Intelligent agricultural system and agricultural greenhouse humidity intelligent control method

Abstract

The invention provides an intelligent agricultural system and an agricultural greenhouse humidity intelligent control method. The method comprises the following steps: firstly, acquiring a humidity sampling data sequence in a specified monitoring time window before a current regulation and control period in an agricultural greenhouse, and calculating a humidity change trend index; comparing the humidity change trend index with a dynamic trend threshold value, distributing a weight for each humidity sampling point, and performing weighted fusion based on the weights to obtain a weighted fusion result; and then generating a humidity regulation and control reference value of the current regulation and control period based on the weighted fusion result, the current real-time humidity and the external environment parameters. And finally, the deviation between the target humidity of the crop growth model and the current real-time humidity and the humidity regulation and control reference value are input into a controller to generate a dynamic regulation and control instruction, and operation parameters of a humidity execution mechanism are regulated according to the dynamic regulation and control instruction. According to the scheme, the robustness of accurate regulation and control of the humidity of the agricultural greenhouse can be improved.

Description

Technical Field

[0001] This application relates to the field of smart agriculture technology, and more specifically, to a smart agriculture system and a method for intelligent humidity control in agricultural greenhouses. Background Technology

[0002] In smart agriculture systems, humidity control in agricultural greenhouses is a core component of precise regulation of the crop growth environment. It is of great significance for improving photosynthetic efficiency, reducing the incidence of pests and diseases, and ensuring yield and quality. Traditional humidity control methods mostly focus on threshold judgment of instantaneous humidity data and simple on / off control, lacking detailed perception and dynamic response capabilities for historical humidity change trends. This makes it difficult to support early suppression of overshoot oscillations and targeted parameter optimization. With the diversification of crop types and the increasing complexity of the external environment, refined humidity trend analysis and highly timely adaptive regulation capabilities are gradually becoming core requirements of smart agriculture systems.

[0003] Current mainstream humidity control methods for agricultural greenhouses generally suffer from static design issues, making it difficult to adapt to the ever-changing historical trends and external disturbances within the greenhouse. This is especially true in scenarios involving crop growth stage transitions or frequent wind speed fluctuations, where the nonlinear dynamic correlation between humidity sampling sequences and real-time deviations is difficult to extract effectively. This results in a lack of responsiveness to trend overshooting. The root cause of this deficiency lies in the fact that traditional methods fail to introduce an adaptive weight allocation mechanism that reflects the dynamic coupling between historical trends and the current state, lacking strategies to strengthen recent key sampling points and suppress noise sampling points. Consequently, these methods limit the system's ability to analyze humidity fluctuations and its control stability. Therefore, how to achieve adaptive fusion constraints of different sampling features in the humidity sampling sequence to improve the robustness of precise humidity control in agricultural greenhouses has become a challenge for the industry. Summary of the Invention

[0004] This application provides a smart agriculture system and a method for intelligent humidity control in agricultural greenhouses, which can improve the robustness of precise humidity control in agricultural greenhouses.

[0005] In a first aspect, this application provides a method for intelligent humidity control in agricultural greenhouses, comprising: Collect humidity sampling data sequences inside agricultural greenhouses within a specified monitoring time window prior to the current control cycle, and calculate humidity change trend indicators based on the humidity sampling data sequences; The humidity change trend index is compared with the dynamic trend threshold, and weights are assigned to each humidity sampling point within the monitoring time window based on the comparison results. Based on the weights, the humidity sampling values ​​of each humidity sampling point within the monitoring time window are weighted and fused to obtain the weighted fusion result. Based on the weighted fusion result, the current real-time humidity, and the collected external environmental parameters, a humidity control reference value for the current control cycle is generated. The deviation between the target humidity set by the crop growth model and the current real-time humidity, as well as the humidity control reference value of the current control cycle, are input into the controller to generate dynamic control instructions. Based on the dynamic control instructions, the operating parameters of the humidity actuator in the greenhouse are dynamically adjusted to achieve precise humidity control.

[0006] In some embodiments, collecting humidity sampling data sequences inside agricultural greenhouses within a specified monitoring time window prior to the current control cycle, and calculating humidity change trend indicators based on the humidity sampling data sequences specifically includes: Humidity sampling values ​​are collected from humidity sensors installed inside agricultural greenhouses within a specified monitoring time window at predetermined sampling intervals, and then arranged in chronological order to obtain a humidity sampling data sequence. Traverse the humidity sampling data sequence and calculate the humidity difference between adjacent humidity sampling values ​​for each pair to obtain the humidity difference sequence; A dynamic sliding window moving average filter, which is associated with the monitoring window length, is applied to the humidity difference sequence, and the filtered output is used as an indicator of humidity change trend. The length of the specified monitoring time window is dynamically adjusted based on the inertia of greenhouse humidity changes and the response lag time of the actuator to ensure data timeliness.

[0007] In some embodiments, comparing the humidity change trend index with a dynamic trend threshold and assigning weights to each humidity sampling point within the monitoring time window based on the comparison result specifically includes: The humidity change trend index is compared with the dynamic trend threshold. If the humidity change trend index is greater than the dynamic trend threshold, an initial weight distribution that increases as the sampling timestamp approaches the current time is assigned to the humidity sampling points within the monitoring time window. Otherwise, an initial weight distribution that decreases as the sampling timestamp approaches the current time is assigned. Based on a dynamic time decay coefficient that is negatively correlated with the duration of the monitoring window, the time decay weight of each humidity sampling point is calculated. The initial weight of each humidity sampling point is multiplied by its time decay weight to obtain its unnormalized weight; The unnormalized weights of all humidity sampling points are normalized, and the result is used as the final weight assigned to each humidity sampling point.

[0008] In some embodiments, the weighted fusion of humidity sampling values ​​at each humidity sampling point within the monitoring time window based on the weights, to obtain the weighted fusion result, specifically includes: Iterate through each humidity sampling point within the monitoring time window, multiply each humidity sampling value by the corresponding assigned weight, and obtain a weighted set of humidity sampling values. Perform an accumulation and summation operation on the weighted humidity sample value set, and simultaneously calculate the sum of each weight; Divide the sum of the accumulated values ​​by the sum of the weights, and output the quotient as the weighted fusion result.

[0009] In some embodiments, generating a humidity control reference value for the current control cycle based on the weighted fusion result, the current real-time humidity, and the collected external environmental parameters specifically includes: Read the current real-time humidity output from the humidity sensor; Using the current real-time humidity as a baseline value, the humidity change acceleration index is calculated based on the weighted fusion result. Collect the external temperature, external relative humidity, and wind speed values ​​at the start of the current control cycle, input the external temperature, external relative humidity, and wind speed values ​​into a preset environmental correction table, and output the environmental correction coefficient. The current real-time humidity is added to the humidity change acceleration index, and then multiplied by the environmental correction coefficient to output the humidity control reference value for the current control cycle.

[0010] In some embodiments, the controller is a proportional-integral-derivative controller. The deviation between the target humidity set by the crop growth model and the current real-time humidity, and the humidity control reference value for the current control cycle are input into the controller to generate dynamic control commands, specifically including: Read the target humidity corresponding to the current crop type from the crop growth model output module, and subtract the current real-time humidity from the target humidity to obtain the target humidity deviation; The target humidity deviation and the humidity control reference value of the current control cycle are spliced ​​together in a predetermined order to form a control input signal, which is then input to a proportional-integral-derivative controller. The proportional-integral-derivative controller performs the following computational steps to generate dynamic control commands: The predefined actuator lag time parameter and controlled object model are read, and the control signal output by the proportional-integral-derivative controller is divided into two paths: one path is directly sent to the humidity actuator; the other path is input to a signal processing unit. This signal processing unit first simulates humidity changes through the controlled object model, and then performs a pure lag operation on the simulated signal with a time delay equal to the lag time parameter. The output signal of the signal processing unit is obtained as the estimated humidity output. The current actual humidity of the greenhouse is collected, and the difference between the estimated humidity output and the current actual humidity is calculated as the estimated compensation signal. The target humidity deviation is converted into a corresponding deviation electrical signal, and the deviation electrical signal is superimposed on the estimated compensation signal by an adder to generate a comprehensive control signal. The integrated control signal is sequentially input into the proportional, integral, and derivative components of the controller for calculation. The results of the proportional, integral, and differential operations are summed, and the summation result is output as a dynamic control command.

[0011] In some embodiments, dynamically adjusting the operating parameters of the humidity actuator inside the greenhouse according to the dynamic control command specifically includes: Analyze the dynamic control commands and extract the target operating frequency of the humidity actuator; The compensated operating duration is calculated based on the target operating frequency, the predefined actuator lag time parameter, and the required humidity adjustment amount. That is, the humidity adjustment rate of the actuator is first determined according to the target operating frequency, and then the required humidity adjustment amount is divided by the adjustment rate and the actuator lag time is subtracted to obtain the compensated operating duration. The target operating frequency is sent to the frequency control unit of the humidity actuator, which then drives the humidity actuator to operate at the target operating frequency. Start the duration timer of the humidity actuator. When the timer reaches the compensated operating duration, send a stop command to the frequency control unit to stop the operation of the humidity actuator.

[0012] Secondly, this application provides a smart agriculture system, including an intelligent humidity control unit for agricultural greenhouses, wherein the intelligent humidity control unit specifically includes: The data acquisition module is used to collect humidity sampling data sequences inside agricultural greenhouses within a specified monitoring time window prior to the current control cycle, and to calculate humidity change trend indicators based on the humidity sampling data sequences. The processing module is used to compare the humidity change trend index with the dynamic trend threshold, and assign weights to each humidity sampling point within the monitoring time window based on the comparison result. The processing module is also used to perform weighted fusion of the humidity sampling values ​​of each humidity sampling point within the monitoring time window based on the weight, so as to obtain a weighted fusion result; The processing module is also used to generate a humidity control reference value for the current control cycle based on the weighted fusion result, the current real-time humidity, and the collected external environmental parameters; The execution module is used to input the deviation between the target humidity set by the crop growth model and the current real-time humidity, as well as the humidity control reference value of the current control cycle, into the controller to generate dynamic control instructions. Based on the dynamic control instructions, the operating parameters of the humidity actuator in the greenhouse are dynamically adjusted to achieve precise humidity control.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described intelligent humidity control method for agricultural greenhouses.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent humidity control method for agricultural greenhouses.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: This application describes an intelligent humidity control process in agricultural greenhouses: It collects humidity sampling data sequences within a specified monitoring time window prior to the current control cycle inside the greenhouse; calculates a humidity change trend index based on the humidity sampling data sequence; compares the humidity change trend index with a dynamic trend threshold; assigns weights to each humidity sampling point within the monitoring time window based on the comparison results; performs weighted fusion on the humidity sampling values ​​of each humidity sampling point within the monitoring time window based on the weights to obtain a weighted fusion result; generates a humidity control reference value for the current control cycle based on the weighted fusion result, the current real-time humidity, and collected external environmental parameters; inputs the deviation between the target humidity set by the crop growth model and the current real-time humidity, along with the humidity control reference value for the current control cycle, into the controller to generate a dynamic control command; and dynamically adjusts the operating parameters of the humidity actuator inside the greenhouse according to the dynamic control command to achieve precise humidity control.

[0016] Therefore, this application achieves robustness in precise humidity control of agricultural greenhouses through adaptive weight allocation and control reference value generation based on humidity change trend perception. First, by collecting historical humidity sampling data sequences within a specified monitoring time window and calculating humidity change trend indicators, it can capture dynamic humidity change characteristics at a fine-grained level, establishing a trend-driven humidity data expression mechanism and enhancing the system's ability to perceive historical fluctuations from the source. Second, by assigning weights based on the comparison results of trend indicators and dynamic trend thresholds, it achieves adaptive enhancement of recent key sampling points and suppression of noise points, overcoming the hyperparameter dependence problem of traditional simple weighted averaging and ensuring that the weighted fusion results are both timely and accurate. Then, by fusing the weighted results, current real-time humidity, and external environmental parameters to generate humidity control reference values, it can predict humidity trends and offset external interference, providing a reliable benchmark consistent with crop growth models. Finally, the target deviation and humidity control reference values ​​are input into the controller to generate dynamic control commands, enabling precise adjustment of the actuators. This mechanism, through the dynamic coupling of historical trends and real-time states, constructs an adaptive path for humidity control, effectively suppressing overshoot oscillations and slow response. In summary, the proposed solution can achieve robust and precise humidity control in agricultural greenhouses, thereby providing a stable microenvironment for crop growth. Attached Figure Description

[0017] Figure 1 This is an exemplary flowchart of an intelligent humidity control method for agricultural greenhouses according to some embodiments of this application; Figure 2 This is an exemplary flowchart illustrating the allocation of weights to each humidity sampling point according to some embodiments of this application; Figure 3 This is a schematic diagram of the structure of an intelligent humidity control unit for agricultural greenhouses, as shown in some embodiments of this application. Figure 4 This is a schematic diagram of the structure of a computer device for implementing an intelligent humidity control method for agricultural greenhouses, according to some embodiments of this application. Detailed Implementation

[0018] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0019] Referring to Figure 1, which is an exemplary flowchart of an intelligent humidity control method for agricultural greenhouses according to some embodiments of this application, the method mainly includes the following steps: In step S101, a humidity sampling data sequence is collected inside the agricultural greenhouse within a specified monitoring time window prior to the current control cycle, and a humidity change trend index is calculated based on the humidity sampling data sequence.

[0020] The designated monitoring time window refers to the historical interval used to capture dynamic changes in humidity before the start of the current control cycle. For example, in a tomato greenhouse, the control cycle is set to 15 to 25 minutes based on crop transpiration and the inertia of the ventilation system; the predetermined sampling interval is 5 to 10 seconds to ensure the capture of transient fluctuations.

[0021] In some embodiments, humidity sampling values ​​collected at predetermined sampling intervals within a specified monitoring time window are obtained from humidity sensors deployed inside the agricultural greenhouse. The humidity sampling data sequence is obtained by arranging the sampling data in chronological order. Specifically, humidity sensors can be deployed at three heights: the top, middle, and crop canopy of the greenhouse. Humidity sampling values ​​at each height are collected at predetermined sampling intervals to form a humidity sampling data sequence matrix associated with timestamps. For each row in the matrix (i.e., the same timestamp), the arithmetic mean of the humidity values ​​at the top, middle, and canopy is calculated to obtain the single comprehensive humidity sampling value corresponding to that timestamp, and finally, the humidity sampling data sequence is obtained. Traverse the humidity sampling data sequence and calculate the humidity difference between adjacent humidity sampling values ​​one by one to obtain a humidity difference sequence. In specific implementation, start from the first point of the sequence and calculate the humidity difference between each sampling point and its subsequent sampling points one by one to form a complete difference sequence. A dynamic sliding window moving average filter, associated with the monitoring window length, is applied to the humidity difference sequence. The filtered output is used as an indicator of humidity change trend. Specifically, the number of difference points in the dynamic sliding window is set to about 10% of the total number of difference points in the monitoring window (e.g., if there are 120 difference points in the monitoring window, the sliding window is set to 12 difference points). The window slides point by point with the starting position as the center, and the arithmetic mean of all differences in the window is calculated. The filtered value at the end of the sequence is taken as the final indicator of humidity change trend.

[0022] It should be noted that, through the above method, the absolute value change of humidity can be converted into the change of humidity over time. A positive difference indicates that the humidity has increased from the previous moment to the next moment; a negative difference indicates that the humidity has decreased. The larger the absolute value of the difference, the more drastic the change in humidity within that time interval; the smaller the absolute value, the more gradual the change. After sliding window filtering, the noise can be further smoothed, thereby extracting the overall trend of humidity change, which will not be elaborated here.

[0023] The length of the specified monitoring time window is dynamically adjusted based on the inertia of humidity changes in the agricultural greenhouse and the response lag time of the actuator to ensure data timeliness. Specifically, the greenhouse humidity time constant (i.e., the time it takes for the humidity to reach a steady-state value of 63.2%) and the lag time of the actuator from the instruction to the humidity change are statistically analyzed using historical control data. The window length is dynamically set to a predetermined multiple of the maximum value of the two, for example, 1.5 to 2 times. This is only an example and is not intended to limit the scope of this application.

[0024] In step S102, the humidity change trend index is compared with the dynamic trend threshold, and weights are assigned to each humidity sampling point within the monitoring time window based on the comparison result.

[0025] Among them, the humidity change trend index is the average humidity change within a single sampling interval. The dynamic trend threshold is preset according to the crop type and season. When presetting this threshold, the upper limit of the allowable humidity change rate in this scenario can be determined first based on the current crop type and season. Then, combined with the fixed predetermined sampling interval, the upper limit of the rate is converted into the allowable humidity change within a single sampling interval, so that the two can be compared.

[0026] refer to Figure 2 This diagram is an exemplary flowchart for assigning weights to each humidity sampling point. In practice, the humidity change trend index is compared with a dynamic trend threshold. Based on the comparison result, weights are assigned to each humidity sampling point within the monitoring time window in the following manner: In step S1021, the humidity change trend index is compared with the dynamic trend threshold. If the humidity change trend index is greater than the dynamic trend threshold, an initial weight distribution that increases as the sampling timestamp approaches the current time is assigned to the humidity sampling points within the monitoring time window. Otherwise, an initial weight distribution that decreases as the sampling timestamp approaches the current time is assigned. In step S1022, based on the dynamic time decay coefficient which is negatively correlated with the monitoring window duration, the time decay weight of each humidity sampling point is calculated. The longer the monitoring window duration, the smaller the time decay coefficient; the shorter the monitoring window duration, the larger the time decay coefficient. The reference time decay coefficient can be calibrated to 0.5 for example, based on a 15-minute monitoring window. In step S1023, the initial weight of each humidity sampling point is multiplied by its time decay weight to obtain its unnormalized weight; In step S1024, the unnormalized weights of all humidity sampling points are normalized, and the processing result is used as the final weight assigned to each humidity sampling point.

[0027] It should be noted that the core principle of the weighting rule in the above method is to give higher weight to the more valuable sampling data, so that the subsequent weighted fusion result can accurately match the current humidity dynamics and provide a reliable basis for greenhouse humidity control. The specific principle is as follows: The humidity change trend index reflects the degree of drastic change in current humidity. The dynamic trend threshold is a judgment standard for stable / drastic humidity based on crop type and season: If the humidity change trend index is greater than the dynamic trend threshold, it means that the current humidity is changing rapidly. In this case, the recent sampling data (i.e., the data closest to the current time) can better reflect the latest trend (i.e., the old data can no longer keep up with the rapid changes). Therefore, the initial weight increases as the sampling time approaches the current time, so that the recent data dominates the fusion result and avoids the old data from misleading the control. If the humidity change trend index is less than or equal to the dynamic trend threshold, it indicates that the current humidity change is gradual and is within a stable range. At this time, earlier sampling data (i.e., far from the current time) can also reflect the overall stable trend (short-term data may be affected by occasional fluctuations). Therefore, the initial weight decreases as the sampling time approaches the present, allowing long-term data to participate in the decision-making process and avoiding interference from short-term fluctuations in the control judgment.

[0028] In addition, the principle of calculating the time decay weight in the above method is as follows: the dynamic time decay coefficient is a quantitative standard for data timeliness. Based on the time difference between each sampling point and the current control cycle, the larger the time difference, the older the data. The time decay weight is calculated by combining the time difference with the dynamic time decay coefficient. Essentially, the weight of old data is reduced to avoid interference from outdated data, and the weight of recent data is higher. That is, the reference value of the latest humidity dynamics is retained, and a reasonable time dimension is added to the subsequent weight allocation.

[0029] The principle behind multiplying the initial weight by the time decay weight to obtain the unnormalized weight is as follows: the initial weight is determined by comparing the humidity change trend index with the dynamic trend threshold, representing the importance of the data in the humidity trend dimension; multiplying it by the time decay weight in the time dimension combines the importance of the trend with the timeliness of the data, so that sampling points that are both in line with the current trend and relatively new receive a higher unnormalized weight, ensuring that such high-value data are given priority in subsequent fusion, which will not be elaborated here.

[0030] The principle behind the final weighting process after normalization is as follows: the sum of unnormalized weights may not be 1, and directly using them for weighted fusion of humidity sampling values ​​will lead to deviations in the results, such as some weights being too large and excessively dominating the fusion result. By normalizing, the final sum of the weights of all sampling points is made to be 1. At this time, each weight can accurately reflect the contribution ratio of the corresponding humidity sampling point in the overall fusion, ensuring that the subsequent weighted fusion result can accurately integrate the value of each data and provide a reliable basis for generating humidity control reference values.

[0031] In step S103, the humidity sampling values ​​of each humidity sampling point within the monitoring time window are weighted and fused based on the weights to obtain a weighted fusion result.

[0032] In some embodiments, the humidity sampling values ​​of each humidity sampling point within the monitoring time window are weighted and fused based on the weights to obtain the weighted fusion result. This can be done in the following manner: Iterate through each humidity sampling point within the monitoring time window, multiply each humidity sampling value by the corresponding assigned weight, and obtain a weighted set of humidity sampling values. Perform an accumulation and summation operation on the weighted humidity sample value set, and simultaneously calculate the sum of each weight; Divide the sum of the accumulated values ​​by the sum of the weights, and output the quotient as the weighted fusion result.

[0033] It should be noted that in the above implementation method, the accumulated weighted humidity sampling value is the sum of the weighted influence of all humidity sampling points to obtain the total weighted contribution of the overall humidity information; the calculation of the weighted sum provides a benchmark for subsequent proportional conversion, ensuring that the proportion of each sampling point in the total contribution can be accurately separated; and the fusion result obtained by dividing the accumulated result by the weighted sum distributes the total weighted contribution according to the weighted sum. The final quotient can balance the influence of each humidity sampling point, highlighting the dominant role of high-weight data while taking into account the basic information of low-weight data, avoiding the result deviation caused by individual extreme values ​​or low-value data. This quotient can be directly used as the core basis for reflecting the overall humidity situation within the monitoring window, which will not be elaborated here.

[0034] In step S104, a humidity control reference value for the current control cycle is generated based on the weighted fusion result, the current real-time humidity, and the collected external environmental parameters.

[0035] Among them, the humidity control reference value is a specific humidity target value used to guide the operation of greenhouse humidity control equipment. External environmental parameters can be collected by meteorological stations outside the greenhouse, including external temperature, external relative humidity and wind speed.

[0036] In some embodiments, the current real-time humidity can be collected by a humidity sensor. The current real-time humidity is used as a reference value. The humidity change acceleration index is calculated based on the weighted fusion result. Specifically, the instantaneous humidity change rate of each adjacent sampling point in the monitoring window is first calculated (the humidity difference between adjacent sampling points is divided by a fixed sampling interval). Then, the humidity change acceleration index is obtained by dividing the first and last difference of the instantaneous humidity change rate by the total number of instantaneous change rates minus one. In addition, the external temperature, external relative humidity and wind speed values ​​at the start of the current control cycle are collected, and the environmental correction coefficients corresponding to the external temperature, external relative humidity and wind speed values ​​are obtained according to the environmental correction table. For example, the environmental correction coefficient corresponding to an external temperature of 20-25℃, a wind speed of less than 2m / s and an external relative humidity of more than 60% (the humidity inside the shed is prone to be too high in this scenario, and the humidity needs to be slightly suppressed) is 0.95. This is only an example and is not intended to limit the specific application of this application. The current real-time humidity is added to the humidity change acceleration index, and then multiplied by the environmental correction coefficient to output the humidity control reference value for the current control cycle.

[0037] It should be noted that the calculation of the humidity change acceleration index in the above implementation method is to predict the speed and direction of future humidity changes, such as accelerating upward or downward, so that the control reference value can respond to the trend in advance, rather than just passively adapting to the current state. The collection of external temperature, relative humidity, and wind speed and the generation of environmental correction coefficients are because the external environment directly affects the humidity inside the greenhouse. For example, high temperature and high wind speed can easily lead to a decrease in humidity inside the greenhouse. The environmental correction coefficient can offset the interference of the external environment on the humidity inside the agricultural greenhouse, making the humidity control reference value more accurate. Adding the current real-time humidity and the humidity change acceleration index can reflect the current state plus trend adjustment. Multiplying it by the environmental correction coefficient can adapt to the influence of the external environment. The final humidity control reference value takes into account both current and future changes and fits the actual external environment. It can guide the control equipment to accurately maintain the humidity inside the greenhouse to meet the needs of crops. This will not be elaborated further here.

[0038] In step S105, the deviation between the target humidity set by the crop growth model and the current real-time humidity, and the humidity control reference value of the current control cycle are input into the controller to generate a dynamic control command. The operating parameters of the humidity actuator in the greenhouse are dynamically adjusted according to the dynamic control command to achieve precise humidity control.

[0039] The controller can be a proportional-integral-derivative controller, but other types of controllers are also possible in practice. This is only an example and is not intended to limit the scope of this application. Furthermore, crop growth models, as mature existing technologies, can be categorized based on two main dimensions: model construction principles and applicable crop range. The crop growth model in this application can adopt a multi-crop general hybrid model, which is based on mechanism-driven principles and integrates data-driven logic, such as the SPACSYS model. This is not a specific limitation and is merely an example.

[0040] In some embodiments, the deviation between the target humidity set by the crop growth model and the current real-time humidity, and the humidity control reference value of the current control cycle are input into the controller to generate dynamic control commands. This can be achieved in the following manner: Read the target humidity corresponding to the current crop type from the crop growth model output module, and subtract the current real-time humidity from the target humidity to obtain the target humidity deviation; The target humidity deviation and the humidity control reference value of the current control cycle are spliced ​​together in a predetermined order to form a control input signal, which is then input to a proportional-integral-derivative controller. The proportional-integral-derivative controller performs the following computational steps to generate dynamic control commands: Read the predefined actuator lag time parameters and the controlled object model (e.g., a first-order inertial model), and split the controller output signal into two paths: one path is directly sent to the humidity actuator; the other path is input to a signal processing unit, which first simulates humidity changes through the controlled object model, and then performs a pure lag operation on the simulated signal with a time delay equal to the lag time parameters, and obtains the output signal of the signal processing unit as the estimated humidity output; collect the current actual humidity of the greenhouse, and calculate the difference between the estimated humidity output and the current actual humidity as the estimated compensation signal; The target humidity deviation is converted into a corresponding deviation electrical signal, and the deviation electrical signal is superimposed on the estimated compensation signal by an adder to generate a comprehensive control signal. The integrated control signal is sequentially input into the proportional, integral, and derivative components of the controller for calculation. The results of the proportional, integral, and differential operations are summed, and the summation result is output as a dynamic control command.

[0041] It should be noted that the core function of the signal processing unit in the above implementation is to first simulate humidity changes through the controlled object model, and then apply a pure lag operation. The time delay of the pure lag operation is completely consistent with the read lag time parameter of the actuator. This simulates the lag characteristics of the actuator when it actually moves and the response to changes in greenhouse humidity. Finally, the difference between the estimated humidity output and the current actual humidity output by the signal processing unit is the estimated compensation signal used to compensate for the lag effect of the actuator. This can help the subsequent control link adapt to the lag effect in advance and avoid control deviation. In specific implementation, the signal processing unit can, for example, use an STC12LE5408 microcontroller. This model of microcontroller is often used as a humidity signal processing unit in the humidity monitoring scenario of agricultural greenhouses. This is only an example and is not intended to limit the specific application.

[0042] In some embodiments, the dynamic adjustment of the operating parameters of the humidity actuator inside the greenhouse according to the dynamic control command can be achieved in the following manner: Analyze the dynamic control commands and extract the target operating frequency of the humidity actuator; Based on the target operating frequency, the predefined actuator lag time parameter, and the required humidity adjustment amount (i.e., the absolute value of the current real-time humidity and the humidity control reference value), the compensated operating duration is calculated as follows: First, the humidity adjustment rate of the actuator is determined according to the target operating frequency (for example, the higher the fan frequency, the faster the dehumidification rate). Then, the required humidity adjustment amount is divided by the adjustment rate, and the actuator lag time is subtracted to avoid excessive adjustment due to lag. The compensated operating duration is obtained. If the calculation result is negative, the actuator lag time is taken as the minimum operating duration. The target operating frequency is sent to the frequency control unit of the humidity actuator, which then drives the humidity actuator to operate at the target operating frequency. Start the duration timer of the humidity actuator. When the timer reaches the compensated operating duration, send a stop command to the frequency control unit to stop the operation of the humidity actuator.

[0043] It should be noted that the target operating frequency in the above implementation method is the preset operating frequency required by the humidity actuator to reach the humidity control reference value, which can be output by the proportional-integral-derivative controller in combination with the control requirements; the predefined actuator lag time parameter can be a fixed value obtained in advance through experimental testing or equipment manual, which is used to quantify the impact of lag on the operating effect; the required humidity adjustment amount is the difference between the current humidity and the control target, which is the core basis for calculating the operating duration.

[0044] It should also be noted that in the above implementation, while the humidity actuator starts operating according to the control command, the system simultaneously starts a duration timer. The initial state of this timer is zero. After starting, it begins to accumulate the actual running time of the actuator in real time, ensuring that the timing is completely synchronized with the operating state of the actuator, laying the foundation for subsequent precise control. During the operation of the timer, the current accumulated time is continuously compared with the compensated running duration in real time: if the accumulated time has not reached the compensated running duration, the timer continues to count, and the actuator maintains its current operating state; if the accumulated time reaches the compensated running duration, the timer immediately triggers the stop signal generation mechanism. The generated stop signal is sent to the frequency control unit (i.e., the core unit responsible for controlling the operating frequency and start / stop state of the actuator in the prior art) in the form of an instruction. After receiving the instruction, the frequency control unit immediately sends a stop control signal to the humidity actuator, and the actuator stops working after receiving the signal. The entire process, through precise timing of the timer and threshold triggering, ensures that the runtime of the actuator matches the compensation requirements exactly, offsetting the lag effect and enabling the greenhouse humidity to stably approach the control reference value. This avoids problems such as excessive control (e.g., dehumidification for too long leading to excessively low humidity) or insufficient control (e.g., insufficient humidification time leading to unmet humidity standards), which will not be elaborated further here.

[0045] Furthermore, in another aspect of this application, in some embodiments, this application provides a smart agriculture system, which includes an intelligent humidity control unit for agricultural greenhouses, as referenced. Figure 3 The figure is a schematic diagram of the structure of an intelligent humidity control unit for agricultural greenhouses according to some embodiments of this application. The intelligent humidity control unit specifically includes: a data acquisition module 201, a processing module 202, and an execution module 203. The acquisition module 201 described in this application is mainly used to acquire humidity sampling data sequence inside the agricultural greenhouse within a specified monitoring time window before the current control cycle, and to calculate humidity change trend index based on the humidity sampling data sequence. It can integrate a multi-point sensor data bus and a filter processor. Processing module 202, as described in this application, is mainly used to compare the humidity change trend index with the dynamic trend threshold, and to assign weights to each humidity sampling point within the monitoring time window based on the comparison result; In a specific implementation, the processing module 202 is also used to perform weighted fusion of the humidity sampling values ​​of each humidity sampling point within the monitoring time window based on the weight, so as to obtain a weighted fusion result; In addition, the processing module 202 is also used to generate a humidity control reference value for the current control cycle based on the weighted fusion result, the current real-time humidity, and the collected external environmental parameters. It may include an environmental correction table lookup unit and an acceleration calculation unit. The execution module 203, as described in this application, is mainly used to input the deviation between the target humidity set by the crop growth model and the current real-time humidity, as well as the humidity control reference value of the current control cycle, into the controller to generate dynamic control instructions. According to the dynamic control instructions, the operating parameters of the humidity actuator in the greenhouse can be dynamically adjusted to achieve precise humidity control. Internally, a PID controller, a Smith predictor, and a frequency conversion drive interface can be integrated.

[0046] In addition, this application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described intelligent humidity control method for agricultural greenhouses.

[0047] In some embodiments, reference Figure 4 The figure is a schematic diagram of the structure of a computer device for implementing an intelligent humidity control method for agricultural greenhouses, according to some embodiments of this application. The methods in the above embodiments can... Figure 4 The computer device shown is used to implement this, and the computer device includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0048] The processor 301 may be a general-purpose central processing unit, an application-specific integrated circuit, or one or more executions for controlling the intelligent humidity control method for agricultural greenhouses in this application.

[0049] The communication bus 302 may include a path for transmitting information between the aforementioned components.

[0050] The memory 303 may be a read-only memory or other type of static storage device, a random access memory or other type of dynamic storage device, or an electrically erasable programmable read-only memory, a read-only optical disc or other optical disc storage, a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via a communication bus 302. The memory 303 may also be integrated with the processor 301.

[0051] The memory 303 stores program code for executing the scheme of this application, and its execution is controlled by the processor 301. The processor 301 executes the program code stored in the memory 303. The program code may include one or more software modules. In the above embodiments, the calculation, weight allocation, and PID control of the humidity change trend index can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0052] Communication interface 304 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, wireless access networks, wireless local area networks, etc.

[0053] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core processor or a multi-core processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0054] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld computer, a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0055] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent humidity control method for agricultural greenhouses.

[0056] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0057] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of the invention. Therefore, if these modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. An intelligent agricultural greenhouse humidity control method, characterized in that, The method comprises the following steps: Collecting a humidity sampling data sequence of the inside of the agricultural greenhouse in a specified monitoring time window before the current control period, calculating a humidity change trend index based on the humidity sampling data sequence; Comparing the humidity change trend index with a dynamic trend threshold, and assigning weights to each humidity sampling point in the monitoring time window according to the comparison result; Based on the weight, the humidity sampling values of each humidity sampling point in the monitoring time window are weighted and fused to obtain a weighted fusion result; Based on the weighted fusion result, the current real-time humidity, and the collected external environment parameters, a humidity control reference value of the current control period is generated; The deviation of the target humidity set by the crop growth model from the current real-time humidity and the humidity control reference value of the current control period are input into the controller to generate a dynamic control instruction, and the running parameters of the humidity execution mechanism in the greenhouse are dynamically adjusted according to the dynamic control instruction to realize precise humidity control.

2. The method of claim 1, wherein, Collecting a humidity sampling data sequence of the inside of the agricultural greenhouse in a specified monitoring time window before the current control period, calculating a humidity change trend index based on the humidity sampling data sequence specifically comprises: Reading the humidity sampling values collected at a predetermined sampling interval in the specified monitoring time window from the humidity sensors arranged in the agricultural greenhouse, and arranging the humidity sampling values in the order of collection time to obtain a humidity sampling data sequence; Iterating through the humidity sampling data sequence, calculating the humidity difference value of each pair of adjacent humidity sampling values to obtain a humidity difference value sequence; Performing a dynamic sliding window moving average filtering on the humidity difference value sequence associated with the length of the monitoring window to obtain the filtering output result as the humidity change trend index; Wherein, the length of the specified monitoring time window is dynamically adjusted according to the greenhouse humidity change inertia and the execution mechanism response lag time to ensure data timeliness.

3. The method of claim 1, wherein, Comparing the humidity change trend index with a dynamic trend threshold, and assigning weights to each humidity sampling point in the monitoring time window according to the comparison result specifically comprises: Numerically comparing the humidity change trend index with the dynamic trend threshold, if the humidity change trend index value is greater than the dynamic trend threshold, assigning an initial weight distribution to the humidity sampling points in the monitoring time window, which increases with the proximity of the sampling timestamp to the current time, otherwise, assigning an initial weight distribution which decreases with the proximity of the sampling timestamp to the current time; Based on the dynamic time decay coefficient negatively related to the length of the monitoring window, calculating the time decay weight of each humidity sampling point; Multiplying the initial weight of each humidity sampling point with its time decay weight to obtain its unnormalized weight; Normalizing the unnormalized weights of all humidity sampling points, and taking the processing result as the final assigned weight of each humidity sampling point.

4. The method of claim 1, wherein, Based on the weight, the humidity sampling values of each humidity sampling point in the monitoring time window are weighted and fused to obtain a weighted fusion result, which specifically comprises: Iterating through each humidity sampling point in the monitoring time window, multiplying each humidity sampling value by the corresponding assigned weight to obtain a set of weighted humidity sampling values; Performing an accumulation sum operation on the set of weighted humidity sampling values, and simultaneously calculating the accumulation sum of each weight; Dividing the accumulation sum result by the accumulation sum of the weights, and outputting the quotient as the weighted fusion result.

5. The method of claim 1, wherein, Generating the humidity control reference value of the current control period based on the weighted fusion result, the current real-time humidity, and the collected external environment parameters specifically comprises: reading the current real-time humidity output by the humidity sensor at the current time; taking the current real-time humidity as a reference value, and calculating a humidity change acceleration index based on the weighted fusion result; collecting the external temperature value, the external relative humidity value, and the wind speed value at the starting time of the current control period, inputting the external temperature value, the external relative humidity value, and the wind speed value into a preset environment correction table, and outputting an environment correction coefficient; adding the current real-time humidity and the humidity change acceleration index, and then multiplying the result by the environment correction coefficient to output the humidity control reference value of the current control period.

6. The method of claim 1, wherein, The controller is a proportional-integral-derivative controller, and the target humidity set by the crop growth model, the deviation of the current real-time humidity, and the humidity control reference value of the current control period are input into the controller to generate a dynamic control instruction, which specifically comprises: reading the target humidity corresponding to the current crop type from the crop growth model output module, and subtracting the current real-time humidity from the target humidity to obtain a target humidity deviation; splicing the target humidity deviation and the humidity control reference value of the current control period in a predetermined order to form a control input signal, and inputting the control input signal into the proportional-integral-derivative controller; The proportional-integral-derivative controller performs the following operation steps to generate a dynamic control instruction: reading a predefined actuator lag time parameter and a controlled object model, and dividing the control signal output by the proportional-integral-derivative controller into two paths: one path is directly sent to the humidity actuator; the other path is input into a signal processing unit, which first simulates humidity change through the controlled object model, and then performs a pure lag operation with a time delay equal to the lag time parameter on the simulated signal to obtain an output signal of the signal processing unit as a predicted humidity output, collects the current actual humidity of the greenhouse, and calculates the difference between the predicted humidity output and the current actual humidity as a predicted compensation signal; converting the target humidity deviation into a corresponding deviation electric signal, and inputting the deviation electric signal and the predicted compensation signal into an adder to generate a comprehensive control signal; inputting the comprehensive control signal into the proportional, integral, and differential links of the controller in sequence for operation; summing the operation results of the proportional, integral, and differential links, and outputting the summing result as a dynamic control instruction.

7. The method of claim 1, wherein, According to the dynamic control instruction, the running parameters of the humidity actuator in the greenhouse are dynamically adjusted, which specifically comprises: parsing the dynamic control instruction to extract the target running frequency of the humidity actuator; calculating the compensated running duration based on the target running frequency, the predefined actuator lag time parameter, and the required humidity adjustment amount, i.e., first determining the humidity adjustment rate of the actuator according to the target running frequency, then dividing the required humidity adjustment amount by the adjustment rate, and subtracting the actuator lag time to obtain the compensated running duration; sending the target running frequency to the frequency control unit of the humidity actuator to drive the humidity actuator to work at the target running frequency; The duration timer of the humidity actuator is started, and when the timer counts to the compensated operation duration, a stop instruction is sent to the frequency control unit to stop the operation of the humidity actuator.

8. A smart agriculture system comprising an agricultural greenhouse humidity intelligent control unit, characterized in that, The agricultural greenhouse humidity intelligent control unit specifically comprises: The acquisition module is configured to acquire a humidity sampling data sequence of the agricultural greenhouse in a specified monitoring time window before a current control period, and calculate a humidity change trend index based on the humidity sampling data sequence; The processing module is configured to compare the humidity change trend index with a dynamic trend threshold, and assign weights to each humidity sampling point in the monitoring time window according to a comparison result; The processing module is further configured to perform weighted fusion on humidity sampling values of each humidity sampling point in the monitoring time window based on the weights, and obtain a weighted fusion result; The processing module is further configured to generate a humidity control reference value of the current control period based on the weighted fusion result, a current real-time humidity, and the acquired external environment parameters; The execution module is configured to input a deviation between a target humidity set by a crop growth model and the current real-time humidity, and the humidity control reference value of the current control period into a controller, generate a dynamic control instruction, and dynamically adjust an operation parameter of a humidity actuator in the greenhouse according to the dynamic control instruction to achieve precise humidity control.

9. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the agricultural greenhouse humidity intelligent control method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the agricultural greenhouse humidity intelligent control method according to any one of claims 1 to 7.

Images