data, and generate an irrigation allocation plan in combination with the crop water demand model.

[0117] In one embodiment, the real-time irrigation allocation decision module 14 is further configured to:

[0118] Obtain the grid coordinate information of crops and divide them into different irrigation units according to the terrain and hydraulic characteristics;

[0119] Obtain real-time soil moisture data for different irrigation units and construct a coupled soil moisture response matrix during the growing season in combination with the crop water demand model;

[0120] Acquire real-time multi-source meteorological data and water source flow data for different irrigation units, and combine them with the coupled soil moisture response matrix during the growth period to calculate the synergistic constraint conditions for plant growth water requirement thresholds and maximizing water-saving benefits;

[0121] Based on collaborative constraints, a multi-objective optimization learning algorithm is used to generate personalized irrigation plans.

[0122] In one embodiment, the real-time irrigation allocation decision module 14 is further configured to:

[0123] The soil moisture dynamic assessment matrix is constructed using the following formula:

[0124]

[0125] Where d∈{1, 2, 3} is the index of soil stratification, where 1, 2, and 3 correspond to the soil layer, root zone layer, and subsoil layer, respectively. represents the deep dynamic weight, represents the sensitivity coefficient of the dth soil layer during the growth period, κ represents the adjustment parameter, represents the soil moisture response coefficient, τ p Indicates the number of days in the current reproductive period, T max and T min Respectively represent the preset maximum and minimum time ranges of the reproductive period p, S d，t represents the actual soil moisture value of the dth layer at time t, μ d represents the base soil moisture value of the dth layer, g represents the crop growth function, φ p represents the threshold soil moisture of the growing period p, DTW(·) represents the dynamic time warping operator, represents the actual moisture content sequence of the dth layer from the initial moment to time t, Represents the standard soil moisture process template for growing period p, represents the process stability factor, represents the change in soil moisture in the dth layer within the time interval Δ, Indicates the critical soil moisture change threshold during the growth period p.

[0126] In one embodiment, the real-time irrigation allocation decision module 14 is further configured to construct a multi-objective optimization learning algorithm using the following mathematical formula:

[0127] Objective function:

[0128]

[0129] in:

[0130]

[0131] Constraints:

[0132]

[0133] D i，t ≤x i，

[0134]

[0135] In the above formula, x=(x i，t ) represents the irrigation decision matrix, x i，t represents the irrigation amount of the i-th irrigation unit in time window t, M i，t represents the soil moisture dynamic evaluation matrix, represents the dynamic water demand threshold, τ p represents the response coefficient of the reproductive period, represents the soil moisture sensitivity coefficient, Q max，t represents the upper limit of water source flow in time window t, DTW(·) represents the dynamic time warping operator, γ∈(0,1) represents the nonlinear adjustment factor of water-saving benefit, represents the process stability factor, f(S, x) represents the soil moisture dynamic evaluation model, Indicates the critical soil moisture change threshold during the growth period p.

[0136] In one embodiment, Figure 4 As shown, the device further includes a feedback optimization module 15, which is used to:

[0137] Generate irrigation control instructions based on personalized irrigation plans;

[0138] The dynamic data variational analysis algorithm is used to analyze the real-time changes in multi-source meteorological data, soil moisture data, and water source flow data within the irrigation unit when the irrigation control command is executed, and an irrigation status monitoring data set is generated;

[0139] Based on the irrigation status monitoring dataset, a closed-loop feedback mechanism is used to update the irrigation parameters corresponding to the personalized irrigation plan and generate optimized irrigation control instructions.

[0140] In one embodiment, the closed-loop feedback mechanism in the feedback optimization module 15 includes:

[0141] Build a corresponding system characteristic compensation model based on the irrigation system type, and calculate the deviation between the actual irrigation and the preset plan based on the irrigation status monitoring data set;

[0142] Based on the deviation, a multi-agent reinforcement learning algorithm is used to generate irrigation parameter compensation values. The compensation values include but are not limited to pipe network flow adjustment and irrigation duration correction values.

[0143] The compensation value is mapped to the execution parameter space of the personalized irrigation plan in real time to generate an anti-interference optimized irrigation plan.

[0144] In one embodiment, the multi-source meteorological data fusion module 11 is further configured to:

[0145] Acquire historical multi-source meteorological data and use a fusion algorithm based on Kalman filtering to perform fusion processing to obtain historical meteorological fusion data; meteorological factors of historical multi-source meteorological data include but are not limited to temperature, humidity, wind speed, wind direction, light intensity, and rainfall;

[0146] Based on the meteorological factors of meteorological data, a dynamic meteorological weight distribution model is constructed using the analytic hierarchy process;

[0147] The dynamic meteorological weight distribution model is used to optimize and adjust the historical meteorological fusion data to obtain the historical meteorological dataset.

[0148] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the water-saving irrigation control method integrating meteorological data as described above are implemented.

[0149] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0150] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0151] The above-described embodiments merely represent several implementation methods of the embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the concept of the embodiments of the present application, and these modifications and improvements fall within the scope of protection of the embodiments of the present application.

Claims

1. A water-saving irrigation control method integrating meteorological data, characterized in that: The method comprises: Obtain historical multi-source meteorological data and fuse them to obtain a historical meteorological dataset; Using the historical meteorological data set and a pre-established crop water requirement characteristic database for matching and calculation, a water requirement table for crops under different meteorological conditions is generated; Obtaining historical soil moisture data and combining it with the water demand table to generate a crop water demand model; Real-time multi-source meteorological data, soil moisture data, and water source flow data are obtained, and combined with the crop water demand model to generate an irrigation allocation plan.

2. The method according to claim 1, characterized in that The acquisition of real-time multi-source meteorological data, soil moisture data, and water source flow data, and combining them with the crop water demand model to generate an irrigation allocation plan, includes: Obtain the grid coordinate information of crops and divide them into different irrigation units according to the terrain and hydraulic characteristics; Acquire real-time soil moisture data of different irrigation units, and construct a coupled soil moisture response matrix during the growth period in combination with the crop water demand model; Acquire real-time multi-source meteorological data and water source flow data of different irrigation units, and calculate the synergistic constraint conditions of plant growth water requirement threshold and water-saving benefit maximization based on the growth period coupled soil moisture response matrix; Based on the collaborative constraints, a multi-objective optimization learning algorithm is used to generate a personalized irrigation plan.

3. The method according to claim 2, characterized in that The construction of the soil moisture dynamic assessment matrix includes: The soil moisture dynamic assessment matrix is constructed using the following formula: Where d∈{1, 2, 3} is the index of soil stratification, where 1, 2, and 3 correspond to the soil layer, root zone layer, and subsoil layer, respectively. represents the deep dynamic weight, represents the sensitivity coefficient of the dth soil layer during the growth period, κ represents the adjustment parameter, represents the soil moisture response coefficient, τ p Indicates the number of days in the current reproductive period, T max and T min Respectively represent the preset maximum and minimum time ranges of the reproductive period p, S d，t represents the actual soil moisture value of the dth layer at time t, μ d represents the base soil moisture value of the dth layer, g represents the crop growth function, φ p represents the threshold soil moisture of the growing period p, DTW(·) represents the dynamic time warping operator, represents the actual moisture content sequence of the dth layer from the initial moment to time t, Represents the standard soil moisture process template for growing period p, represents the process stability factor, represents the change in soil moisture in the dth layer within the time interval Δ, Indicates the critical soil moisture change threshold during the growth period p.

4. The method according to claim 2, characterized in that The multi-objective optimization learning algorithm adopts the following mathematical formula: Objective function: in: Constraints: In the above formula, x=(x i，t ) represents the irrigation decision matrix, x i，t represents the irrigation amount of the i-th irrigation unit in time window t, M i，t represents the soil moisture dynamic evaluation matrix, represents the dynamic water demand threshold, τ p represents the response coefficient of the growth period, represents the soil moisture sensitivity coefficient, Q max，t represents the upper limit of water source flow in time window t, DTW(·) represents the dynamic time warping operator, γ∈(0,1) represents the nonlinear adjustment factor of water-saving benefit, represents the process stability factor, f(S, x) represents the soil moisture dynamic evaluation model, Indicates the critical soil moisture change threshold during the growth period p.

5. The method according to claim 2, characterized in that The method further comprises: generating irrigation control instructions according to the personalized irrigation plan; Analyzing the real-time changes of multi-source meteorological data, soil moisture data, and water source flow data in the irrigation unit when the irrigation control instruction is executed using a dynamic data variation analysis algorithm, and generating an irrigation status monitoring data set; Based on the irrigation status monitoring data set, a closed-loop feedback mechanism is used to update the irrigation parameters corresponding to the personalized irrigation plan, and an optimized irrigation control instruction is generated.

6. The method according to claim 5, characterized in that The closed-loop feedback mechanism includes: Constructing a corresponding system characteristic compensation model based on the irrigation system type, and calculating the deviation between the actual irrigation and the preset plan based on the irrigation status monitoring data set; Based on the deviation, a multi-agent reinforcement learning algorithm is used to generate an irrigation parameter compensation value, wherein the compensation value includes but is not limited to a pipe network flow adjustment amount and an irrigation duration correction value; The compensation value is mapped to the execution parameter space of the personalized irrigation plan in real time to generate an anti-interference optimized irrigation plan.

7. The method according to claim 1, characterized in that The acquisition of historical multi-source meteorological data and the fusion of the historical meteorological data set include: Acquire historical multi-source meteorological data and perform fusion processing using a fusion algorithm based on Kalman filtering to obtain historical meteorological fusion data; meteorological factors of the historical multi-source meteorological data include but are not limited to temperature, humidity, wind speed, wind direction, light intensity, and rainfall; Based on the meteorological factors of the meteorological data, a dynamic meteorological weight distribution model is constructed using the analytic hierarchy process; The dynamic meteorological weight distribution model is used to optimize and adjust the historical meteorological fusion data to obtain a historical meteorological data set.

8. A water-saving irrigation control device integrating meteorological data, characterized in that: The device comprises: Multi-source meteorological data fusion module, used to obtain historical multi-source meteorological data and fuse them to obtain historical meteorological data sets; A crop water requirement characteristic matching module is used to match and calculate the historical meteorological data set with a pre-established crop water requirement characteristic database to generate a water requirement table for crops under different meteorological conditions; A crop water demand modeling module is used to obtain historical soil moisture data and generate a crop water demand model in combination with the water demand table; The real-time irrigation allocation decision module is used to obtain real-time multi-source meteorological data, soil moisture data and water source flow data, and generate an irrigation allocation plan in combination with the crop water demand model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images