Intelligent orchard soil moisture content real-time monitoring and early warning system

Abstract

The invention discloses an intelligent orchard soil moisture content real-time monitoring and early warning system. A multi-depth sensor node network is deployed through a low-power wide area network architecture, real-time soil moisture content data covering a main distribution layer of a fruit tree root system is periodically collected, and industrial information processing is performed on the soil moisture content data, synchronous meteorological data and soil attribute data to generate a standard data set; the data set is combined with the fruit tree growth stage, the soil texture and future weather forecast to be input into a pre-trained water demand dynamic evaluation model to output a dynamic water shortage level signal, and finally a precise regulation and control instruction is generated according to the signal and an irrigation strategy rule; the problems that multi-point interpolation or single depth monitoring deviates from actual root system moisture distribution, power consumption and transmission stability are difficult to balance, an early warning model is not combined with growth stages and meteorological information, so that accuracy is low, and monitoring and irrigation decision linkage timeliness is poor are solved. The accuracy of orchard soil moisture content monitoring and early warning and the timeliness of irrigation regulation and control are improved.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture monitoring and early warning technology, and in particular to a smart orchard soil moisture real-time monitoring and early warning system. Background Technology

[0002] With the rapid development of smart agriculture, precision management of orchards places higher demands on soil moisture monitoring. Current technologies for soil moisture data acquisition mainly rely on multi-point interpolation or single depth sensors, leading to discrepancies between monitoring results and the actual water distribution within the fruit tree root system. Furthermore, balancing power consumption and real-time data transmission stability is difficult during high-frequency continuous monitoring, especially in large-scale orchard deployments where signal coverage blind spots and node failures significantly impact monitoring continuity. At the early warning mechanism level, current systems often employ fixed threshold judgments or simple linear models, failing to effectively integrate dynamic adjustments based on the water requirements of fruit trees at different growth stages, weather forecasts, and soil texture differences. This results in low early warning accuracy and frequent false alarms and missed alarms. In addition, there is a lack of effective linkage between traditional monitoring data and irrigation decisions, and the timeliness of converting early warning information into control instructions is insufficient, making it difficult to support the needs of refined water and fertilizer management.

[0003] Therefore, there is an urgent need to provide a technical solution to address the above problems. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a smart orchard soil moisture real-time monitoring and early warning system. The technical solution of this system is as follows:

[0005] The data acquisition module is used to periodically collect real-time soil moisture data covering the main distribution layer of fruit tree roots through a multi-depth sensor node network deployed based on a low-power wide area network architecture.

[0006] The data processing module is used to perform industrial information processing on the real-time soil moisture data, meteorological data and soil attribute data that are synchronized with the real-time soil moisture data in time and space, and generate a standard soil moisture dataset containing the characteristics of root zone water distribution.

[0007] The dynamic early warning module is used to input the standard soil moisture dataset, the current fruit tree growth stage identifier, soil texture classification information and future weather forecast data into the pre-trained fruit tree water demand dynamic assessment model, and output a dynamic water shortage level signal associated with the specific irrigation area.

[0008] The intelligent decision-making module is used to generate precise control instructions for specific irrigation execution terminals based on the dynamic water shortage level signal and preset irrigation strategy rules.

[0009] Furthermore, the industrial information processing performed by the data processing module includes a spatiotemporal kriging interpolation process, which is used to complete the spatially missing data caused by node failure and the time-disrupted data caused by communication delay in the multi-depth sensor node network.

[0010] Furthermore, the multi-depth sensor node network is configured with an adaptive scheduling strategy, which dynamically adjusts the data acquisition frequency of the periodic collection based on the precipitation probability in the future weather forecast data.

[0011] Furthermore, the dynamic assessment model for fruit tree water demand is an ensemble learning model, which integrates the outputs of a first sub-model that takes the standard soil moisture dataset as input and a second sub-model that takes the future weather forecast data as input.

[0012] Furthermore, the first sub-model is a convolutional neural network used to extract spatial correlation features of the real-time soil moisture data between different soil layers; the second sub-model is a time series prediction network used to process the temporal changes of the future weather forecast data.

[0013] Furthermore, the preset irrigation strategy rule is a multi-objective optimization rule, which simultaneously considers the dynamic water shortage level signal, the preset total irrigation water consumption constraint, and the evaporation force parameter in the future weather forecast data.

[0014] Furthermore, the system also includes a visualization interaction module, which is used to receive and integrate the output data of the dynamic early warning module and the intelligent decision-making module to generate a comprehensive situation map that includes the spatiotemporal distribution of soil moisture, early warning areas, and recommended irrigation schemes.

[0015] Furthermore, the system also includes an offline training module, which uses historical data to train the dynamic assessment model of fruit tree water demand. The loss function used during training is:

[0016]

[0017] in, Indicates the loss value; Indicates the total number of training samples; Indicates the first The true water requirement label for each sample; This indicates that the dynamic assessment model for water demand of fruit trees is applicable to the first... Predicted water demand for each sample; This represents the number of spatially adjacent sample pairs in the standard soil moisture dataset; Indicates the first The gradient of predicted water demand for spatially adjacent samples; represents the weighting coefficients of the spatial smoothing regularization term.

[0018] Furthermore, the dynamic early warning module incorporates a soil moisture stress index when calculating the dynamic water shortage level signal. As the core criterion, the soil moisture stress index is calculated using the following formula:

[0019]

[0020] in, Indicates the soil moisture stress index; This represents the current effective water content of the root zone, calculated based on the standard soil moisture dataset and the soil texture classification information. The critical water requirement threshold associated with the current fruit tree growth stage identifier; For sensitivity parameters; This represents the actual evapotranspiration calculated based on recent environmental data. This represents the potential evapotranspiration from the aforementioned future weather forecast data; It is an exponential function.

[0021] Furthermore, the mathematical expression of the multi-objective optimization rule is to minimize the objective function. :

[0022]

[0023] The constraints are:

[0024]

[0025] in, This represents the value of the objective function that needs to be minimized. The predicted total irrigation water consumption is a decision variable. (Irrigation duration) (The initial soil moisture content characterized by the standard soil moisture dataset) and A function of (the predicted evapotranspiration from the aforementioned future weather forecast data); The ideal water consumption to meet the requirements for releasing the dynamic water shortage level signal; Represents the square of the Euclidean norm; This is the duration penalty coefficient; This refers to the maximum permissible duration for a single irrigation session; The total amount of irrigation water is constrained.

[0026] The technical solution of this invention deploys a multi-depth sensor node network through a low-power wide-area network architecture to periodically collect real-time soil moisture data covering the main distribution layer of fruit tree roots. The soil moisture data, synchronous meteorological data, and soil attribute data are processed using industrial information technology to generate a standard dataset. This dataset is then combined with the fruit tree growth stage, soil texture, and future weather forecasts to input into a pre-trained dynamic water demand assessment model, which outputs a dynamic water shortage level signal. Finally, precise control instructions are generated based on the signal and irrigation strategy rules. This solution solves the problems of discrepancies between multi-point interpolation or single-depth monitoring and the actual root water distribution, the difficulty in balancing power consumption and transmission stability, low accuracy of early warning models that do not incorporate growth stage and meteorological information, and poor timeliness of monitoring and irrigation decision-making linkage. It improves the accuracy of orchard soil moisture monitoring and early warning, and the timeliness of irrigation control.

[0027] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0029] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0030] Figure 1 This is a schematic diagram of an embodiment of a smart orchard soil moisture real-time monitoring and early warning system according to the present invention. Detailed Implementation

[0031] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0032] Figure 1 This diagram illustrates a structural schematic of an embodiment of a smart orchard soil moisture real-time monitoring and early warning system provided by the present invention. Figure 1 As shown, the system includes:

[0033] The data acquisition module 110 is used to periodically collect real-time soil moisture data covering the main distribution layer of fruit tree roots through a multi-depth sensor node network deployed based on a low-power wide area network architecture.

[0034] Among them, the low-power wide-area network architecture refers to a wireless network communication system that supports long-distance, low-speed communication and extremely low power consumption of terminal devices; for example, deploying a network built using NB-IoT technology in an orchard, connecting all sensors in the orchard and allowing the sensors to send data to a remote server with extremely low power consumption. A multi-depth sensor node network refers to a monitoring network composed of multiple independent sensing units, each integrating a set of sensors capable of measuring soil parameters at different vertical depths; for example, arranging multiple nodes within the canopy projection area of ​​fruit trees, with each node having soil moisture sensors installed at depths of 20cm, 40cm, and 60cm, collectively forming a monitoring network. Covering the main root distribution layer of fruit trees means that the vertical monitoring range of the sensor network covers the main soil layer depth where the target fruit tree species' roots actively absorb water; for example, for mature fruit trees, sensors are arranged at depths of 20cm to 60cm to monitor the soil conditions in most areas where fibrous roots are distributed. Real-time soil moisture data refers to physical quantity data reflecting soil moisture status collected and reported in real time by sensor nodes; for example, at 10:00 am, a node in an orchard collected and reported that the soil volume moisture content at a depth of 20cm was 18.5% and at a depth of 40cm was 21.2%.

[0035] The data processing module 120 is used to perform industrial information processing on the real-time soil moisture data, meteorological data and soil attribute data that are synchronized with the real-time soil moisture data in time and space, and generate a standard soil moisture dataset containing the root zone water distribution characteristics.

[0036] Meteorological data refers to atmospheric environmental data that is synchronized with soil moisture data in terms of collection time and geographical location; for example, air temperature, humidity, wind speed, and solar radiation data recorded simultaneously and at the same location as the moisture data. Soil property data refers to inherent parameter data characterizing the physicochemical properties of the soil at the monitoring point; for example, data on soil texture, bulk density, field water holding capacity, and wilting coefficient at orchard monitoring points obtained through prior laboratory analysis. Standard soil moisture dataset refers to a standardized dataset generated after cleaning, formatting, spatiotemporal alignment, and fusion of the original collected data; for example, a set of soil moisture data for each point and layer, represented in a two-dimensional matrix, generated by uniformly calibrating, removing outliers, and filling in missing values ​​for moisture content data collected at all nodes in the orchard at a certain time. The root zone moisture distribution characteristics refer to indicators extracted from standard soil moisture datasets that can quantitatively describe the vertical and horizontal variation patterns of moisture in the root zone; for example, the calculated average moisture content of the fruit tree root zone, the moisture content gradient of each soil layer, and the coefficient of variation of moisture in the horizontal direction.

[0037] The dynamic early warning module 130 is used to input the standard soil moisture dataset, the current fruit tree growth stage identifier, soil texture classification information and future weather forecast data into the pre-trained fruit tree water demand dynamic assessment model, and output a dynamic water shortage level signal associated with the specific irrigation area.

[0038] Among these, the current fruit tree growth stage identifier refers to a code or label used to uniquely identify a specific stage in the fruit tree's growth cycle; for example, stage codes such as "budding and leaf unfolding stage," "flowering and fruit setting stage," "fruit enlargement stage," or "flower bud differentiation stage" set for fruit trees based on phenological observations. Soil texture classification information refers to soil type information classified according to soil particle composition; for example, information classifying the soil in a certain area of ​​the orchard as "loam" or "sandy loam" according to international soil texture classification standards. Future weather forecast data refers to weather condition data predicted for a period of time in the future; for example, hourly temperature, precipitation probability, and potential evapotranspiration forecast data for the orchard location over the next 72 hours obtained from a meteorological department interface. The fruit tree water demand dynamic assessment model refers to a mathematical model established through algorithms that can integrate multi-source input information and output an assessment result of the fruit tree's water demand status; for example, a machine learning model that takes soil moisture, weather, growth stage, and soil texture as inputs and, after training, can output a water shortage index between 0 and 1. A specific irrigation area refers to the smallest management unit in an orchard where irrigation can be independently controlled; for example, a section of fruit trees in the orchard, divided into zone 3, row 5, where a drip irrigation system is individually controlled by a set of solenoid valves. A dynamic water shortage level signal refers to a graded signal output by an assessment model that reflects the severity of water shortage in a specific irrigation area over a specific time period; for example, a "Level 2 water shortage" signal output for "zone 3, row 5" in an orchard indicates moderate water stress.

[0039] The intelligent decision-making module 140 is used to generate precise control instructions for specific irrigation execution terminals based on the dynamic water shortage level signal and preset irrigation strategy rules.

[0040] Irrigation strategy rules refer to a set of pre-defined logical rules used to convert water shortage level signals into specific irrigation operation parameters. For example, one rule might stipulate that when a "Level 2 water shortage" signal is received and there is no effective rainfall forecast for the next 12 hours, the irrigation decision-making process is triggered. Specific irrigation execution terminals refer to physical irrigation equipment installed in the orchard that can receive and execute control commands. For example, the intelligent solenoid valve numbered "VALVE-3-5" that controls the water supply to the drip irrigation pipe in "Row 5, Zone 3" in the orchard. Precise control commands refer to control commands that clearly specify the target, content, and timing of the operation. For example, a command sent to terminal "VALVE-3-5": "Start irrigation at 2:00 AM tomorrow, with a continuous irrigation duration of 45 minutes."

[0041] The technical solution of this embodiment deploys a multi-depth sensor node network through a low-power wide-area network architecture to periodically collect real-time soil moisture data covering the main distribution layer of fruit tree roots. The soil moisture data, synchronous meteorological data, and soil attribute data are processed by industrial information to generate a standard dataset. The dataset is then combined with the fruit tree growth stage, soil texture, and future weather forecasts to input a pre-trained dynamic water demand assessment model to output a dynamic water shortage level signal. Finally, precise control instructions are generated based on the signal and irrigation strategy rules. This solution solves the problems of deviation between multi-point interpolation or single-depth monitoring and the actual root water distribution, difficulty in balancing power consumption and transmission stability, low accuracy of early warning models due to not combining growth stage and meteorological information, and poor timeliness of monitoring and irrigation decision linkage. It improves the accuracy of orchard soil moisture monitoring and early warning and the timeliness of irrigation control.

[0042] In one alternative approach, the industrial information processing performed by the data processing module 120 includes a spatiotemporal kriging interpolation process, which is used to complete spatially missing data caused by node failure and temporal breakpoint data caused by communication delay in the multi-depth sensor node network.

[0043] Spatiotemporal kriging interpolation refers to a mathematical interpolation method based on geostatistics, utilizing both spatial proximity and temporal series correlation to estimate the data values ​​of unsampled points or time periods. For example, when a sensor node in an orchard malfunctions, the soil moisture content of each soil layer at the current time can be estimated using interpolation algorithms based on monitoring data from surrounding nodes over a past period. Spatially missing data refers to monitoring data that cannot be obtained at a specific spatial location due to complete sensor node failure or communication interruption. For example, sensor node "Node-7" in an orchard experienced a hardware failure, resulting in no soil moisture content data being uploaded at any depth throughout the day. Temporally discontinuous data refers to interruptions or gaps in the data sequence at a specific point in time due to communication delays, packet loss, or other reasons. For example, the data reporting of all sensors in an orchard at a certain hour was delayed by 15 minutes due to network congestion, resulting in the missing data record for that hour.

[0044] Among the above-mentioned optional methods, the spatial missing data caused by node failure and the time breakpoint data caused by communication delay are further supplemented by the spatiotemporal kriging interpolation process, which ensures the spatiotemporal continuity and integrity of soil moisture data and provides a reliable foundation for subsequent accurate analysis.

[0045] In one alternative approach, the multi-depth sensor node network is configured with an adaptive scheduling strategy that dynamically adjusts the data acquisition frequency of the periodic data collection based on the precipitation probability in the future weather forecast data.

[0046] The adaptive scheduling strategy refers to a strategy that dynamically changes the data acquisition plan based on external conditions. For example, the control strategy of an orchard sensor network stipulates that when a weather forecast indicating a greater than 60% probability of precipitation in the next 6 hours is received, the data acquisition cycle is automatically extended from 1 hour to 4 hours. Precipitation probability refers to the likelihood of measurable precipitation occurring within a certain future period, usually expressed as a percentage; for example, a weather forecast showing a 70% probability of precipitation in the orchard area between 2 PM and 8 PM. Data acquisition frequency refers to the number of times a sensor node performs data acquisition and reporting per unit of time; for example, the acquisition frequency set in the orchard sensor network under normal mode is to acquire and report data once per hour.

[0047] Among the above-mentioned optional methods, an adaptive scheduling strategy can be further configured to dynamically adjust the sensor acquisition frequency based on the precipitation probability in future weather forecasts. This reduces power consumption during rainfall periods and increases monitoring density during non-rainy periods, achieving a balance between monitoring accuracy and energy consumption.

[0048] In one alternative approach, the dynamic assessment model for fruit tree water demand is an ensemble learning model, which integrates the outputs of a first sub-model that takes the standard soil moisture dataset as input and a second sub-model that takes the future weather forecast data as input.

[0049] An ensemble learning model refers to a meta-model that constructs and combines multiple machine learning models to complete a prediction task. For example, a model for orchard water demand assessment includes a sub-model that processes spatial soil moisture data and a sub-model that processes time-series meteorological data. The final result is obtained by a weighted average of the outputs of the two sub-models. The first sub-model refers to a component model in the ensemble learning model specifically designed to process a particular type of data or feature; for example, a neural network model in the ensemble model specifically responsible for processing standard soil moisture datasets to extract spatial moisture distribution features. The second sub-model refers to a component model in the ensemble learning model specifically designed to process another specific type of data or feature; for example, a recurrent neural network model in the ensemble model specifically responsible for processing future weather forecast data sequences to predict water stress trends.

[0050] Among the above-mentioned optional methods, an ensemble learning model is further adopted to integrate the output results of the standard soil moisture dataset sub-model and the future weather forecast data sub-model. This comprehensive multi-source heterogeneous information improves the accuracy of water demand assessment and the adaptability to different orchard environments.

[0051] In one alternative approach, the first sub-model is a convolutional neural network used to extract spatial correlation features of the real-time soil moisture data between different soil layers; the second sub-model is a time series prediction network used to process the temporal changes of the future weather forecast data.

[0052] Among them, convolutional neural networks refer to: a type of feedforward neural network containing convolutional computation layers and a deep structure, which is adept at processing data with grid-like topology; for example, the network structure used in the first sub-model treats soil moisture data at different depths as different channels and performs convolution in horizontal space to extract the spatial distribution pattern of orchard soil moisture. Spatial correlation features refer to: quantitative features extracted from data that reflect the relationships between different spatial locations; for example, the correlation weights between the water content at a depth of 40cm at a certain point and the water content at a depth of 20cm at four adjacent points (east, west, south, and north) learned from orchard soil moisture data by a convolutional neural network. Time series prediction networks refer to: a type of neural network specifically designed to process data sequences arranged in chronological order and predict future values; for example, the long short-term memory network used in the second sub-model takes the hourly meteorological data sequence of the past 72 hours as input and outputs the comprehensive water demand trend of the orchard for the next 24 hours. Temporal variation refers to the dynamic evolution patterns and dynamic evolution of data indicators over time; for example, the fluctuation process of parameters such as temperature and potential evapotranspiration in future weather forecast data, which first rise and then fall within the next 24 hours.

[0053] In the above-mentioned optional methods, the first sub-model further employs a convolutional neural network to extract the spatial correlation features of soil moisture data between different soil layers, and the second sub-model employs a time series prediction network to process the temporal changes of weather forecasts, thereby enhancing the model's ability to capture spatiotemporal features.

[0054] In one alternative approach, the preset irrigation strategy rule is a multi-objective optimization rule that simultaneously considers the dynamic water shortage level signal, the preset total irrigation water consumption constraint, and the evaporation force parameter in the future weather forecast data.

[0055] Among them, multi-objective optimization rules refer to mathematical rules formulated to simultaneously consider and balance multiple objectives when generating irrigation decisions; for example, when formulating orchard irrigation decisions, mathematical programming rules are required to simultaneously achieve the two objectives of maximizing water shortage relief and minimizing irrigation water consumption. Total irrigation water consumption constraints refer to the maximum allowable water resource volume for irrigation decisions within a specific time period; for example, a management red line setting that the total water consumption for orchard irrigation this week must not exceed 500 m³. Evapotranspiration parameters refer to meteorological elements or indicators derived from them that characterize atmospheric evaporation capacity; for example, potential evapotranspiration in future weather forecast data, which comprehensively reflects the influence of temperature, humidity, wind speed, and radiation on the evaporation process.

[0056] Among the above-mentioned optional methods, the irrigation strategy rules are further set as multi-objective optimization rules, taking into account dynamic water shortage level signals, total irrigation water consumption constraints, and future evaporation parameters, so as to minimize water consumption while ensuring irrigation effect and achieve scientific allocation of water resources.

[0057] In an alternative embodiment, the system further includes a visualization interaction module, which receives and integrates the output data of the dynamic early warning module and the intelligent decision-making module to generate a comprehensive situation map containing the spatiotemporal distribution of soil moisture, the early warning area, and recommended irrigation schemes.

[0058] Among them, the spatiotemporal distribution of soil moisture refers to the changes and patterns of soil moisture status in three-dimensional space and one-dimensional time; for example, the dynamic changes in soil moisture content at different plots and soil depths displayed on the orchard comprehensive situation map over the past week. The warning area refers to the orchard area specially marked on the comprehensive situation map that has reached the predetermined water shortage level and requires attention; for example, the northeast corner of the orchard, highlighted with an orange polygon on the situation map, indicating a level two water shortage. The recommended irrigation plan refers to a plan generated based on the warning and decision-making results, containing specific operational suggestions; for example, a pop-up text box on the situation map suggesting drip irrigation of the warning area at dawn tomorrow for 40 minutes, with an estimated water consumption of 12m³. 3 A comprehensive situation map refers to a comprehensive visual graphical interface that integrates multi-dimensional monitoring data, analysis results, and decision-making recommendations; for example, an electronic map of an orchard displayed on a computer or mobile terminal, overlaid with real-time soil moisture contour lines, warning area polygons, irrigation equipment status icons, and pop-up recommended irrigation plan text.

[0059] In the above-mentioned optional methods, a visualization interaction module is further added to receive and integrate the output data of the dynamic early warning module and the intelligent decision-making module to generate a comprehensive situation map that includes the spatiotemporal distribution of soil moisture, early warning areas and recommended irrigation schemes, providing intuitive information display.

[0060] In an alternative embodiment, the system further includes an offline training module, which trains the dynamic assessment model of fruit tree water demand using historical data, and the loss function used during training is:

[0061]

[0062] in, Indicates the loss value; Indicates the total number of training samples; Indicates the first The true water requirement label for each sample; This indicates that the dynamic assessment model for water demand of fruit trees is applicable to the first... Predicted water demand for each sample; This represents the number of spatially adjacent sample pairs in the standard soil moisture dataset; Indicates the first The gradient of predicted water demand for spatially adjacent samples; represents the weighting coefficients of the spatial smoothing regularization term.

[0063] It should be noted that the expression for the model training loss function combines empirical risk minimization and structural risk minimization. The mean squared error term ensures prediction accuracy, while the spatial smoothing regularization term penalizes the gradient of the predicted values ​​to ensure the continuity of the prediction results in spatial distribution. This expression guides model parameter optimization during offline training to obtain a water demand assessment model that is both accurate in prediction and has a reasonable spatial distribution.

[0064] In the above optional approach, an offline training module is further set up to train the water demand dynamic evaluation model using historical data and adopt a loss function that includes a spatial smoothing regularization term, which limits the gradient of predicted values ​​in adjacent regions and improves the spatial consistency of the model's prediction results.

[0065] In one alternative approach, the dynamic early warning module 130 incorporates a soil moisture stress index when calculating the dynamic water shortage level signal. As the core criterion, the soil moisture stress index is calculated using the following formula:

[0066]

[0067] in, Indicates the soil moisture stress index; This represents the current effective water content of the root zone, calculated based on the standard soil moisture dataset and the soil texture classification information. The critical water requirement threshold associated with the current fruit tree growth stage identifier; For sensitivity parameters; This represents the actual evapotranspiration calculated based on recent environmental data. This represents the potential evapotranspiration from the aforementioned future weather forecast data; It is an exponential function.

[0068] It should be noted that the formula for calculating the soil moisture stress index uses a logistic function to simulate the nonlinear continuous change in water stress experienced by fruit trees as the effective soil water content varies around the critical water demand threshold. This is then multiplied by the ratio of actual to potential evapotranspiration to correct for the influence of external meteorological conditions. This formula quantitatively calculates a stress index that comprehensively reflects soil moisture status and meteorological driving factors, serving as the core criterion for generating dynamic water shortage level signals.

[0069] Among the above-mentioned optional methods, the soil moisture stress index is further introduced as the core criterion in the dynamic early warning module. By comprehensively calculating the effective water content, the water requirement threshold at the growth stage, and the ratio of actual to potential evapotranspiration, the scientificity and sensitivity of the water shortage level judgment are improved.

[0070] In one alternative approach, the mathematical expression of the multi-objective optimization rule is to minimize the objective function. :

[0071]

[0072] The constraints are:

[0073]

[0074] in, This represents the value of the objective function that needs to be minimized. The predicted total irrigation water consumption is a decision variable. (Irrigation duration) (The initial soil moisture content characterized by the standard soil moisture dataset) and A function of (the predicted evapotranspiration from the aforementioned future weather forecast data); The ideal water consumption to meet the requirements for releasing the dynamic water shortage level signal; Represents the square of the Euclidean norm; This is the duration penalty coefficient; This refers to the maximum permissible duration for a single irrigation session; The total amount of irrigation water is constrained.

[0075] It should be noted that the multi-objective optimization objective function establishes a mathematical framework that balances irrigation effectiveness and resource consumption. It pursues irrigation precision by minimizing the deviation between predicted and target water usage, while simultaneously incentivizing efficient irrigation through penalties on irrigation duration. The multi-objective optimization objective function transforms irrigation decisions into a constrained optimization problem, and by solving it, obtains a comprehensive optimal precision control command that satisfies multiple constraints.

[0076] In the above-mentioned optional methods, the multi-objective optimization rules are further expressed mathematically as minimizing the objective function. Under the conditions of satisfying the total irrigation volume constraint and duration constraint, the optimal irrigation scheme is solved, and the comprehensive optimization of water consumption and energy consumption is achieved while ensuring the elimination of dynamic water shortage level.

[0077] Furthermore, the system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the system can be divided into different functional modules according to the actual situation to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0078] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.

[0079] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.

[0080] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A smart orchard soil moisture real-time monitoring and early warning system, characterized in that, The system includes: The data acquisition module is used to periodically collect real-time soil moisture data covering the main distribution layer of fruit tree roots through a multi-depth sensor node network deployed based on a low-power wide area network architecture. The data processing module is used to perform industrial information processing on the real-time soil moisture data, meteorological data and soil attribute data that are synchronized with the real-time soil moisture data in time and space, and generate a standard soil moisture dataset containing the characteristics of root zone water distribution. The dynamic early warning module is used to input the standard soil moisture dataset, the current fruit tree growth stage identifier, soil texture classification information and future weather forecast data into the pre-trained fruit tree water demand dynamic assessment model, and output a dynamic water shortage level signal associated with the specific irrigation area. The intelligent decision-making module is used to generate precise control instructions for specific irrigation execution terminals based on the dynamic water shortage level signal and preset irrigation strategy rules.

2. The smart orchard soil moisture real-time monitoring and early warning system according to claim 1, characterized in that, The industrial information processing performed by the data processing module includes a spatiotemporal kriging interpolation process, which is used to complete the spatial missing data caused by node failure and the time breakpoint data caused by communication delay in the multi-depth sensor node network.

3. The smart orchard soil moisture real-time monitoring and early warning system according to claim 1, characterized in that, The multi-depth sensor node network is configured with an adaptive scheduling strategy, which dynamically adjusts the data acquisition frequency of the periodic data collection based on the precipitation probability in the future weather forecast data.

4. The smart orchard soil moisture real-time monitoring and early warning system according to claim 1, characterized in that, The dynamic assessment model for water demand of fruit trees is an ensemble learning model, which integrates the outputs of a first sub-model that takes the standard soil moisture dataset as input and a second sub-model that takes the future weather forecast data as input.

5. The smart orchard soil moisture real-time monitoring and early warning system according to claim 4, characterized in that, The first sub-model is a convolutional neural network, used to extract the spatial correlation features of the real-time soil moisture data between different soil layers; the second sub-model is a time series prediction network, used to process the temporal changes of the future weather forecast data.

6. The smart orchard soil moisture real-time monitoring and early warning system according to claim 1, characterized in that, The preset irrigation strategy rule is a multi-objective optimization rule, which simultaneously considers the dynamic water shortage level signal, the preset total irrigation water consumption constraint, and the evaporation force parameter in the future weather forecast data.

7. The smart orchard soil moisture real-time monitoring and early warning system according to claim 1, characterized in that, The system also includes a visualization interaction module, which is used to receive and integrate the output data of the dynamic early warning module and the intelligent decision-making module to generate a comprehensive situation map that includes the spatiotemporal distribution of soil moisture, the early warning area, and the recommended irrigation scheme.

8. The smart orchard soil moisture real-time monitoring and early warning system according to claim 4, characterized in that, The system also includes an offline training module, which uses historical data to train the dynamic assessment model of fruit tree water demand. The loss function used during training is: in, Indicates the loss value; Indicates the total number of training samples; Indicates the first The true water requirement label for each sample; This indicates that the dynamic assessment model for water demand of fruit trees is applicable to the first... Predicted water demand for each sample; This represents the number of spatially adjacent sample pairs in the standard soil moisture dataset; Indicates the first The gradient of predicted water demand for spatially adjacent samples; represents the weighting coefficients of the spatial smoothing regularization term.

9. The smart orchard soil moisture real-time monitoring and early warning system according to claim 1, characterized in that, When calculating the dynamic water shortage level signal, the dynamic early warning module incorporates the soil moisture stress index. As the core criterion, the soil moisture stress index is calculated using the following formula: in, Indicates the soil moisture stress index; This represents the current effective water content of the root zone, calculated based on the standard soil moisture dataset and the soil texture classification information. The critical water requirement threshold associated with the current fruit tree growth stage identifier; For sensitivity parameters; This represents the actual evapotranspiration calculated based on recent environmental data. This represents the potential evapotranspiration from the aforementioned future weather forecast data; It is an exponential function.

10. The smart orchard soil moisture real-time monitoring and early warning system according to claim 6, characterized in that, The mathematical expression of the multi-objective optimization rule is to minimize the objective function. : The constraints are: in, This represents the value of the objective function that needs to be minimized. The predicted total irrigation water consumption is a decision variable. (Irrigation duration) (The initial soil moisture content characterized by the standard soil moisture dataset) and A function of (the predicted evapotranspiration from the aforementioned future weather forecast data); The ideal water consumption to meet the requirements for releasing the dynamic water shortage level signal; Represents the square of the Euclidean norm; This is the duration penalty coefficient; This refers to the maximum permissible duration for a single irrigation session; The total amount of irrigation water is constrained.

Images