Agricultural irrigation monitoring method and system based on remote sensing

Abstract

The invention relates to the technical field of agricultural irrigation, and discloses an agricultural irrigation monitoring method based on remote sensing, which comprises the following steps: acquiring soil moisture data through satellite remote sensing, unmanned aerial vehicle inspection and a ground sensor, and establishing a partial differential equation to simulate moisture migration; variable irrigation is optimized by combining an optimal control theory, and water resource distribution is balanced by utilizing a partial differential game model; the optimal irrigation scheme is solved through numerical calculation, irrigation is accurately executed, and adjustment parameters are fed back in real time; the invention further provides an agricultural irrigation monitoring system based on remote sensing. The agricultural irrigation monitoring system comprises a data acquisition module, a soil moisture modeling module, a variable irrigation optimization module, an inter-plot water resource optimization module, a variable irrigation execution module and a feedback adjustment module. The remote sensing monitoring technology of multi-source data fusion is adopted, all-around and high-precision monitoring of farmland soil moisture is achieved, and accurate soil moisture assessment from macroscopic to microscopic is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural irrigation, and specifically provides a method and system for agricultural irrigation monitoring based on remote sensing. Background Art

[0002] Agricultural irrigation is a core link to ensure the stable growth of crops. The rational utilization of water resources is directly related to the yield and quality of agricultural products. However, in the actual production process, the supply of water resources is not always balanced, and the water demand of crops is also affected by soil characteristics, climatic conditions, and growth cycles. Therefore, relying solely on traditional experience or fixed irrigation patterns is difficult to adapt to complex agricultural environments.

[0003] Certain progress has been made in existing agricultural irrigation technologies. Some farmlands have begun to introduce ground sensor networks to monitor soil moisture, thus achieving data-driven irrigation management to a certain extent. In addition, emerging means such as drone inspections can provide image data on crop growth conditions, which helps to assist irrigation decision-making. Some intelligent irrigation systems also adopt fixed-program irrigation patterns and conduct irrigation scheduling based on historical data or statistical laws. These methods have obvious advantages in improving the level of agricultural automation and reducing labor input, and can optimize water resource allocation under specific conditions, improving crop yields and water resource utilization efficiency.

[0004] However, there are still many limitations in the existing technologies, making it difficult to meet the requirements of modern agricultural precision irrigation. First, the layout of ground sensors is limited, and the monitoring range is restricted, making it difficult to cover large areas of farmland. If the layout density is insufficient, it will lead to insufficient data sampling and affect the decision-making accuracy. Second, traditional intelligent irrigation systems often adopt fixed rules or empirical judgments, lacking accurate modeling of real-time soil moisture dynamics and being difficult to optimize the water demand characteristics of different plots. In addition, existing technologies usually only focus on the irrigation optimization of a single plot, ignoring the mutual influence of moisture between plots and being unable to optimize the allocation of water resources as a whole, which will result in water surplus in some areas and water shortage in other areas. Finally, although some irrigation systems can adjust the irrigation volume according to the monitoring data, the feedback mechanism is not flexible enough, making it difficult to adjust the irrigation strategy in real time. Once sudden weather changes or soil moisture anomalies occur, the system cannot respond quickly, resulting in waste of irrigation water resources or crop water shortage. For this reason, technical personnel in this field have proposed a method and system for agricultural irrigation monitoring based on remote sensing to solve the above problems. Summary of the Invention

[0005] Aiming at the deficiencies of the existing technologies, the present invention provides a method and system for agricultural irrigation monitoring based on remote sensing, which solves the limitations of existing agricultural irrigation technologies in aspects such as incomplete data acquisition, insufficient optimization of irrigation strategies, lack of coordination of water resources between plots, and untimely feedback adjustment.

[0006] To achieve the above object, the present invention is implemented through the following technical solutions: A remote sensing-based agricultural irrigation monitoring method includes the following steps: Obtain farmland soil moisture data through satellite remote sensing, unmanned aerial vehicle (UAV) inspections, and ground sensors, and analyze the moisture distribution of the farmland soil based on the obtained data to obtain the spatial and temporal variation characteristics of soil humidity. Establish a dynamic model of farmland soil moisture using partial differential equations to describe the processes of moisture diffusion, infiltration, and transpiration, accurately simulate the moisture migration law under different environmental conditions, and characterize the dynamic variation characteristics of soil moisture. Based on the dynamic model of soil moisture, establish a variable irrigation optimization model in combination with the optimal control theory, calculate the optimal irrigation input according to the water demand characteristics of crops and the soil moisture distribution, and achieve precise control of farmland moisture. Optimize the water resource allocation between plots using a partial differential game model, comprehensively consider the water demand and available water volume in different farmland areas, achieve balanced allocation of water resources between plots, and improve the overall irrigation efficiency. Use numerical calculation methods to solve the optimal irrigation strategy, and obtain the optimal irrigation plan that meets the crop growth requirements through the calculation of the variable irrigation optimization model and the water resource allocation model. Execute precise irrigation according to the calculated optimal variable irrigation plan, and use a feedback mechanism to monitor the changes in farmland moisture in real time, dynamically adjust the irrigation parameters, and ensure the adaptive optimization of the irrigation strategy to achieve precise and efficient agricultural irrigation management.

[0007] Preferably, the obtaining of the farmland soil moisture data includes: Obtain multi-spectral images and thermal infrared images through satellite remote sensing, and invert the soil moisture distribution. Obtain high-resolution soil humidity data through UAV inspections to correct the satellite data. Real-time collect soil humidity, air temperature, wind speed, and light data through ground sensors.

[0008] Preferably, the establishment of the dynamic model of farmland soil moisture includes: Use the Richards equation to describe the soil moisture diffusion process. Use the Penman-Monteith formula to calculate the evapotranspiration. Use the finite difference method to discretize the dynamic equation of soil moisture and construct a time-discrete model.

[0009] Preferably, the establishment of the variable irrigation optimization model includes: Use the Hamilton-Jacobi-Bellman equation to construct an optimal control problem. Constraining the irrigation input with the objective function to ensure optimal soil moisture; Calculating the optimal variable irrigation amount for a single plot through the optimal value function.

[0010] Preferably, the construction of the optimal control problem using the Hamilton-Jacobi-Bellman equation includes: Constructing an objective function to minimize the square of the deviation of soil moisture from the optimal value while minimizing the irrigation input; Describing the soil moisture state using the optimal value function and solving its time derivative; Calculating the irrigation input amount through the optimal solution of the control variable so that the optimal value function satisfies the HJB equation.

[0011] Preferably, the optimization of water resource allocation among plots includes: Modeling the moisture competition relationship between different plots using partial differential games; Calculating the optimal irrigation amount for each plot using the Nash equilibrium method; Adjusting the variable irrigation strategy through the moisture diffusion effect to optimize the overall water resource utilization.

[0012] Preferably, the modeling of the moisture competition relationship between different plots using partial differential games includes: Describing the change of soil moisture in different plots using the state equation; Calculating the moisture diffusion influence factor using the moisture gradient between adjacent plots; Calculating the irrigation input strategy for different plots through variable control.

[0013] Preferably, the numerical calculation method includes: Discretizing the Richards equation using the finite element method to simulate moisture diffusion; Using the reinforcement learning method to approximate the optimal solution of the Hamilton-Jacobi-Bellman equation; Using the iterative optimization method to calculate the Nash equilibrium solution of the partial differential game equation.

[0014] Preferably, the execution of the precise irrigation includes: Generating a variable irrigation prescription map to optimize the irrigation strategy by region; Adjusting the irrigation water volume in combination with the game model between plots to ensure optimal irrigation; Adjusting the irrigation parameters through real-time sensor feedback to optimize the execution accuracy.

[0015] There is also provided a remote sensing-based agricultural irrigation monitoring system, including: A data acquisition module, which is used to obtain satellite remote sensing, UAV inspection, and ground sensor data, and perform fusion processing on the acquired data to form spatio-temporal distribution information of farmland soil moisture; A soil moisture modeling module, which is used to establish a soil moisture dynamic equation based on the moisture information provided by the data acquisition module, describe the processes of moisture diffusion, infiltration, and transpiration, and solve the evolution law of the soil moisture state through numerical calculation methods; A variable irrigation optimization module, which is used to construct an optimal control model based on the moisture dynamic information provided by the soil moisture modeling module, calculate the optimal irrigation input for a single plot to meet the water demand of crops and optimize water resource utilization; An inter-plot water resource optimization module, which is used to optimize the water resource allocation strategy between plots based on the irrigation input of a single plot calculated by the variable irrigation optimization module, combined with the mutual influence of moisture between plots, to achieve optimal water resource regulation within the regional scope; A variable irrigation execution module, which is used to control the sprinkler and drip irrigation system to perform precise irrigation according to the optimal variable irrigation plan calculated by the inter-plot water resource optimization module, and collect farmland moisture data in real time to dynamically adjust irrigation parameters; A feedback adjustment module, which is used to compare the moisture data collected by the variable irrigation execution module with the remote sensing and sensing data provided by the data acquisition module, analyze the irrigation effect, correct the irrigation strategy, and continuously optimize the irrigation plan.

[0016] The present invention provides a remote sensing-based agricultural irrigation monitoring method and system. It has the following beneficial effects: 1. The present invention adopts a remote sensing monitoring technology of multi-source data fusion, achieving all-round and high-precision monitoring of farmland soil moisture. Compared with the monitoring method that only relies on ground sensors in the prior art, it breaks through the problems of limited spatial coverage and insufficient data sampling, and realizes precise soil moisture assessment from macro to micro.

[0017] 2. The present invention uses soil moisture dynamic modeling combined with numerical optimization to effectively predict the trend of moisture change. Compared with the traditional empirical irrigation method, it avoids the problems of lag in human decision-making and inaccurate calculation, improves the scientificity and feasibility of the irrigation plan, and makes water resource utilization more reasonable.

[0018] 3. The present invention realizes refined irrigation scheduling for a single plot and the entire farmland area based on variable irrigation optimization and inter-plot water resource allocation strategies. Compared with the traditional unified irrigation plan, it overcomes the defects of low water resource utilization rate, local over-irrigation or water shortage, and ensures the balanced supply of moisture in different plots.

[0019] 4. The present invention introduces a precise irrigation execution and real-time feedback adjustment mechanism, achieving the intelligent and adaptive adjustment of the irrigation system. Compared with the existing fixed irrigation strategy, it avoids the problem of being unable to adjust the irrigation plan according to real-time environmental changes, making agricultural irrigation more flexible and efficient, while reducing water resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a schematic flow chart of the method of the present invention; Figure 2 is a schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] Please refer to the attached Figure 1 , the embodiments of the present invention provide a remote sensing-based agricultural irrigation monitoring method and system, including the following steps: S1. Obtain farmland soil moisture data through satellite remote sensing, unmanned aerial vehicle (UAV) inspection, and ground sensors, and analyze the moisture distribution of the farmland soil based on the obtained data; Specifically, in this embodiment, step S1 mainly involves obtaining farmland soil moisture data. This step is the primary link in the remote sensing-based agricultural irrigation monitoring method of the present invention, aiming to obtain accurate soil moisture distribution information through multi-source data, providing basic support for subsequent soil moisture dynamic modeling and variable irrigation optimization.

[0023] Generally, the data acquisition process relies on the combination of multiple technical means, including satellite remote sensing, UAV inspection, and ground sensor data. These data are obtained through different sensing technologies, effectively supplementing the requirements of different spatial scales and resolutions, ensuring the spatio-temporal continuity and multi-dimensional accuracy of farmland soil moisture data. Specifically, the data acquisition module obtains the spatial distribution map and dynamic changes of soil moisture by combining the use of remote sensing images, high-precision UAV scanning, and ground real-time measurement.

[0024] The remote sensing satellite first conducts a wide-area scan of the farmland through multi-spectral images and thermal infrared images. These remote sensing images can reflect the humidity changes of the soil. The multi-spectral images can obtain the soil moisture distribution information of the farmland through remote sensing products such as vegetation indices and soil moisture indices. Satellite data usually covers a large area, but its spatial resolution is low, so the macroscopic soil moisture distribution is obtained.

[0025] To solve the problem of relatively low accuracy of remote sensing satellite data, drone inspection is adopted as a supplement. Drone inspection has a high spatial resolution and can obtain more accurate soil moisture data in local areas of farmland. The drone is equipped with a high-resolution humidity sensor to measure the surface soil humidity in real time and transmits the measurement results to the data center for processing through a data transmission module. Drone inspection is usually used in key areas or areas with relatively scarce data to correct errors or missing information in satellite data.

[0026] Specifically, the collection of soil moisture data also relies on the real-time support of ground sensors. Ground sensors include soil humidity sensors, air temperature sensors, wind speed sensors, light sensors, etc. These ground sensors monitor parameters such as soil humidity changes and climate environment, and can provide calibration information for remote sensing data and drone data, and update and feedback the data in real time.

[0027] The soil humidity data obtained through ground sensors can be combined with remote sensing data and drone data for data fusion to eliminate errors in a single data source and improve monitoring accuracy. In addition, ground sensors can also provide real-time data on the microenvironment of farmland, especially the immediate feedback on soil humidity in a small local area, which is crucial for the rapid adjustment of irrigation plans and precise irrigation strategies.

[0028] To further improve the accuracy of data, the system can perform multi-scale and multi-level data fusion on sensor data from different data sources. In the data fusion process, first, the remote sensing satellite data is preliminarily processed to obtain the preliminary humidity distribution of farmland soil; then, the accurate data obtained by the drone is used to correct the large-scale errors in the satellite data; finally, combined with the real-time data obtained by ground sensors, the timeliness and accuracy of soil humidity data are further optimized.

[0029] In addition, in the data acquisition module, the acquired data will also be preprocessed through specific algorithms. For example, the multi-spectral data in remote sensing images is inverted for soil humidity through an inversion algorithm. This process uses a soil humidity inversion model, such as a quadratic curve inversion algorithm, and calculates the soil humidity value of the corresponding area by analyzing the spectral reflectance data in the image. The commonly used formula in this process is: SM = a·R + b; where: SM represents soil humidity; a and b are fitting parameters; R represents the reflectance of the image.

[0030] When using this algorithm, a and b are usually obtained through on-site data calibration. These inversion results will be used as key inputs for remote sensing data and further combined with drone and ground sensor data to jointly provide high-precision soil moisture data.

[0031] Through this multi-dimensional and multi-data-source data collection method, it is possible to ensure that the monitoring data of soil moisture is more accurate and comprehensive, thus providing a more reliable basis for subsequent soil moisture modeling, irrigation optimization, and resource allocation.

[0032] S2. Establish a dynamic model of farmland soil moisture using partial differential equations to describe the processes of water diffusion, infiltration, and transpiration, and accurately simulate the water migration law under different environmental conditions; Specifically, in step S1, through the collaborative work of satellite remote sensing, UAV inspection, and ground sensors, we obtained multi-dimensional data of farmland soil moisture. Based on these data, in step S2, we further accurately described the dynamic changes of soil moisture through modeling means, providing theoretical support for subsequent irrigation optimization decisions. The goal of this step is to reveal the propagation process and influencing factors of water in the soil through dynamic modeling of soil moisture, so as to provide a dynamic model basis for precise irrigation.

[0033] Generally, the dynamic changes of soil moisture are not only affected by water diffusion but also by multiple factors such as evaporation and plant water absorption. Therefore, in this embodiment, partial differential equations (PDEs) are used to describe the dynamic changes of soil moisture. This equation can consider processes such as soil moisture diffusion, infiltration, and transpiration, as well as the influence of crop water absorption on soil moisture. Specifically, the Richards equation is used to describe the processes of soil moisture diffusion and infiltration, while transpiration is calculated through the Penman-Monteith formula. These models together constitute the dynamic equation of soil moisture.

[0034] Specifically, in one possible implementation, the processes of soil moisture diffusion and infiltration are described by the Richards equation. The Richards equation is a non-linear partial differential equation, usually written as: Where: represents the rate of change of soil moisture with time; θ represents the volumetric water content of the soil; t is time; represents the gradient operator; K(θ) is the hydraulic conductivity of the soil; ψ is the soil water potential; S(θ,ψ) represents the water source term, reflecting the crop water absorption process.

[0035] The Richards equation describes the process of soil moisture diffusion and takes into account the flow and infiltration characteristics of water in the soil. Through this equation, the changes of soil moisture with time and space can be accurately calculated.

[0036] As an option, the Penman-Monteith formula is usually used to estimate the transpiration of crops, and its formula is: Where: E is the evapotranspiration; Δ is the slope of the saturated water vapor pressure curve; R n is the net radiation; G is the soil heat flux; ρ a is the air density; c p is the specific heat capacity of air; e s is the saturated water vapor pressure; e a is the actual water vapor pressure; γ is the psychrometric constant; r a is the aerodynamic resistance; r s is the soil surface resistance.

[0037] The application of the Penman-Monteith formula can help us estimate the transpiration and further simulate the process of soil water consumption.

[0038] In this embodiment, the dynamic modeling of soil water is achieved by combining the Richards equation and the Penman-Monteith formula, considering the diffusion, infiltration, and evaporation of water. Through the numerical solution of these equations, the accurate simulation of the dynamic changes of soil water can be realized. Specifically, the finite difference method (FDM) or the finite element method (FEM) can be used to discretize these equations for numerical calculation.

[0039] Specifically, the discretization steps are as follows: for the time term, an explicit or implicit time difference formula is used to discretize it; for the spatial term, grid division is used to discretize the gradient term . Through these discretization operations, discrete equations suitable for numerical calculation can be obtained.

[0040] For example, for a one-dimensional problem, when using the explicit difference method for discretization, the time difference formula is: Where: represents the soil moisture at position i at time n; Δt is the time step; Δx is the spatial step; ψ is the soil water potential; is the soil water diffusivity, related to the soil volumetric water content θ; is the soil water potential at position i at time n; and are the soil water potentials at positions i + 1 and i - 1 respectively, representing the change of water potential in space.

[0041] This formula is used to update the soil water status at each moment.

[0042] In addition, for the simulation of crop transpiration, the evaporation can be dynamically calculated and added to the model as a water consumption term to reflect the actual consumption of soil water.

[0043] In this embodiment, the steps of soil moisture dynamic modeling provide a detailed mathematical description of soil moisture changes, enabling the system to accurately simulate the diffusion and consumption processes of moisture according to different soil characteristics and environmental conditions. The results of this modeling provide a scientific basis for subsequent irrigation optimization and water resource allocation.

[0044] S3. Based on the soil moisture dynamic model, combined with the optimal control theory, establish a variable irrigation optimization model, and calculate the optimal irrigation input according to the water demand characteristics of crops and the soil moisture distribution. Specifically, in the aforementioned steps S1 and S2, soil moisture data has been effectively acquired and modeled, which lays a solid foundation for the subsequent variable irrigation optimization. The core task of step S3 is to establish an accurate variable irrigation optimization model based on the soil moisture dynamic model, combined with the water demand characteristics of crops and the soil moisture distribution. The purpose of this model is to reasonably adjust the irrigation amount according to the soil moisture conditions of different plots, ensure that crops obtain sufficient water, and thus improve the efficiency and benefits of agricultural production.

[0045] Generally, the soil moisture dynamic model provides moisture data with spatio-temporal variations for irrigation optimization, while the variable irrigation optimization model calculates the most suitable irrigation amount based on these data using the optimal control theory to meet the growth needs of crops. Specifically, in this embodiment, the Hamilton-Jacobi-Bellman (HJB) equation in the optimal control theory is used to construct the variable irrigation optimization model, thereby obtaining the optimal irrigation amount for each plot.

[0046] The variable irrigation optimization problem can be described by the optimal control theory. Assuming that the goal of the system is to minimize the difference between the crop water demand and the actual irrigation amount, while minimizing the water resources required for irrigation, for this purpose, the HJB equation is used to construct the control problem in this embodiment. The general form of this equation is: Where: represents the partial derivative of the state function V(θ, t) with respect to time t, describing the change of the system over time; V(θ, t) represents the optimal value function at the soil volumetric water content θ; u is the control variable (i.e., the irrigation amount); is the dynamic model of the system, reflecting the change of soil moisture and the water demand of crops; represents the gradient of the system dynamic model, reflecting the sensitivity of the system state to soil moisture and crop demand; represents the partial derivative of the optimal value function with respect to the soil moisture state, reflecting the impact of moisture change on the objective function.

[0047] This equation represents the optimal control problem. Solving this equation can obtain the optimal irrigation strategy for each moment and each plot.

[0048] Specifically, it can consist of two parts: one part is the difference between crop water demand and soil moisture, representing the deviation between the water required by the crop and the current soil moisture; the other part is the cost function of irrigation input, representing the resource consumption per unit of irrigation water volume. By minimizing this objective function, the optimal irrigation input volume, that is, the irrigation volume for each moment and each plot, can be obtained.

[0049] In practical applications, the optimal value function V(θ,t) usually needs to be approximated by numerical methods, such as using value iteration method or policy iteration method for calculation. In this way, the optimal irrigation strategy for each plot can be obtained, and the irrigation efficiency can be effectively improved.

[0050] As an option, time constraints and spatial constraints of soil moisture can also be introduced during the optimization process to reflect the dynamic changes of soil moisture during irrigation. For example, assume that the soil moisture of a certain plot cannot be lower than a preset value within a specific time range. This constraint can be described by introducing relevant boundary conditions in the optimization model. In addition, the spatial distribution of irrigation volume can also be introduced as a constraint condition into the optimization model to avoid over-concentrated irrigation in a certain area, resulting in resource waste or uneven crop growth.

[0051] Furthermore, the calculation of the irrigation optimization model can be realized by modern numerical optimization techniques such as gradient descent method, simulated annealing or genetic algorithm. These methods can effectively search for the optimal solution and avoid the problem of low computational efficiency caused by over-reliance on traditional optimization methods.

[0052] In some embodiments, the optimization model not only considers the irrigation volume of a single plot, but also can consider the interactive effects between multiple plots. For example, the soil moisture states of adjacent plots will affect each other, so irrigation decisions need to take these interaction effects into account. This problem can be handled by a multivariable optimization model, further improving the accuracy and efficiency of irrigation decisions.

[0053] S4. Optimize the water resource allocation between plots by using a partial differential game model, comprehensively consider the water demands and available water volumes of different farmland areas, and achieve the balanced allocation of water resources between plots; Specifically, in the aforementioned steps S1 to S3, we have obtained the farmland soil moisture data and calculated the optimal irrigation amount for each plot through soil moisture dynamic modeling and variable irrigation optimization models. Next, the main task of step S4 is to optimize the water resource allocation between plots, ensuring the rational allocation of water resources between different plots to achieve the optimal utilization of water resources within the region. This step introduces a partial differential game model to conduct collaborative optimization among multiple plots, considering the water requirements of different plots and their mutual influences, thereby optimizing the overall irrigation strategy.

[0054] Generally, different plots of farmland have different soil types, crop varieties, and water requirements, so their irrigation demands vary. In addition, the diffusion effect of soil moisture causes the water states of adjacent plots to influence each other. Therefore, the irrigation strategy of a single plot cannot meet the water resource optimization requirements within the region. Therefore, in this embodiment, a partial differential game model is adopted to solve this problem and optimize the water resource allocation between plots.

[0055] The partial differential game model solves the optimal water resource allocation strategy between plots by considering the water requirements of multiple plots and their mutual influences. Specifically, the model simulates the diffusion effect of water by describing the water gradient between different plots and combines the irrigation requirements of each plot to calculate the optimal water resource allocation plan.

[0056] Specifically, the basic assumption of the partial differential game model is that each plot is regarded as a participant, and its goal is to optimize its own irrigation amount to maximize water conservation while meeting the crop requirements. Therefore, the irrigation decision of each plot in the game model is restricted by both its own water requirements and the irrigation decisions of adjacent plots. For this reason, this game problem can be transformed into a partial differential equation to describe the interaction of water between plots.

[0057] The basic form of the model is as follows: where: u i (x,t) represents the soil moisture state of the i-th plot at position x and time t; D i is the diffusion coefficient of the soil; represents the spatial gradient; f i (x,t) is the irrigation requirement of the i-th plot.

[0058] By solving this equation, the optimal irrigation amount of each plot at each moment can be obtained.

[0059] In addition, to consider the mutual influence between different plots, the game model needs to introduce the moisture gradient between adjacent plots. Assuming there is moisture exchange or diffusion between plot i and plot j, an interaction term can be introduced to describe this process. Specifically, the moisture flow between plots can be described by the following equation: where: α ij represents the moisture exchange coefficient between plot i and plot j, reflecting the intensity of moisture flow; u i (x,t) represents the soil moisture state of the i-th plot at position x and time t; D i is the diffusion coefficient of the soil; represents the spatial gradient; f i (x,t) is the irrigation demand of the i-th plot; u j (x,t) represents the soil moisture state of the j-th plot at position x and time t.

[0060] Through this equation, the mutual influence of moisture between adjacent plots can be taken into account, and the reasonable allocation of water resources can be achieved.

[0061] As an option, the game model can introduce the concept of Nash equilibrium to solve the optimal water resource allocation. In this case, the irrigation strategy of each plot should be a Nash equilibrium solution, that is, when the irrigation strategies of other plots remain unchanged, each plot chooses its own optimal irrigation amount. By solving the Nash equilibrium solution of this game model, the optimal water resource allocation plan for the entire region can be obtained.

[0062] In the numerical solution process, the finite difference method (FDM) or the finite element method (FEM) can be used to discretize the partial differential equation. By discretizing the model and combining the iterative calculation method, the optimal irrigation amount of each plot can be solved, and the efficient allocation of water resources can be achieved.

[0063] Specifically, the finite difference method approximately solves the partial differential equation by discretizing the continuous space into multiple grid points. By choosing appropriate time steps and space steps, efficient calculation of the numerical solution can be achieved. The finite element method divides the entire region into several small elements, uses interpolation functions to represent the moisture distribution, and obtains the approximate solution of the equation through numerical integration.

[0064] S5. Adopt numerical calculation methods to solve the optimal irrigation strategy. Through the calculation of the variable irrigation optimization model and the water resource allocation model, obtain the optimal irrigation plan that meets the crop growth requirements; Specifically, in the aforementioned steps S1 to S4, we obtained the soil moisture data of the farmland, established a soil moisture dynamic model, and determined the optimal irrigation strategy for each plot through variable irrigation optimization and inter-plot water resource optimization models. Next, step S5 aims to solve the optimization model through numerical calculation methods to further implement the optimal irrigation plan. Through this process, the system can transform the theoretically optimal irrigation strategy into executable operation instructions to control irrigation equipment and achieve precise irrigation.

[0065] Generally, the solution of the optimal irrigation strategy depends on efficient numerical calculation methods, which can quickly converge to the optimal solution under complex mathematical models. Specifically, in this embodiment, by introducing numerical calculation techniques such as the finite element method (FEM), reinforcement learning, and iterative optimization algorithms, the solution of the variable irrigation optimization model and the water resource allocation model is completed to ensure precise and efficient irrigation management.

[0066] The process of numerical calculation includes discretizing the soil moisture dynamic equation (such as the Richards equation) and gradually approaching the optimal solution using numerical integration methods. For example, for the moisture diffusion term and infiltration term in the Richards equation, the finite element method is used to discretize it spatially, and the time term is discretized by the time difference method. Specifically, the finite element method divides the farmland area into multiple finite small units (elements), and constructs interpolation functions on each unit to represent the moisture distribution. Through numerical integration, the soil moisture state at each moment and each unit can be gradually calculated, and the overall soil moisture distribution can be obtained. Formulated as: where: Ω is the farmland area; θ represents the volumetric water content of the soil; ψ is the soil water potential; K(θ) is the hydraulic conductivity; φ i is the interpolation function; S(θ,ψ) is the water source term, which reflects the water absorption process of the crop; represents the gradient operator.

[0067] Through the finite element method, the continuous soil moisture equation can be discretized into a set of linear equations, thereby obtaining the numerical solution of the moisture distribution.

[0068] As an option, the reinforcement learning method can further accelerate the solution of the optimal irrigation strategy. Specifically, by modeling the irrigation process as a Markov decision process (MDP), at each moment, the system decides the amount of irrigation water according to the current soil moisture state and the water demand of the crop. The system adjusts the irrigation strategy by continuously interacting with the environment and gradually learns the optimal irrigation decision in multiple iterations. Through the Q-learning or deep Q-network (DQN) algorithm in reinforcement learning, the system can automatically optimize the irrigation strategy through the reward feedback mechanism without an explicit model.

[0069] Specifically, the Q-learning algorithm calculates the expected reward for performing a certain action in each state by continuously updating the Q-value function. The Q-value update formula is as follows: Q(s,a1) = Q(s,a1) + α(r + γmax a Q(s ′ ,a1 ′ ); Where: s represents the current state; a1 is the selected action (i.e., irrigation water volume); r is the immediate reward; γ is the discount factor; α is the learning rate; s ′ is the new state after performing the action; a1 ′ is the optimal action in the new state.

[0070] Through this process, the system can continuously update the Q-value and learn the optimal irrigation decision.

[0071] In addition, for the problem of water resource allocation among multiple plots, numerical calculations can also be solved by iterative optimization algorithms. Specifically, optimization algorithms such as the gradient descent method, simulated annealing, or genetic algorithm are used, combined with the known soil moisture status and crop requirements, to solve for the optimal irrigation input. These optimization algorithms can solve the optimal solution of minimizing or maximizing the objective function (such as water resource usage, crop water requirements, etc.) under constraint conditions. Through these numerical calculation techniques, precise adjustment of irrigation amounts between plots can be achieved, avoiding waste of water resources.

[0072] In some embodiments, the optimization algorithm can adjust the step size according to the actual requirements of different plots to ensure the convergence and computational efficiency of the calculation results. Through multiple iterations, the system can gradually approach the optimal solution and dynamically adjust the irrigation plan.

[0073] S6. Perform precise irrigation according to the calculated optimal variable irrigation plan, and use the feedback mechanism to monitor the change of farmland moisture in real time and dynamically adjust the irrigation parameters.

[0074] Specifically, in the previous steps S1 to S5, we have completed the acquisition of farmland soil moisture, dynamic modeling, variable irrigation optimization, and water resource allocation. Next, the core task of step S6 is to convert the optimal irrigation plan into specific operation instructions, perform precise irrigation, and adjust the irrigation parameters according to real-time feedback to ensure the accuracy of water supply and the optimization of irrigation effects.

[0075] Generally, the implementation of precise irrigation relies on the precise control of irrigation systems (such as sprinkler irrigation, drip irrigation, etc.). Specifically, in this embodiment, the precise irrigation execution module will dynamically control the irrigation equipment according to the irrigation amount obtained from the optimization calculation in the previous steps, and adjust the irrigation parameters in real time through a feedback mechanism to ensure that the crops receive appropriate amounts of water at each growth stage.

[0076] The process of precise irrigation execution first generates an irrigation prescription map containing the irrigation requirements of each plot based on the output results of the irrigation optimization model. This prescription map specifies the irrigation amounts for different time periods for each plot, and these amounts are dynamically adjusted according to the changes in crop water requirements and soil moisture. The irrigation prescription map not only considers the requirements of individual plots but also the irrigation requirements of surrounding plots and their impact on water, so the allocation of irrigation water volume and time is based on the overall optimization plan to achieve the rational allocation of water resources.

[0077] Specifically, when implementing precise irrigation, first obtain real-time data of soil moisture through the sensor network and compare it with the previous moisture model. Assume that at this time, the soil moisture of a certain plot is low and immediate irrigation is required. The irrigation equipment (such as a sprinkler irrigation system) will be started according to the prescription map instructions, and the opening and closing timing and flow rate of each nozzle will be precisely controlled to ensure the accurate delivery of the required water volume.

[0078] During the execution process, the irrigation parameters will be dynamically adjusted according to real-time feedback. For example, assume that the feedback from the soil moisture monitoring sensor of a certain plot shows that the moisture content is close to the preset value. The irrigation equipment will make fine adjustments by adjusting the spraying time or water flow rate. This feedback mechanism ensures the accuracy and timeliness of water supply through data interaction with the soil moisture monitoring system. Specifically, the soil moisture data can be adjusted in real time through the following formula: Where: Δθ(t) is the change in soil moisture; I(t) represents the irrigation input; E(t) is the evaporation and crop water absorption; t is the time variable, representing the moment of irrigation or moisture change; represents the integral of the difference between the irrigation input and the evaporation / crop water absorption from time 0 to time t, reflecting the total change in soil moisture.

[0079] By comparing the expected value and the actual value of soil moisture, the system can dynamically adjust the irrigation amount and irrigation time to ensure that the soil moisture remains within the most suitable range.

[0080] As an option, the irrigation execution module can also integrate an automated control system to achieve remote connection with the sprinkler irrigation equipment through wireless communication technology. The system can receive and transmit irrigation instructions in real time. Whether it is manually adjusted in the central control system or directly scheduled through an automated algorithm, the irrigation equipment can timely adjust the water volume according to the instructed irrigation requirements.

[0081] In addition, in some embodiments, in order to improve irrigation efficiency and save water resources, the irrigation execution module can also monitor environmental parameters, such as temperature, humidity, wind speed, etc., through sensors. These environmental parameters directly affect the irrigation effect. Combining this information, the system can further optimize the irrigation strategy. For example, when the wind speed is high, the system can automatically reduce the sprinkling water volume to avoid water loss; while when the temperature is high, the system may increase the irrigation volume to supplement the evaporation loss.

[0082] In the process of implementing precision irrigation, the dynamic adjustment mechanism is crucial. The system can adjust the irrigation water volume and time in real time according to the feedback signal, so that the soil moisture can be maintained within the optimal range, avoiding excessive or insufficient water supply and ensuring the healthy growth of crops. For example, when the soil humidity approaches the threshold, the irrigation equipment will reduce the water output or pause irrigation to prevent excessive water accumulation.

[0083] In some embodiments, the system can also perform differentiated irrigation according to the water demand characteristics of different crops. For example, the system can customize irrigation for different crops through different irrigation strategies according to the water requirements of different crops. This process not only considers the soil moisture condition, but also involves multiple factors such as the growth stage of the crops and the climate conditions.

[0084] A remote sensing-based agricultural irrigation monitoring system described below can be correspondingly referred to the remote sensing-based agricultural irrigation monitoring method described above.

[0085] Please refer to the appendix Figure 2 , a remote sensing-based agricultural irrigation monitoring system, comprising: A data acquisition module, configured to obtain satellite remote sensing, unmanned aerial vehicle inspection, and ground sensor data, and perform fusion processing on the obtained data; A soil moisture modeling module, configured to establish a soil moisture dynamic equation based on the moisture information provided by the data acquisition module, describe the processes of moisture diffusion, infiltration, and transpiration, and solve the evolution law of the soil moisture state through numerical calculation methods; A variable irrigation optimization module, configured to construct an optimal control model based on the moisture dynamic information provided by the soil moisture modeling module, and calculate the optimal irrigation input for a single plot; Inter - plot water resource optimization module, which is used to optimize the water resource allocation strategy between plots based on the single - plot irrigation input calculated by the variable irrigation optimization module and combining the mutual influence of moisture between plots; Variable irrigation execution module, which is used to control the sprinkler and drip irrigation system to perform precise irrigation according to the optimal variable irrigation plan calculated by the inter - plot water resource optimization module, and collect farmland moisture data in real - time to dynamically adjust irrigation parameters; Feedback adjustment module, which is used to compare the moisture data collected by the variable irrigation execution module with the remote sensing and sensing data provided by the data collection module, analyze the irrigation effect, correct the irrigation strategy, and continuously optimize the irrigation plan.

[0086] Specifically, the data collection module is used to obtain and fuse multi - source data, including satellite remote sensing images, UAV inspection data, and information collected by ground sensors. This module can not only obtain large - scale farmland moisture distribution information, but also supplement detailed data through high - resolution UAV inspections, and at the same time combine the precise measurements of ground sensors to form a multi - scale, high - spatio - temporal resolution soil moisture information set. In addition, this module also supports data pre - processing, such as noise removal, data interpolation, and multi - source information fusion, to improve data quality and provide reliable data support for subsequent modeling and optimization.

[0087] Soil moisture modeling module, which is used to establish a soil moisture dynamic equation based on the soil moisture information provided by the data collection module to describe the processes of moisture diffusion, infiltration, and transpiration. This module uses numerical calculation methods, such as the finite element method (FEM) or the finite difference method (FDM), to solve the soil moisture equation to obtain the spatio - temporal evolution law of soil moisture status. At the same time, this module can combine meteorological data, crop transpiration requirements, and soil characteristic parameters to correct the moisture migration process to ensure the accuracy and applicability of model calculations. In some embodiments, this module also supports soil moisture prediction based on deep learning to enhance the modeling ability in complex environments.

[0088] Variable irrigation optimization module, which is used to construct an optimal control model based on the moisture dynamic information provided by the soil moisture modeling module and calculate the optimal irrigation input for a single plot. This module uses dynamic optimization algorithms, such as reinforcement learning, gradient descent method, or Lagrangian optimization method, to minimize water resource consumption while ensuring the optimal growth environment for crops. This module can dynamically adjust the optimization strategy according to the crop growth stage, soil moisture changes, and historical irrigation data to make irrigation decisions more intelligent. In some embodiments, this module can also combine with a crop growth model to further optimize the irrigation strategy to meet the water demand characteristics of different crops.

[0089] The inter-plot water resource optimization module is used to further optimize the overall water resource allocation strategy based on the optimal irrigation input of a single plot calculated by the variable irrigation optimization module. Considering the water interaction between plots, this module adopts global optimization methods such as distributed optimal control, genetic algorithm or simulated annealing algorithm to maximize irrigation efficiency and balance water resource use. This module can dynamically evaluate the soil water transfer between plots, reasonably adjust the irrigation input of each plot, and ensure the balance and sustainability of overall water resource utilization. In addition, this module can also combine meteorological forecast information to adjust the irrigation plan in advance to adapt to future precipitation and evapotranspiration changes.

[0090] The variable irrigation execution module is used to accurately control irrigation equipment such as sprinklers and drip irrigation systems according to the optimal variable irrigation plan calculated by the inter-plot water resource optimization module, so as to achieve intelligent and precise irrigation execution. This module adopts automatic control technology and conducts data interaction with the intelligent control terminal through a wireless communication network to achieve remote control and automatic adjustment. This module can real-time monitor the water status of farmland and dynamically adjust the irrigation water volume, irrigation duration and irrigation frequency according to the feedback information. In some embodiments, this module can also combine the real-time data of field sensors to achieve closed-loop control, ensure the accurate execution of the irrigation plan, and reduce water resource waste.

[0091] The feedback adjustment module is used to compare and analyze based on the water data collected by the variable irrigation execution module and combined with the remote sensing and sensing data provided by the data collection module to evaluate the irrigation effect and optimize the irrigation strategy. This module adopts machine learning and data analysis methods to evaluate the soil water change after irrigation, identify abnormal situations, and automatically adjust the subsequent irrigation plan. In addition, this module can also continuously optimize the irrigation model through long-term data accumulation to improve the adaptive ability of the system. In some embodiments, this module can predict the future soil water change trend based on time series analysis and pre-adjust the irrigation plan to improve agricultural water use efficiency.

[0092] The system of this embodiment can be used to execute the above method embodiment, and its principle and technical effect are similar, which will not be elaborated here.

[0093] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote sensing-based agricultural irrigation monitoring method, characterized in that Including the following steps: Obtain farmland soil moisture data through satellite remote sensing, drone inspection, and ground sensors, and analyze the moisture distribution of farmland soil based on the acquired data; Establish a dynamic model of farmland soil moisture using partial differential equations to describe the processes of moisture diffusion, infiltration, and transpiration, and accurately simulate the moisture migration law under different environmental conditions; Based on the dynamic model of soil moisture, establish a variable irrigation optimization model in combination with the optimal control theory, and calculate the optimal irrigation input according to the water demand characteristics of crops and the soil moisture distribution; Optimize the water resource allocation between plots using a partial differential game model, comprehensively consider the water demand and available water volume in different farmland areas, and achieve the balanced allocation of water resources between plots; Use numerical calculation methods to solve the optimal irrigation strategy, and obtain the optimal irrigation plan that meets the crop growth requirements through the calculation of the variable irrigation optimization model and the water resource allocation model; Execute precise irrigation according to the calculated optimal variable irrigation plan, and use the feedback mechanism to monitor the changes in farmland moisture in real time and dynamically adjust the irrigation parameters.

2. The method for monitoring agricultural irrigation based on remote sensing according to claim 1, characterized in that, The obtaining of farmland soil moisture data includes: Obtain multi-spectral images and thermal infrared images through satellite remote sensing, and invert the soil moisture distribution; Obtain high-resolution soil humidity data through drone inspection to correct satellite data; Real-time collect soil humidity, air temperature, wind speed, and light data through ground sensors.

3. The method for monitoring agricultural irrigation based on remote sensing according to claim 1, characterized in that, The establishment of the dynamic model of farmland soil moisture includes: Use the Richards equation to describe the soil moisture diffusion process; Use the Penman-Monteith formula to calculate the evapotranspiration; Discretize the dynamic equation of soil moisture using the finite difference method to construct a time-discrete model.

4. The method for monitoring agricultural irrigation based on remote sensing according to claim 1, wherein The establishment of the variable irrigation optimization model includes: Use the Hamilton-Jacobi-Bellman equation to construct an optimal control problem; Use the objective function to constrain the irrigation input amount to ensure the optimal soil moisture; Calculate the optimal variable irrigation amount of a single plot through the optimal value function.

5. The method for monitoring agricultural irrigation based on remote sensing according to claim 4, characterized in that, The use of the Hamilton-Jacobi-Bellman equation to construct an optimal control problem includes: Construct an objective function to minimize the square of the deviation of soil moisture from the optimal value and minimize the irrigation input at the same time; Use the optimal value function to describe the soil moisture state and solve its time derivative; Calculate the irrigation input amount through the optimal solution of the control variable to make the optimal value function satisfy the HJB equation.

6. The method for monitoring agricultural irrigation based on remote sensing according to claim 1, wherein The optimization of water resource allocation between plots includes: Use partial differential game to model the moisture competition relationship between different plots; Use the Nash equilibrium method to calculate the optimal irrigation amount of each plot; Adjust the variable irrigation strategy through the moisture diffusion effect to optimize the overall water resource utilization.

7. The method for monitoring agricultural irrigation based on remote sensing according to claim 6, characterized in that, The use of partial differential game to model the moisture competition relationship between different plots includes: Use the state equation to describe the soil moisture change of different plots; Calculate the moisture diffusion influence factor using the moisture gradient between adjacent plots; Calculate the irrigation input strategy of different plots through variable control.

8. The method for monitoring agricultural irrigation based on remote sensing according to claim 1, wherein The numerical calculation method includes: Use the finite element method to discretize the Richards equation to simulate moisture diffusion; Use the reinforcement learning method to approximate the optimal solution of the Hamilton-Jacobi-Bellman equation; Use the iterative optimization method to calculate the Nash equilibrium solution of the partial differential game equation.

9. A method for monitoring agricultural irrigation based on remote sensing according to claim 1, characterized in that, The implementation of the precise irrigation includes: Generate a variable irrigation prescription map and optimize the irrigation strategy by region; Adjust the irrigation water volume in combination with the inter-plot game model to ensure optimal irrigation; Adjust the irrigation parameters through real-time sensor feedback to optimize the execution accuracy.

10. A remote sensing-based agricultural irrigation monitoring system, which is applied to the remote sensing-based agricultural irrigation monitoring method described in any one of claims 1-9, and is characterized in that, It includes: A data acquisition module, which is used to obtain satellite remote sensing, UAV inspection, and ground sensor data, and perform fusion processing on the acquired data; A soil moisture modeling module, which is used to establish a soil moisture dynamic equation based on the moisture information provided by the data acquisition module, describe the processes of moisture diffusion, infiltration, and transpiration, and solve the evolution law of the soil moisture state through numerical calculation methods; A variable irrigation optimization module, which is used to construct an optimal control model based on the moisture dynamic information provided by the soil moisture modeling module and calculate the optimal irrigation input for a single plot; An inter-plot water resource optimization module, which is used to optimize the inter-plot water resource allocation strategy based on the single-plot irrigation input calculated by the variable irrigation optimization module and the mutual influence of moisture between plots; A variable irrigation execution module, which is used to control the sprinkler and drip irrigation system to perform precise irrigation according to the optimal variable irrigation plan calculated by the inter-plot water resource optimization module, and collect the farmland moisture data in real time to dynamically adjust the irrigation parameters; A feedback adjustment module, which is used to compare the moisture data collected by the variable irrigation execution module with the remote sensing and sensing data provided by the data acquisition module, analyze the irrigation effect, correct the irrigation strategy, and continuously optimize the irrigation plan.

Images