Irrigation volume control method and system for prune cultivation

Abstract

The invention relates to an irrigation amount control method and system for prune cultivation. The method comprises the steps of performing timestamp calibration and dynamic weight distribution processing according to soil data, meteorological data and transpiration data to obtain a multi-modal data set; based on the multi-modal data set, microclimate abnormal event detection and soil humidity correction processing are carried out, and anti-interference soil water content data are obtained; according to the anti-interference soil moisture content data and preset prune growth stage parameters, carrying out root system layered water absorption dynamic modeling treatment to obtain a deep root system moisture demand prediction value; according to the deep root system moisture demand prediction value and the transpiration data, irrigation amount dynamic planning processing is carried out, and the target irrigation amount is obtained. According to the method, through multi-source data fusion, root system water absorption modeling processing and dynamic irrigation planning, the target irrigation amount of prune cultivation can be accurately determined, the accuracy and scientificity of irrigation management are remarkably improved, and powerful guarantee is provided for high-quality growth of prunes and efficient utilization of water resources.

Description

Technical Field

[0001] The present invention belongs to the technical field of agriculture, and particularly relates to a method and system for controlling irrigation amount in plum cultivation. Background Technique

[0002] In modern agricultural production, precise irrigation is crucial for improving crop yield and quality and saving water resources. As a fruit tree with high economic value, plums have different water requirements at different growth stages, and it is particularly important to reasonably control the irrigation amount. Traditional plum irrigation methods usually rely on empirical judgment or simple soil moisture detection for irrigation. For example, growers often operate according to fixed irrigation frequencies and water volumes during the germination period, flowering period, fruit expansion period, etc. of plums based on historical planting experience, or detect the surface soil moisture through soil moisture sensors and irrigate when the moisture is below a certain threshold. However, due to the large differences in climate and soil conditions in different regions, fixed irrigation strategies are difficult to adapt to the changing environment, easily causing insufficient or excessive irrigation, affecting plum growth and fruit quality. And making irrigation decisions only based on the surface soil moisture ignores the influence of deep soil moisture conditions, meteorological factors, etc. on soil water evaporation and plum transpiration, cannot meet the actual water requirements of plums, and may also lead to water resource waste. Summary of the Invention

[0003] Based on this, in view of the above technical problems, it is necessary to provide a method and system for controlling irrigation amount in plum cultivation to further improve the irrigation precision of plums and enhance the water resource utilization efficiency.

[0004] In the first aspect, the present application provides a method for controlling irrigation amount in plum cultivation, including:

[0005] Performing timestamp calibration and dynamic weight allocation processing on soil data, meteorological data, and transpiration data to obtain a time-synchronized multi-modal data set. The transpiration data includes plant transpiration rate and leaf surface temperature, the soil data includes soil moisture content and soil conductivity, and the meteorological data includes rainfall, light intensity, ambient temperature, and wind speed;

[0006] Based on the multi-modal data set, performing microclimate anomaly event detection and soil moisture correction processing to obtain anti-interference soil moisture content data;

[0007] According to the anti-interference soil moisture content data and preset plum growth stage parameters, performing deep root layer water absorption dynamic modeling processing to obtain a predicted value of deep root layer water demand. The preset plum growth stage parameters include root layer water absorption coefficients and water demand thresholds corresponding to the germination period, flowering period, and fruit expansion period;

[0008] Based on the predicted values of deep root water requirements and transpiration data, dynamic programming of irrigation amounts is carried out to obtain the target irrigation amount.

[0009] In one embodiment, based on soil data, meteorological data, and transpiration data, timestamp calibration and dynamic weight allocation processing are performed to obtain a time-synchronized multi-modal dataset, including:

[0010] Based on soil data, meteorological data, and transpiration data, the sampling frequency differences are calculated, and according to the sampling frequency differences, the timestamps are aligned through an interpolation algorithm to generate an initial time-axis dataset;

[0011] According to the pre-stored transmission delay time differences of each sensor and the sensor type coefficients, dynamic weight calculation processing is performed to obtain the dynamic weights of the soil data, meteorological data, and transpiration data;

[0012] Through the dynamic weights, weighted fusion processing is performed on the soil data, meteorological data, and transpiration data to generate an intermediate dataset;

[0013] According to the transmission delay time differences of each sensor, time offset reverse compensation processing is performed on the intermediate dataset to obtain a corrected dataset;

[0014] When the sliding window variance is greater than a preset threshold, the abnormal data points in the corrected dataset are corrected according to the median filtering algorithm to generate a multi-modal dataset.

[0015] In one embodiment, based on the multi-modal dataset, microclimate anomaly event detection and soil moisture correction processing are performed to obtain soil water content data with anti-interference, including:

[0016] According to the first rainfall amount and the second rainfall amount of adjacent weather stations in the meteorological data, the rainfall intensity gradient vector is calculated;

[0017] Based on the rainfall intensity gradient vector, the microclimate anomaly index is calculated by combining the wind speed and the ambient temperature through the following formula:

[0018]

[0019] Where, is the rainfall intensity gradient vector, α is the temperature influence coefficient, T base is the reference temperature, T env is the ambient temperature;

[0020] When the microclimate anomaly index exceeds the preset anomaly threshold, the soil moisture data is extracted from the multi-modal dataset;

[0021] Based on the Bayesian posterior probability model, soil moisture correction processing is performed on the soil moisture data, and verification is carried out by using the sliding window variance to obtain the corrected soil moisture data;

[0022] Combine the corrected soil moisture data with the transpiration data to generate soil water content data with anti-interference.

[0023] In one embodiment, based on the soil water content data with anti-interference and the preset plum growth stage parameters, perform dynamic modeling processing of root system layered water absorption to obtain the predicted value of deep root system water demand, including:

[0024] Perform layered processing on the soil water content data with anti-interference according to the preset soil layer parameters to obtain surface soil moisture and deep soil moisture;

[0025] Based on the preset plum growth stage parameters, combine with the Richards water migration equation to construct a dynamic model of root system layered water absorption, calculate the water absorption of surface roots and deep roots, and obtain the predicted value of root system water absorption:

[0026] According to the deviation between the plant transpiration rate in the transpiration data and the predicted value of root system water absorption, perform dynamic update processing of the root system coefficient through the LSTM network to obtain the updated root system layered water absorption coefficient;

[0027] When the deep soil moisture is less than the preset deep moisture threshold, increase the deep water absorption weight according to the preset linear compensation rule to obtain the corrected deep root system water absorption;

[0028] Calculate the predicted value of deep root system water demand according to the corrected deep root system water absorption and the water demand threshold.

[0029] In one embodiment, the calculation formula of the predicted value of root system water absorption is:

[0030]

[0031] Among them, β k and γ k are the pre-stored root system layered water absorption coefficients of the k-th layer, S h,min is the water demand threshold, is the soil moisture of the k-th layer.

[0032] In one embodiment, through the predicted value of deep root system water demand and the transpiration data, perform dynamic planning processing of irrigation amount to obtain the target irrigation amount, including:

[0033] Calculate the basic irrigation amount according to the predicted value of deep root system water demand and the plant transpiration rate in the transpiration data through the following formula:

[0034]

[0035] Among them, λ1 and λ2 are preset weight factors, is the surface soil moisture, Q demand is the predicted value of deep root water demand, T r is the plant transpiration rate;

[0036] According to the deviation between the leaf surface temperature and the ambient temperature, calculate the transpiration attenuation factor through the Sigmoid function;

[0037] Based on the transpiration attenuation factor, dynamically adjust the basic irrigation amount to obtain the corrected irrigation amount;

[0038] Obtain the meteorological prediction data of the irrigation area, and combine it with the corrected irrigation amount, and optimize it through the dynamic programming algorithm to obtain the target irrigation amount.

[0039] In one embodiment, the method further includes:

[0040] Generate an irrigation instruction based on the target irrigation amount, and the irrigation instruction is used to control the solenoid valve to perform an irrigation operation on the target area;

[0041] Obtain the irrigated soil data and crop growth data of the target area after irrigation, and combine the preset crop growth benchmark parameters and the target soil moisture to calculate the irrigation effect deviation index. The crop growth data includes the leaf surface temperature after irrigation and the stem flow rate after irrigation;

[0042] When the irrigation effect deviation index exceeds the irrigation threshold, it is determined that the irrigation effect does not meet the standard, and any one of the irrigation adjustment operations is performed to generate an updated solenoid valve control instruction. Among them, the irrigation adjustment operation includes generating a compensation irrigation amount according to the deviation between the irrigated soil data and the target soil moisture in proportion; when the conductivity in the irrigated soil data exceeds the safety threshold, a warning instruction is generated.

[0043] In a second aspect, the present application also provides an irrigation amount control system for Prunus domestica cultivation, including:

[0044] A multi-source data processing module, which is used to perform timestamp calibration and dynamic weight allocation processing according to soil data, meteorological data and transpiration data to obtain a time-synchronized multi-modal data set. The transpiration data includes plant transpiration rate and leaf surface temperature, the soil data includes soil moisture content and soil conductivity, and the meteorological data includes rainfall, light intensity, ambient temperature and wind speed;

[0045] A microclimate anomaly correction module, which is used to perform microclimate anomaly event detection and soil moisture correction processing based on the multi-modal data set to obtain anti-interference soil moisture content data;

[0046] A moisture demand prediction module, which is used to perform dynamic modeling processing of root layer water absorption according to the interference-resistant soil water content data and preset plum growth stage parameters, so as to obtain the predicted value of the deep root system water demand. The preset plum growth stage parameters include the root layer water absorption coefficients and water demand thresholds corresponding to the germination period, flowering period, and fruit swelling period.

[0047] A target irrigation control module, which is used to perform dynamic planning processing of the irrigation amount through the predicted value of the deep root system water demand and transpiration data to obtain the target irrigation amount.

[0048] In a third aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps in the first aspect are implemented.

[0049] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the first aspect are implemented.

[0050] For the above irrigation amount control method and system for plum cultivation, through timestamp calibration and dynamic weight allocation of soil data, meteorological data, and transpiration data, multi-source information reflecting soil conditions, meteorological changes, and crop physiological states is comprehensively integrated, providing a comprehensive and accurate data basis for subsequent plum irrigation decision-making. Secondly, through microclimate anomaly event detection and soil humidity correction based on a multi-modal data set, data fluctuations caused by factors such as meteorological mutations can be effectively avoided, and thus the true crop water demand situation can be more accurately reflected, providing a solid foundation for subsequent dynamic modeling of root water absorption. Moreover, this method constructs a dynamic model of root layer water absorption based on the interference-resistant soil water content data and preset plum growth stage parameters, accurately predicts the deep root system water demand, fully considers the differences in root distribution and water absorption characteristics of plums at different growth stages, and realizes accurate matching of plum growth requirements. Finally, this method performs dynamic planning of the irrigation amount through the predicted value of the deep root system water demand and transpiration data, and can adjust the irrigation amount in real time to ensure the scientific nature and timeliness of irrigation decision-making.

[0051] Compared with traditional plum irrigation methods, this method significantly improves the accuracy, scientific nature, and timeliness of plum irrigation amount control through multi-source data integration, environmental interference correction, root layer water absorption modeling, and dynamic planning adjustment, providing a strong technical guarantee for efficient plum cultivation and rational utilization of water resources. Description of the Drawings

[0052] To more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the following will briefly introduce the drawings required for use in the description of the embodiments or the related art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0053] Figure 1 Flowchart of an irrigation amount control method for plum cultivation provided for an exemplary embodiment of the present invention;

[0054] Figure 2 Schematic structural diagram of an irrigation amount control system for plum cultivation provided for an exemplary embodiment of the present invention. Detailed implementation manners

[0055] In order to make the objectives, technical solutions and advantages of the present application clearer, the following further details the present application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] In one embodiment, as Figure 1 shown, an irrigation amount control method for plum cultivation is provided. In this embodiment, the application of this method to a terminal is used as an example for illustration. It can be understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0057] S101: According to soil data, meteorological data and transpiration data, perform timestamp calibration and dynamic weight allocation processing to obtain a time-synchronized multi-modal data set. The transpiration data includes plant transpiration rate and leaf surface temperature, the soil data includes soil moisture content and soil conductivity, and the meteorological data includes rainfall, light intensity, ambient temperature and wind speed.

[0058] Specifically, in the actual prune planting environment, soil data, meteorological data, and transpiration data reflect the growth environment and physiological state of prunes from different perspectives. Among them, soil moisture content intuitively reflects the amount of water in the soil, and soil conductivity reflects the content of substances such as salts in the soil that affect the water absorption of prune roots. Precipitation in meteorological data directly increases soil moisture, while light intensity, ambient temperature, and wind speed affect water evaporation and the transpiration of prunes. In addition, plant transpiration rate and leaf surface temperature are important indicators of the water metabolism of prunes themselves. However, these data usually come from different sensors or monitoring devices, and there may be deviations in their acquisition times. For example, a soil moisture sensor may record data every 10 minutes, while the ambient temperature data of a weather station may be recorded every 5 minutes. The deviation in acquisition frequency will cause errors during data fusion and affect the accuracy of irrigation decisions. Therefore, time stamp calibration can be performed on these three types of data, that is, by aligning the time stamps of different data sources to a unified time reference to eliminate time deviation. And, in order to reflect the importance of different data in irrigation decisions, a dynamic weight allocation mechanism is also introduced in this embodiment. For example, in the dry season, the real-time state of soil moisture and the water demand of plants are more critical, that is, the weight of precipitation data can be reduced, while the weights of soil moisture content and transpiration rate are increased accordingly.

[0059] S102: Based on the multi-modal data set, perform microclimate anomaly event detection and soil moisture correction processing to obtain interference-resistant soil water content data.

[0060] Specifically, in the natural environment, the growth of prunes is affected by various microclimate factors, such as sudden heavy rain, strong wind, or extreme temperature changes. Schematically, a microclimate state model can be constructed based on meteorological data such as precipitation, light intensity, temperature, and wind speed to identify normal and abnormal microclimate states. An abnormal microclimate state may cause abnormal fluctuations in soil moisture data. For example, after heavy rain, soil moisture may rise sharply in a short period of time, while strong wind may cause the water on the soil surface to evaporate rapidly, resulting in distorted humidity data. Therefore, data such as changes in soil conductivity and abnormal fluctuations in plant transpiration rate in multi-modal data can be used, combined with a preset soil moisture correction model, to correct the soil moisture data, remove the interference caused by microclimate anomalies, and obtain interference-resistant soil water content data that can more truly reflect the actual water content of the soil and has practical guiding significance for the water supply of prune roots.

[0061] S103: According to the interference-resistant soil water content data and preset prune growth stage parameters, perform dynamic modeling of root layer water absorption to obtain the predicted value of deep root water demand. The preset prune growth stage parameters include the root layer water absorption coefficients and water demand thresholds corresponding to the germination period, flowering period, and fruit expansion period.

[0062] Specifically, the root system distribution and water absorption capacity of plums vary greatly at different growth stages. During the budding stage, the roots are relatively shallow, and the water absorption range is mainly concentrated in the shallow layer of the soil. However, as the growth progresses into the flowering stage and the fruit swelling stage, the roots gradually extend into the deep soil layer, and the water absorption capacity and demand of the deep roots also change continuously. Among the preset plum growth stage parameters, the root layer water absorption coefficients corresponding to the budding stage, flowering stage, and fruit swelling stage reflect the strength of the water absorption capacity of the roots at different soil depths in different growth stages. The water requirement threshold can be determined based on the physiological characteristics of plums at each growth stage and past planting experience, and it is the bottom line for ensuring the normal growth of plums in terms of water demand. According to the soil moisture data with anti-interference, combined with the preset growth stage parameters, models such as the Richards water migration equation are used to dynamically model the water absorption of roots at different soil depths. For example, by calculating the soil water gradient, root distribution density, and root water absorption coefficient of different soil layers, the water absorption amounts of the surface roots and deep roots can be simulated, thereby obtaining the predicted value of root water absorption. This predicted value can accurately reflect the actual water demand of the plum roots under the current soil moisture conditions and plum growth stage, thus providing a key data basis for determining the subsequent irrigation amount.

[0063] S104: Through the predicted value of the water demand of the deep roots and the transpiration data, perform dynamic planning processing of the irrigation amount to obtain the target irrigation amount.

[0064] Specifically, after obtaining the predicted value of the water demand of the deep roots, the overall water demand situation of plums can be further determined in combination with the transpiration data. The transpiration rate of plants reflects the current water consumption rate of plants, and the leaf surface temperature can indirectly reflect the water stress state of plants. For example, when the transpiration rate is high and the leaf surface temperature rises abnormally, it indicates that the plant may be in a water deficit state and needs to be replenished with water in a timely manner. Therefore, based on the predicted value of the water demand of the deep roots and the transpiration data, the optimal irrigation amount can be calculated through an optimization algorithm. For example, in dry weather with a high transpiration rate, the irrigation amount can be increased, while in periods with more rainfall, even if the water demand of the deep roots is high, the irrigation amount can be appropriately reduced to avoid wasting water resources. Through calculation, the target irrigation amount corresponding to plums can be obtained. This target irrigation amount not only takes into account the growth needs of plums but also takes into account environmental factors and the rational use of water resources, realizing the scientific and refined irrigation.

[0065] In the above irrigation volume control method for plum cultivation, through timestamp calibration and dynamic weight allocation processing, soil data, meteorological data, and transpiration data are multi-modally fused to obtain a time-synchronized multi-modal dataset, which makes up for the deficiency of insufficient utilization of single-modal data and provides a rich data basis for subsequent irrigation control decisions. Secondly, through the detection of microclimate abnormal events and soil moisture correction processing based on the multi-modal dataset, this method can not only improve the accuracy of irrigation decisions but also enhance the robustness against microclimate disturbances, avoiding irrigation decision errors caused by local abnormal weather events such as sudden heavy rain. Moreover, through the dynamic modeling of root-layer water absorption, according to the anti-interference soil water content data and the preset plum growth stage parameters, the water demand of deep roots is predicted, comprehensively considering the differential water absorption characteristics of plum roots at different growth stages, and can more accurately control the irrigation volume to avoid the adverse effects of over-irrigation or under-irrigation on plum growth. Finally, through the dynamic programming of irrigation volume based on the predicted value of deep root water demand and transpiration data, the irrigation volume can be dynamically adjusted according to the real-time water demand of plums and environmental conditions, improving the water resource utilization efficiency and further enhancing the scientificity and timeliness of irrigation decisions.

[0066] In an exemplary embodiment, according to soil data, meteorological data, and transpiration data, timestamp calibration and dynamic weight allocation processing are performed to obtain a time-synchronized multi-modal dataset, including:

[0067] Based on soil data, meteorological data, and transpiration data, the sampling frequency difference is calculated, and according to the sampling frequency difference, the timestamps are aligned through an interpolation algorithm to generate an initial time-axis dataset;

[0068] According to the pre-stored transmission delay time differences and sensor type coefficients of each sensor, dynamic weight calculation processing is performed to obtain the dynamic weights of soil data, meteorological data, and transpiration data;

[0069] Through the dynamic weights, weighted fusion processing is performed on soil data, meteorological data, and transpiration data to generate an intermediate dataset;

[0070] According to the transmission delay time differences of each sensor, time offset reverse compensation processing is performed on the intermediate dataset to obtain a corrected dataset;

[0071] When the sliding window variance is greater than a preset threshold, abnormal data points in the corrected dataset are corrected according to the median filtering algorithm to generate a multi-modal dataset.

[0072] Specifically, for devices such as soil moisture sensors, weather stations, and plant transpiration rate sensors, their data collection frequencies often vary. By calculating the sampling frequency differences of each data source and analyzing the sampling intervals of different sensors, the time unit to be aligned can be determined. Subsequently, an interpolation algorithm can be used for processing. The interpolation algorithm can estimate data values at other time points based on known data points. For example, for soil data with a lower sampling frequency, at the time nodes of meteorological data, methods such as linear interpolation and spline interpolation can be used to calculate the soil data values at corresponding moments, thereby generating an initial time-axis dataset so that all data can be analyzed at the same time scale. Moreover, the importance of different types of data in irrigation decision-making is not fixed, but changes with the growth stage of plums and environmental conditions.

[0073] In addition, there is a certain time delay from when the sensors collect data to when the data is transmitted to the system, and the delay times of different sensors are different. To further reflect this difference, the dynamic weights of each type of data can be calculated based on the pre-stored transmission delay time differences of each sensor and the sensor type coefficients. The sensor type coefficient reflects the importance degree of each type of data in different scenarios. For example, during a drought period, the data of plant transpiration rate and soil moisture content are more crucial for judging the water demand of plums, and their sensor type coefficients can be increased accordingly; while during frequent rainfall, the coefficient of rainfall data can be increased. Through these parameters, considering factors such as the current growth stage of plums and environmental conditions, the dynamic weights of soil data, meteorological data, and transpiration data can be calculated to quantify the relative importance of different data in irrigation decision-making. After obtaining the dynamic weights, the soil data, meteorological data, and transpiration data can be weighted and fused to generate an intermediate dataset.

[0074] Specifically, due to the time delay in data transmission by each sensor, there is a deviation between the timestamp of the data in the intermediate dataset and the actual occurrence time. Therefore, in order to restore the true time sequence of the data, the intermediate dataset can be processed by time offset reverse compensation according to the pre-stored time delay differences of each sensor. For example, if the transmission delay of a certain soil sensor is 2 minutes, the data timestamp of this sensor can be adjusted forward by 2 minutes to make the time information of the data consistent with the actual acquisition time, obtaining a corrected dataset to ensure that subsequent analysis is based on accurate time logic. However, there may still be abnormal data points in the corrected dataset due to factors such as sensor failures and environmental interference. Therefore, the stability of the data can be further judged by calculating the variance of the sliding window. The sliding window refers to a continuous data interval selected from the time series data. When the variance of the data within the window is greater than the preset threshold, it indicates that the data within this interval fluctuates abnormally and there may be incorrect data. Subsequently, the median filtering algorithm can be used to correct the abnormal data points, that is, the value of the abnormal data point is replaced with the median of the data within the sliding window, which can effectively suppress the influence of noise and outliers and retain the true change trend of the data, and finally generate a high-quality multi-modal dataset to provide reliable data support for subsequent microclimate anomaly detection, soil moisture correction, and irrigation volume planning, etc.

[0075] In an exemplary embodiment, based on the multi-modal dataset, microclimate anomaly event detection and soil moisture correction processing are performed to obtain interference-resistant soil water content data, including:

[0076] Calculate the rainfall intensity gradient vector according to the first rainfall and the second rainfall of adjacent weather stations in the meteorological data;

[0077] Based on the rainfall intensity gradient vector, calculate the microclimate anomaly index by combining the wind speed and the ambient temperature through the following formula:

[0078]

[0079] Where, is the rainfall intensity gradient vector, α is the temperature influence coefficient, T base is the reference temperature, T env is the ambient temperature;

[0080] When the microclimate anomaly index exceeds the preset anomaly threshold, extract the soil moisture data from the multi-modal dataset;

[0081] Based on the Bayesian posterior probability model, perform soil moisture correction processing on the soil moisture data and verify it by using the variance of the sliding window to obtain the corrected soil moisture data;

[0082] Combine the corrected soil moisture data with the transpiration data to generate soil water content data resistant to interference.

[0083] Specifically, in the farmland environment, the spatial distribution of rainfall is not uniform, and the rainfall difference between adjacent areas can reflect the diffusion trend and intensity change of rainfall. The first rainfall and the second rainfall of adjacent weather stations in the meteorological data can be obtained, and combined with the spatial coordinate information of each weather station, the rainfall intensity gradient vector can be calculated. This vector contains the change rate of rainfall in the horizontal direction, and its magnitude and direction can intuitively show the increasing and decreasing trend of rainfall intensity in space. For example, if the numerical value of the rainfall intensity gradient vector in a certain direction is large, it means that the rainfall changes violently in that direction, and there may be a situation where a strong rainfall belt moves. In the natural environment, wind speed and temperature have important effects on the diffusion, evaporation of rainfall and the stability of microclimate. For example, strong wind weather will accelerate the diffusion of rainfall and enhance water evaporation at the same time; high temperature environment will exacerbate the loss of soil moisture and change the humidity conditions of the microclimate. Therefore, after obtaining the rainfall intensity gradient vector, the microclimate anomaly index can be calculated by the above formula in combination with wind speed and environmental temperature data. Among them, the temperature influence coefficient and the reference temperature in the formula can be determined according to the climate characteristics and historical meteorological data of the target area. Through calculation, the microclimate anomaly index can be obtained. This index can comprehensively reflect the degree to which the current microclimate state deviates from the normal situation. The larger the index value, the more serious the microclimate anomaly situation.

[0084] Schematically, when the calculated microclimate anomaly index exceeds the preset anomaly threshold, it indicates that the current microclimate state is abnormal, and this anomaly is very likely to have a significant impact on soil moisture. At this time, it is necessary to correct the soil moisture data in the multimodal dataset. Among them, the Bayesian posterior probability model is a data analysis method based on the principles of probability theory and statistics, which can combine prior knowledge and newly obtained data to calculate the posterior probability of an event occurring. During the process of soil moisture correction, historical soil moisture data, meteorological data, and relevant environmental parameters can be used as prior knowledge to construct a Bayesian model. After extracting the soil moisture data under the current abnormal state, the model can update the prior probability based on this data, calculate the posterior probability distribution of soil moisture under the current microclimate anomaly conditions, and thus correct the soil moisture data. For example, after a heavy rainfall anomaly event occurs, the model can comprehensively consider factors such as rainfall amount, rainfall duration, soil type, etc., and combine the variation law of soil moisture in similar situations in historical data to reasonably adjust the measured soil moisture data to make it closer to the actual soil water content. And after the correction is completed, the sliding window variance can be used to verify the corrected soil moisture data, that is, by calculating the degree of dispersion of data within a certain time window to judge the stability of the data. If the variance is within a reasonable range, it indicates that the corrected data has a high credibility; if the variance is too large, it indicates that there are still abnormal fluctuations in the data and further inspection and correction are required.

[0085] Specifically, after being corrected by the Bayesian posterior probability model and verified by the sliding window variance, relatively reliable corrected soil moisture data can be obtained. Subsequently, integrating the corrected soil moisture data with the transpiration data can comprehensively present the water information of the plum growth environment from two perspectives of soil supply and plant demand, and finally generate anti-interference soil water content data. This data excludes the interference caused by microclimate anomalies and provides accurate and reliable basic data support for subsequent root water absorption modeling and irrigation volume planning, which can ensure that irrigation decisions can accurately match the actual water demand of plums.

[0086] In an exemplary embodiment, based on the anti-interference soil water content data and the preset plum growth stage parameters, dynamic modeling of root layer-by-layer water absorption is performed to obtain the predicted value of deep root water demand, including:

[0087] Perform layer-by-layer processing on the anti-interference soil water content data according to the preset soil layer parameters to obtain the surface soil moisture and deep soil moisture;

[0088] Based on the preset plum growth stage parameters, combined with the Richards water migration equation, construct a dynamic model of root layer-by-layer water absorption, calculate the surface root water absorption amount and deep root water absorption amount, and obtain the predicted value of root water absorption amount:

[0089] According to the deviation between the plant transpiration rate and the predicted value of root water absorption in the transpiration data, the root coefficient is dynamically updated through an LSTM network to obtain the updated root layer water absorption coefficient.

[0090] When the soil humidity in the deep layer is less than the preset deep humidity threshold, the deep water absorption weight is increased according to the preset linear compensation rule to obtain the corrected deep root water absorption.

[0091] According to the corrected deep root water absorption and the water demand threshold, the predicted value of the deep root water demand is calculated.

[0092] Specifically, there are significant differences in the water distribution of the soil in the vertical direction, and the water supply capacities of soil layers at different depths to the roots of Prunus domestica are different. Therefore, according to the preset soil layer parameters, such as dividing 0 - 30 cm into the surface layer and 30 - 100 cm into the deep layer, the soil water content data resistant to interference can be calculated layer by layer. Schematically, the surface soil humidity and the deep soil humidity can be extracted respectively by methods such as integration or weighted average. For example, the surface soil is more directly affected by environmental factors such as rainfall and evaporation, and the water content changes faster; while the deep soil water is relatively stable. Through this layer-by-layer processing, the water states of different soil layers can be analyzed independently, providing basic data for subsequent root water absorption modeling. And during different growth stages of Prunus domestica, such as the budding stage, flowering stage, and fruit swelling stage, there are significant differences in the root distribution density, activity, and water demand. Therefore, based on the preset Prunus domestica growth stage parameters, combined with the Richards water migration equation, considering the influence of factors such as soil texture, water content, and pore structure on water conduction, a dynamic model of root layer water absorption is constructed. In this model, the root layer water absorption coefficient reflects the relative water absorption capacities of the surface and deep roots at different growth stages. For example, during the budding stage, the shallow roots develop faster and the surface root water absorption coefficient is relatively high; while during the fruit swelling stage, to meet the large water demand of the plant, the deep root water absorption coefficient can be increased accordingly. By substituting parameters such as the soil layer humidity and the root layer water absorption coefficient into the equation, the surface root water absorption and the deep root water absorption can be calculated respectively, and then the predicted value of the overall root water absorption can be obtained. This predicted value reflects the theoretically water absorption amount of the Prunus domestica roots under the current soil water conditions and growth stage.

[0093] In an exemplary embodiment, the calculation formula for the predicted value of root water absorption is:

[0094]

[0095] where β k and γ k are the pre-stored root layer water absorption coefficients of the kth layer, S h,min is the water demand threshold, is the soil moisture of the k-th layer.

[0096] In addition, during the actual planting process, the water absorption capacity of the roots of Prunus domestica will be dynamically affected by environmental factors and its own physiological state, and it is difficult for the pre-stored root layer water absorption coefficients to fully cover all changing situations. Therefore, a long short-term memory network can be introduced for dynamic updating of the root coefficients. This long short-term memory network is suitable for processing time series data and can learn the long-term dependencies in the data. Schematically, the plant transpiration rate in the transpiration data can be used as a key indicator reflecting the actual water demand of Prunus domestica, and compared with the predicted value of the root water absorption amount, and the deviation between the two is calculated. This deviation reflects the difference between the model prediction value and the actual situation. Information such as historical deviation data, soil moisture data, and meteorological data can be input into this network, and by training to learn the rules in the data, the root layer water absorption coefficients can be automatically adjusted. For example, when it is continuously monitored that the transpiration rate is higher than the predicted value, it indicates that the actual water demand is greater than the model prediction. This network can appropriately increase the root water absorption coefficient to make the model more in line with the actual situation and enhance the accuracy and adaptability of the prediction.

[0097] Specifically, the deep soil is an important water source reserve for Prunus domestica under drought conditions. When it is detected that the deep soil moisture is less than the preset deep moisture threshold, it indicates that the deep soil is in a water shortage state, which may cause stress to the growth of Prunus domestica. Therefore, the deep water absorption weight can be increased according to the preset linear compensation rule. This rule can be set based on the physiological characteristics of Prunus domestica and historical experimental data. For example, when the deep soil moisture decreases by a certain proportion, the deep root water absorption weight is increased by a fixed multiple, so as to preferentially meet the water demand of the deep roots. Through this process, the root water absorption distribution strategy can be dynamically adjusted, the survival ability of Prunus domestica in the drought environment can be enhanced, and the corrected deep root water absorption amount can be obtained. Subsequently, in combination with the preset water demand threshold, considering factors such as the variety characteristics, growth stage, and target yield of Prunus domestica, the predicted value of the deep root water demand can be calculated. And by comparing and analyzing the corrected deep root water absorption amount with the water demand threshold, the current water deficit degree of the deep roots can be quantified, so as to determine the amount of water that needs to be supplemented. This predicted value provides a direct basis for the dynamic planning of the subsequent irrigation amount, ensuring that the irrigation decision can accurately meet the water demand of the deep roots of Prunus domestica, avoiding water resource waste or insufficient irrigation, and realizing efficient water-saving precision irrigation management.

[0098] In an exemplary embodiment, through the predicted value of the deep root water demand and the transpiration data, dynamic planning processing of the irrigation amount is performed to obtain the target irrigation amount, including:

[0099] According to the predicted value of the deep root water demand and the plant transpiration rate in the transpiration data, the basic irrigation amount is calculated through the following formula:

[0100]

[0101] where λ1 and λ2 are preset weight factors, is the surface soil moisture, Q demand is the predicted value of the water demand of deep roots, T r is the plant transpiration rate;

[0102] According to the deviation between the leaf surface temperature and the ambient temperature, the transpiration attenuation factor is calculated through the Sigmoid function;

[0103] Based on the transpiration attenuation factor, the basic irrigation amount is dynamically adjusted to obtain the corrected irrigation amount;

[0104] Obtain the meteorological prediction data of the irrigation area, and combine it with the corrected irrigation amount, and optimize it through the dynamic programming algorithm to obtain the target irrigation amount.

[0105] Specifically, the predicted value of the water demand of deep roots reflects the water demand of the roots of Prunus domestica L. in the current soil conditions and growth stages to obtain water from the deep soil. The plant transpiration rate reflects the speed at which Prunus domestica L. loses water through leaf transpiration. Based on the predicted value of the water demand of deep roots and the plant transpiration rate, the surface soil moisture and preset weight factors can be introduced and comprehensively considered through the above formula to calculate the basic irrigation amount. Among them, the preset weight factor is a parameter preset according to factors such as the growth stage of Prunus domestica L. and soil type, and is used to adjust the importance of different data in the calculation of the irrigation amount. For example, during the fruit swelling period of Prunus domestica L., it is more sensitive to water demand. At this time, the weight factor of the predicted value of the water demand of deep roots can be appropriately increased; while in the clay area with strong water retention capacity of the soil, the weight of the surface soil moisture in the calculation can be reduced. The calculated basic irrigation amount initially balances the water demand of the roots of Prunus domestica L. and the water consumption caused by transpiration, providing a benchmark value for subsequent more refined adjustment of the irrigation amount. And the deviation between the leaf surface temperature and the ambient temperature is an important indicator to measure the degree of influence of the environment on the transpiration of Prunus domestica L. When the leaf surface temperature is higher than the ambient temperature, it indicates that the transpiration of Prunus domestica L. is stronger and the water loss is faster; on the contrary, the transpiration is weaker. The Sigmoid function is a mathematical function commonly used to map data between 0 and 1, with a smooth curve characteristic, and can well simulate the non-linear relationship between environmental factors and transpiration.

[0106] Schematically, based on the deviation between the leaf surface temperature and the ambient temperature, it can be substituted into the Sigmoid function for calculation to obtain the transpiration attenuation factor. The closer the value of this factor is to 1, the stronger the promoting effect of the environmental conditions on the transpiration of plums, and the faster the water loss of plums; the closer the value is to 0, the stronger the inhibitory effect of the environment on transpiration, and the relatively slower the water loss. Through this transpiration attenuation factor, the influence degree of the current environmental conditions on the transpiration of plums can be quantified, providing a quantitative basis for the dynamic adjustment of the irrigation amount.

[0107] Specifically, this transpiration attenuation factor can dynamically adjust the basic irrigation amount. For example, when the transpiration attenuation factor is large, it means that the plums lose water quickly in the current environment and more water needs to be supplemented, so the basic irrigation amount can be increased accordingly; conversely, when the transpiration attenuation factor is small, the basic irrigation amount can be appropriately reduced. By operating the basic irrigation amount with the transpiration attenuation factor, the corrected irrigation amount is obtained. This corrected irrigation amount fully considers the influence of the current environmental factors on the transpiration of plums. Compared with the basic irrigation amount, it better fits the water demand of plums in the actual environment, making the determination of the irrigation amount more scientific and reasonable. And in order to make the irrigation decision more forward-looking, meteorological prediction data of the irrigation area can be obtained, that is, information such as rainfall, light intensity, temperature, wind speed, etc. in the next period of time. This meteorological prediction data can reflect the influence of future environmental changes on soil water evaporation, rainfall replenishment, and the transpiration of plums.

[0108] Specifically, by combining the corrected irrigation amount with the meteorological prediction data, the dynamic programming algorithm can be used for optimization. Among them, the dynamic programming algorithm is a method that decomposes the original problem into relatively simple sub-problems and obtains the optimal solution of the original problem by solving the optimal solutions of the sub-problems. In the process of optimizing the irrigation amount, this algorithm can comprehensively consider the changes in meteorological conditions at different future time points and the water demand of plums at each time point to adjust the corrected irrigation amount. For example, if the meteorological prediction shows that there will be rainfall in the next few days, the current irrigation amount can be appropriately reduced to avoid waste caused by excessive water; if it is predicted that there will be continuous high temperature and drought in the future, the irrigation amount can be increased to meet the water demand of plums. Through the optimization of this dynamic programming algorithm, the target irrigation amount is finally obtained. This target irrigation amount not only considers the current physiological needs and environmental status of plums, but also prospectively combines future meteorological changes, realizing the precise dynamic planning of the irrigation amount, and can maximize the utilization efficiency of water resources while ensuring the growth of plums.

[0109] In an exemplary embodiment, the method further includes:

[0110] Based on the target irrigation amount, an irrigation instruction is generated, and the irrigation instruction is used to control the solenoid valve to perform an irrigation operation on the target area;

[0111] Obtain the irrigated soil data and crop growth data of the target area after irrigation, and calculate the irrigation effect deviation index in combination with the preset crop growth benchmark parameters and the target soil humidity. The crop growth data includes the leaf surface temperature after irrigation and the stem flow rate after irrigation;

[0112] When the irrigation effect deviation index exceeds the irrigation threshold, it is determined that the irrigation effect does not meet the standard, and any one of the irrigation adjustment operations is executed to generate an updated solenoid valve control instruction. Among them, the irrigation adjustment operation includes generating a compensatory irrigation amount according to the deviation between the irrigated soil data and the target soil humidity in proportion; when the conductivity in the irrigated soil data exceeds the safety threshold, a warning instruction is generated.

[0113] Specifically, the target irrigation amount can be converted into an executable irrigation instruction to control the opening and closing of the solenoid valve. As the core execution component of the irrigation system, after receiving the instruction, the solenoid valve can accurately adjust the water flow rate and irrigation duration according to the target irrigation amount. For example, when the target irrigation amount is 10 cubic meters, the opening time of the solenoid valve can be calculated based on the sprinkler flow rate parameter of the irrigation area to ensure that 10 cubic meters of water is provided for the target area during this period. And after the irrigation is completed, the irrigated soil data and crop growth data of the target area can be collected in a timely manner. The irrigated soil data and crop growth data intuitively reflect the actual effect after irrigation from two levels of soil moisture status and crop physiological response. Among them, the soil moisture content directly reflects the water replenishment situation of the soil after irrigation. By comparing it with the preset target soil humidity, it can be judged whether the soil moisture reaches the ideal state; the soil conductivity can reflect the soil salt status, but too high conductivity may affect the water absorption of Prunus domestica L. roots. The leaf surface temperature can reflect the water stress degree of Prunus domestica L. If the irrigation is insufficient, the leaves may heat up due to water shortage; the stem flow rate is a key index to measure the water transport speed in plants, and its numerical change can intuitively reflect the water absorption and utilization efficiency of crops. This data can be comprehensively compared with the preset crop growth benchmark parameters such as the ideal leaf surface temperature range and stem flow rate standard value at different growth stages and the target soil humidity, and the irrigation effect deviation index can be calculated through a specific algorithm. This index can reflect the deviation degree between the irrigation effect and the ideal state. The larger the value, the greater the gap between the irrigation effect and the expected target. When the irrigation effect deviation index exceeds the preset irrigation threshold, it can be determined that the current irrigation effect does not meet the standard, and then the irrigation adjustment operation is executed.

[0114] Schematically, if the irrigation effect deviation index mainly stems from the soil moisture not reaching the target humidity, the compensatory irrigation volume can be calculated according to the deviation between the irrigated soil data and the target soil humidity at a preset ratio. For example, if the target soil humidity is 25%, the measured value after irrigation is only 20%, and the preset ratio is to supplement 5 cubic meters of water for every 1% deviation, then the compensatory irrigation volume can be calculated as 25 cubic meters, and an updated solenoid valve control instruction can be generated for secondary irrigation until the soil humidity approaches the target value. When the conductivity in the irrigated soil data exceeds the safety threshold, it indicates that the soil salinity may be too high, posing a risk to the growth of Prunus domestica. At this time, a warning instruction can be immediately generated and sent to the management personnel by means of text messages, APP push, or pop-up windows on the management platform, etc., and specific information such as the specific value of the soil conductivity and the degree of exceeding the standard is provided synchronously. The management personnel can take targeted measures according to this instruction, such as leaching and desalting or adjusting the irrigation water quality, to prevent damage to Prunus domestica caused by soil salinization.

[0115] Based on the same inventive concept, as Figure 2 shown, an irrigation volume control system 200 for Prunus domestica cultivation is further provided in an embodiment of the present application. The implementation solution provided by this system to solve the problem is similar to the implementation solution described in the above method. Therefore, the specific limitations in one or more embodiments of the irrigation volume control system for Prunus domestica cultivation provided below can refer to the limitations on the irrigation volume control method for Prunus domestica cultivation in the above text, and will not be repeated here. The system includes:

[0116] A multi-source data processing module 201, configured to perform timestamp calibration and dynamic weight allocation processing according to soil data, meteorological data, and transpiration data to obtain a time-synchronized multi-modal data set. The transpiration data includes plant transpiration rate and leaf surface temperature, the soil data includes soil water content and soil conductivity, and the meteorological data includes rainfall, light intensity, ambient temperature, and wind speed;

[0117] A microclimate anomaly correction module 202, configured to perform microclimate anomaly event detection and soil humidity correction processing based on the multi-modal data set to obtain anti-interference soil water content data;

[0118] A water demand prediction module 203, configured to perform dynamic modeling processing of root layer-by-layer water absorption according to the anti-interference soil water content data and preset Prunus domestica growth stage parameters to obtain a predicted value of the deep root water demand. The preset Prunus domestica growth stage parameters include the root layer-by-layer water absorption coefficients and water demand thresholds corresponding to the germination period, flowering period, and fruit swelling period;

[0119] A target irrigation control module 204, configured to perform dynamic planning processing of the irrigation volume through the predicted value of the deep root water demand and the transpiration data to obtain the target irrigation volume.

[0120] In the above system, the multi-source data processing module 201 processes the soil data, meteorological data, and transpiration data through timestamp calibration and dynamic weight allocation to obtain a time-synchronized multi-modal data set, avoiding the limitations of a single data source in traditional irrigation methods, and thus providing a rich data basis for subsequent irrigation decisions. The microclimate anomaly correction module 202 then performs microclimate anomaly event detection and soil moisture correction processing based on the multi-modal data set, which can effectively identify and correct abnormal interferences in the soil moisture data, effectively solving the problem of inaccurate irrigation caused by data noise or abnormal events, and thus providing more reliable soil moisture information for subsequent irrigation decisions. The water demand prediction module 203 performs dynamic modeling processing of root layer-by-layer water absorption according to the interference-resistant soil water content data and the preset plum growth stage parameters, and can accurately predict the water demand of deep roots to obtain the predicted value of the water demand of deep roots. The target irrigation control module 204 can then perform dynamic planning processing of the irrigation amount through the predicted value of the water demand of deep roots and the transpiration data, and adjust the irrigation amount in real time to make the irrigation amount fit the real-time water demand situation of the plum, further ensuring the scientificity and timeliness of the irrigation decision.

[0121] In an exemplary embodiment, the present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of an irrigation amount control method for plum cultivation according to the present application are implemented. A multi-core processor is preferably used to improve the parallel processing ability of the system. Memory: Provide sufficient temporary storage space to support the operation of the program and the processing of data. The memory capacity should be large enough to accommodate a large amount of supply information and computing tasks.

[0122] In an exemplary embodiment, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of an irrigation amount control method for plum cultivation according to the present application are implemented. The computer-readable storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), solid-state drive (SSD, Solid State Drives), or optical disc, etc. Among them, the random access memory may include resistive random access memory (ReRAM, Resistance Random Access Memory) and dynamic random access memory (DRAM, Dynamic Random Access Memory).

[0123] The above-described embodiments merely represent several implementation manners of the embodiments of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent for the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the embodiments of the present application.

Claims

1. A method for controlling the irrigation amount in plum cultivation, characterized in that The method includes: Performing timestamp calibration and dynamic weight allocation processing based on soil data, meteorological data, and transpiration data to obtain a time-synchronized multimodal dataset. The transpiration data includes plant transpiration rate and leaf surface temperature. The soil data includes soil water content and soil conductivity. The meteorological data includes rainfall, light intensity, ambient temperature, and wind speed; Based on the multimodal dataset, performing microclimate anomaly event detection and soil moisture correction processing to obtain soil water content data with anti-interference; According to the soil water content data with anti-interference and preset plum growth stage parameters, performing dynamic modeling processing of root layer-by-layer water absorption to obtain a predicted value of deep root water demand. The preset plum growth stage parameters include root layer-by-layer water absorption coefficients and water demand thresholds corresponding to the germination period, flowering period, and fruit swelling period; Through the predicted value of deep root water demand and the transpiration data, performing dynamic planning processing of irrigation amount to obtain the target irrigation amount.

2. The method according to claim 1, wherein The performing timestamp calibration and dynamic weight allocation processing based on soil data, meteorological data, and transpiration data to obtain a time-synchronized multimodal dataset includes: Based on the soil data, the meteorological data, and the transpiration data, calculating the sampling frequency difference, and according to the sampling frequency difference, aligning timestamps through an interpolation algorithm to generate an initial time axis dataset; According to the pre-stored transmission delay time differences between sensors and sensor type coefficients, performing dynamic weight calculation processing to obtain the dynamic weights of the soil data, the meteorological data, and the transpiration data; Through the dynamic weights, performing weighted fusion processing on the soil data, the meteorological data, and the transpiration data to generate an intermediate dataset; According to the transmission delay time differences between sensors, performing time offset reverse compensation processing on the intermediate dataset to obtain a corrected dataset; When the variance of the sliding window is greater than a preset threshold, correcting abnormal data points in the corrected dataset according to the median filtering algorithm to generate the multimodal dataset.

3. The method according to claim 1, wherein The performing microclimate anomaly event detection and soil moisture correction processing based on the multimodal dataset to obtain soil water content data with anti-interference includes: Calculating a rainfall intensity gradient vector according to the first rainfall and the second rainfall of adjacent weather stations in the meteorological data; Based on the rainfall intensity gradient vector, calculating a microclimate anomaly index by combining the wind speed and the ambient temperature through the following formula: Among them, is the rainfall intensity gradient vector, α is the temperature influence coefficient, and T base is the reference temperature, and T env is the ambient temperature; When the microclimate anomaly index exceeds a preset anomaly threshold, extracting soil moisture data from the multimodal dataset; Based on the Bayesian posterior probability model, performing soil moisture correction processing on the soil moisture data and verifying it by using the variance of the sliding window to obtain the corrected soil moisture data; Combining the corrected soil moisture data with the transpiration data to generate soil water content data with anti-interference.

4. The method according to claim 1, wherein The performing dynamic modeling processing of root layer-by-layer water absorption according to the soil water content data with anti-interference and preset plum growth stage parameters to obtain a predicted value of deep root water demand includes: Stratify the interference-resistant soil moisture data according to the preset soil stratification parameters to obtain surface soil moisture and deep soil moisture; Based on the preset plum growth stage parameters, combined with the Richards water migration equation, construct a dynamic model of root layer-by-layer water absorption, calculate the water absorption of surface roots and deep roots, and obtain the predicted value of root water absorption: According to the deviation between the plant transpiration rate in the transpiration data and the predicted value of root water absorption, perform dynamic update processing of the root coefficient through the LSTM network to obtain the updated root layer-by-layer water absorption coefficient; When the deep soil moisture is less than the preset deep moisture threshold, increase the deep water absorption weight according to the preset linear compensation rule to obtain the corrected deep root water absorption; Calculate the predicted value of the deep root water demand according to the corrected deep root water absorption and the water demand threshold; 5. The method according to claim 4, characterized in that, The calculation formula for the predicted value of root water absorption is: where β k and γ k are the pre-stored root system water absorption coefficients of the k-th layer, S h,min is the water requirement threshold, is the soil moisture of the k-th layer.

6. The method according to claim 1, characterized in that, Through the predicted value of the deep root water demand and the transpiration data, perform dynamic planning processing of the irrigation amount to obtain the target irrigation amount, including: Calculate the basic irrigation amount according to the predicted value of the deep root water demand and the plant transpiration rate in the transpiration data through the following formula: where λ1 and λ2 are preset weight factors, is the surface soil moisture, Q demand is the predicted value of the deep root water demand, T r is the transpiration rate of the plant; Calculate the transpiration attenuation factor through the Sigmoid function according to the deviation between the leaf surface temperature and the ambient temperature; Based on the transpiration attenuation factor, dynamically adjust the basic irrigation amount to obtain the corrected irrigation amount; Obtain the meteorological prediction data of the irrigation area, and combine it with the corrected irrigation amount, and optimize it through the dynamic programming algorithm to obtain the target irrigation amount; 7. The method according to claim 1, wherein The method further includes: Generate an irrigation instruction based on the target irrigation amount, and the irrigation instruction is used to control the solenoid valve to perform irrigation operations on the target area; Obtain the irrigation soil data and crop growth data of the target area after irrigation, and combine the preset crop growth benchmark parameters and the target soil humidity to calculate the irrigation effect deviation index. The crop growth data includes the leaf surface temperature after irrigation and the stem flow rate after irrigation; When the irrigation effect deviation index exceeds the irrigation threshold, it is determined that the irrigation effect does not meet the standard, and any one of the irrigation adjustment operations is performed to generate an updated solenoid valve control instruction. Among them, the irrigation adjustment operation includes generating a compensation irrigation amount according to the deviation between the irrigation soil data and the target soil humidity in proportion; when the conductivity in the irrigation soil data exceeds the safety threshold, a warning instruction is generated; 8. An irrigation volume control system for Prunus domestica cultivation, characterized in that, The system includes: A multi-source data processing module, which is used to perform timestamp calibration and dynamic weight allocation processing on soil data, meteorological data and transpiration data to obtain a time-synchronized multi-modal data set. The transpiration data includes plant transpiration rate and leaf surface temperature, the soil data includes soil moisture content and soil conductivity, and the meteorological data includes rainfall, light intensity, ambient temperature and wind speed; A microclimate anomaly correction module, which is used to detect microclimate anomaly events and correct soil moisture based on the multi-modal data set to obtain interference-resistant soil moisture data; A moisture demand prediction module, configured to perform dynamic modeling processing on the root layer water absorption according to the anti-interference soil water content data and preset plum growth stage parameters, so as to obtain a predicted value of the deep root system water demand, wherein the preset plum growth stage parameters include the root layer water absorption coefficients and water demand thresholds corresponding to the germination period, flowering period, and fruit expansion period; A target irrigation control module, configured to perform dynamic planning processing on the irrigation amount through the predicted value of the deep root system water demand and the transpiration data, so as to obtain a target irrigation amount.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images