Water and fertilizer integrated irrigation control method for wheat planting

Abstract

The invention relates to the technical field of intelligent agriculture, and discloses a water and fertilizer integrated irrigation control method for wheat planting, and the method comprises the following steps: deploying a composite physiological sensor (comprising a high-sensitivity piezoelectric acoustic probe and an ion selective microelectrode array) to collect acoustic and ion signals reflecting moisture and nutrient states, extracting feature vectors, and calculating a water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water and fertilizer integrated water; the method comprises the following steps: calculating physiological rhythm synchronism to determine a dynamic decision weight so as to eliminate abnormal data, weighting and fusing feature vectors of normal sentinel plants, constructing partitioned comprehensive physiological features, decoding the features into a specific physiological state by utilizing a classification model, and finally, according to the state, applying a coupling control logic of water stress priority, so as to obtain a comprehensive physiological state of the sentinel plants. And generating and executing a water and fertilizer control instruction. By directly monitoring the physiological status of crops and introducing a self-checking mechanism, accurate and self-adaptive closed-loop water and fertilizer management is realized.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture technology, specifically to a method for integrated water and fertilizer irrigation control in wheat cultivation. Background Technology

[0002] As a major global food crop, wheat's yield and quality are crucial to ensuring food security. During wheat growth, water and fertilizer management is a core factor determining its yield. Scientific and efficient water and fertilizer management can not only significantly improve crop yield and resource utilization efficiency, but also effectively avoid water waste and environmental problems such as soil and groundwater pollution caused by excessive irrigation and fertilization.

[0003] Traditional irrigation and fertilization methods rely heavily on agronomic experience or fixed irrigation cycles, which is relatively extensive and difficult to precisely match the actual needs of crops. In recent years, some automated control methods based on environmental parameters have emerged, such as estimating crop evapotranspiration by deploying soil moisture sensors or combining weather station data to guide irrigation decisions. However, these existing environmental parameter-based technologies have a fundamental limitation: they monitor the external environment of crop growth (such as soil or air), rather than the physiological state of the crop itself. This indirect measurement method leads to a significant lag in water and fertilizer decisions, failing to respond in real time to the internal physiological stresses of crops caused by environmental changes or disease stress. Furthermore, due to the high spatial heterogeneity of soil physicochemical properties in the field, readings from a single or small number of environmental sensors often fail to accurately represent the average condition of the entire management area, significantly reducing the uniformity and effectiveness of water and fertilizer management. More importantly, existing technologies are mostly focused on water management, lacking direct and effective online monitoring capabilities for the dynamic needs of crops for key nutrients such as nitrogen and potassium at different growth stages, making true integrated water and fertilizer precision control difficult to achieve. At the same time, these systems generally lack a self-verification mechanism for the reliability of front-end sensor data. Once the sensor malfunctions or the data is abnormal, it may lead to incorrect irrigation or fertilization decisions, which may have an adverse effect on crop growth. Summary of the Invention

[0004] The purpose of this invention is to provide an integrated water and fertilizer irrigation control method for wheat cultivation, which aims to solve the technical problem of inaccurate water and fertilizer application caused by the inability to accurately and timely obtain the physiological needs of crops through indirect parameters of the external environment in the existing technology.

[0005] To achieve the above objectives, the present invention provides a method for integrated water and fertilizer irrigation control in wheat cultivation, the method comprising the following steps:

[0006] S1. In one or more preset management zones, collect in real time the acoustic signal of xylem sap flow and the net rhizosphere ion flow rate of at least one sentinel plant in each management zone.

[0007] S2. Determine the crop physiological state of the management zone based on the acoustic signal of the xylem sap flow and the net flow rate of rhizosphere ions;

[0008] S3. Generate and execute water and fertilizer integrated irrigation control instructions for the management zone based on the crop physiological state.

[0009] Preferably, step S2 specifically includes:

[0010] Acoustic feature vectors are extracted from the acoustic signal of the xylem fluid flow, and ion flow feature vectors are extracted from the net rhizosphere ion flow rate.

[0011] The acoustic feature vector and the ion flow feature vector are combined to determine the physiological state of the crop.

[0012] Furthermore, the acoustic feature vector includes at least one of the following: total signal energy, embolic event energy percentage, frequency centroid, and embolic event count;

[0013] The ion flow characteristic vector includes at least one of the following: average absorption rate, peak absorption rate, and flow direction stability.

[0014] Preferably, the method further includes:

[0015] Before step S2, the dynamic decision weight of each sentinel plant is calculated for each sentinel plant in each management partition.

[0016] The step of determining the crop physiological state of the management zone in step S2 is based on the dynamic decision weights, which are used to weight the acoustic signals of xylem sap flow and the net rhizosphere ion flow rate of each sentinel plant.

[0017] In one specific embodiment, the step of calculating the dynamic decision weights of each sentinel plant includes:

[0018] Establish a baseline for population physiological rhythms;

[0019] Calculate the synchronicity score between the physiological rhythm of each sentinel plant and the baseline of the population physiological rhythm;

[0020] The dynamic decision weights are determined based on the synchronicity score.

[0021] Furthermore, the synchronicity score is obtained by calculating the normalized cross-correlation function between the physiological rhythm signal of the sentinel plant and the baseline physiological rhythm signal of the population. The synchronicity score (W) sync,i It is determined by the following formula:

[0022]

[0023] Among them, R i,base (τ) is the normalized cross-correlation function, calculated by the following formula:

[0024]

[0025] In the formula: i is the index of the sentinel plant; E[·] is the mathematical expectation calculation; S i (t) represents the time series of physiological rhythm signals of sentinel plant i; S base (t) represents the time series of the baseline signal of the population's physiological rhythm; τ represents the time delay; μ i and μ base Signal S i (t) and S base The mean of (t); σ i and σ base Signal S i (t) and S base The standard deviation of (t).

[0026] Furthermore, when the synchronization score of a sentinel plant is lower than a preset synchronization threshold, its dynamic decision weight is set to zero, and it is marked as an abnormal sentinel plant. The dynamic decision weight (W) i It is determined by the following formula:

[0027]

[0028] In the formula: W i θ represents the dynamic decision weights for sentinel plant i. sync This is the preset synchronization threshold.

[0029] In a specific embodiment, the step of determining the crop physiological state of the management zone specifically includes: performing a weighted average of the acoustic feature vector and ion flow feature vector extracted from each sentinel plant according to the dynamic decision weight to obtain the comprehensive physiological feature vector of the management zone; and inputting the comprehensive physiological feature vector into a preset classification model to output the crop physiological state.

[0030] Preferably, in step S3, when the determined crop physiological state is a combination of water stress and nutrient demand, the fertigation control command prioritizes the execution of the irrigation command, and after the water stress is relieved, the fertilization command is then determined based on the new crop physiological state.

[0031] In one specific embodiment, in step S1, the xylem fluid acoustic signal is acquired by a high-sensitivity piezo-acoustic probe deployed on the stem of the sentinel plant; the net rhizosphere ion velocity is measured by an ion-selective microelectrode array deployed in the rhizosphere of the sentinel plant.

[0032] This invention provides a method for integrated water and fertilizer irrigation control in wheat cultivation. It has the following beneficial effects:

[0033] 1. This invention achieves direct and synchronous monitoring of wheat plants' water stress status and specific nutrient requirements by combining real-time acoustic signals of xylem sap flow and net rhizosphere ion flow rate from sentinel plants, and by combining these two different physiological signals to determine the crop's physiological state. This overcomes the lag and uncertainty caused by relying on indirect parameters such as soil and weather, making water and fertilizer control decisions more accurate and timely.

[0034] 2. This invention establishes a systematic self-verification mechanism by setting a baseline for the physiological rhythms of the population and calculating the synchronicity score of each sentinel plant with this baseline to determine dynamic decision-making weights. This mechanism can automatically identify sentinel plants that generate abnormal data due to local diseases or sensor malfunctions, and reduce or eliminate the impact of their data on the decision-making process, thereby significantly improving the robustness and reliability of the entire monitoring and control system and ensuring the accuracy of physiological state assessment in the management zones.

[0035] 3. In the control command generation step, this invention sets up control logic that prioritizes the execution of irrigation commands in response to the concurrent physiological state of water stress and nutrient demand. This strategy aligns with the physiological principle that plants have limited nutrient absorption capacity under water stress, avoiding ineffective fertilization when plants are water-deficient. This not only improves the utilization efficiency of water and fertilizer resources but also prevents the risk of secondary soil salinization that may result from ineffective fertilization. Attached Figure Description

[0036] Figure 1 This is a structural block diagram of the system of the present invention;

[0037] Figure 2 This is a flowchart of the multimodal physiological signal processing and characterization process of the present invention;

[0038] Figure 3 This is a flowchart of the dynamic weight allocation mechanism of the present invention;

[0039] Figure 4 This is a flowchart of the partitioned physiological state decoding process of the present invention;

[0040] Figure 5 This is a flowchart illustrating the generation and execution of integrated water and fertilizer irrigation control commands according to the present invention.

[0041] Among them, 10 is the sensing and execution unit; 11 is the composite physiological sensor; 111 is the high-sensitivity piezo-acoustic probe; 112 is the ion-selective microelectrode array; 12 is the zoned data acquisition module; 20 is the central processing server; 30 is the water and fertilizer execution device; 31 is the zoned water and fertilizer controller; 32 is the solenoid valve; and 33 is the fertilizer injection pump. Detailed Implementation

[0042] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] See attached document Figure 1 The present invention provides a system for the integrated water and fertilizer irrigation control method for wheat planting. The system may include: one or more sensing and execution units 10 deployed in the field, a central processing server 20, and one or more water and fertilizer execution devices 30.

[0044] The sensing and execution unit 10 is used to collect crop physiological signals and receive control commands. The central processing server 20 is used to process physiological signals, determine the crop physiological state, and generate control commands. The water and fertilizer execution device 30 is used to execute specific water and fertilizer application operations according to the control commands.

[0045] In one specific embodiment, before proceeding with the method, the target wheat field is first divided into several independent management zones based on prior geographical and soil information of the farmland. The division is based on, but is not limited to, elevation and slope data obtained through a digital elevation model, and soil texture and organic matter content distribution data obtained through soil surveys or sensor networks. Each management zone is equipped with an independent sensing execution unit 10 and a water and fertilizer execution device 30.

[0046] See attached document Figure 1 At least one sensing execution unit 10 is deployed in each management zone. The sensing execution unit 10 includes a composite physiological sensor 11 and a zone data acquisition module 12. In each management zone, at least one wheat plant with a representative growth state is selected as a sentinel plant, and the composite physiological sensor 11 is deployed on the sentinel plant.

[0047] The composite physiological sensor 11 includes a highly sensitive piezoacoustic probe 111 and an ion-selective microelectrode array 112. The highly sensitive piezoacoustic probe 111 is non-invasively fixed to the epidermis at the base of the sentinel plant stem using an acoustic coupling agent and a fixing clamp, and is used to collect acoustic vibration signals generated by sap flow within the xylem vessels of wheat. The ion-selective microelectrode array 112 is inserted into the rhizosphere soil surrounding the roots of the sentinel plant, with its sensing end face adjacent to the root surface, and is used to measure specific nutrient ions (e.g., NO3-) on the root surface. - NH4 + K + The concentration gradient of ).

[0048] The partition data acquisition module 12 is connected to the composite physiological sensor 11 via wired or wireless means. It is responsible for receiving and processing the raw electrical signals generated by the high-sensitivity piezoacoustic probe 111 and the ion-selective microelectrode array 112, and packaging them into data frames. The partition data acquisition module 12 periodically sends the data frames to the central processing server 20 via wireless communication (e.g., LoRaWAN or NB-IoT protocol).

[0049] In one specific embodiment, each electrode of the ion-selective microelectrode array 112 is modified with an ion-selective membrane that selectively responds to a specific ion. For example, for measuring nitrate ions (NO3-). - The electrode, whose selective membrane material can be a polyvinyl chloride (PVC) membrane based on tetradecylammonium bromide (TDDA), is used to measure potassium ions (K). + The electrodes, whose selective membrane material can be a valine-based PVC membrane, are also included in the array. The array also includes a reference electrode to provide a stable potential reference.

[0050] The central processing server 20 receives data from all partition data acquisition modules 12 and performs subsequent data analysis and decision generation. After generating an integrated water and fertilizer irrigation control command, the central processing server 20 sends the command to the water and fertilizer execution device 30 corresponding to the target management partition.

[0051] The water and fertilizer execution device 30 includes a zone water and fertilizer controller 31, a solenoid valve 32 connected to the main irrigation pipeline, and a fertilizer injection pump 33 connected to a fertilizer storage tank. The zone water and fertilizer controller 31 communicates with the central processing server 20 and receives control commands. The zone water and fertilizer controller 31 is electrically connected to the solenoid valve 32 and the fertilizer injection pump 33 respectively, and precisely controls the opening duration of the solenoid valve 32 to regulate the irrigation water volume and the running time of the fertilizer injection pump 33 to regulate the fertilizer application rate according to the command content.

[0052] See attached document Figure 2After acquiring the raw signal through the composite physiological sensor 11, the method of the present invention further includes the acquisition and characterization of multimodal physiological signals.

[0053] In one specific embodiment, for processing the acoustic signal of the xylem fluid flow, the partition data acquisition module 12 samples and quantizes the continuous analog signal output by the high-sensitivity piezoacoustic probe 111 at a preset sampling rate (e.g., 1 MHz) to obtain a discrete original acoustic signal time series A. i (t). After receiving the time series, the central processing server 20 first performs denoising processing on it. This denoising processing uses the wavelet transform method; specifically, a mother wavelet (e.g., the Daubechies 5 wavelet) is selected to denoise the original acoustic signal time series A. i (t is decomposed into multiple layers, and then a preset soft threshold or hard threshold is applied to the wavelet coefficients obtained from each layer of decomposition. Finally, the denoised signal is reconstructed through inverse wavelet transform.)

[0054] Subsequently, the central processing server 20 extracts a set of acoustic feature vectors F based on the denoised signal to characterize the water stress state. A,i The acoustic feature vector F A,i Includes one or more of the following characteristics: total signal energy E t otal,i, is calculated using the following formula:

[0055] E total,i =∫ T [A′ i (t)] 2 dt;

[0056] Among them, A′ i (t) represents the denoised signal, and T represents the integration time window.

[0057] Embolism event count N e,i The calculation method is as follows: in the denoised time-domain signal A′ i In (t), the signal pulse is identified by the peak detection algorithm. The pulse that is identified, whose amplitude exceeds the first preset threshold and whose duration is within the preset time range, is counted as an embolism event.

[0058] For processing the net rhizosphere ion flow rate, the ion-selective microelectrode array 112 is configured to measure the concentration of a specific ion at two different distances along the normal direction of the root surface of the sentinel plant, thereby obtaining the concentration gradient of that ion. Based on this concentration gradient, the central processing server 20 calculates the net ion flow rate J according to Fick's first law. ion,i (t), its calculation formula is:

[0059]

[0060] Among them, D ion The effective diffusion coefficient of this specific ion in the rhizosphere medium is a constant pre-calibrated based on parameters such as soil type and temperature. The calculation result J... ion,i A negative value of (t) indicates that ions flow from the rhizosphere into the root system, while a positive value indicates that ions flow from the root system into the rhizosphere.

[0061] Based on the calculated net ion flow rate time series, the central processing server 20 further extracts a set of ion flow feature vectors F to characterize the nutrient absorption status of the plant. E,i The ion current feature vector F E,i Includes one or more of the following features:

[0062] Average absorption rate It is within an observation time window T obs The average net flow velocity within.

[0063] Peak absorption rate It is within the observation time window T obs Internal ion net flow rate J ion,i The minimum value of (t) is the maximum inflow rate.

[0064] Flow stability S dir,i It is defined as a time window T during observation. obs The ratio of the total duration of negative net ion flow rate (inflow into the root system) to the total duration of the observation time window.

[0065] Finally, the central processing server 20 will process the acoustic feature vector F A,i and ion current eigenvector F E,i The features are combined to form a combined physiological feature vector for subsequent analysis and processing.

[0066] In the acoustic feature vector F A,i and ion current eigenvector F E,i Before combining the two vectors into a unified combined physiological feature vector, it is preferable to first normalize each feature component in the two vectors. For example, the Z-score normalization method can be used to transform each feature component into a distribution with a mean of 0 and a standard deviation of 1. This normalization step aims to eliminate the weight imbalance problem in subsequent model training caused by excessive differences in the units and numerical ranges between different features.

[0067] See attached document Figure 3 The method of the present invention further includes a dynamic weight allocation step based on population physiological rhythms to quantitatively evaluate and verify the data credibility of each sentinel plant.

[0068] In one specific embodiment, this step is performed by the central processing server 20. First, a population circadian rhythm baseline S needs to be established for comparison. base (t). Specifically, a signal that reflects the diurnal rhythmic changes in water activity of the plant is selected as the physiological rhythm signal. For example, one of the acoustic features extracted in Part II—the total signal energy E. total,i The 24-hour time series. During the system initialization phase, or during a confirmed non-stress period (e.g., after a full irrigation and within a period of mild environmental conditions), the central processing server 20 collects the circadian rhythm signals S of all sentinel plants within the management zone. i (t), and calculate its arithmetic mean, using this mean sequence as the initial population physiological rhythm baseline S. base (t). To adapt to the changes in physiological characteristics of wheat at different growth stages, the baseline physiological rhythm S of this population... base (t) can be set to be recalibrated and updated at the beginning of each major growth stage (e.g., greening, jointing, grain filling).

[0069] During system operation, the central processing server 20 periodically (e.g., hourly) calculates the physiological rhythm signal S of each sentinel plant i within the partition. i (t) and the established population circadian rhythm baseline S base The synchronicity score W of (t) symc,i The synchronicity score is obtained by calculating the normalized cross-correlation function between the two time series. The reason for using the normalized cross-correlation function is that its calculation result is not sensitive to the absolute amplitude and DC bias of the signals, but only focuses on the similarity of the two signals in waveform and the degree of alignment in phase.

[0070] To increase the baseline of population physiological rhythms S base To enhance the robustness of (t), an iterative elimination method can be employed when calculating the initial arithmetic mean. Specifically, first, the preliminary mean and standard deviation of the physiological rhythm signals of all sentinel plants are calculated. Then, signals deviating from the preliminary mean by more than twice the standard deviation are considered initial anomalous signals and temporarily eliminated. The mean is then recalculated based on the remaining signals, and this is used as the final baseline for the population's physiological rhythms. This method can effectively reduce the impact of individual anomalous plants on the accuracy of the baseline during initial calibration.

[0071] Synchronization score W sync,i The calculation process is as follows: First, calculate the physiological rhythm signal S of sentinel plant i. i (t) and the baseline of population physiological rhythm S base (t) Normalized cross-correlation function R under different time delays τ i,base (τ):

[0072]

[0073] In the formula, E[·] represents the mathematical expectation; μ i and μ base Signal S i (t) and S base The mean of (t); σ i and σ base Signal S i (t) and S base The standard deviation of (t).

[0074] Then, within the preset time delay τ, the normalized cross-correlation function R is taken. i,base The maximum value of (τ) is taken as the final synchronization score W. sync,i :

[0075]

[0076] Result in a synchronization score W sync,i Subsequently, the central processing server 20 determines the dynamic decision weight W for each sentinel plant i according to the following rules. i :

[0077]

[0078] Where, θ sync It is a preset synchronization threshold (e.g., 0.6) stored in the central processing server 20.

[0079] If a sentinel plant scores W in synchronicity... symc,i Not less than the synchronization threshold θ symc If the physiological rhythm of the plant is synchronized with that of the population, the data is considered reliable, and its synchronicity score is directly used as its dynamic decision weight W. i If the synchronicity score W sync,i Below the synchronization threshold θ symc If the condition is abnormal, the plant is considered to be in an abnormal state (possibly due to local pests or diseases, sensor hardware failure, or loose installation). Its dynamic decision weight W i The weight is set to zero. Simultaneously, the central processing server 20 marks the sentinel plant as an anomalous sentinel plant and generates an alarm message sent to the user terminal, prompting an on-site inspection. By resetting the weight of the anomalous sentinel plant to zero, its subsequently collected data will not participate in the next weighted fusion calculation, thus achieving automatic removal of anomalous data.

[0080] See attached document Figure 4 After obtaining the dynamic decision weights of each sentinel plant, the method of the present invention further includes the steps of weighted fusion and decoding of the physiological state of the partition.

[0081] In one specific embodiment, this step is performed by the central processing server 20. For each management partition z, the central processing server 20 first constructs a comprehensive physiological feature vector V for that partition. z Specifically, for all dynamic decision weights W within partition z i Sentinel plants with a value greater than zero are used to extract the acoustic feature vector F from the second part. A,i and ion current eigenvector F E,i The data is spliced ​​together to form a combined physiological feature vector [F]. A,i ,F E,i Then, a weighted average is calculated for all these combined physiological feature vectors, using the following formula:

[0082]

[0083] in, W represents all dynamic decision weights within the management partition z. i A set of sentinel plants with values ​​greater than zero. This step yields the comprehensive physiological feature vector V. z It integrates physiological information from all trusted sentinel plants within the partition and weights them according to their synchronization with the population rhythm, enabling it to more accurately characterize the overall physiological state of the management partition.

[0084] Subsequently, the central processing server 20 will construct the comprehensive physiological feature vector V z The data is input into a pre-trained classification model M to decode the crop physiological state C of the current management zone. state The classification model M can be a support vector machine (SVM) model, a random forest model, or a deep neural network model. This model is pre-stored in the memory of the central processing server 20.

[0085] Training the classification model M is an offline process. Specifically, different known treatments are applied to sample wheat plants under controlled conditions, causing them to enter different physiological states, such as: ample water and fertilizer, mild water stress, moderate water stress, nitrogen deficiency, and potassium deficiency. During each stable physiological state, physiological signals are collected and combined physiological feature vectors are extracted. Each feature vector is labeled with its corresponding known physiological state code. The classification model M is trained using a training dataset composed of these state-labeled feature vectors until the model's classification accuracy reaches a preset standard.

[0086] In actual operation, the central processing server 20 will calculate the comprehensive physiological feature vector V in real time. zAs input, the trained classification model M is invoked for computation, and the model outputs a discrete state code C. state This output represents the crop's physiological state within the management zone. For example, the output state code can be defined as C0 representing a water and fertilizer abundance state, C1 representing a mild water stress state, C2 representing a moderate water stress state, C3 representing a nitrogen requirement state, and C4 representing a potassium requirement state. This output will serve as the direct basis for generating specific control instructions in the next step.

[0087] See attached document Figure 5 After decoding the crop physiological state of the management zone, the method of the present invention further includes the steps of generating and executing integrated water and fertilizer irrigation control instructions.

[0088] In one specific embodiment, this step is performed by the central processing server 20. The central processing server 20 internally stores a preset decision rule base. This decision rule base establishes a set of rules based on crop physiological state encoding C. state The mapping relationship to specific equipment control commands. When the crop physiological state code C from the fourth part is received... state Then, the central processing server 20 queries the decision rule base to generate the corresponding water and fertilizer integrated irrigation control instructions.

[0089] The mapping rules included in the decision rule base specifically include: if C state If the water and fertilizer sufficiency state is coded as C0, then no operation instruction is generated; if C state If the mild water stress state is coded as C1, a low-volume irrigation instruction is generated, which includes opening the target zone solenoid valve 32 for a first preset duration T. irr1 The control parameters; if C state If the moderate water stress state is coded as C2, a moderate irrigation instruction is generated, which includes opening the solenoid valve 32 for a second preset duration T. irr2 The control parameters, where T irr2 >T irr1 If C state If nitrogen demand state is coded as C3, a nitrogen fertilization instruction is generated, which includes opening solenoid valve 32 for a third preset duration T. irr3 And the fertilizer pump 33 connected to the nitrogen fertilizer source operates for a first fertilization duration T. fert1 The control parameters.

[0090] Furthermore, the control parameters defined in the decision rule base (such as irrigation duration T) irr1 T irr2The baseline value (etc.) is not fixed but can be dynamically adjusted based on external environmental parameters. For example, the central processing server 20 can also access meteorological data. When generating irrigation instructions, the baseline irrigation duration can be corrected based on the reference crop evapotranspiration (ET0) for the day, using the following formula: T′ irr =T irr ×(1+k·(ET 0,current -ET 0,avg ), where k is the correction factor, ET 0,current ET is the ET0 value for the day. 0,avg This is the historical average ET0 value for the same period. This measure allows irrigation amounts to better match actual water evaporation needs.

[0091] This method further incorporates specific processing logic for complex physiological states. When the crop physiological state encoding output by the classification model M is concurrent water stress and nutrient demand (e.g., simultaneously detecting moderate water stress and nitrogen demand), the control logic within the central processing server 20 will prioritize water stress. Specifically, during execution, the system first generates and issues only a moderate irrigation command (i.e., only including the opening duration T of the solenoid valve 32). irr2 (Parameters). After the irrigation command is executed, the system enters a waiting and re-evaluation cycle and does not immediately execute the fertilization operation. During this cycle, the system continuously collects and analyzes the physiological signals of the managed area. Only when the results of a new round of physiological state decoding indicate that the water stress state of the area has been relieved (e.g., the state code is updated to a water and fertilizer sufficiency state C0 or a simple nitrogen demand state C3) will the system decide whether to generate and execute a nitrogen fertilization command based on this new state code.

[0092] The central processing server 20 encapsulates the final generated control instructions into a data packet, which contains a unique identifier for the target management partition and specific control parameters. This data packet is sent via a wireless network to the partition water and fertilizer controller 31 corresponding to the target management partition. After receiving and parsing the data packet, the partition water and fertilizer controller 31 sends drive signals to the solenoid valve 32 and / or the fertilizer injection pump 33 through its output interface according to the control parameters therein, in order to complete the set irrigation or fertilization operation.

[0093] After a control command is executed, the entire system does not terminate but seamlessly enters the next monitoring and control cycle. The composite physiological sensor 11 continues to collect physiological signals from the sentinel plants and uploads the data to the central processing server 20 for a new round of feature extraction, state decoding, and decision-making. This closed-loop feedback mechanism ensures that water and fertilizer management measures can be adaptively adjusted according to the real-time dynamic changes in the crop's physiological state.

[0094] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A water and fertilizer integrated irrigation control method for wheat cultivation, characterized in that, Includes the following steps: S1. In one or more preset management zones, collect in real time the acoustic signal of xylem sap flow and the net rhizosphere ion flow rate of at least one sentinel plant in each management zone. S2. Determine the crop physiological state of the management zone based on the acoustic signal of the xylem sap flow and the net flow rate of rhizosphere ions; S3. Generate and execute water and fertilizer integrated irrigation control instructions for the management zone based on the crop physiological state.

2. The water and fertilizer integrated irrigation control method for wheat planting according to claim 1, characterized in that, The S2 step specifically includes: Acoustic feature vectors are extracted from the acoustic signal of the xylem fluid flow, and ion flow feature vectors are extracted from the net rhizosphere ion flow rate. The acoustic feature vector and the ion flow feature vector are combined to determine the physiological state of the crop.

3. The water and fertilizer integrated irrigation control method for wheat planting according to claim 2, characterized in that, The acoustic feature vector includes at least one of the following: total signal energy, energy percentage of embolism events, frequency centroid, and embolism event count; The ion flow characteristic vector includes at least one of the following: average absorption rate, peak absorption rate, and flow direction stability.

4. The water and fertilizer integrated irrigation control method for wheat planting according to claim 1, characterized in that, Also includes: Before step S2, the dynamic decision weight of each sentinel plant is calculated for each sentinel plant in each management partition. The step of determining the crop physiological state of the management zone in step S2 is based on the dynamic decision weights, which are used to weight the acoustic signals of xylem sap flow and the net rhizosphere ion flow rate of each sentinel plant.

5. The water and fertilizer integrated irrigation control method for wheat planting according to claim 4, characterized in that, The steps for calculating the dynamic decision weights of each sentinel plant include: Establish a baseline for population physiological rhythms; Calculate the synchronicity score between the physiological rhythm of each sentinel plant and the baseline of the population physiological rhythm; The dynamic decision weights are determined based on the synchronicity score.

6. The water and fertilizer integrated irrigation control method for wheat planting according to claim 5, characterized in that, The synchronicity score is obtained by calculating the normalized cross-correlation function between the physiological rhythm signal of the sentinel plant and the baseline physiological rhythm signal of the population.

7. The method for integrated water and fertilizer irrigation control in wheat cultivation according to claim 5 or 6, characterized in that, When the synchronization score of a sentinel plant is lower than the preset synchronization threshold, its dynamic decision weight is set to zero, and it is marked as an abnormal sentinel plant.

8. The water and fertilizer integrated irrigation control method for wheat planting according to claim 4, characterized in that, The specific steps for determining the crop physiological state of the management zone are as follows: Based on the dynamic decision weights, a weighted average is performed on the acoustic feature vectors and ion current feature vectors extracted from each sentinel plant to obtain the comprehensive physiological feature vector of the management zone; and, The comprehensive physiological feature vector is input into a preset classification model to output the physiological state of the crop.

9. The water and fertilizer integrated irrigation control method for wheat planting according to claim 1, characterized in that, In step S3, when the determined crop physiological state is a combination of water stress and nutrient demand, the integrated water and fertilizer irrigation control command prioritizes the execution of the irrigation command. After the water stress is relieved, the decision on whether to execute the fertilization command is made based on the new crop physiological state.

10. The method for integrated water and fertilizer irrigation control in wheat cultivation according to claim 1, characterized in that, In step S1, the acoustic signal of the xylem fluid flow is collected by a high-sensitivity piezoelectric acoustic probe placed on the stem of the sentinel plant. The net rhizosphere ion flow rate was measured by an ion-selective microelectrode array deployed in the rhizosphere of the sentinel plant.

Images