An intelligent control method and system for high-pressure atomizing irrigation and a storage medium thereof

Abstract

The application relates to an intelligent control method and system of high-pressure atomization irrigation and a storage medium thereof, and aims to solve the problem that traditional water and fertilizer management of high-value crops cannot realize real-time sensing of internal physiological states of crops, leading to inconsistent quality. The intelligent control method divides an irrigation area into multiple management partitions, randomly selects a plurality of sacrifice samples in each management partition, destructively samples the sacrifice samples, obtains true value data of crops through physiological and biochemical detection, establishes and dynamically updates an irrigation rule library based on the true value data, generates variable control instructions for each management partition in combination with real-time environmental data, and finally controls a high-pressure atomization sprinkler irrigation system to accurately execute. The application also provides a system and a storage medium for realizing the method, realizes accurate regulation and control starting from the real demand of crops, effectively improves the consistency of crop quality and the resource utilization efficiency, and has self-learning and predictive prevention and control capabilities.

Description

Technical Field

[0001] This application relates to the field of irrigation technology, and in particular to an intelligent control method, system and storage medium for high-pressure atomized irrigation. Background Technology

[0002] The cultivation of high-value crops requires a balance between traditional improvements and modern technological innovations. Relying on modern agricultural planting techniques ensures stable yields and crop quality. Their utilization value is diverse and efficient, which is conducive to their promotion and application.

[0003] Due to the differences between high-value crops and ordinary crops, high-value crops have higher requirements for growth conditions such as soil conditions, temperature, humidity, and light. This requires more refined technical management. For example, in the cultivation of Chinese medicinal herbs, it is necessary to accurately control the ratio of nitrogen, phosphorus, and potassium to ensure the content of effective ingredients. At the same time, it is also necessary to avoid problems such as root burn and reduction of effective ingredient content caused by excessive fertilization and watering.

[0004] In the cultivation of high-value crops, the quality of high-value crops is improved by scientifically regulating water and fertilizer. For example, in terms of water control, soil moisture sensors and crop canopy temperature sensors are used to monitor soil moisture content and crop water requirements in real time, and intelligent irrigation systems such as drip irrigation and sprinkler irrigation are used in conjunction with this. As for fertilizer control, soil nutrient content is first tested using soil testing and formula fertilization technology. Personalized fertilization plans are then customized based on the different needs of crops for nitrogen, phosphorus, potassium and trace elements at different growth stages (such as the root and stem enlargement stage of Chinese medicinal herbs and the flower bud differentiation stage of fruit trees). At the same time, efficient fertilization methods such as slow-release fertilizer, water-soluble fertilizer or foliar fertilizer are used to supply nutrients to crops.

[0005] However, in practical applications, the inventors discovered that even with the adoption of the aforementioned scientific crop cultivation methods and the implementation of intelligent control over processes such as crop fertilization and irrigation, the content of nutrients in mature crops obtained through the use of preferentially customized watering and fertilization lifecycle management methods varies significantly across different planting batches due to deviations in the genetic characteristics of the crop parent, the influence of the crop's later growth environment, and human interference factors during the actual planting stage of high-value crops. This affects product quality (for example, the nutrient content of high-value crops obtained through scientific cultivation does not deviate significantly from that of conventional crops), which is not conducive to improving crop quality. Summary of the Invention

[0006] The purpose of this invention is to improve the quality of the crops grown. This application provides an intelligent control method, system and storage medium for high-pressure atomized irrigation.

[0007] To achieve the above objectives, the intelligent control method, system, and storage medium for high-pressure atomized irrigation provided in this application adopt the following technical solution:

[0008] Firstly, this application discloses an intelligent control method for high-pressure atomized irrigation.

[0009] The irrigation area was divided into multiple management zones, and a number of crops were randomly selected in each management zone to be defined as sacrificial samples.

[0010] The sacrificial samples were destructively sampled according to a preset sampling sequence to obtain crop tissue samples. The crop tissue samples were then subjected to fragmented physiological and biochemical index detection to obtain at least one true value data characterizing the crop's water status, nutrient status, or health status.

[0011] Establish an irrigation rule base, define the mapping relationship between the range of true value data and irrigation control instructions through the established irrigation rule base, and obtain true value data based on the current sampling time series to generate or update personalized irrigation rules for each management partition;

[0012] Based on the latest personalized irrigation rules and combined with real-time environmental data obtained from environmental sensors, variable control instructions for watering, fertilization, drug administration or pest control are generated for each management zone.

[0013] Control the high-pressure atomized sprinkler system to execute the generated variable control commands.

[0014] Preferred methods for establishing an irrigation rule base include:

[0015] Combine the current sampled ground truth data with historical sampled data and historical irrigation operation records.

[0016] The corresponding environmental data are fused and standardized to construct a spatiotemporal correlated dataset;

[0017] An association rule learning algorithm is used to mine spatiotemporal association datasets and generate candidate rules with ground truth data and environmental data as conditions and irrigation control instructions as results.

[0018] Based on historical sampling data and historical irrigation operation records, the success probability of each candidate rule achieving the expected effect after execution is calculated, and the success probability is used as the confidence level of the association rule learning algorithm.

[0019] Candidate rules with confidence levels higher than a preset threshold are added to the irrigation rule base, and the applicable crop growth period and environmental conditions range are marked for each candidate rule.

[0020] Preferably, the variable fertilization control instructions include:

[0021] Based on the true data, the physiological and biochemical test results of the sacrificial sample are obtained, and the measured content values ​​of several target nutrients in the sacrificial sample are extracted.

[0022] The measured content values ​​of several target nutrients are compared one by one with the current optimal nutrient target range for the species and growth stage to obtain the surplus or deficit of each nutrient. The surplus or deficit includes missing items and surplus items.

[0023] Input the profit and loss ratio into the nutrient trade-off optimization model to obtain the optimal fertilizer formula and application rate;

[0024] Variable fertilization control instructions are generated based on the optimal fertilizer formula and application rate.

[0025] Preferably, the variable-rate insecticidal control instructions include:

[0026] The canopy multispectral images of crops are automatically acquired by image acquisition devices deployed in the management area. The canopy multispectral images are processed based on the target detection model to identify pest types and mark the coordinates of concentrated pest occurrence points in the images.

[0027] The coordinate information of the centralized points is integrated with the GIS geographic information of the management area to generate a digital map of pest distribution in the management area, and the pest types and severity levels with different coordinate information are marked on the digital map of pest distribution.

[0028] Based on the digital map of pest distribution, plan the operation path for the high-pressure mist irrigation system;

[0029] According to the operating path, the high-pressure atomizing irrigation system is controlled to turn on the nozzles when it is located at the current pest concentration point and turn off the nozzles when it is located at a non-pest concentration point.

[0030] Preferably, it also includes predictive pest control methods, specifically including:

[0031] A distributed network of microenvironment sensors is deployed within the management zone to continuously monitor and acquire microenvironmental data such as temperature, humidity, light intensity, and volatile organic compound concentration inside the crop canopy.

[0032] By correlating microenvironment data with historical pest occurrence records and using time series prediction models, the probability of pest occurrence and hotspot areas in a specific future time period are predicted.

[0033] When the predicted probability of pest outbreaks exceeds a preset risk threshold, a preventative variable application instruction is automatically generated.

[0034] Preventative pesticides were sprayed in predicted pest hotspots using a high-pressure atomizing irrigation system.

[0035] Preferably, after the high-pressure atomizing irrigation system executes the variable control command, it also includes:

[0036] After the variable control command is executed, the sacrificial sample is sampled again in the same management partition and the true value data of the sacrificial sample is obtained in the next sampling sequence to obtain the verified data;

[0037] The relative improvement rate η of the key indicators between the verified data and the true data before the execution of the variable control instructions is calculated using the following formula:

[0038] η=( - ) / ×100%

[0039] in, To verify the measured value of a certain indicator in the subsequent data, These are the measured values ​​of the corresponding indicators in the true data;

[0040] The calculated improvement rate η is compared with the preset expected improvement rate threshold in the personalized irrigation rules. , Compare;

[0041] If η < If the decision fails to achieve the expected results, the system will automatically impose a confidence penalty on the personalized irrigation rule that triggered the decision and mark it as a rule to be optimized.

[0042] like ≤η≤ If so, the confidence level of the personalized irrigation rule is maintained;

[0043] If η> If the decision-making effect is significant, the confidence level of the personalized irrigation rule will be improved.

[0044] For rules that need optimization, the system initiates a reinforcement learning process, storing the corresponding entire decision-making process data as a failure case in a specific dataset. This data drives the irrigation rule base to prioritize learning such cases when mining candidate rules in the next iteration.

[0045] Preferably, it also includes cross-cycle dynamic optimization strategies aimed at the ultimate quality goal, including:

[0046] Ultimate quality target setting: At the beginning of the crop growth cycle, set a clear ultimate quality target vector for the current planting batch. =[ , ,..., ],in , Quantitative indicators such as the concentration of specific functional components;

[0047] Reverse path planning: Utilizing a digital twin model of crop growth to vectorize the ultimate quality target By decomposing the data in reverse to each critical reproductive stage, the target range of intermediate physiological states that each reproductive stage i needs to reach is calculated. =[ , ];

[0048] Real-time tracking and correction: After sacrificial sampling is performed at each reproductive period i, the obtained ground truth data is compared with the target interval of the intermediate physiological state for that period. Perform comparisons and generate variable control instructions;

[0049] The objective function F of the variable control command is defined as minimizing the weighted sum of squares deviation between the current state and the target state, i.e.

[0050] F=

[0051] Where j represents different physiological and biochemical indicators. The importance weight of the corresponding indicator j;

[0052] The implementation of cross-cycle dynamic optimization strategies enables the crop growth trajectory to eventually converge to the ultimate quality target vector. .

[0053] Preferably, soil sampling and analysis are also required:

[0054] Soil samples were collected from the rhizosphere region of the sacrificial sample while the sacrificial sample was being destructively sampled.

[0055] Metagenomic sequencing analysis was performed on soil samples from the rhizosphere region to detect the abundance of functional genes of pathogens in the soil and to analyze the structure of the soil microbial community.

[0056] Establish a soil-borne disease occurrence risk index R, calculated using the following formula:

[0057]

[0058] in, For the abundance of genes of specific pathogens,

[0059] For the abundance of beneficial microbial genes,

[0060] These are stress factors calculated from soil moisture and temperature.

[0061] When the risk index R exceeds the safety threshold, it is determined that there is a high risk of soil-borne diseases in the management area, and the system generates a variable drug administration instruction.

[0062] The high-pressure atomizing sprinkler system is controlled to deliver functional water-soluble pesticides with soil disinfection capabilities.

[0063] Liquid is applied to the rhizosphere of crops to regulate the soil microecology.

[0064] Secondly, this application discloses a high-pressure atomizing irrigation system, which is applied to the first method.

[0065] The intelligent control method for high-pressure atomized irrigation described above includes:

[0066] The first module is used to randomly select a number of crops within each management zone and define them as sacrificial samples.

[0067] The second module is used to perform destructive sampling of the sacrificial sample according to a preset sampling sequence to obtain crop tissue samples, and to perform fragmented physiological and biochemical index detection on the crop tissue samples to obtain at least one true value data characterizing the crop's water status, nutrient status or health status.

[0068] The third module is used to establish an irrigation rule base. The established irrigation rule base defines the mapping relationship between the range of true value data and irrigation control instructions, and obtains true value data based on the current sampling time sequence to generate or update personalized irrigation rules for each management partition.

[0069] The fourth module is used to generate variable control instructions for each management zone based on the latest personalized irrigation rules and real-time environmental data obtained by environmental sensors, including variable watering, variable fertilization, variable drug administration, or variable pest control.

[0070] The fifth module is used to control the high-pressure atomizing irrigation system to execute the generated variable control commands.

[0071] Thirdly, this application discloses a storage medium storing data that can be loaded by a processor.

[0072] And execute the computer program of the intelligent control method for high-pressure atomized irrigation described in the first aspect above.

[0073] Compared with the prior art, the present invention provides an intelligent control method, system and storage medium for high-pressure atomized irrigation, which has the following beneficial effects:

[0074] 1. By introducing destructive sampling of "sacrifice samples" and obtaining true data, the transformation from "indirect inference based on external environment or crop phenotypic characteristics" to "direct perception based on the actual physiological and biochemical state of crops" has been realized. This can directly respond to the most essential water, fertilizer and pesticide needs of crops, thereby significantly improving the quality consistency and resource utilization efficiency of high-value crops.

[0075] 2. Through association rule mining and confidence assessment, the system can continuously learn from the successes and failures of historical decisions, constantly optimize and adjust control strategies, effectively overcome the problems of poor adaptability and easy obsolescence of traditional preset models, and ensure the accuracy and robustness of long-term use.

[0076] 3. Through a precise spatial matching mechanism, the overlap between the application area and the pest distribution area is increased to over 90%, while reducing the pesticide coverage area in non-target areas by 60%-75%, thus achieving simultaneous optimization of pesticide utilization and pest control rate. Attached Figure Description

[0077] Figure 1 This is a flowchart illustrating the execution steps of an intelligent control method for high-pressure atomized irrigation according to an embodiment of this application.

[0078] Figure 2 This is a flowchart illustrating the method for establishing an irrigation rule base in an intelligent control method for high-pressure atomized irrigation according to an embodiment of this application.

[0079] Figure 3 This is a flowchart of a variable fertilization control instruction method for an intelligent control method of high-pressure atomized irrigation according to an embodiment of this application.

[0080] Figure 4 This is a flowchart illustrating the method for generating and executing variable insecticidal control commands in an intelligent control method for high-pressure atomized irrigation according to an embodiment of this application.

[0081] Figure 5 This is a flowchart of a pest prediction and control method for a smart control method of high-pressure atomized irrigation according to an embodiment of this application. Detailed Implementation

[0082] The following is in conjunction with the appendix Figure 1-5 The technical solutions in this application are clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application. It should be noted that similar reference numerals and letters in the following drawings indicate similar items; therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0083] In traditional high-value crop cultivation systems, static water and fertilizer management strategies based on preset thresholds struggle to adapt to non-linear changes in crop growth, causing plant physiological and biochemical indicators to deviate from their optimal range. Fixed control models, reliant on historical experience data, cannot perceive the crop's actual needs in real time, hindering nutrient absorption efficiency and the synthesis pathways of effective components. Environmental parameters collected by sensor networks are decoupled from the actual physiological state of the crop, and traditional non-destructive testing methods, limited by the complexity of the canopy structure, struggle to obtain accurate metabolic data at the tissue level, resulting in water and fertilizer decisions lagging behind changes in the crop's actual needs.

[0084] For example, in a medicinal herb cultivation base, a monitoring system constructed using soil moisture sensors and canopy temperature probes, combined with pre-set drip irrigation control logic, is used for water and fertilizer management. Irrigation is triggered when soil moisture content reaches a set threshold, and standard formula water-soluble fertilizer is applied according to the growth stage. In actual operation, it was found that the content of ferulic acid, an effective component, fluctuated within ±30% during the rhizome enlargement stage of Angelica sinensis. Detection showed that the nitrate nitrogen metabolism rate differed by 2.8 times among different plants. Traditional systems failed to capture the differences in nutrient absorption efficiency among plants in a timely manner, and continuous application of homogenized fertilizers led to ion antagonism in some plants. The correlation between canopy temperature monitoring data and leaf stomatal conductance remained at only 0.65, failing to accurately reflect the degree of water stress and causing errors in irrigation decisions. Destructive testing is usually conducted after problems have emerged, resulting in a 14-21 day delay in corrective measures.

[0085] Among the aforementioned problems, the harm to high-value crops lies in the fact that their effective components cannot achieve a qualitative leap during maturity. When faced with improper management, crops will experience nutritional imbalances, leading to deviations in metabolic pathways and a decrease in the synthesis efficiency of target compounds. Furthermore, improper water management may induce abnormal accumulation of secondary metabolites. The superposition of such problems prevents the cultivated crops from highlighting their "high-value" characteristics.

[0086] Therefore, in the first aspect, this application discloses an intelligent control method for high-pressure atomized irrigation.

[0087] Reference Figure 1 A smart control method for high-pressure atomized irrigation includes the following steps:

[0088] S1. Divide the irrigation area into multiple management zones, and randomly select a number of crops in each management zone as sacrificial samples.

[0089] S2. Destructively sample the sacrificial sample according to the preset sampling sequence to obtain crop tissue samples, and perform fragmented physiological and biochemical index detection on the crop tissue samples to obtain at least one true value data characterizing the crop's water status, nutrient status or health status.

[0090] S3. Establish an irrigation rule base, define the mapping relationship between the true value data range and irrigation control instructions through the established irrigation rule base, and obtain true value data based on the current sampling time sequence to generate or update personalized irrigation rules for each management partition.

[0091] S4. Based on the latest personalized irrigation rules and combined with real-time environmental data obtained from environmental sensors, generate variable control instructions for variable watering, variable fertilization, variable drug administration, or variable pest control for each management zone.

[0092] S5. Control the high-pressure atomized sprinkler system to execute the generated variable control commands.

[0093] The “sacrifice sample” refers to a number of plants randomly selected within the management zone. The selected plants are in the same growth environment as other plants in the management zone and are random. The sacrificial plants are physically obtained through destructive sampling, and the true physiological and biochemical data of the crop tissue are directly obtained.

[0094] Meanwhile, the true data mentioned includes, but is not limited to, moisture status data, nutrient status data, and health and stress data.

[0095] As a preferred embodiment, the implementation method of the solution in this application is as follows:

[0096] First, the 100-mu irrigation area is divided into 10 management zones of 10 mu each;

[0097] Secondly, the sampling sequence can be flexibly adjusted according to the type of plant and its growth status, such as a time interval of three days, four days, or several hours. Plants in the current management zone are randomly selected as sacrificial samples, and the selected samples are placed in liquid nitrogen for quick-freezing and preservation.

[0098] Furthermore, the frozen leaf samples were crushed. A ball mill was used to grind the leaves into a fine powder, and indicators such as chlorophyll, soluble sugars, and nitrate reductase were extracted. The contents of these indicators were measured using instruments such as spectrophotometer and high-performance liquid chromatography to obtain true-value data characterizing the crop's water and nutrient status.

[0099] Next, an irrigation rule base is established. Machine learning algorithms, such as decision trees or random forests, are used to train a mapping model between the range of ground truth data and irrigation control commands. For example, when the relative leaf water content is below 75%, an instruction to increase irrigation is triggered; when nitrate reductase activity is below 0.5 μmol NO... 2- When the nitrogen fertilizer application rate is / g·h, an instruction to increase the nitrogen fertilizer application rate is triggered.

[0100] Therefore, the ground truth data obtained in the current sampling period is used to generate or update personalized irrigation rules for each management zone. If a significant change in the crop growth status of a management zone is detected, the irrigation rule parameters for that management zone are adjusted accordingly.

[0101] Specifically, by combining real-time environmental data acquired from temperature and humidity sensors, light sensors, and other sensors installed in each management zone, variable control instructions are generated based on the latest personalized irrigation rules. For example, when the soil moisture in a management zone is below a threshold and there is no rainfall, a variable watering instruction is generated for that management zone.

[0102] Finally, the high-pressure atomizing sprinkler system executes the generated variable control commands. Based on these commands, the system adjusts the opening time, water volume, and atomized particle size of the sprinklers in different management zones to achieve precision irrigation. Simultaneously, variable-rate fertilization is achieved by adding water-soluble fertilizers of different formulations to the irrigation water.

[0103] In some of the solutions mentioned above in this application, an irrigation method based on truth data and an irrigation rule base to generate variable control instructions was proposed. However, in the process of establishing the irrigation rule base, there are problems such as insufficient integration of historical sampling data and historical irrigation operation records with real-time environmental data, lack of dynamic adaptability in rule generation, and inability to quantify and verify the effect of rule execution.

[0104] In response, this application further proposes a method for establishing an irrigation rule base, referring to... Figure 2 This includes the following steps:

[0105] S31. Combine the current sampled true data with historical sampled data and historical irrigation operations.

[0106] The environmental data corresponding to the records are fused and standardized to construct a spatiotemporally correlated dataset;

[0107] S32. An association rule learning algorithm is used to mine the spatiotemporal association dataset to generate candidate rules with ground truth data and environmental data as conditions and irrigation control instructions as results.

[0108] S33. Calculate the success probability of each candidate rule achieving the expected effect after execution based on historical sampling data and historical irrigation operation records, and use the success probability as the confidence level of the association rule learning algorithm;

[0109] S34. Add candidate rules with confidence levels higher than the preset threshold to the irrigation rule base, and mark the applicable crop growth period and environmental conditions range for each candidate rule.

[0110] Specifically, in the stage of constructing the spatiotemporally correlated dataset, data cleaning and format conversion are used to eliminate the dimensional differences between different sensor data. For example, soil moisture percentage and temperature in degrees Celsius are uniformly converted into standardized Z-values. During candidate rule mining, feature cross-validation techniques are used to generate multi-dimensional condition combinations, such as combining light intensity ranges with air humidity thresholds to form composite condition terms. A time decay factor is introduced when calculating confidence levels; for example, historical cases from three months ago are assigned a weight of 0.8, while cases from within the past week are assigned a weight of 1.2, to enhance the timeliness of the irrigation rule base. In the candidate rule labeling stage, a decision tree model automatically identifies the boundary conditions under which rules take effect; for example, an irrigation rule becomes invalid when the diurnal temperature range exceeds 8°C. The rule base is continuously optimized through dynamically updated datasets; for example, newly added sampling data each week triggers incremental learning of the irrigation rule base, enabling irrigation control instructions to adapt to the gradual changes in crop growth stages and fluctuations in environmental parameters, thereby improving the stability of high-value crop nutrient content.

[0111] In the process of generating variable fertilization control instructions based on true data to achieve precision fertilization, relying solely on the mapping relationship between the true data range and irrigation instructions cannot dynamically analyze the degree of difference between the actual nutrient requirements of the crop and the target state. This results in the lack of a quantitative compensation mechanism for the surplus or deficit of different nutrient components in the generation of fertilization formulas, making it difficult to achieve synergistic optimization of multiple nutrient elements. Therefore, referring to... Figure 3 This application further proposes generating variable fertilization control instructions, including the following steps:

[0112] S41. Based on the true data, obtain the physiological and biochemical test results of the sacrificial sample, and extract the measured content values ​​of several target nutrients in the sacrificial sample.

[0113] S42. Compare the measured content values ​​of several target nutrients with the current optimal nutrient target range for the species and growth period one by one to obtain the surplus or deficit of each nutrient. The surplus or deficit includes missing items and surplus items.

[0114] S43. Input the profit and loss into the nutrient trade-off optimization model to obtain the optimal fertilizer formula and application rate;

[0115] S44. Generate variable fertilization control instructions based on the optimal fertilizer formula and application rate.

[0116] The measured content values ​​of target nutrients were obtained through fragmented physiological and biochemical testing, including but not limited to macronutrients such as nitrogen, phosphorus, and potassium, and micronutrients such as iron, zinc, and boron. The optimal nutrient target range was determined jointly using a crop variety database and a growth stage model. For example, during the root and stem enlargement stage of medicinal herbs, the target range for potassium was set at 2.3 mg / g-2.8 mg / g dry weight. The nutrient trade-off optimization model employed a multi-objective programming algorithm, where missing terms corresponded to compensation constraints, and excess terms corresponded to inhibition constraints. The objective function was set as a weighted combination of minimizing overall nutrient deviation and minimizing fertilizer cost. During fertilizer formulation generation, when the deficiency of a certain element exceeded a threshold, the compensation mechanism for that element was triggered first. For example, when the potassium deficiency exceeded the lower limit of the target range by 15%, a water-soluble fertilizer with a potassium content ≥45% was automatically matched as the base material.

[0117] Specifically, after obtaining crop tissue samples through destructive sampling, the physiological and biochemical test results are determined using inductively coupled plasma mass spectrometry (ICP-MS). When comparing the measured data with the optimal target range, an elemental deviation matrix is ​​established, where the relative deviation value for each element is calculated independently. For example, when the measured value of magnesium is 0.8 mg / g and the target range is 1.0 mg / g-1.2 mg / g, the missing amount is quantified as -20%. After receiving the surplus / deficit data, the nutrient trade-off optimization model first models elemental interactions. For example, the calcium-magnesium antagonism coefficient is set to 0.7 to ensure that the compensation scheme does not trigger elemental antagonism. When generating the fertilizer formula, if a phosphorus surplus and nitrogen deficiency are detected, a compound fertilizer with a nitrogen-to-phosphorus ratio ≥3:1 is automatically selected, and the application rate is calculated to ensure that the nitrogen supply reaches 110%-120% of the missing amount. When the final variable fertilization instruction is executed through a high-pressure atomization system, the fertilizer solution concentration is dynamically adjusted according to the application rate. For example, a nitrogen application rate of 5 kg per hectare corresponds to a nitrogen concentration of 0.5% in the atomized solution. This scheme establishes a multi-element independent compensation mechanism to ensure that fertilization decisions are precisely matched with the actual needs of crops.

[0118] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0119] Based on true data, the physiological and biochemical test results of the sacrificial sample were obtained, and the measured content values ​​of target nutrients such as nitrogen, phosphorus, potassium, calcium, and magnesium were extracted from the sacrificial sample. For example, the contents of potassium, calcium, and magnesium were determined by atomic absorption spectrophotometry, the nitrogen content was determined by the Kjeldahl method, and the phosphorus content was determined by the molybdenum blue colorimetric method.

[0120] The measured values ​​of target nutrients such as nitrogen, phosphorus, potassium, calcium, and magnesium are compared one by one with the optimal nutrient target range for the current crop type and growth stage. For example, for tomato plants during the flowering and fruiting stage, the measured leaf nitrogen content of 3.5% is compared with the optimal target range of 4.0%-5.0%, indicating a nitrogen deficiency of 0.5%-1.5%; the measured leaf phosphorus content of 0.6% is compared with the optimal target range of 0.3%-0.5%, indicating a phosphorus surplus of 0.1%-0.3%. In this way, the surplus or deficit of each nutrient is determined.

[0121] By inputting the profit and loss ratio into the nutrient trade-off optimization model, the optimal fertilizer formula and application rate are obtained. This model considers the antagonistic and synergistic effects between various nutrient elements, as well as the plant's absorption efficiency for different elements, and calculates the optimal formula that balances the needs of each element through a linear programming algorithm.

[0122] Variable fertilization control instructions are generated based on the optimal fertilizer formula and application rate. These instructions include information such as fertilizer type, ratio, application rate, and application time, and can be directly used to control the operation of intelligent fertilization equipment.

[0123] As a preferred embodiment, the specific implementation of the solution in this application is as follows:

[0124] The method for establishing an irrigation rule base includes the following steps:

[0125] First, the ground truth data obtained from current sampling is fused and standardized with historical sampling data and environmental data corresponding to historical irrigation operation records to construct a spatiotemporally correlated dataset. For example, environmental data such as soil moisture, temperature, and light intensity at different time points can be temporally aligned and spatially matched with ground truth data such as crop water content and chlorophyll content at the corresponding time points to form a unified data format.

[0126] Secondly, association rule learning algorithms are employed to mine the spatiotemporal association dataset, generating candidate rules with ground truth data and environmental data as condition terms and irrigation control instructions as result terms. For example, the Apriori algorithm or FP-Growth algorithm can be used to discover frequent itemsets in the data and generate association rules based on support and confidence thresholds.

[0127] Next, based on historical sampling data and historical irrigation operation records, the success probability of achieving the expected effect after each candidate rule is executed is calculated, and this success probability is used as the confidence score of the association rule learning algorithm. For example, the success probability can be calculated by counting the number of times each candidate rule is executed in historical sampling data and historical irrigation operation records, as well as the number of times crop physiological indicators improve after execution.

[0128] Finally, candidate rules with confidence scores higher than a preset threshold are added to the irrigation rule base, and each candidate rule is labeled with the applicable crop growth stage and environmental condition range. For example, rules with confidence scores greater than 0.8 can be included in the irrigation rule base, and based on the data characteristics used when generating the candidate rule, the candidate rule can be labeled as applicable to a specific growth stage and a specific temperature and humidity range for a certain crop.

[0129] Through the above technical solution, this application achieves dynamic optimization and adaptive updating of the irrigation rule base. By integrating multi-source heterogeneous data, the irrigation rule base can comprehensively reflect the complex relationship between crop growth and environmental factors. Furthermore, through verification and confidence mechanisms using historical sampling data and historical irrigation operation records, the reliability and effectiveness of candidate rules are guaranteed.

[0130] Furthermore, the marking of applicable conditions ensures that candidate rules are triggered in appropriate scenarios, thereby improving the accuracy of irrigation decisions.

[0131] Furthermore, refer to Figure 4 Generate variable-based pest control instructions, including:

[0132] S401. Automatically acquire crop canopy multispectral images through image acquisition devices deployed in the management area, process the canopy multispectral images based on the target detection model, identify pest types, and mark the coordinate information of concentrated pest occurrence points in the images;

[0133] S402. Integrate the coordinate information of the centralized points with the GIS geographic information of the management area to generate a digital map of pest distribution in the management area, and mark the pest types and severity levels with different coordinate information on the digital map of pest distribution.

[0134] S403. Based on the digital map of pest distribution, plan the operation path for the high-pressure atomized irrigation system;

[0135] S404. According to the operating path, control the high-pressure atomizing irrigation system to turn on the nozzles when it is located at the current pest concentration point, and turn off the nozzles when it is located at a non-pest concentration point.

[0136] Among them, the multispectral image acquisition device can be set as a camera with near-infrared band capture capability, such as using a sensor array with a wavelength range of 450nm-900nm, to enhance the recognizability of pest characteristics.

[0137] The target detection model can adopt a deep learning architecture based on convolutional neural networks. During training, the input includes a labeled dataset containing different pest types and severity levels, and the output layer can include pest category confidence and bounding box coordinate parameters.

[0138] During the generation of the digital map of pest distribution, the conversion accuracy of the geographic coordinate system can be controlled within ±0.5 meters to ensure that the nozzle positioning error does not exceed the allowable threshold for operation.

[0139] The sprinkler control logic uses a coordinate matching mechanism. When the irrigation system moves to the preset trigger range of the target point, such as a circular area with a radius of ±1.5 meters from the center point coordinates, the corresponding sprinkler group starts the application program.

[0140] Specifically, crop canopy multispectral images enhance the reflectance differences between insects and healthy tissues through specific wavelength combinations. For example, they can identify chlorophyll aberration regions in the 680nm band and detect insect egg aggregation characteristics in the 850nm band.

[0141] Furthermore, the target detection model performs pixel-by-pixel analysis on the image and outputs a pest heat map containing coordinate information, such as detecting pest spots with a diameter greater than 2 mm at a resolution of one square centimeter.

[0142] Correspondingly, in the GIS geographic information fusion stage, the pest coordinates are mapped to the geographic coordinate system of the management area to form a spatial distribution map with elevation data. For example, the UTM projection coordinate system is used to ensure the accuracy of planar distance calculation.

[0143] Meanwhile, the path planning module automatically generates zigzag or spiral traversal trajectories based on the pest distribution density. For example, continuous spraying mode is activated when the distance between pest points is less than 5 meters, and skipping mode is switched when the distance is greater than 10 meters. The nozzle opening and closing control is matched with preset coordinates through a real-time positioning system. When the equipment moves to within ±0.3 meters directly above the target point, the solenoid valve is triggered to open.

[0144] This solution uses a precise spatial matching mechanism to increase the overlap between the application area and the pest distribution area to over 90%, while reducing the pesticide coverage area in non-target areas by 60%-75%, thus achieving simultaneous optimization of pesticide utilization and pest control rate.

[0145] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0146] Multiple high-resolution multispectral cameras are deployed within the management zone. Each camera automatically captures a canopy image every two hours, with a resolution of 4000x3000 pixels. The acquired images are transmitted in real time to a central processing server via a wireless network.

[0147] A pre-trained deep learning object detection model, based on the YOLOv5 architecture and specifically trained for common crop pests, runs on the server. The model processes the received images, identifies common pest types, and marks the specific coordinates of the pest occurrences within the images.

[0148] The system integrates the identified pest coordinates with the GIS (Geographic Information System) data of the management zone. The GIS system pre-stores spatial data such as topography and soil type for the management zone. The resulting data forms a digital map of pest distribution, using a raster data structure with a spatial resolution of 0.5 meters. On the digital map, different colors represent different pest types, and the intensity of the color indicates the severity of the infestation.

[0149] Based on the generated digital map, the system plans the optimal operating path for the high-pressure atomizing irrigation equipment. The path planning employs an improved ant colony algorithm, with the optimization objective of covering all pest locations and minimizing the total travel distance. The planning results are output as a sequence of coordinate points, serving as navigation instructions for the equipment.

[0150] The high-pressure atomizing irrigation system operates along a planned path and is equipped with a GPS positioning module to obtain its current location coordinates in real time. When the GPS coordinates match the coordinates of a pest location, the control system sends an activation command to the corresponding nozzle, and the nozzle begins spraying pesticides. When leaving the pest location, the corresponding nozzle automatically shuts off. Sprinklers in non-pest areas remain closed at all times.

[0151] In some of the solutions mentioned above in this application, a technical means of generating variable pest control instructions based on a digital map of pest distribution is proposed. However, this solution can only passively kill pests that have already occurred, and cannot predict risk areas and carry out preventive interventions before pest outbreaks, resulting in a lag in actual control and potentially causing irreversible damage to crops.

[0152] In response, this application further proposes a predictive pest control method, referring to... Figure 5 Specifically, it includes:

[0153] S4011. Distribute a microenvironment sensor network within the management zone to continuously monitor and acquire microenvironment data such as temperature, humidity, light intensity, and volatile organic compound concentration inside the crop canopy.

[0154] S4012. Correlate microenvironment data with historical pest occurrence records, and use time series prediction models to predict the probability and hotspot areas of pest occurrence in a specific future time period.

[0155] S4013. When the predicted probability of pest occurrence exceeds the preset risk threshold, a preventive variable application instruction is automatically generated.

[0156] S4014. Spray preventative pesticide solutions in predicted pest hotspots using a high-pressure atomizing irrigation system.

[0157] The microenvironment sensor network was deployed at a density of 3-5 nodes per square meter within the crop canopy at a height of 0.5 to 1.2 meters. Monitoring parameters included a temperature range of 10℃-40℃, relative humidity of 30%RH-95%RH, and light intensity of 0 μmol·m⁻². -2 ·s -1 -1500 μmol·m -2 ·s -1 The concentration of volatile organic compounds is 0 ppb-500 ppb.

[0158] The time series forecasting model adopts a long short-term memory network architecture. The input layer receives the environmental data sequence of the past 72 hours, and the output layer generates the predicted value of the probability of pest occurrence in the next 24 hours. The prediction accuracy error is controlled within ±5%.

[0159] The risk threshold is set to trigger a prevention command when the probability value exceeds 65%, and the concentration of the spray solution is dynamically adjusted to 0.1%-0.5% emulsifier solution based on the predicted pest type.

[0160] Specifically, crop canopy microenvironment data is uploaded to the central processor every 5 minutes via a wireless transmission protocol and spatiotemporally matched with a stored pest history database. When the temperature is between 25℃ and 32℃ and the humidity remains above 80% for 6 consecutive hours, the system automatically marks it as a high-risk environmental combination.

[0161] The predictive model analyzes the rate of change of environmental parameters; for example, when the concentration of volatile organic compounds rises by more than 200 ppb within 3 hours, it determines that the incubation period of the pest has begun. High-pressure atomizing nozzles spray the pesticide solution at a pressure of 0.2 MPa in the predicted hotspot areas, with a spray coverage radius set to 1.5 meters to ensure complete saturation of the potential egg distribution area.

[0162] After a preventative pesticide application instruction is generated, the system automatically calls upon the corresponding irrigation equipment at the coordinates in the GIS map to complete pesticide coverage 24-48 hours before the pest outbreak. This predictive mechanism forms a dual protection system with the existing pest identification system. Even when the image recognition module does not detect pests, the predictive model can still initiate control measures in advance based on environmental anomalies, reducing crop damage rates to below 3%.

[0163] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0164] A distributed microenvironment sensor network is deployed within the management zone. A sensor node can be placed every 10 meters inside the crop canopy, with each node containing a temperature sensor, humidity sensor, light intensity sensor, and volatile organic compound (VOC) concentration sensor. The sensors collect data every 5 minutes and transmit it to the central control system via a wireless network.

[0165] The system continuously monitors and acquires microenvironmental data within the crop canopy. The central control system receives and stores data from all sensor nodes, forming a time-series database. The data includes temperature (°C), relative humidity (%), light intensity (lux), and volatile organic compound (Pb) concentration.

[0166] The system performs correlation analysis between microenvironment data and historical pest occurrence records. It retrieves pest occurrence records from the past three years, including pest type, occurrence time, and location. Machine learning algorithms, such as random forests or support vector machines, are used to establish a correlation model between microenvironment parameters and pest occurrence probabilities.

[0167] Time series forecasting models are used to predict the probability of pest outbreaks and hotspot areas within a specific future time period. Deep learning algorithms, such as Long Short-Term Memory (LSTM) networks, are employed to predict the probability of pest outbreaks in each region over the next seven days based on historical data and current microenvironmental parameters. Simultaneously, a pest risk heatmap is generated using spatial interpolation algorithms.

[0168] When the predicted probability of pest occurrence exceeds a preset risk threshold, a preventative pesticide application instruction is automatically generated. The risk threshold is set at 60%. If the probability of pest occurrence in a certain area exceeds this threshold within the next 7 days, the system automatically triggers the preventative pesticide application process. Based on different pest types and predicted probabilities, the system selects the appropriate pesticide type and concentration.

[0169] Preventative pesticides are sprayed in predicted pest hotspots using a high-pressure atomizing irrigation system. The system controls high-pressure atomizing nozzles to spray directionally in the predicted high-risk areas. During spraying, the system adjusts the nozzle angle and pressure in real time to ensure uniform coverage of the target area.

[0170] In some of the solutions mentioned above in this application, a method for generating variable control instructions based on personalized irrigation rules is proposed to achieve precision irrigation. However, during the rule execution process, due to factors such as crop growth dynamics, environmental disturbances, and insufficient initial experience of the rule base, the actual execution effect of some irrigation rules may deviate from expectations. In this regard, this application further proposes that after the high-pressure atomized irrigation system executes variable control instructions, it also includes a decision effectiveness quantification verification and adaptive learning stage.

[0171] After the variable control command is executed, the sacrificial sample is sampled again in the same management partition and the true value data of the sacrificial sample is obtained in the next sampling sequence to obtain the verified data;

[0172] The relative improvement rate η of the key indicators between the verified data and the true data before the execution of the variable control instructions is calculated using the following formula:

[0173] η=( - ) / ×100%

[0174] in, To verify the measured value of a certain indicator in the subsequent data, These are the measured values ​​of the corresponding indicators in the true data;

[0175] The calculated improvement rate η is compared with the preset expected improvement rate threshold in the personalized irrigation rules. , Compare;

[0176] If η < If the decision fails to achieve the expected results, the system will automatically impose a confidence penalty on the personalized irrigation rule that triggered the decision and mark it as a rule to be optimized.

[0177] like ≤η≤ If so, the confidence level of the personalized irrigation rule is maintained;

[0178] If η> If the decision-making effect is significant, the confidence level of the personalized irrigation rule will be improved.

[0179] For rules that need optimization, the system initiates a reinforcement learning process, storing the corresponding entire decision-making process data as a failure case in a specific dataset. This data drives the irrigation rule base to prioritize learning such cases when mining candidate rules in the next iteration.

[0180] In the data acquisition phase, a time-continuous sampling design is used to ensure the causal relationship between the validation data and the original decision. For example, sampling is performed at fixed time intervals after the execution of variable control instructions. Destructive sampling can be performed using whole-plant sampling or specific organ removal.

[0181] Specifically, in the verification phase after the execution of variable control instructions, continuous sampling within the same management zone avoids the interference of spatial heterogeneity on the evaluation results. True-value data obtained through destructive sampling can reflect the true changes in the internal physiological state of the crop; for example, water transport efficiency can be verified through root xylem sap composition analysis. The relative improvement rate calculation transforms multidimensional indicators into a single quantitative parameter; for example, the comprehensive improvement value is calculated by weighting nitrogen uptake rate and transpiration efficiency. The confidence penalty mechanism reduces the priority of inefficient rules in the decision tree; for example, rules with a confidence level below 60% are removed from the list applicable to the current growth stage. The reinforcement learning process optimizes the rule generation algorithm using failure cases; for example, negative sample weighting coefficients are added in association rule mining, enabling new rules to avoid historical erroneous decision-making patterns. Through a closed-loop feedback mechanism, the system can dynamically eliminate ineffective rules and strengthen efficient rules; for example, rules with a confidence level consistently above 80% for three consecutive growth cycles will be marked as preferred strategies.

[0182] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0183] After the high-pressure atomized irrigation system executes variable control commands, it performs quantitative verification of decision effectiveness and adaptive learning. First, in the next sampling time after the variable control command execution, sacrificial sampling and ground truth data acquisition are performed again on the same management zone to obtain verified data. Next, the relative improvement rate η between the verified data and the ground truth data before the variable control command execution is calculated for key indicators.

[0184] Furthermore, the calculated improvement rate η is compared with the preset expected improvement rate threshold in the personalized irrigation rules. , Comparisons can be made. Specifically, settings can be configured. 5%, It is 20%. When η At 5%, the decision is deemed to have failed to meet expectations. The system automatically applies a confidence penalty to the rule that triggered this decision, for example, reducing its confidence by 0.1 and marking it as a rule to be optimized. When 5% ≤ η ≤ 20%, the rule's confidence remains unchanged. When η At 20%, the decision-making effect is considered significant, and the confidence level of the rule is increased, for example, by increasing its confidence level by 0.1.

[0185] In some of the solutions mentioned above in this application, although the method of dynamically adjusting irrigation rules based on real-time sampling data achieves precise water and fertilizer management, the lack of a cross-growth period collaborative control mechanism based on the ultimate quality goal results in uncontrollable deviations in core indicators such as the concentration of key functional components between different planting batches.

[0186] In response, this application further proposes a cross-cycle dynamic optimization strategy oriented towards the ultimate quality goal, which includes three stages: setting the ultimate quality goal, reverse path planning, and real-time tracking and correction.

[0187] Among them, the ultimate quality target setting stage sets a clear quantitative indicator vector for the current planting batch at the beginning of the crop growth cycle. For example, in the planting of Chinese medicinal herbs, saponin content and polysaccharide concentration can be used as target vector elements.

[0188] In the reverse path planning stage, the ultimate goal is decomposed into each growth stage using a digital twin model. Specifically, the chlorophyll content range during the flowering stage can be set as [2.3 mg / g, 2.8 mg / g], and the soluble solids content range during the fruit enlargement stage can be set as [12%, 15%].

[0189] In the real-time tracking and correction stage, true data is obtained through destructive sampling. For example, the measured value of anthocyanin concentration during the fruit color-changing period is 1.2 mg / L. After comparing it with the target range [1.5 mg / L, 1.8 mg / L], a control instruction is generated.

[0190] Specifically, an ultimate quality target vector is set at the early stage of crop transplanting. For example, the target for saponin content in ginseng cultivation is set at 4.2% and polysaccharide content at 28%. Through the reverse transmission of the digital twin model, the target for maturity is decomposed into the root activity target [35μg / g·h-40μg / g·h] for the seedling stage and the leaf nitrogen content target [3.5%-4.0%] for the growth stage.

[0191] At each growth stage, the actual data obtained from destructive sampling is compared with the decomposed target. When the root activity of the seedling stage is detected to be only 30 μg / g·h, the system automatically calculates the weighted squared deviation, where the root activity weight coefficient is set to 0.6 and the leaf area index weight is set to 0.4.

[0192] Based on the deviation calculation results, variable control instructions for increasing phosphorus and potassium fertilizer application are generated, and the nutrient supply in the rhizosphere microzone is precisely adjusted through a high-pressure atomization system. This cross-cycle optimization mechanism ensures that local regulation at each growth stage always targets the ultimate quality goal. In ginseng cultivation practice, the fluctuation range of saponin content between different batches can be reduced from ±15% in traditional methods to within ±5%.

[0193] As a preferred embodiment, the solution of this application is specifically implemented as follows:

[0194] At the start of the crop growth cycle, set the ultimate quality target vector for the current planting batch. =[ , ,..., For example, for a certain Chinese medicinal herb, a setting can be made. =[Active ingredient A content 80mg / g, active ingredient B content 50mg / g, dry matter content 30%].

[0195] Using a digital twin model of crop growth, The process is decomposed in reverse to the critical growth stages. Assuming the crop has three critical growth stages, the target range of intermediate physiological states for each growth stage i is calculated using the model. .For example:

[0196] =[Chlorophyll content 40SPAD-50SPAD, plant height 20cm-25cm];

[0197] =[Root activity 500μg / (g·h)-600μg / (g·h), stem diameter 1.5cm-2cm];

[0198] =[Photosynthetic rate 15μmol / ( ·s)-20μmol / ( •s), leaf area index 4-5.

[0199] After sampling the sacrificial offerings during each reproductive period, obtain ground truth data and... Comparison was performed. Assume the sampling results during growth period 2 are: root activity 480 μg / (g·h), stem diameter 1.8 cm.

[0200] According to F= Construct the objective function F: F = in , The importance weights are root vitality and stem diameter, respectively.

[0201] Based on the calculation results of F, the system generates variable control instructions, such as increasing the application rate of root growth promoters and adjusting the water-fertilizer ratio. Through continuous real-time tracking and correction, the crop growth trajectory eventually converges to... .

[0202] In some of the solutions mentioned above in this application, a method for generating variable control instructions based on physiological and biochemical detection and environmental data of sacrificial samples was proposed to achieve precision irrigation and pest control. However, in this process, the risk of soil-borne diseases caused by abnormal soil microbial community structure was not considered. Traditional solutions rely only on the above-ground crop status and environmental data, which cannot identify potential diseases caused by soil pathogens, resulting in irreversible damage to crops due to root diseases and affecting the achievement of the ultimate quality goal.

[0203] In response, this application further proposes collecting soil samples from the rhizosphere region of the sacrificial sample while simultaneously performing destructive sampling; conducting metagenomic sequencing analysis on the rhizosphere soil samples to detect the abundance of functional genes of pathogens in the soil and analyze the soil microbial community structure; and establishing a soil-borne disease occurrence risk index R, calculated using the following formula:

[0204] ;

[0205] in, ( ) represents the gene abundance of a specific pathogen.

[0206] ( ) represents the abundance of beneficial microbial genes.

[0207] ( The stress factor is calculated from soil moisture and temperature.

[0208] When the risk index R exceeds the safety threshold, the management zone is determined to have a high risk of soil-borne diseases, and the system generates a variable drug administration instruction; the high-pressure atomized sprinkler irrigation system is controlled to apply a functional aqueous solution with soil disinfection function to the crop rhizosphere area to regulate the soil microecology.

[0209] The collection of rhizosphere soil samples and the destructive sampling of sacrificial samples were carried out simultaneously, ensuring spatiotemporal consistency between aboveground crop physiological data and underground microbial community data. Metagenomic sequencing analysis was performed using the Illumina NovaSeq platform for paired-end sequencing at a sequencing depth of 10M reads / sample. Species annotation was performed using the Kraken2 algorithm, and the gene abundance ratio of pathogens to beneficial microorganisms was calculated.

[0210] The safety threshold for the risk index R can be set in the range of 2.5-3.0. When soil moisture is higher than 85% of field capacity and the temperature is between 25-30℃, the stress factor... The value is 1.2-1.5. The functional aqueous solution may contain a compound preparation of 0.1% amino oligosaccharide and 0.05% Bacillus subtilis, which is formed into droplets with a particle size of 5-10 μm by high pressure atomization and penetrates into the rhizosphere soil.

[0211] Specifically, during the destructive sampling process, a rhizosphere soil sample with a radius of 15 cm and a depth of 20 cm was collected using a ring-shaped soil sampler, centered on the main root of the sacrificial sample, to ensure the spatial representativeness of the microbial community data.

[0212] After bioinformatics analysis of the metagenomic sequencing results, the sum of gene abundances of pathogens such as Fusarium and Phytophthora was used as the basis for... The sum of gene abundance of beneficial bacteria such as Trichoderma and Pseudomonas was used as the basis for... .

[0213] Stress factors Based on real-time data from soil moisture and temperature sensors, the stress factor is automatically adjusted to 1.3 when the average humidity is ≥80% and the temperature is ≥28℃ for three consecutive days. When the R value exceeds the threshold, the system triggers a precise application command, controlling the high-pressure atomizing nozzle to spray the pesticide solution at a pressure of 0.8MPa onto the target area. The application rate is dynamically adjusted according to the extent to which the R value exceeds the threshold. The dosage for application is When R=4.0, it increases to 80 ml / m 2This approach utilizes microbial activity detection and dynamic risk models to implement targeted interventions during the incubation period of soil-borne diseases. Compared to traditional periodic application methods, it reduces pesticide usage by 40%-60% while maintaining the abundance of beneficial soil microorganisms at no less than 85% of the initial value.

[0214] As a preferred embodiment, the specific implementation of the scheme in this application is as follows: while the crop selected as the sacrificial sample is subjected to destructive sampling, rhizosphere soil samples are collected within 0 mm to 5 mm from the surface of the main root system, and the sample weight is controlled between 50 g and 100 g. Metagenomic sequencing is performed using the Illumina NovaSeq sequencing platform.

[0215] By comparing with the KEGG database, the gene sequences of *Fusarium oxysporum* and *Rhizoctonia solani* in soil samples were identified, and their abundance percentages relative to the total microbial community were calculated. Simultaneously, the abundance of beneficial bacteria genes in *Trichoderma harzianum* and *Pseudomonas fluorescens* was detected. In the stress factor calculation module, when soil moisture exceeds 85% of field capacity for three consecutive days and the temperature is between 25℃ and 30℃, the stress factor value is set to 1.5. The threshold for the risk index R is set to 2.0. When the calculated value exceeds this threshold, the system automatically generates a mixing instruction containing 0.1% niclosamide and 0.05% *Bacillus subtilis* soluble powder. This powder is then sprayed at specific points within a 15 cm radius of the target plant's rhizosphere using a high-pressure atomizer, with a single application rate controlled at 200 ml / m³. .

[0216] Secondly, this application discloses a high-pressure atomizing irrigation system applied to an intelligent control method for high-pressure atomizing irrigation as described in the first aspect, comprising:

[0217] The first module is used to randomly select a number of crops within each management zone and define them as sacrificial samples.

[0218] The second module is used to perform destructive sampling of the sacrificial sample according to a preset sampling sequence to obtain crop tissue samples, and to perform fragmented physiological and biochemical index detection on the crop tissue samples to obtain at least one true value data characterizing the crop's water status, nutrient status or health status.

[0219] The third module is used to establish an irrigation rule base. The established irrigation rule base defines the mapping relationship between the range of true value data and irrigation control instructions, and obtains true value data based on the current sampling time sequence to generate or update personalized irrigation rules for each management partition.

[0220] The fourth module is used to generate variable control instructions for each management zone based on the latest personalized irrigation rules and real-time environmental data obtained by environmental sensors, including variable watering, variable fertilization, variable drug administration, or variable pest control.

[0221] The fifth module is used to control the high-pressure atomizing irrigation system to execute the generated variable control commands.

[0222] The first module randomly selects several crops within each management zone as sacrificial samples. The random selection algorithm combines spatial grid partitioning with Monte Carlo sampling to ensure that the samples cover areas with different light intensity gradients and soil moisture conditions.

[0223] The second module performs destructive sampling operations using a robotic arm. The sampling time is set before the crop's morning transpiration begins. After sampling, the crop tissue samples are flash-frozen in liquid nitrogen and then transferred to the detection chamber. Microwave-assisted disruption technology is used to achieve efficient cell wall lysis. The built-in spectral analysis unit in the detection chamber can simultaneously measure the physiological and biochemical indicators of the sacrificial samples.

[0224] The irrigation rule base in the third module adopts a dynamic update mechanism based on confidence. When the mapping relationship between the newly acquired true data and the existing rule base deviates by more than 15%, the rule reconstruction process is triggered.

[0225] The variable control instruction generation unit of the fourth module integrates a fuzzy logic controller, which quantifies the temperature, humidity, and light intensity parameters in the real-time environmental data into membership functions of 0-1, and performs weighted fusion with the basic instructions output by the irrigation rule base.

[0226] The fifth module communicates with the high-pressure atomizing nozzle array via the CAN bus protocol. Each nozzle is equipped with an independent solenoid valve and pressure sensor, enabling zoned control with a response accuracy of 0.1 seconds.

[0227] Through the above technical solution, this application achieves precise closed-loop management of high-value crops, solving the problem of control command lag caused by the low modularity of traditional irrigation systems. The synergistic effect of ground truth data obtained through destructive sampling and a dynamic rule base shortens the water and fertilizer regulation response time to within 2 hours, effectively maintaining the batch stability of crop effective component content. The data linkage mechanism between modules overcomes the interference of environmental variables, increasing the accuracy of irrigation decisions to over 93%, while reducing ineffective pesticide application by up to 40%, ensuring the uniformity of crop quality indicators.

[0228] Thirdly, this application discloses a storage medium storing a computer program that can be loaded by a processor and executed as an intelligent control method for high-pressure atomized irrigation.

[0229] The storage medium embeds the intelligent control method into an executable program, forming a stable platform for algorithm execution. This medium is compatible with various control method versions that incorporate dynamic selection of destructive sampling data, ground truth data fusion analysis, and self-learning mechanisms for the irrigation rule base. Examples include versions supporting integrated pest predictive control modules or cross-cycle dynamic optimization strategies. During program execution, each sampling data point, rule update record, and execution effect feedback is fully recorded, forming a traceable data processing chain and providing fundamental data support for subsequent algorithm iterations. By programmatically executing the irrigation decision-making process, continuous optimization of the irrigation rule base is achieved. For instance, after each rule confidence adjustment, a reinforcement learning process is automatically triggered to update the candidate rule set.

[0230] In a preferred embodiment, the storage medium uses an embedded flash memory chip with UFS 3.1 specification, which is connected to the ARM Cortex-A72 architecture processor via a PCIe interface. The computer program is stored in the medium as a binary executable file, containing a sequence of instructions for the intelligent control method of high-pressure atomized irrigation described in the first aspect. This sequence of instructions is divided into a data acquisition thread, a rule generation thread, and an execution control thread according to a preset priority. When the program is loaded into the processor's L3 cache, the destructive sampling timing table, truth data fusion algorithm parameters, and irrigation rule confidence thresholds are loaded into the working memory via memory mapping. During program execution, the storage medium continuously records the difference data packets before and after each rule base update. These difference data packets contain the logical expressions of newly added rules, the hash values ​​of eliminated rules, and statistical information on the number of rule triggers. If the program detects that the currently loaded control method version includes a pest predictive control module, it automatically calls the microenvironment sensor data stream parsing interface; if it detects that a cross-cycle dynamic optimization strategy is activated, it initiates the digital twin model parameter synchronization process.

[0231] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A smart control method for high-pressure atomized irrigation, characterized in that: The irrigation area was divided into multiple management zones, and a number of crops were randomly selected in each management zone to be defined as sacrificial samples. The sacrificial samples were destructively sampled according to a preset sampling sequence to obtain crop tissue samples. The crop tissue samples were then subjected to fragmented physiological and biochemical index detection to obtain at least one true value data characterizing the crop's water status, nutrient status, or health status. Establish an irrigation rule base, define the mapping relationship between the range of true value data and irrigation control instructions through the established irrigation rule base, and obtain true value data based on the current sampling time series to generate or update personalized irrigation rules for each management partition; Based on the latest personalized irrigation rules and combined with real-time environmental data obtained from environmental sensors, variable control instructions for watering, fertilization, drug administration or pest control are generated for each management zone. Control the high-pressure atomized sprinkler system to execute the generated variable control commands; It also includes cross-cycle dynamic optimization strategies aimed at achieving ultimate quality goals: Ultimate quality target setting: At the beginning of the crop growth cycle, set a clear ultimate quality target vector for the current planting batch. ,in , Quantitative indicators such as the concentration of specific functional components; Reverse path planning: Utilizing a digital twin model of crop growth to vectorize the ultimate quality target By decomposing the data in reverse to each critical reproductive stage, the target range of intermediate physiological states that each reproductive stage i needs to reach is calculated. ; Real-time tracking and correction: After sacrificial sampling is performed at each reproductive period i, the obtained ground truth data is compared with the target interval of the intermediate physiological state for that period. Perform comparisons and generate variable control instructions; The objective function F of the variable control command is defined as minimizing the weighted sum of squares deviation between the current state and the target state, i.e. ， in, Representing different physiological and biochemical indicators, For corresponding indicators Importance weights; The implementation of cross-cycle dynamic optimization strategies enables the crop growth trajectory to eventually converge to the ultimate quality target vector. .

2. The intelligent control method for high-pressure atomized irrigation according to claim 1, characterized in that, The method for establishing the irrigation rule base includes: Combine the current sampled ground truth data with historical sampled data and historical irrigation operation records. The corresponding environmental data are fused and standardized to construct a spatiotemporal correlated dataset; An association rule learning algorithm is used to mine spatiotemporal association datasets and generate candidate rules with ground truth data and environmental data as conditions and irrigation control instructions as results. Based on historical sampling data and historical irrigation operation records, the success probability of each candidate rule achieving the expected effect after execution is calculated, and the success probability is used as the confidence level of the association rule learning algorithm. Candidate rules with confidence levels higher than a preset threshold are added to the irrigation rule base, and the applicable crop growth period and environmental conditions range are marked for each candidate rule.

3. The intelligent control method for high-pressure atomized irrigation according to claim 2, characterized in that, The variable fertilization control instructions include: Based on the true data, the physiological and biochemical test results of the sacrificial sample are obtained, and the measured content values ​​of several target nutrients in the sacrificial sample are extracted. The measured content values ​​of several target nutrients are compared one by one with the optimal nutrient target range for the current crop type and growth stage to obtain the surplus or deficit of each nutrient. The surplus or deficit includes missing items and excess items. Input the profit and loss ratio into the nutrient trade-off optimization model to obtain the optimal fertilizer formula and application rate; Variable fertilization control instructions are generated based on the optimal fertilizer formula and application rate.

4. The intelligent control method for high-pressure atomized irrigation according to claim 1, characterized in that, The generated variable-based pest control command includes: The canopy multispectral images of crops are automatically acquired by image acquisition devices deployed in the management area. The canopy multispectral images are processed based on the target detection model to identify pest types and mark the coordinates of concentrated pest occurrence points in the images. The coordinate information of the centralized points is integrated with the GIS geographic information of the management area to generate a digital map of pest distribution in the management area, and the pest types and severity levels with different coordinate information are marked on the digital map of pest distribution. Based on the digital map of pest distribution, plan the operation path for the high-pressure mist irrigation system; According to the operating path, the high-pressure atomizing irrigation system is controlled to turn on the nozzles when it is located at the current pest concentration point and turn off the nozzles when it is located at a non-pest concentration point.

5. The intelligent control method for high-pressure atomized irrigation according to claim 4, characterized in that, It also includes predictive pest control methods, specifically including: A distributed network of microenvironment sensors is deployed within the management zone to continuously monitor and acquire microenvironmental data such as temperature, humidity, light intensity, and volatile organic compound concentration inside the crop canopy. By correlating microenvironment data with historical pest occurrence records and using time series prediction models, the probability of pest occurrence and hotspot areas in a specific future time period are predicted. When the predicted probability of pest occurrence exceeds the preset risk threshold, a preventative variable application instruction is automatically generated. Preventative pesticides were sprayed in predicted pest hotspots using a high-pressure atomizing irrigation system.

6. The intelligent control method for high-pressure atomized irrigation according to claim 1 is characterized in that, After the high-pressure atomizing irrigation system executes the variable control command, it also includes: After the variable control command is executed, the sacrificial sample is sampled again in the same management partition and the true value data of the sacrificial sample is obtained in the next sampling sequence to obtain the verified data; The relative improvement rate η of the key indicators between the verified data and the true data before the execution of the variable control instructions is calculated using the following formula: ， in, To verify the measured value of a certain indicator in the subsequent data, These are the measured values ​​of the corresponding indicators in the true data; The calculated improvement rate η is compared with the preset expected improvement rate threshold in the personalized irrigation rules. Compare; like If the decision fails to achieve the expected results, the system will automatically impose a confidence penalty on the personalized irrigation rule that triggered the decision and mark it as a rule to be optimized. like If so, the confidence level of the personalized irrigation rule is maintained; like If the decision-making effect is significant, the confidence level of the personalized irrigation rule will be improved. For rules that need optimization, the system initiates a reinforcement learning process, storing the corresponding entire decision-making process data as a failure case in a specific dataset. This data drives the irrigation rule base to prioritize learning such cases when mining candidate rules in the next iteration.

7. The intelligent control method for high-pressure atomized irrigation according to claim 5, characterized in that, Soil sampling and analysis are also required: Soil samples were collected from the rhizosphere region of the sacrificial sample while the sacrificial sample was being destructively sampled. Metagenomic sequencing analysis was performed on soil samples from the rhizosphere region to detect the abundance of functional genes of pathogens in the soil and to analyze the structure of the soil microbial community. Establish a soil-borne disease occurrence risk index R, calculated using the following formula: ， in, For the abundance of genes of specific pathogens, For the abundance of beneficial microbial genes, These are stress factors calculated from soil moisture and temperature. When the risk index R exceeds the safety threshold, it is determined that there is a high risk of soil-borne diseases in the management area, and the system generates a variable drug administration instruction. The high-pressure atomized sprinkler system is controlled to apply a functional aqueous solution with soil disinfection function to the rhizosphere of crops in order to regulate the soil microecology.

8. A high-pressure atomizing irrigation system, applied to the intelligent control method for high-pressure atomizing irrigation as described in any one of claims 1-7, characterized in that, include: The first module is used to randomly select a number of crops within each management zone and define them as sacrificial samples. The second module is used to perform destructive sampling of the sacrificial sample according to a preset sampling sequence to obtain crop tissue samples, and to perform fragmented physiological and biochemical index detection on the crop tissue samples to obtain at least one true value data characterizing the crop's water status, nutrient status or health status. The third module is used to establish an irrigation rule base. The established irrigation rule base defines the mapping relationship between the range of true value data and irrigation control instructions. Based on the current sampling time series, true value data is obtained to generate or update personalized irrigation rules for each management partition. The fourth module is used to generate variable control instructions for each management zone based on the latest personalized irrigation rules and real-time environmental data obtained by environmental sensors, including variable watering, variable fertilization, variable drug administration, or variable pest control. The fifth module is used to control the high-pressure atomizing irrigation system to execute the generated variable control commands.

9. A storage medium, characterized in that, The computer program contains a computer program that can be loaded by a processor and executed as described in any one of claims 1-7 for intelligent control of high-pressure atomized irrigation.

Images