A method for intelligent regulation of soil micro-ecology in mung bean planting

Abstract

The present application relates to the technical field of data processing, and especially relates to a kind of mung bean planting in soil micro-ecological intelligent regulation and control method, the method in the process of mung bean planting, obtain the multidimensional monitoring data of current monitoring moment;According to the multidimensional monitoring data of historical monitoring moment and the most suitable growth environment parameter, obtain the comprehensive environment risk coefficient of current monitoring moment, the soil ecological response characteristic vector of current monitoring moment is formed by the difference between the comprehensive environment risk coefficient and the carbon dioxide release rate of current monitoring moment;The soil ecological response characteristic vector of all historical monitoring moment is clustered, and the soil ecological response characteristic vector threshold value is obtained;According to the difference between soil ecological response characteristic vector and soil ecological response characteristic vector threshold value, state deviation is obtained, abnormal early warning regulation and control is carried out to the soil micro-ecological of mung bean planting, the paradigm shift from static threshold alarm to dynamic trend early warning is realized, and the regulation and control decision-making ability is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for intelligent regulation of soil microecology in mung bean cultivation. Background Technology

[0002] In the current modern agricultural system focused on precision, green sustainability, and high-quality cultivation, precise regulation of the soil microecology in mung bean planting has become a key factor in improving product quality. Currently, a real-time monitoring and control technology system centered on IoT sensors has been established in this field. At the decision-making level, existing practices mainly rely on multi-parameter rule-based judgment methods with fixed thresholds. This involves collecting environmental parameters such as soil temperature, humidity, and pH, and comparing them one by one with existing static safety thresholds for logical judgment. The advantages of this method are its intuitive and clear judgment logic, mature and reliable sensor technology, and relatively controllable system deployment and maintenance costs. It can effectively guarantee basic production safety when environmental conditions are relatively stable and the crop growth stage is singular. With the development of smart agriculture, multi-parameter rule-based judgment methods with fixed thresholds have been combined with scheduled inspections and standardized agricultural operations, forming a standardized operating process from environmental monitoring to fixed responses, laying the foundation for the standardization and partial automation of production management.

[0003] However, the multi-parameter rule judgment method based on fixed thresholds is only simple and feasible under stable environmental and single-stage growth conditions. Its underlying static diagnostic model shows fundamental inadequacy when facing the systematic state drift caused by crop growth, seasonal changes and soil evolution in the field environment. It cannot dynamically assess the health status of the micro-ecology, and it is difficult to capture the gradual risk evolution in a timely manner, nor can it provide accurate basis for differentiated gradient regulation, resulting in lagging and extensive regulation decisions.

[0004] Therefore, how to enhance the ability of real-time monitoring and control methods based on fixed thresholds to adapt to complex environmental dynamic changes during the soil microecological regulation process of mung bean cultivation, and thus improve the efficiency of real-time monitoring and control, has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a method for intelligent regulation of soil microecology in mung bean cultivation, in order to solve the problem of how to enhance the ability of real-time monitoring and regulation methods based on fixed thresholds to adapt to dynamic changes in complex environments during the soil microecology regulation process of mung bean cultivation, thereby improving the efficiency of real-time monitoring and regulation.

[0006] This invention provides a method for intelligent regulation of soil microecology in mung bean cultivation, which includes the following steps: During the mung bean planting process, multidimensional monitoring data at the current monitoring time is obtained. The multidimensional monitoring data includes soil temperature, volumetric water content, pH, and carbon dioxide release rate, wherein the carbon dioxide release rate is the normalized data. Based on the multidimensional monitoring data and optimal growth environment parameters of historical monitoring times, the multidimensional monitoring data of the current monitoring time are standardized and weighted and fused to obtain the comprehensive environmental risk coefficient of the current monitoring time. The comprehensive environmental risk coefficient and the carbon dioxide release rate of the current monitoring time are combined to form the soil ecological response feature vector of the current monitoring time. Obtain the soil ecological response feature vector for each historical monitoring time before the current monitoring time, cluster the soil ecological response feature vectors of all historical monitoring times, and obtain the soil ecological response feature vector threshold corresponding to the current monitoring time to characterize the critical transition of soil ecology in mung bean planting. Based on the difference between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector, the state offset of the soil micro-ecological environment at the current monitoring time is quantified to obtain the state offset degree. Based on the state offset degree, anomaly early warning and regulation of the soil micro-ecology for mung bean planting is carried out.

[0007] Preferably, the step of standardizing and weighting the multidimensional monitoring data at the current monitoring time based on multidimensional monitoring data and optimal growth environment parameters from historical monitoring times to obtain the comprehensive environmental risk coefficient at the current monitoring time includes: The optimal growth environment parameters include optimal soil temperature, optimal volumetric water content, and optimal pH. Calculate the absolute value of the temperature difference between the soil temperature at each historical monitoring time and the optimal soil temperature to obtain a sequence of absolute temperature difference values. Take the 99th percentile of the absolute temperature difference value sequence as the maximum temperature deviation. Calculate the absolute value of the humidity difference between the soil temperature at the current monitoring time and the optimal soil temperature, and record it as the current temperature deviation. Use the maximum temperature deviation to normalize the current temperature deviation to obtain the normalized value of the current temperature deviation. Based on the optimal volumetric water content and the volumetric water content at each historical monitoring time, the normalized value of the current volumetric water content is obtained; based on the optimal pH and the pH at each historical monitoring time, the normalized value of the current pH is obtained; the normalized value of the current temperature deviation, the normalized value of the current volumetric water content, and the normalized value of the current pH are weighted and summed to obtain the comprehensive environmental risk coefficient at the current monitoring time.

[0008] Preferably, before performing a weighted summation of the current temperature deviation normalized value, the current volumetric water content normalized value, and the current pH normalized value, the following steps are included: Based on the soil temperature, volumetric water content, and pH at each historical monitoring moment, the coefficients of variation for soil temperature, volumetric water content, and pH are obtained respectively. The sum of the coefficients of variation for soil temperature, volumetric water content, and pH is calculated. The ratios of the coefficients of variation for soil temperature, volumetric water content, and pH to the sum are calculated respectively, and these ratios are used as the soil temperature weight, volumetric water content weight, and pH weight.

[0009] Preferably, the step of clustering the soil ecological response feature vectors of all historical monitoring times to obtain the soil ecological response feature vector threshold corresponding to the current monitoring time, used to characterize the critical transition of soil ecology in mung bean cultivation, includes: A two-dimensional clustering space was constructed using a comprehensive environmental risk coefficient and carbon dioxide emission rate. Soil ecological response feature vectors from all historical monitoring times were mapped into this two-dimensional clustering space. The K-means clustering algorithm was used to cluster the data points in the two-dimensional clustering space, resulting in three clusters. The cluster center of each cluster was obtained. In the two-dimensional clustering space, the Euclidean distance between each cluster center and a preset data point was calculated. The preset data point refers to the soil ecological response feature vector corresponding to the most abnormal soil ecology. All Euclidean distances were sorted in ascending order to obtain the cluster center corresponding to the second Euclidean distance, which was denoted as the target cluster center. The soil ecological response feature vector corresponding to the target cluster center was used as the threshold soil ecological response feature vector for the current monitoring time, used to characterize the critical transition of soil ecology in mung bean planting.

[0010] Preferably, the step of quantifying the state shift of the soil micro-ecological environment at the current monitoring time based on the difference between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector to obtain the state shift degree includes: Calculate the difference between the comprehensive environmental risk coefficient and the carbon dioxide release rate between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector. Then, perform a weighted difference on the difference between the comprehensive environmental risk coefficient and the difference between the carbon dioxide release rate to obtain the offset at the current monitoring time. Obtain the saturation boundary value of the offset, and use the ratio between the offset at the current monitoring time and the saturation boundary value of the offset as the independent variable of the hyperbolic tangent function to obtain the state offset at the current monitoring time.

[0011] Preferably, obtaining the saturation boundary value of the offset includes: Obtain the absolute value of the offset at each historical monitoring moment, and take the 95th percentile of all absolute values ​​as the saturation boundary value of the offset.

[0012] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: In this invention, a quantitative fusion mechanism for environmental risk is constructed by standardizing and weighting multidimensional and heterogeneous environmental parameters collected in real time. This results in a measurable comprehensive environmental state index that transcends judgment based on a single parameter, namely, the comprehensive environmental risk coefficient. This coefficient compresses complex environmental states into a continuous scalar representing the overall stress level, overcoming the limitations of traditional methods that rely on single or isolated static environmental thresholds for judgment. Furthermore, microbial activity (carbon dioxide release rate) is introduced as a key biological indicator, which, together with the comprehensive environmental risk coefficient, constructs a two-dimensional feature space. By clustering data from historical monitoring times, adaptive thresholds are obtained, resulting in a soil ecological response feature vector threshold. This threshold is compared with the soil ecological response feature vector at the current monitoring time, transforming the previously difficult-to-balance benchmark vector (soil ecological response feature vector threshold) into an interpretable and operable decision indicator (state offset). This provides a quantitative basis for whether to regulate and the urgency of regulation, thereby determining whether the current state is normal. This achieves a paradigm shift from static threshold alarms to dynamic trend warnings, enabling the system to adapt to environmental changes, accurately identify risk levels, and continuously optimize decisions. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart of a method for intelligent regulation of soil microecology in mung bean cultivation, provided in Embodiment 1 of the present invention. Detailed Implementation

[0015] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0016] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0017] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0018] See Figure 1 This is a flowchart of a method for intelligent regulation of soil microecology in mung bean cultivation provided in Embodiment 1 of the present invention. Figure 1 As shown, the method may include: Step S101: During the mung bean planting process, acquire multidimensional monitoring data at the current monitoring time. The multidimensional monitoring data includes soil temperature, volumetric water content, pH, and carbon dioxide release rate, wherein the carbon dioxide release rate is normalized data.

[0019] By deploying a dedicated IoT sensor network, key environmental parameters and representative indicators of microbial activity in the rhizosphere microecology of mung beans are precisely collected during the mung bean planting process. This multi-source, heterogeneous data is then cleaned and aligned to form standardized time-series data for subsequent monitoring of the planting soil microecological environment. Specifically, soil temperature and humidity sensors, pH sensors, and carbon dioxide microflux sensors are deployed in the mung bean rhizosphere region. At fixed collection intervals (e.g., 10 minutes), soil temperature, volumetric water content (i.e., humidity), pH value, and carbon dioxide release rate are collected at each monitoring moment to ensure spatiotemporal consistency of the data.

[0020] Next, the received data streams are automatically cleaned to obtain multi-dimensional monitoring data for each monitoring moment. The multi-dimensional monitoring data includes soil temperature, volumetric water content, pH, and carbon dioxide release rate. The carbon dioxide release rate is the normalized data. The automatic cleaning includes, but is not limited to: setting thresholds based on the physical quantities of the sensors and the rationality of the signals to remove obviously erroneous data; using a strict timestamp matching method to align the data streams of all sensors on a unified time axis to ensure complete alignment of data at each monitoring moment; filling in the missing data due to transmission loss or short-term failures using linear interpolation of data from adjacent monitoring moments; applying a moving average filtering method to the original signal to suppress high-frequency random noise, improving data quality without changing the trend, and forming a complete and clean time-series dataset; and performing dimensionless processing on the collected carbon dioxide release rate (e.g., normalization using the Norm function) to map the carbon dioxide release rate to the [0, 1] interval to eliminate dimensional differences.

[0021] This allows us to obtain multi-dimensional monitoring data at the current monitoring moment, which can then be used for subsequent data analysis.

[0022] Step S102: Based on the multidimensional monitoring data and optimal growth environment parameters of historical monitoring times, the multidimensional monitoring data of the current monitoring time are standardized and weighted and fused to obtain the comprehensive environmental risk coefficient of the current monitoring time. The comprehensive environmental risk coefficient and the carbon dioxide release rate of the current monitoring time are combined to form the soil ecological response feature vector of the current monitoring time.

[0023] Existing methods for real-time monitoring and regulation of micro-ecology using fixed thresholds suffer from a core flaw: the pre-set static rules cannot adapt to systemic state shifts caused by crop growth, seasonal changes, and field management, leading to delayed and inefficient regulatory decisions. Therefore, in this invention, to address the problem of the fixed threshold method's sluggish response to dynamic environmental changes, a quantitative fusion mechanism for environmental risk is established. This mechanism compares real-time collected soil temperature, volumetric water content, and pH with known optimal growth environment parameters for beneficial microorganisms. Through normalization and weight allocation, a risk coefficient that comprehensively characterizes the degree of environmental stress is calculated.

[0024] Specifically, based on multidimensional monitoring data and optimal growth environment parameters from historical monitoring times, the multidimensional monitoring data for the current monitoring time are standardized and weighted and fused to obtain the comprehensive environmental risk coefficient for the current monitoring time. The optimal growth environment parameters for mung bean cultivation are obtained, including optimal soil temperature, optimal volumetric moisture content, and optimal pH. Specifically, the optimal soil temperature for beneficial microorganisms is 25-28 degrees Celsius, the optimal volumetric moisture content is 60%-70%, and the optimal pH is 6.5-7.5. Actual values ​​can be set according to environmental requirements. The absolute value of the temperature difference between the soil temperature at each historical monitoring time and the optimal soil temperature is calculated to obtain a sequence of absolute temperature difference values. The 99th percentile of this sequence is taken as the maximum temperature deviation. The absolute value of the humidity difference between the soil temperature at the current monitoring time and the optimal soil temperature is calculated and recorded as the current temperature deviation. Using the maximum temperature deviation, the current temperature deviation is normalized to obtain the normalized value of the current temperature deviation. Similarly, based on the optimal volumetric water content and the volumetric water content at each historical monitoring time, the normalized value of the current volumetric water content is obtained; based on the optimal pH and the pH at each historical monitoring time, the normalized value of the current pH is obtained; the weighted sum of the normalized value of the current temperature deviation, the normalized value of the current volumetric water content, and the normalized value of the current pH is obtained to obtain the comprehensive environmental risk coefficient at the current monitoring time.

[0025] In one embodiment, the formula for calculating the comprehensive environmental risk coefficient at the current monitoring time is:

[0026] in, This represents the overall environmental risk coefficient at the current monitoring time. This indicates the soil temperature at the current monitoring moment. This indicates the pH level at the current monitoring time. This indicates the volumetric water content at the current monitoring time. Indicates the optimal soil temperature. Indicates the optimal acidity or alkalinity. Indicates the optimal volumetric water content. Indicates the maximum temperature deviation. This indicates the maximum volumetric moisture content deviation. Indicates the maximum pH deviation, | represents the absolute value sign. Indicates soil temperature weighting. Indicates the weight of volumetric water content. This indicates the acidity / alkalinity weight.

[0027] It should be noted that the risk level of the micro-ecological environment at the current monitoring time is quantified by using the difference between the collected real-time data and the optimal environmental parameters for beneficial microorganisms. The larger the absolute value in the above formula, the greater the risk. By compressing the complex environmental conditions (soil temperature, volumetric water content, pH) into a continuous scalar (comprehensive environmental risk coefficient) that characterizes the overall stress level, the multidimensional information is reduced to a one-dimensional scalar. This provides a measurable comprehensive environmental risk coefficient that goes beyond judgment based on a single parameter, thus providing a core input feature with high information density for subsequent quantitative comparison. This method is more accurate and robust than the traditional method of comparing single data with a threshold.

[0028] Among them, the soil temperature weight, volumetric water content weight, and pH weight are obtained through the coefficient of variation. The coefficient of variation measures the stability of the data. Unstable data is a continuous source of stress and interference, and is the main risk source that needs to be prioritized for control. The more unstable the data, the larger the coefficient of variation, and the greater the corresponding weight. Therefore, before performing a weighted summation of the current normalized values ​​of temperature deviation, current normalized values ​​of volumetric water content, and current normalized values ​​of pH, the soil temperature weight, volumetric water content weight, and pH weight are obtained based on the soil temperature, volumetric water content, and pH at each historical monitoring time. The specific method for obtaining these weights is as follows: Based on the soil temperature, volumetric water content, and pH at each historical monitoring moment, the coefficients of variation for soil temperature, volumetric water content, and pH are obtained respectively. The sum of the coefficients of variation for soil temperature, volumetric water content, and pH is calculated. The ratios of the coefficients of variation for soil temperature, volumetric water content, and pH to the sum are calculated respectively, and these ratios are used as the soil temperature weight, volumetric water content weight, and pH weight.

[0029] In one embodiment, the formulas for calculating the soil temperature weight, volumetric water content weight, and pH weight are as follows:

[0030]

[0031]

[0032] in, Indicates soil temperature weighting. Indicates the weight of volumetric water content. Indicates acidity / alkalinity weight. Indicates the coefficient of variation of soil temperature. This represents the coefficient of variation of volumetric water content. This represents the coefficient of variation of pH.

[0033] Thus, based on the method for obtaining the comprehensive environmental risk coefficient at the current monitoring time, the comprehensive environmental risk coefficient at each monitoring time can be obtained. Then, the comprehensive environmental risk coefficient and the carbon dioxide release rate at the current monitoring time can be combined to form the soil ecological response feature vector at the current monitoring time. Similarly, the soil ecological response feature vector at each historical monitoring time before the current monitoring time can be obtained.

[0034] Step S103: Obtain the soil ecological response feature vector of each historical monitoring time before the current monitoring time, cluster the soil ecological response feature vectors of all historical monitoring times, and obtain the soil ecological response feature vector threshold corresponding to the current monitoring time to characterize the critical transformation of the soil ecology of mung bean planting.

[0035] Although the comprehensive environmental risk coefficient can effectively quantify environmental stress, due to the differences among microorganisms, a single comprehensive environmental risk coefficient is still insufficient to accurately define the health status of the soil micro-ecology for mung bean cultivation. Therefore, in this embodiment of the invention, microbial activity (carbon dioxide release rate) is introduced and combined with the comprehensive environmental risk coefficient at each historical monitoring time to form a two-dimensional soil ecological response feature vector, thereby obtaining the soil ecological response feature vector at the current monitoring time and each historical monitoring time before it.

[0036] First, a two-dimensional clustering space is constructed based on the comprehensive environmental risk coefficient and carbon dioxide release rate. The soil ecological response feature vectors at all historical monitoring times are mapped into the two-dimensional clustering space. The K-means clustering algorithm is used to cluster the data points in the two-dimensional clustering space, resulting in three clusters, i.e., K=3, representing the safe state, the risk state, and the transitional state between the two, specifically the critical transition state from safe to risk. The K-means clustering algorithm is an existing technology and will not be elaborated here.

[0037] Then, since the comprehensive environmental risk coefficient and carbon dioxide release rate are both in the range of 0-1 in the two-dimensional cluster space, in this embodiment of the invention, the preset data point in the two-dimensional cluster space is (1,0), which represents the theoretical data point corresponding to the worst soil micro-ecology for mung bean cultivation, that is, the environmental stress is the greatest and the microbial activity is the lowest, which is the extreme case of the greatest pressure and the worst resilience in an ecological environment. Furthermore, since the soil ecological response feature vector at historical monitoring times can characterize three micro-ecological modes—safe, transitional, and risky—and clustering algorithms can effectively distinguish these three modes, the cluster center of each cluster is obtained. In the two-dimensional clustering space, the Euclidean distance between each cluster center and a preset data point is calculated. The preset data point refers to the soil ecological response feature vector corresponding to the most abnormal soil ecology. The farther away from the preset data point, the better the corresponding micro-ecology. The cluster with the farthest Euclidean distance is the safe cluster, the cluster with the closest Euclidean distance is the risky cluster, and the cluster with a medium Euclidean distance is the transitional cluster. Therefore, all Euclidean distances are sorted in ascending order to obtain the cluster center corresponding to the second Euclidean distance, which is denoted as the target cluster center. The soil ecological response feature vector corresponding to the target cluster center is used as the threshold soil ecological response feature vector for characterizing the critical transition of mung bean planting soil ecology at the current monitoring time, denoted as... As time goes on, the threshold of the soil ecological response feature vector will be dynamically adjusted adaptively with the addition of new data points.

[0038] Step S104: Based on the difference between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector, the state offset of the soil micro-ecological environment at the current monitoring time is quantified to obtain the state offset degree. Based on the state offset degree, the soil micro-ecology of mung bean planting is subjected to abnormal early warning and regulation.

[0039] The dynamic threshold (soil ecological response feature vector threshold) obtained by clustering historical data can achieve intelligent regulation if it simultaneously provides signals indicating either danger or safety (e.g., a high comprehensive environmental risk coefficient and low microbial activity, indicating a worse micro-ecological state, and vice versa). However, if the signals are opposite (e.g., a high comprehensive environmental risk coefficient and high microbial activity, making it impossible to determine the state of the microorganisms), the current state cannot be determined. Therefore, in this embodiment of the invention, a discriminant value with clear direction is constructed to determine the state and degree. This transforms the originally difficult-to-balance benchmark vector (soil ecological response feature vector threshold) into an interpretable and operable decision indicator, directly providing a quantitative basis for whether to regulate and the urgency of regulation, thereby determining whether the current microbial state is normal.

[0040] Specifically, based on the difference between the soil ecological response feature vector at the current monitoring time and the soil ecological response feature vector threshold, the state offset of the soil micro-ecological environment at the current monitoring time is quantified to obtain the state offset degree: the difference between the comprehensive environmental risk coefficient and the difference between the soil ecological response feature vector at the current monitoring time and the soil ecological response feature vector threshold are calculated, and the difference between the comprehensive environmental risk coefficient difference and the difference between the carbon dioxide release rate is weighted and subtracted to obtain the offset degree at the current monitoring time; the saturation boundary value of the offset degree is obtained, and the ratio between the offset degree at the current monitoring time and the saturation boundary value of the offset degree is used as the independent variable of the hyperbolic tangent function to obtain the state offset degree at the current monitoring time.

[0041] In one embodiment, the formula for calculating the state offset at the current monitoring time is:

[0042] in, Indicates the state offset at the current monitoring moment. Represents the tangent function of a hyperbola. This represents the first weighting coefficient. This represents the second weighting coefficient. This represents the threshold of the soil ecological response feature vector. represents the soil ecological response feature vector at the current monitoring time t, and M represents the saturation boundary value of the offset.

[0043] It should be noted that, It is the environmental risk coefficient offset term. A positive value indicates that it is higher than the historical critical level, which promotes the micro-ecology of the mung bean planting soil to be more risky, while a negative value promotes the micro-ecology of the mung bean planting soil to be more safe. This is the offset term for microbial activity. A positive value indicates a level higher than the historical critical level, pushing the microecological environment of mung bean planting soil towards risk, while a negative value pushes it towards safety. Because the comprehensive environmental risk coefficient and the microbial activity index do not have an equal impact on the safety of the mung bean planting soil's microecology, and the environmental stress represented by the comprehensive environmental risk index is the primary factor, it is given a higher weight in this scheme. , The specifics can be adjusted according to the actual situation. This is the original theoretical offset, formulated as the comprehensive environmental risk offset minus the biological activity offset. A positive value is obtained when the environmental risk is positive and the microbial activity risk is negative, indicating environmental deterioration. Similarly, other sign cases conform to logical requirements. Overall, a positive calculated state offset value indicates a bias towards danger, while a negative value indicates a bias towards safety. Theoretically, it can take on infinitely large values, making it overly sensitive to extreme values. Therefore, to reduce sensitivity to extreme values, the original theoretical offset is normalized using the saturation boundary value M of the offset. Scaling to a scale primarily based on M ensures sensitivity to normal, small changes while saturating for extreme values, while also utilizing the hyperbolic tangent function. The obtained normalized offset is mapped to the range (-1, 1). The magnitude of the value is the degree of offset. A positive sign indicates a bias towards danger, while a negative sign indicates a bias towards safety.

[0044] The method for obtaining the saturation boundary value M of the offset is as follows: obtain the absolute value of the offset at each historical monitoring time, and take the 95th percentile of all absolute values ​​as the saturation boundary value of the offset.

[0045] This yields the state offset S at the current monitoring moment. The sign indicates the state: positive values ​​indicate a tendency towards danger, while negative values ​​indicate a tendency towards safety. The absolute value indicates the degree of offset; the larger the absolute value, the more extreme the offset. Furthermore, based on the state offset, anomaly warning and regulation of the soil micro-ecology for mung bean cultivation at the current monitoring moment are implemented. This is existing technology, and will be briefly described here: (1) Real-time status assessment and control level determination: Read the value of status deviation S in real time, and determine the status abnormality warning level and control level according to the preset status deviation threshold range, such as: Level 1 warning (record abnormal data), Level 2 warning (trace threshold check) or Level 3 warning (clearly start intervention).

[0046] (2) Matching of control strategies: Generate corresponding control methods according to the warning level. If it is a level 1 warning, focus on observation and do not control it first. If it is a level 2 warning, carry out automatic control, such as starting the drip irrigation system and replenishing the target area with water at 150% of the benchmark amount, or applying the standard dose of microbial growth promoter through the fertilization system. If it is a level 3 warning, carry out mixed control, such as increasing the water benchmark or increasing the dose of growth promoter, and arrange for manual analysis to be carried out at the same time.

[0047] (3) Execution of control instructions and tracking of effects: The central control system sends the generated control prescription to the corresponding intelligent actuators (such as intelligent water and fertilizer integrated machines, drones, and bacterial application equipment). After the actuators complete their operations, they send a feedback signal to the system. The system then enters an enhanced monitoring period, collecting data at a higher frequency to track the change trajectory of the S value, thereby evaluating the initial effect of the control measures in real time.

[0048] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for intelligent regulation of soil microecology in mung bean cultivation, characterized in that, The method includes: During the mung bean planting process, multidimensional monitoring data at the current monitoring time is obtained. The multidimensional monitoring data includes soil temperature, volumetric water content, pH, and carbon dioxide release rate, wherein the carbon dioxide release rate is the normalized data. Based on the multidimensional monitoring data and optimal growth environment parameters of historical monitoring times, the multidimensional monitoring data of the current monitoring time are standardized and weighted and fused to obtain the comprehensive environmental risk coefficient of the current monitoring time. The comprehensive environmental risk coefficient and the carbon dioxide release rate of the current monitoring time are combined to form the soil ecological response feature vector of the current monitoring time. Obtain the soil ecological response feature vector for each historical monitoring time before the current monitoring time, cluster the soil ecological response feature vectors of all historical monitoring times, and obtain the soil ecological response feature vector threshold corresponding to the current monitoring time to characterize the critical transition of soil ecology in mung bean planting. Based on the difference between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector, the state offset of the soil micro-ecological environment at the current monitoring time is quantified to obtain the state offset degree. Based on the state offset degree, anomaly early warning and regulation of the soil micro-ecology for mung bean planting is carried out.

2. The method for intelligent regulation of soil microecology in mung bean cultivation according to claim 1, characterized in that, The process involves standardizing and weighting the multidimensional monitoring data at the current monitoring time based on multidimensional monitoring data from historical monitoring times and optimal growth environment parameters to obtain the comprehensive environmental risk coefficient for the current monitoring time, including: The optimal growth environment parameters include optimal soil temperature, optimal volumetric water content, and optimal pH. Calculate the absolute value of the temperature difference between the soil temperature at each historical monitoring time and the optimal soil temperature to obtain a sequence of absolute temperature difference values. Take the 99th percentile of the absolute temperature difference value sequence as the maximum temperature deviation. Calculate the absolute value of the humidity difference between the soil temperature at the current monitoring time and the optimal soil temperature, and record it as the current temperature deviation. Use the maximum temperature deviation to normalize the current temperature deviation to obtain the normalized value of the current temperature deviation. Based on the optimal volumetric water content and the volumetric water content at each historical monitoring time, the normalized value of the current volumetric water content is obtained; based on the optimal pH and the pH at each historical monitoring time, the normalized value of the current pH is obtained; the normalized value of the current temperature deviation, the normalized value of the current volumetric water content, and the normalized value of the current pH are weighted and summed to obtain the comprehensive environmental risk coefficient at the current monitoring time.

3. The method for intelligent regulation of soil microecology in mung bean cultivation according to claim 2, characterized in that, Before performing a weighted summation of the current temperature deviation normalized value, the current volumetric water content normalized value, and the current pH normalized value, the following steps are included: Based on the soil temperature, volumetric water content, and pH at each historical monitoring moment, the coefficients of variation for soil temperature, volumetric water content, and pH are obtained respectively. The sum of the coefficients of variation for soil temperature, volumetric water content, and pH is calculated. The ratios of the coefficients of variation for soil temperature, volumetric water content, and pH to the sum are calculated respectively, and these ratios are used as the soil temperature weight, volumetric water content weight, and pH weight.

4. The method for intelligent regulation of soil microecology in mung bean cultivation according to claim 1, characterized in that, The process of clustering the soil ecological response feature vectors at all historical monitoring times to obtain the threshold of the soil ecological response feature vector corresponding to the current monitoring time, used to characterize the critical transition of soil ecology in mung bean cultivation, includes: A two-dimensional clustering space was constructed using a comprehensive environmental risk coefficient and carbon dioxide emission rate. Soil ecological response feature vectors from all historical monitoring times were mapped into this two-dimensional clustering space. The K-means clustering algorithm was used to cluster the data points in the two-dimensional clustering space, resulting in three clusters. The cluster center of each cluster was obtained. In the two-dimensional clustering space, the Euclidean distance between each cluster center and a preset data point was calculated. The preset data point refers to the soil ecological response feature vector corresponding to the most abnormal soil ecology. All Euclidean distances were sorted in ascending order to obtain the cluster center corresponding to the second Euclidean distance, which was denoted as the target cluster center. The soil ecological response feature vector corresponding to the target cluster center was used as the threshold soil ecological response feature vector for the current monitoring time, used to characterize the critical transition of soil ecology in mung bean planting.

5. The method for intelligent regulation of soil microecology in mung bean cultivation according to claim 1, characterized in that, The step of quantifying the state shift of the soil micro-ecological environment at the current monitoring time based on the difference between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector, to obtain the state shift degree, includes: Calculate the difference between the comprehensive environmental risk coefficient and the carbon dioxide release rate between the soil ecological response feature vector at the current monitoring time and the threshold of the soil ecological response feature vector. Then, perform a weighted difference on the difference between the comprehensive environmental risk coefficient and the difference between the carbon dioxide release rate to obtain the offset at the current monitoring time. Obtain the saturation boundary value of the offset, and use the ratio between the offset at the current monitoring time and the saturation boundary value of the offset as the independent variable of the hyperbolic tangent function to obtain the state offset at the current monitoring time.

6. The method for intelligent regulation of soil microecology in mung bean cultivation according to claim 5, characterized in that, The process of obtaining the saturation boundary value of the offset includes: Obtain the absolute value of the offset at each historical monitoring moment, and take the 95th percentile of all absolute values ​​as the saturation boundary value of the offset.

Images