System and method for intelligently regulating and controlling growth environment of fruit tree growth state

Abstract

The invention relates to the technical field of agricultural environment intelligent regulation and control, in particular to a growth environment intelligent regulation and control system and method for the growth state of fruit trees, and the method comprises the steps: periodically collecting the multi-dimensional parameters of soil, atmosphere, illumination, trunks, fruits, leaves and the like through a sensing node; and extracting amplitude transition, a fluctuation mode and an interval comparison relation from the time sequence data to quantify an environment disturbance factor, and further synthesizing the comprehensive environment disturbance intensity of each node. And calculating spatial correlation and time sequence characteristics of the intensity to obtain an environment coupling degree index. Meanwhile, a collaborative evolution track of multiple physiological parameters is analyzed, and a physiological stress index is calculated. And fusing the two indexes, generating an environment regulation and control strategy correction amount aiming at each node, iteratively updating a strategy database according to the environment regulation and control strategy correction amount, and driving an execution mechanism to carry out regulation. According to the invention, fine identification of environment dynamic disturbance and closed-loop coupling regulation and control of environment change and plant physiological response are realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent agricultural environmental control technology, and in particular to an intelligent control system and method for the growth environment of fruit trees. Background Technology

[0002] Current environmental control technologies for fruit tree cultivation mainly rely on independent monitoring and threshold control of environmental parameters such as soil temperature and humidity, air temperature and humidity, and light intensity. Control strategies are based on comparing instantaneous sensor readings or simple statistical values ​​within a preset period with fixed thresholds to trigger operations such as irrigation, supplemental lighting, or ventilation. This method treats environmental parameters as static or isolated variables, failing to effectively analyze their dynamic changes and intrinsic relationships over time. The actual characteristics of environmental disturbances, such as the intensity, speed, and duration of changes, as well as the coordinated fluctuations between different parameters, cannot be fully identified and quantified by existing methods. This results in control responses often lagging behind actual dynamic needs or unnecessary interventions in non-critical fluctuations.

[0003] Current technologies typically treat environmental monitoring and fruit tree physiological status monitoring as two relatively independent processes. Regulatory decisions rarely consider the spatial correlation characteristics of single-point environmental fluctuations across the entire planting area, making it difficult to assess the potential impact of local interventions on the overall microclimate. Judgments of fruit tree growth status often rely on single or a few physiological indicators, lacking the ability to integrate and synergistically analyze multi-dimensional physiological signals such as trunk micro-changes, diurnal fruit expansion and contraction, and leaf spectral reflectance. A closed-loop feedback relationship has not been established between environmental control objectives and the real-time, comprehensive physiological stress state of fruit trees, causing regulatory actions to tend towards maintaining fixed environmental parameter targets rather than dynamically adapting to the overall growth load and health needs of the fruit trees. A smart regulation method is needed that can deeply analyze dynamic environmental disturbance patterns and achieve close coupling between environmental factors and plant physiological responses. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and to propose an intelligent control system and method for the growth environment of fruit trees.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for intelligent regulation of the growth environment of fruit trees, comprising: Based on sensing nodes deployed in the growth environment, soil parameters, light parameters, atmospheric parameters, trunk dimension changes, fruit size changes, and leaf radiation parameters are periodically recorded as a growth parameter sequence. The environmental disturbance factors for each sensing period are extracted based on the growth parameter sequence. The dynamic features include the amplitude transition relationship between adjacent sampling points, the overall fluctuation pattern of the sequence, and the horizontal comparison relationship between subsequence intervals. Based on the correlation between the environmental disturbance factors and multiple parameter types, the comprehensive environmental disturbance intensity of each sensing node in each sensing cycle is synthesized. Based on the spatial correlation between the comprehensive environmental disturbance intensities of each sensing node, the environmental fluctuation characteristic value of each sensing node is calculated. Based on the temporal distribution characteristics of the environmental fluctuation feature values, an environmental coupling index is generated for each sensing node. The co-evolutionary trajectory of the changes in trunk dimension, fruit size, and leaf radiation parameters was analyzed, and the physiological stress index of each physiological sensing node was calculated. By integrating the environmental coupling index and the physiological stress index, the correction amount of the environmental regulation strategy for each sensing node is derived. The environmental control strategy correction amount is used to iteratively update the environmental control strategy database for the next sensing cycle, and the actuator is driven to adjust the fruit tree growth environment based on the updated environmental control strategy database.

[0006] Preferably, the method of periodically recording soil parameters, light parameters, atmospheric parameters, trunk dimensional changes, fruit size changes, and leaf radiation parameters based on sensing nodes deployed in the growth environment as a growth parameter sequence specifically includes: At the start of each sensing cycle, all sensing nodes deployed in the fruit tree growth environment are activated; Each sensing node synchronously collects soil moisture data, soil pH data, light intensity data, light duration data, atmospheric temperature data, atmospheric humidity data, trunk diameter micro-change data, fruit transverse and longitudinal diameter data, and leaf multispectral reflectance data according to a preset sampling frequency. The various types of data collected at each sampling time are arranged in chronological order and aligned with the timestamps; Missing values ​​are detected in the sorted data, and missing values ​​are filled using linear interpolation of data from adjacent time points. The filled data is grouped according to parameter type to form soil parameter sequence, light parameter sequence, atmospheric parameter sequence, trunk dimension change sequence, fruit size change sequence, and leaf radiation parameter sequence, which together constitute the growth parameter sequence of the sensing node in the current sensing cycle.

[0007] Preferably, the step of extracting environmental disturbance factors for each sensing cycle based on the growth parameter sequence includes: For each sensing node's environmental parameter sequence in each sensing cycle, identify all sampling points in the sequence that meet the preset inflection point conditions, and mark the sampling points as feature inflection points. For each of the aforementioned characteristic inflection points, the absolute value of the difference between the parameter value of the characteristic inflection point and the parameter value of the immediately preceding moment is calculated as the instantaneous disturbance amplitude of the characteristic inflection point. For each of the aforementioned feature inflection points, analyze the subsequences extending forward and backward by a specific time span centered on it, calculate the arithmetic mean of all parameter values ​​of the forward subsequence and the arithmetic mean of all parameter values ​​of the backward subsequence respectively, and take the absolute value of the difference between the two arithmetic means as the trend offset of the feature inflection point. Extract the first-order difference sequence of the entire environmental parameter sequence, calculate the statistical distribution characteristics of the first-order difference sequence, and determine the baseline stability measure of the environmental parameter sequence based on the statistical distribution characteristics. By combining the instantaneous disturbance amplitude, the trend offset, and the baseline stability measure, the environmental disturbance factor of each characteristic inflection point is obtained through weighted function calculation.

[0008] Preferably, determining the baseline stability measure of the environmental parameter sequence based on the statistical distribution characteristics specifically involves: Calculate the standard deviation of the first-order difference sequence of the environmental parameter sequence, compare the standard deviation with the maximum standard deviation of the first-order difference sequence of the environmental parameters in all historical sensing periods, and use the reciprocal of the comparison result as the baseline stability measure.

[0009] Preferably, the synthesis of the comprehensive environmental disturbance intensity of each sensing node in each sensing cycle includes: Within a single sensing cycle, all sampling times are traversed. When a certain sampling time is identified as a characteristic inflection point in the soil parameter sequence, the illumination parameter sequence, and the atmospheric parameter sequence, the sampling time is marked as a cooperative perturbation time. For each of the cooperative disturbance moments, the environmental disturbance factors corresponding to the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence at the cooperative disturbance moment are obtained respectively. The geometric mean of the environmental disturbance factors corresponding to the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence is calculated, and the geometric mean is used as the comprehensive environmental disturbance intensity of the sensing node at this cooperative disturbance moment in the sensing period.

[0010] Preferably, the calculation of the environmental fluctuation characteristic values ​​of each sensing node includes: Using the comprehensive environmental disturbance intensity of all sensing nodes in the same sensing period as the sample set, the sample set is divided into several clusters by applying an unsupervised classification method. Determine the cluster to which the comprehensive environmental disturbance intensity of each sensing node belongs, and denot it as the sensing node's membership cluster; For each sensing node at each time of coordinated disturbance, data of soil parameters, illumination parameters and atmospheric parameters within a specific time window before and after the time of coordinated disturbance are extracted to form local time-series segments of soil parameters, illumination parameters and atmospheric parameters. The morphological difference degree between the local time series segments of various environmental parameters of the sensing node and the local time series segments of the same environmental parameters of all other sensing nodes in the cluster is calculated. The average value of all morphological difference degrees is obtained to obtain the environmental heterogeneity index of the sensing node at this cooperative disturbance moment. By combining the comprehensive environmental disturbance intensity of the sensing node at this time of coordinated disturbance, the size of its cluster, and the environmental heterogeneity index, the environmental fluctuation characteristic value of the sensing node at this time of coordinated disturbance is calculated through a nonlinear mapping relationship.

[0011] Preferably, the generation of the environmental coupling index for each sensing node specifically includes: For a series of environmental fluctuation characteristic values ​​obtained by each sensing node in each sensing cycle, dynamic segmentation technology is applied to find the critical value that can distinguish the environmental fluctuation characteristic value sequence into high fluctuation range and low fluctuation range. Cooperative disturbance moments with environmental fluctuation characteristic values ​​greater than or equal to the critical value are classified as strong disturbance moments, and cooperative disturbance moments with environmental fluctuation characteristic values ​​less than the critical value are classified as weak disturbance moments. The number of strong disturbance moments and the number of weak disturbance moments for each sensing node in each sensing period are counted, and the ratio of the number of strong disturbance moments to the total number of coordinated disturbance moments is calculated as the proportion of strong environmental disturbances. Calculate the sum of environmental fluctuation characteristic values ​​at all times of strong disturbance and the sum of environmental fluctuation characteristic values ​​at all times of weak disturbance, and calculate the absolute value of the difference between the two sums as the environmental fluctuation polarity difference; Multiply the proportion of strong environmental disturbances by the difference in polarity of environmental fluctuations, and then divide by the average value of the environmental fluctuation characteristics at the time of weak disturbances. The result is defined as the environmental coupling degree index of the sensing node in the sensing period.

[0012] Preferably, the calculation of the physiological stress index of each physiological sensing node specifically includes: For each physiological sensing node in each sensing cycle, the trunk dimension change sequence, the fruit size change sequence, and the leaf radiation parameter sequence are respectively subjected to moving average filtering to obtain smoothed trunk dimension reference sequence, fruit size reference sequence, and leaf radiation reference sequence. Select a subset of times that are classified as strong perturbations from all cooperative perturbation times; For each of the strong disturbance moments, the instantaneous change slopes of the smoothed trunk dimension reference sequence, fruit size reference sequence, and leaf radiation reference sequence at the strong disturbance moment are calculated respectively. By combining the three instantaneous slope changes at the same strong disturbance moment into a vector, the composite physiological change vector at the strong disturbance moment is obtained. Calculate the average magnitude of the composite physiological change vector at all strong disturbance moments within the sensing period, and use the average value after logarithmic transformation as the physiological stress index of the physiological sensing node in the sensing period.

[0013] Preferably, the derivation of the environmental control strategy correction amount for each sensing node specifically includes: The environmental coupling index and the physiological stress index are normalized to a closed interval of zero to one, respectively, to obtain the normalized environmental coupling index and the normalized physiological stress index. Assign preset fusion weights to the normalized environmental coupling index and the normalized physiological stress index; The normalized environmental coupling index and the normalized physiological stress index are linearly weighted and summed according to the assigned fusion weights to obtain a comprehensive correction factor. The comprehensive correction factor is multiplied by the preset basic control strategy matrix, and the product is used as the environmental control strategy correction amount of the sensing node in the sensing cycle.

[0014] Preferably, the present invention also includes an intelligent control system for the growth environment of fruit trees, the system including a storage device, a processing device, and a set of instructions stored in the storage device and loaded and executed by the processing device. When the processing device executes the set of instructions, it implements the entire operation process of the intelligent control method for the growth environment of fruit trees as described above.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: By extracting amplitude transitions between adjacent sampling points, overall fluctuation patterns, and horizontal comparisons between subsequence intervals from continuous sequences of parameters such as soil, light, and atmosphere, environmental disturbances are quantified into dynamic factors encompassing intensity, pattern, and stage characteristics. This method can accurately identify different types of environmental disturbance events, such as sudden sunshine and rain, persistent haze, and periods of high temperatures, surpassing the limitations of judgments based on static thresholds or averages. It provides a refined data foundation for subsequent analysis to characterize the dynamic properties of the environment, achieving a deep perception and feature characterization of environmental instability.

[0016] Based on the correlation between environmental disturbance factors across multiple parameter types, a comprehensive disturbance intensity is synthesized. Then, the environmental coupling index is calculated through its spatial correlation and temporal distribution. This index reveals the degree of correlation between single-point environmental fluctuations and the overall regional microclimate pattern. The physiological stress index is calculated by analyzing the co-evolutionary trajectories of trunk dimension, fruit size, and leaf radiation parameters. This index reflects the intensity of the integrated physiological response of fruit trees to environmental changes. By integrating the environmental coupling degree and physiological stress index, the correction amount of the regulatory strategy is derived. This ensures that the generated regulatory instructions not only aim to mitigate environmental fluctuations but also directly respond to the real-time physiological state and load of the fruit trees. This dual-dimensional fusion mechanism facilitates a shift in regulatory logic from "environmental parameter compliance" to "environment-physiological system synergistic optimization," driving implementing agencies to conduct more predictive and adaptive interventions. Attached Figure Description

[0017] Figure 1 The flowchart shows the intelligent control method for the growth environment of fruit trees according to the present invention. Figure 2 A flowchart for recording growth parameter sequences; Figure 3 A flowchart for extracting environmental disturbance factors; Figure 4 A bar chart showing the average intensity of comprehensive environmental disturbances at each sensing node; Figure 5 Heatmaps of physiological stress indices for each sensory node at different growth stages. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1This paper presents an intelligent method for regulating the growth environment of fruit trees, which is based on a network of sensing nodes deployed in the fruit tree's growth environment. The method uses these sensing nodes to periodically record soil parameters, light parameters, atmospheric parameters, trunk dimensional changes, fruit size changes, and leaf radiation parameters, forming a sequence of growth parameters reflecting the temporal changes in environmental and physiological states. From this sequence, dynamic features characterizing abnormal environmental changes are further extracted. These features include amplitude transitions between adjacent sampling points, overall sequence fluctuation patterns, and horizontal contrast relationships between sub-sequence intervals; these features are defined as environmental disturbance factors. By analyzing the correlation between environmental disturbance factors and different parameter types such as soil, light, and atmosphere, the comprehensive environmental disturbance intensity of each sensing node in each sensing cycle is synthesized. Based on the spatial distribution relationship among the comprehensive environmental disturbance intensities of all sensing nodes, the environmental fluctuation characteristic value of each sensing node is calculated. Further analysis of the distribution characteristics of the environmental fluctuation characteristic values ​​on the time axis generates an environmental coupling index to quantify the correlation between environmental changes and node locations. This study analyzes the co-evolutionary trajectories of three physiological parameters—trunk dimensional changes, fruit size changes, and leaf radiation parameters—with environmental changes, and calculates the physiological stress index, which characterizes the intensity of the fruit tree's physiological response. By integrating the environmental coupling index and the physiological stress index, personalized environmental regulation strategy corrections for each sensing node's microenvironment are derived. These corrections are used to iteratively update the existing environmental regulation strategy database, and the updated strategies drive actuators such as irrigation, supplemental lighting, ventilation, and shading, achieving precise and adaptive regulation of the fruit tree's growth environment.

[0021] In one embodiment of the present invention, see [reference] Figure 2 At the start of each sensing cycle, the system activates all sensing nodes deployed in the fruit tree's growing environment. These sensing nodes are pre-configured at fixed locations within the orchard monitoring area, and immediately enter a data acquisition ready state upon activation. Each sensing node synchronously collects soil moisture data, soil pH data, light intensity data, light duration data, atmospheric temperature data, atmospheric humidity data, trunk diameter micro-change data, fruit transverse and longitudinal diameter data, and leaf multispectral reflectance data according to a preset sampling frequency. The sampling frequency is uniformly set by the central controller to ensure that all sensing nodes perform the acquisition action at the same time. Soil moisture data is acquired through capacitive sensors, soil pH data through ion-selective electrode sensors, light intensity data through photosensors, light duration data through photoperiod calculation modules, atmospheric temperature data through thermocouple sensors, atmospheric humidity data through capacitive humidity sensors, trunk diameter micro-change data through stem flow meters or micro-displacement sensors, fruit transverse and longitudinal diameter data through image recognition modules or laser rangefinders, and leaf multispectral reflectance data through multispectral imagers.

[0022] In practice, the various types of data collected at each sampling time are arranged chronologically and aligned with timestamps provided by the global positioning system module or network time protocol to ensure data consistency. The arranged data is stored in matrix form, with rows corresponding to sampling times and columns corresponding to different parameter types. Missing value detection is performed on the arranged data, and missing values ​​are filled using linear interpolation of adjacent time points. Missing value detection is accomplished by scanning for null values ​​or outlier identifiers in the data matrix. The linear interpolation method calculates the value of missing points based on adjacent valid data points, expressed by the formula:

[0023] in: Indicates time The filled data value Indicates time Valid data values, Indicates time Valid data values, Indicates the time when data points are missing. This indicates the valid time point preceding the missing point. This indicates a valid time point following the missing point.

[0024] In some embodiments, the filled data is grouped according to parameter type to form soil parameter sequences, light parameter sequences, atmospheric parameter sequences, trunk dimension variation sequences, fruit size variation sequences, and leaf radiation parameter sequences. The soil parameter sequences include soil moisture data sequences and soil pH data sequences, the light parameter sequences include light intensity data sequences and light duration data sequences, the atmospheric parameter sequences include atmospheric temperature data sequences and atmospheric humidity data sequences, the trunk dimension variation sequences include trunk diameter micro-variation data sequences, the fruit size variation sequences include fruit transverse diameter data sequences and fruit longitudinal diameter data sequences, and the leaf radiation parameter sequences include leaf multispectral reflectance data sequences. These sequences together constitute the growth parameter sequences of the sensing node in the current sensing cycle.

[0025] Optionally, the sampling frequency of the sensing nodes can be dynamically adjusted according to the growth stage of the fruit trees, for example, increasing the sampling frequency during the flowering period and decreasing it during the dormant period. Linear interpolation is only applicable when valid data exists before and after the missing point; if the missing point is located at the beginning or end of the sequence, nearest neighbor interpolation is used or it is ignored directly. In specific implementations, the timestamp alignment process includes calibrating the internal clocks of the sensing nodes and ensuring that all node clocks are consistent through a network synchronization protocol. Optionally, missing value detection also includes identifying outliers that exceed a reasonable range and marking them as missing values ​​for imputation.

[0026] In one embodiment of the present invention, see [reference] Figure 3 For each sensing node's environmental parameter sequences across different sensing cycles, the system identifies all sampling points in the sequence that satisfy preset inflection point conditions. These sampling points are marked as feature inflection points. The preset inflection point conditions can be defined as a change in the sign of the first derivative of the environmental parameter sequence at the sampling point, or a change in the parameter value of the sampling point relative to its immediate preceding and following sampling points simultaneously exceeding a preset absolute value threshold. For each sampling point marked as a feature inflection point, the absolute value of the difference between the parameter value of the feature inflection point and the parameter value at its immediate preceding moment is calculated. This absolute value is defined as the instantaneous perturbation amplitude of the feature inflection point, reflecting the degree of abrupt change in parameters between adjacent sampling points. For each feature inflection point, subsequences extending forward and backward by specific time spans are analyzed, centered on the feature inflection point. The forward and backward time spans can be equal or unequal. The arithmetic mean of all parameter values ​​in the forward subsequence and the arithmetic mean of all parameter values ​​in the backward subsequence are calculated separately. The absolute value of the difference between the two arithmetic means is defined as the trend offset of the feature inflection point, which characterizes the difference in the overall level of the parameter before and after the feature inflection point. The first-order difference sequence of the entire environmental parameter sequence is extracted. The first-order difference sequence is composed of the value of each sampling point in the environmental parameter sequence minus the value of its previous sampling point. The statistical distribution characteristics of the first-order difference sequence are calculated, including the standard deviation, skewness, and kurtosis of the first-order difference sequence. Based on the statistical distribution characteristics, the baseline stability measure of the environmental parameter sequence is determined.

[0027] In some embodiments, the baseline stability measure of the environmental parameter sequence is determined based on statistical distribution characteristics. Specifically, this involves calculating the standard deviation of the first-order difference sequence of the environmental parameter sequence, comparing the standard deviation of the first-order difference sequence with the maximum standard deviation of the first-order difference sequence of the environmental parameters across all historical sensing periods, and using the reciprocal of the comparison result as the baseline stability measure. The calculation formula is:

[0028] in: This represents the standard deviation of the first-order difference sequence of the environmental parameter sequence within the current sensing period. This represents the maximum standard deviation of the first-order difference sequence of the environmental parameter across all historical sensing periods. The environmental disturbance factor for each feature inflection point is obtained through a weighted function calculation, which combines the instantaneous disturbance amplitude, trend offset, and baseline stability measure of the environmental parameter sequence. The weighted function assigns weight coefficients to the instantaneous disturbance amplitude, trend offset, and baseline stability measure, respectively. , and Optionally, the weighting function can be designed as a linear form, with environmental disturbance factors... ,in: Represents the instantaneous disturbance amplitude. Represents the trend offset. This represents a baseline stability measure. In some embodiments, the threshold for the magnitude of change in the preset inflection point condition can be set differently based on the type of environmental parameter; for example, the threshold for the magnitude of change in atmospheric temperature can be set to one degree Celsius.

[0029] In one embodiment of the present invention, the baseline stability measure of the environmental parameter sequence is calculated by specifically calculating the standard deviation of the first-order difference sequence of the environmental parameter sequence. The first-order difference sequence of the environmental parameter sequence is obtained by subtracting the parameter value of the previous sampling time from the parameter value at each sampling time in the sequence. The standard deviation of the first-order difference sequence is calculated to measure the dispersion of parameter changes within the current sensing period. The system retrieves the standard deviation records of the first-order difference sequences corresponding to this type of environmental parameter in all historical completed sensing periods, finds the maximum standard deviation from the historical records, compares the standard deviation of the first-order difference sequence of the current sensing period with the historical maximum value, and uses the reciprocal of the comparison result as the baseline stability measure of the environmental parameter sequence in the current sensing period. The calculation formula for the baseline stability measure is as follows. for:

[0030] in: This represents the standard deviation of the first-order difference sequence of the environmental parameter sequence within the current sensing period. This represents the maximum standard deviation of the first-order difference sequence of the environmental parameter across all historical sensing cycles. This historical maximum value is dynamically updated with each iteration of the sensing cycle. To synthesize the comprehensive environmental disturbance intensity of each sensing node across all sensing cycles, it is necessary to traverse all sampling times within a single sensing cycle. The system compares the cases where each sampling time is marked as a characteristic inflection point in the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence. When a sampling time is identified as a characteristic inflection point in all three sequences, it is marked as a cooperative disturbance time. The cooperative disturbance time indicates the moment when the three environmental parameters—soil, illumination, and atmosphere—change significantly simultaneously.

[0031] In some embodiments, for each marked cooperative disturbance time, the environmental disturbance factors corresponding to the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence at the cooperative disturbance time are obtained respectively. The environmental disturbance factors are quantified values ​​pre-calculated based on the characteristic inflection points of each parameter sequence. The geometric mean of the environmental disturbance factors corresponding to the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence is calculated, and the calculated geometric mean is used as the comprehensive environmental disturbance intensity of the sensing node at this cooperative disturbance time in the sensing period. The calculation formula is:

[0032] in: This represents the environmental disturbance factor of the soil parameter sequence at the time of cooperative disturbance. This represents the environmental perturbation factor of the illumination parameter sequence at the time of cooperative perturbation. This represents the environmental disturbance factor of the atmospheric parameter sequence at the moment of cooperative disturbance. Optionally, if a sampling moment is identified as a characteristic inflection point only in the two types of environmental parameter sequences, then this sampling moment is not marked as a cooperative disturbance moment and is not included in the calculation of the comprehensive environmental disturbance intensity.

[0033] In some embodiments, the maximum standard deviation of the first-order difference sequence across all historical sensing cycles is stored in a dynamically updated database. Each time a sensing cycle ends and calculations are completed, the system compares the current cycle's standard deviation with the historical maximum value, and updates the historical maximum value as necessary. Optionally, for newly deployed sensing nodes, when historical data is lacking in the initial stage, an initial standard deviation can be preset. The baseline stability measure can be compared with historical maximum values ​​to ensure comparability of assessment results across time scales. The comprehensive environmental disturbance intensity, achieved by fusing multi-source disturbance information through the geometric mean, can reflect the overall synergistic effect of environmental disturbances.

[0034] In one embodiment of the present invention, calculating the environmental fluctuation characteristic values ​​of each sensing node and generating an environmental coupling index involves spatial clustering analysis and temporal distribution characteristic analysis of the comprehensive environmental disturbance intensity. To calculate the environmental fluctuation characteristic values ​​of each sensing node, the comprehensive environmental disturbance intensity of all sensing nodes in the same sensing period is first used as a sample set. An unsupervised classification method is applied to divide the sample set into several clusters. Each sample in the sample set corresponds to the statistical characteristics of the comprehensive environmental disturbance intensity set of a sensing node at all cooperative disturbance times within the current sensing period, such as the mean or maximum value. The cluster to which the comprehensive environmental disturbance intensity of each sensing node belongs is determined; this cluster is the sensing node's membership cluster, and each sensing node is assigned a cluster label. For each sensing node at each cooperative disturbance time, data on soil parameters, illumination parameters, and atmospheric parameters within a specific time window before and after the cooperative disturbance time are extracted, constituting local temporal segments of soil parameters, illumination parameters, and atmospheric parameters. The length of the time window can be set to include five sampling points before and after the cooperative disturbance time. The morphological differences between local time-series segments of various environmental parameters of a sensing node and local time-series segments of similar environmental parameters of all other sensing nodes in its cluster are calculated. The morphological differences can be calculated using a dynamic time warping algorithm. The arithmetic mean of all morphological differences is then obtained to obtain the environmental heterogeneity index of the sensing node at this moment of cooperative disturbance. The environmental heterogeneity index measures the degree of deviation of the node's environmental parameter change pattern from the overall pattern of its cluster at a specific moment. Combining the comprehensive environmental disturbance intensity of the sensing node at this moment of cooperative disturbance, the size of its cluster, and the environmental heterogeneity index, the environmental fluctuation characteristic value of the sensing node at this moment of cooperative disturbance is calculated through a nonlinear mapping relationship. An example of the nonlinear mapping relationship is as follows:

[0035] in: Indicates the characteristic value of environmental fluctuations. This represents the overall environmental disturbance intensity of the sensing nodes at the current moment of coordinated disturbance. This indicates the size of the cluster to which the sensing node belongs, i.e., the number of sensing nodes contained in that cluster. This represents the environmental heterogeneity index of the sensing nodes at the current moment of cooperative disturbance. See Table 1.

[0036] Table 1: Cluster Division Table of Comprehensive Environmental Disturbance Intensity Samples ; To generate the environmental coupling index for each sensing node, it is necessary to use dynamic segmentation techniques to identify the critical values ​​that distinguish the environmental fluctuation feature value sequence into high-fluctuation and low-fluctuation intervals, based on a series of environmental fluctuation feature values ​​obtained by each sensing node within each sensing cycle. Dynamic segmentation techniques can employ methods based on histogram valley detection or the Otsu thresholding method. Cooperative disturbance moments with environmental fluctuation feature values ​​greater than or equal to the critical value are classified as strong disturbance moments, while those with environmental fluctuation feature values ​​less than the critical value are classified as weak disturbance moments. The number of strong disturbance moments and the number of weak disturbance moments for each sensing node within each sensing cycle are counted, and the ratio of the number of strong disturbance moments to the total number of cooperative disturbance moments is calculated. This ratio is taken as the proportion of strong environmental disturbances. ,in: This indicates the number of moments of strong disturbance. This represents the total number of times of coordinated disturbance. The sum of the environmental fluctuation eigenvalues ​​at all strong disturbance times and the sum of the environmental fluctuation eigenvalues ​​at all weak disturbance times are calculated. The absolute value of the difference between the two sums is taken as the environmental fluctuation polarity difference. ,in: This represents the sum of environmental fluctuation eigenvalues ​​at all times of strong disturbance. This represents the sum of environmental fluctuation characteristic values ​​at all weak disturbance times. Multiplying the proportion of strong environmental disturbances by the difference in environmental fluctuation polarity, and then dividing by the arithmetic mean of the environmental fluctuation characteristic values ​​at weak disturbance times, the result is defined as the environmental coupling index of the sensing node during the sensing period. ,in: This represents the arithmetic mean of the environmental fluctuation characteristic values ​​at the moment of weak disturbance. In some embodiments, the size of the cluster to which it belongs. Normalization can be performed before calculation to prevent excessive influence of large scale differences on the calculation of environmental fluctuation characteristic values. Optionally, the calculation of morphological heterogeneity can be performed only on local time-series segments of soil and atmospheric parameters to reduce computational complexity. It can be understood that environmental heterogeneity indicators reflect the local differences between a node and other nodes within the cluster.

[0037] See Figure 4This is a bar chart showing the average intensity of environmental disturbances across the five sensing nodes during the fruit-setting period. The disturbance intensity of Cluster_A is approximately twice that of Cluster_B, indicating significant differences in environmental stability among the different nodes. The high-disturbance cluster (Cluster_A) contains three nodes, while the low-disturbance cluster (Cluster_B) contains two nodes, reflecting the uneven spatial distribution of environmental disturbances within the orchard. High-disturbance nodes (such as Node_005) may require priority in implementing environmental control strategies to mitigate the impact of environmental fluctuations on tree growth. This clustering result can be used to subsequently calculate environmental fluctuation characteristic values ​​and environmental coupling index, providing data support for precise control.

[0038] In one embodiment of the present invention, calculating the physiological stress index of each physiological sensing node and deriving the environmental regulation strategy correction amount for each sensing node is the core process of connecting environmental fluctuations and physiological responses to generate regulatory instructions. Calculating the physiological stress index of each physiological sensing node specifically includes performing a moving average filter on the trunk dimension change sequence, fruit size change sequence, and leaf radiation parameter sequence for each physiological sensing node in each sensing cycle. The window length of the moving average filter can be set to five sampling points, resulting in smoothed trunk dimension reference sequences, fruit size reference sequences, and leaf radiation reference sequences. The smoothing process is used to eliminate measurement noise and highlight the change trend. A subset of strongly disturbed moments is selected from all cooperative disturbance moments. The classification of strongly disturbed moments is based on the dynamic segmentation results from environmental fluctuation feature values. For each selected strongly disturbed moment, the instantaneous change slope of the smoothed trunk dimension reference sequence, fruit size reference sequence, and leaf radiation reference sequence at the strongly disturbed moment is calculated. The instantaneous change slope is approximated by calculating the midpoint difference of the line segment formed by one sampling point before and after the strongly disturbed moment. By vector synthesis of the three instantaneous slope changes at the same strong disturbance moment, a composite physiological change vector at the strong disturbance moment is obtained. This composite physiological change vector is a three-dimensional vector, with its components corresponding to the instantaneous slope changes of trunk dimension, fruit size, and leaf radiation, respectively. The arithmetic mean of the magnitudes of the composite physiological change vectors at all strong disturbance moments within the perception period is calculated. This arithmetic mean is then transformed using a logarithmic transformation with the natural constant as its base, and this transformed value is used as the physiological stress index of the physiological perception node within the perception period. :

[0039] in: This represents the total number of strong disturbance moments within the sensing period. Indicates the first The instantaneous change slope of the trunk dimension baseline sequence at each strong perturbation moment Indicates the first The slope of the instantaneous change in the fruit size baseline sequence at a time of strong perturbation. Indicates the first The instantaneous change slope of the blade radiation reference sequence at each strong disturbance moment. This represents the natural logarithm function.

[0040] The correction amount for environmental regulation strategies at each sensing node is derived. Specifically, this involves normalizing the environmental coupling index and the physiological stress index to a closed interval of zero to one, resulting in normalized environmental coupling index and normalized physiological stress index. The normalization process uses a minimum-maximum scaling method, calculating the minimum and maximum values ​​based on the corresponding index values ​​of all sensing nodes within the current sensing cycle. Preset fusion weights are assigned to the normalized environmental coupling index and the normalized physiological stress index. These weights are stored as coefficients in the system configuration file; for example, the fusion weight for the normalized environmental coupling index can be set to 0.4, and the fusion weight for the normalized physiological stress index can be set to 0.6. A linear weighted sum is then performed on the normalized environmental coupling index and the normalized physiological stress index according to the assigned fusion weights to obtain a comprehensive correction factor. :

[0041] in: Represents the normalized environmental coupling index. Represents the normalized physiological stress index. The fusion weights represent the normalized environmental coupling index. This represents the fusion weight of the normalized physiological stress index. The comprehensive correction factor is multiplied by a preset baseline regulation strategy matrix; the product serves as the environmental regulation strategy correction for the sensing node during the sensing cycle. The baseline regulation strategy matrix defines the benchmark regulation parameter vectors for different actuators such as irrigation, lighting, and ventilation. In some embodiments, the instantaneous slope can be calculated using a more precise method, such as fitting a quadratic polynomial to three sampling points near the moment of strong disturbance and then taking the derivative. Optionally, the fruit size benchmark sequence may be calculated by smoothing the transverse diameter sequence and the longitudinal diameter sequence separately and then merging them into a comprehensive size sequence. It can be understood that the physiological stress index integrates the intensity information of multidimensional physiological responses through vector magnitude and logarithmic transformation.

[0042] See Figure 5This is a heatmap of the physiological stress index of various sensing nodes at different growth stages. The color depth visually reflects the spatiotemporal distribution characteristics of physiological stress in fruit trees. Node_004 exhibits generally higher stress levels than other nodes throughout the entire cycle, possibly due to a more unstable microenvironment or weaker resistance in the fruit tree itself. The overall stress levels during maturity and fruit expansion are higher than those during fruit setting and budding, indicating a more significant physiological response to environmental fluctuations during fruit development. High-stress nodes (such as Node_004) require increased environmental regulation during high-stress stages (maturity and fruit expansion), such as increasing irrigation frequency and adjusting light intensity, to reduce physiological stress. The stable performance of low-stress nodes (such as Node_002) during fruit setting can serve as a benchmark for optimizing regulation strategies for other nodes.

[0043] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for intelligent regulation of the growth environment of fruit trees, characterized in that, The method includes: Based on sensing nodes deployed in the growth environment, soil parameters, light parameters, atmospheric parameters, trunk dimension changes, fruit size changes, and leaf radiation parameters are periodically recorded as a growth parameter sequence. The environmental disturbance factors for each sensing period are extracted based on the growth parameter sequence. The dynamic features include the amplitude transition relationship between adjacent sampling points, the overall fluctuation pattern of the sequence, and the horizontal comparison relationship between subsequence intervals. Based on the correlation between the environmental disturbance factors and multiple parameter types, the comprehensive environmental disturbance intensity of each sensing node in each sensing cycle is synthesized. Based on the spatial correlation between the comprehensive environmental disturbance intensities of each sensing node, the environmental fluctuation characteristic value of each sensing node is calculated. Based on the temporal distribution characteristics of the environmental fluctuation feature values, an environmental coupling index is generated for each sensing node. The co-evolutionary trajectory of the changes in trunk dimension, fruit size, and leaf radiation parameters was analyzed, and the physiological stress index of each physiological sensing node was calculated. By integrating the environmental coupling index and the physiological stress index, the correction amount of the environmental regulation strategy for each sensing node is derived. The environmental control strategy correction amount is used to iteratively update the environmental control strategy database for the next sensing cycle, and the actuator is driven to adjust the fruit tree growth environment based on the updated environmental control strategy database.

2. The intelligent control method for the growth environment of fruit trees as described in claim 1, characterized in that, The method, based on sensing nodes deployed in the growth environment, periodically records soil parameters, light parameters, atmospheric parameters, trunk dimension changes, fruit size changes, and leaf radiation parameters as a growth parameter sequence, specifically including: At the start of each sensing cycle, all sensing nodes deployed in the fruit tree growth environment are activated; Each sensing node synchronously collects soil moisture data, soil pH data, light intensity data, light duration data, atmospheric temperature data, atmospheric humidity data, trunk diameter micro-change data, fruit transverse and longitudinal diameter data, and leaf multispectral reflectance data according to a preset sampling frequency. The various types of data collected at each sampling time are arranged in chronological order and aligned with the timestamps; Missing values ​​are detected in the sorted data, and missing values ​​are filled using linear interpolation of data from adjacent time points. The filled data is grouped according to parameter type to form soil parameter sequence, light parameter sequence, atmospheric parameter sequence, trunk dimension change sequence, fruit size change sequence, and leaf radiation parameter sequence, which together constitute the growth parameter sequence of the sensing node in the current sensing cycle.

3. The intelligent control method for the growth environment of fruit trees as described in claim 2, characterized in that, The step of extracting environmental disturbance factors for each sensing cycle based on the growth parameter sequence includes: For each sensing node's environmental parameter sequence in each sensing cycle, identify all sampling points in the sequence that meet the preset inflection point conditions, and mark the sampling points as feature inflection points. For each of the aforementioned characteristic inflection points, the absolute value of the difference between the parameter value of the characteristic inflection point and the parameter value of the immediately preceding moment is calculated as the instantaneous disturbance amplitude of the characteristic inflection point. For each of the aforementioned feature inflection points, analyze the subsequences extending forward and backward by a specific time span centered on it, calculate the arithmetic mean of all parameter values ​​of the forward subsequence and the arithmetic mean of all parameter values ​​of the backward subsequence respectively, and take the absolute value of the difference between the two arithmetic means as the trend offset of the feature inflection point. Extract the first-order difference sequence of the entire environmental parameter sequence, calculate the statistical distribution characteristics of the first-order difference sequence, and determine the baseline stability measure of the environmental parameter sequence based on the statistical distribution characteristics. By combining the instantaneous disturbance amplitude, the trend offset, and the baseline stability measure, the environmental disturbance factor of each characteristic inflection point is obtained through weighted function calculation.

4. The intelligent control method for the growth environment of fruit trees as described in claim 3, characterized in that, The determination of the baseline stability measure of the environmental parameter sequence based on the statistical distribution characteristics specifically includes: Calculate the standard deviation of the first-order difference sequence of the environmental parameter sequence, compare the standard deviation with the maximum standard deviation of the first-order difference sequence of the environmental parameters in all historical sensing periods, and use the reciprocal of the comparison result as the baseline stability measure.

5. The intelligent control method for the growth environment of fruit trees as described in claim 4, characterized in that, The synthesized comprehensive environmental disturbance intensity of each sensing node in each sensing cycle includes: Within a single sensing cycle, all sampling times are traversed. When a certain sampling time is identified as a characteristic inflection point in the soil parameter sequence, the illumination parameter sequence, and the atmospheric parameter sequence, the sampling time is marked as a cooperative perturbation time. For each of the cooperative disturbance moments, the environmental disturbance factors corresponding to the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence at the cooperative disturbance moment are obtained respectively. The geometric mean of the environmental disturbance factors corresponding to the soil parameter sequence, illumination parameter sequence, and atmospheric parameter sequence is calculated, and the geometric mean is used as the comprehensive environmental disturbance intensity of the sensing node at this cooperative disturbance moment in the sensing period.

6. The intelligent control method for the growth environment of fruit trees as described in claim 5, characterized in that, The calculation of environmental fluctuation characteristic values ​​for each sensing node includes: Using the comprehensive environmental disturbance intensity of all sensing nodes in the same sensing period as the sample set, the sample set is divided into several clusters by applying an unsupervised classification method. Determine the cluster to which the comprehensive environmental disturbance intensity of each sensing node belongs, and denot it as the sensing node's membership cluster; For each sensing node at each time of coordinated disturbance, data of soil parameters, illumination parameters and atmospheric parameters within a specific time window before and after the time of coordinated disturbance are extracted to form local time-series segments of soil parameters, illumination parameters and atmospheric parameters. The morphological difference degree between the local time series segments of various environmental parameters of the sensing node and the local time series segments of the same environmental parameters of all other sensing nodes in the cluster is calculated. The average value of all morphological difference degrees is obtained to obtain the environmental heterogeneity index of the sensing node at this cooperative disturbance moment. By combining the comprehensive environmental disturbance intensity of the sensing node at this time of coordinated disturbance, the size of its cluster, and the environmental heterogeneity index, the environmental fluctuation characteristic value of the sensing node at this time of coordinated disturbance is calculated through a nonlinear mapping relationship.

7. The intelligent control method for the growth environment of fruit trees as described in claim 6, characterized in that, The generation of the environmental coupling degree index for each sensing node specifically includes: For a series of environmental fluctuation characteristic values ​​obtained by each sensing node in each sensing cycle, dynamic segmentation technology is applied to find the critical value that can distinguish the environmental fluctuation characteristic value sequence into high fluctuation range and low fluctuation range. Cooperative disturbance moments with environmental fluctuation characteristic values ​​greater than or equal to the critical value are classified as strong disturbance moments, and cooperative disturbance moments with environmental fluctuation characteristic values ​​less than the critical value are classified as weak disturbance moments. The number of strong disturbance moments and the number of weak disturbance moments for each sensing node in each sensing period are counted, and the ratio of the number of strong disturbance moments to the total number of coordinated disturbance moments is calculated as the proportion of strong environmental disturbances. Calculate the sum of environmental fluctuation characteristic values ​​at all times of strong disturbance and the sum of environmental fluctuation characteristic values ​​at all times of weak disturbance, and calculate the absolute value of the difference between the two sums as the environmental fluctuation polarity difference; Multiply the proportion of strong environmental disturbances by the difference in polarity of environmental fluctuations, and then divide by the average value of environmental fluctuation characteristics at the time of weak disturbances. The result is defined as the environmental coupling index of the sensing node in the sensing period.

8. The intelligent control method for the growth environment of fruit trees as described in claim 7, characterized in that, The calculation of the physiological stress index of each physiological sensing node specifically includes: For each physiological sensing node in each sensing cycle, the trunk dimension change sequence, the fruit size change sequence, and the leaf radiation parameter sequence are respectively subjected to moving average filtering to obtain smoothed trunk dimension reference sequence, fruit size reference sequence, and leaf radiation reference sequence. Select a subset of times that are classified as strong perturbations from all cooperative perturbation times; For each of the strong disturbance moments, the instantaneous change slopes of the smoothed trunk dimension reference sequence, fruit size reference sequence, and leaf radiation reference sequence at the strong disturbance moment are calculated respectively. By combining the three instantaneous slope changes at the same strong disturbance moment into a vector, the composite physiological change vector at the strong disturbance moment is obtained. Calculate the average magnitude of the composite physiological change vector at all strong disturbance moments within the sensing period, and use the average value after logarithmic transformation as the physiological stress index of the physiological sensing node in the sensing period.

9. The intelligent control method for the growth environment of fruit trees as described in claim 8, characterized in that, The derivation of the environmental control strategy correction amount for each sensing node specifically includes: The environmental coupling index and the physiological stress index are normalized to a closed interval of zero to one, respectively, to obtain the normalized environmental coupling index and the normalized physiological stress index. Assign preset fusion weights to the normalized environmental coupling index and the normalized physiological stress index; The normalized environmental coupling index and the normalized physiological stress index are linearly weighted and summed according to the assigned fusion weights to obtain a comprehensive correction factor. The comprehensive correction factor is multiplied by the preset basic control strategy matrix, and the product is used as the environmental control strategy correction amount of the sensing node in the sensing cycle.

10. An intelligent control system for the growth environment of fruit trees, comprising a storage device, a processing device, and a set of instructions stored in the storage device and loaded and executed by the processing device, characterized in that, When the processing device executes the set of instructions, it implements the entire operation process of the intelligent control method for the growth environment of fruit trees as described in any one of claims 1 to 9.

Images