Plant growth environment monitoring and intelligent regulation and control method and system

Abstract

The invention relates to the technical field of plant intelligent planting, and discloses a plant growth environment monitoring and intelligent regulation and control method and system, and the method comprises the steps: data collection: correspondingly collecting environment parameter data of a plurality of environment factors of a plant growth region through a plurality of sensors; data preprocessing: performing data fusion on the environment parameter data to obtain fused environment data; parameter threshold judgment: performing parameter threshold judgment on the basis of the fused environment data according to a suitable target environment parameter range of plant growth to obtain threshold judgment results corresponding to the plurality of environment factors; intelligent regulation and control: analyzing an environment regulation and control strategy based on a threshold judgment result to obtain a target environment regulation and control mechanism; and intelligent early warning: processing a threshold judgment result through an early warning regulation and control network, dividing early warning levels, and obtaining an early warning regulation and control strategy based on the early warning levels. According to the invention, plant growth environment parameters can be monitored in real time, and parameter data can be automatically regulated and controlled to be within a suitable range.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent plant cultivation, and more specifically, it relates to a method and system for monitoring and intelligent regulation of plant growth environment. Background Art

[0002] The quality of the plant growth environment directly affects the growth quality of plants. For example, soil, climate conditions, irrigation water source, air quality, etc. are all important environmental factors affecting plant quality. Taking Dendrobium huoshanense as an example, as an important Chinese herbal medicine, it has strict regionality and selectivity for the ecological environment. The quality of the environment directly affects the quality of Dendrobium, and the quality of Dendrobium not only directly affects the pharmacological effects and human health, but also involves the quality and safety of traditional Chinese medicine preparations.

[0003] Therefore, in order to ensure the quality and high yield of the Dendrobium industry, it is particularly important to accurately and timely obtain the growth information of Dendrobium and conduct fine management at the appropriate time and in the accurate way and method; establishing a fine monitoring system for the Dendrobium growth environment has become an urgent requirement for the development of modern precision agriculture. Summary of the Invention

[0004] The present invention provides a method and system for monitoring and intelligent regulation of plant growth environment to solve the technical problems mentioned in the related art.

[0005] In a first aspect, the present invention provides a method for monitoring and intelligent regulation of plant growth environment, including:

[0006] Data acquisition: Collecting environmental parameter data of multiple environmental factors in the plant growth area through multiple sensors;

[0007] Data preprocessing: Performing data fusion on the environmental parameter data to obtain fused environmental data;

[0008] Parameter threshold judgment: Based on the fused environmental data and according to the suitable target environmental parameter range for plant growth, performing parameter threshold judgment to obtain threshold judgment results corresponding to multiple environmental factors;

[0009] Intelligent regulation: Analyzing the environmental regulation strategy based on the threshold judgment results to obtain a target environmental regulation mechanism;

[0010] Intelligent warning: Processing the threshold judgment results through a warning regulation network, dividing the warning levels, and obtaining a warning regulation strategy based on the warning levels.

[0011] As a further optimized solution of the present invention, the environmental parameter data includes environmental light intensity value L(t), soil water content value S(t), environmental temperature value T(t), and environmental humidity value H(t), where t is the sampling time;

[0012] The steps of collecting environmental parameter data of multiple environmental factors in the plant growth area include: collecting the environmental light intensity value L(t) through a light intensity sensor, with the unit of lux; collecting the soil water content S(t) through a soil humidity sensor, expressed as the volume water content (%); collecting the environmental temperature T(t) through an air temperature sensor, with the unit of °C; collecting the environmental humidity H(t) through an air humidity sensor, with the unit of %RH.

[0013] As a further optimization scheme of the present invention, the steps of performing data fusion on the environmental parameter data to obtain fused environmental data include:

[0014] Performing denoising processing on the collected environmental parameter data through a wavelet transform model to obtain a denoised data sequence;

[0015] Processing the environmental parameter data in the historical period through a neural network prediction model to obtain the predicted value at the current moment;

[0016] Fusing and processing the denoised data sequence value and the predicted value at the current moment through an ensemble learning model to obtain fused environmental data.

[0017] As a further optimization scheme of the present invention, the wavelet transform model includes: performing 3-layer decomposition on the original data using the db4 wavelet basis function to obtain approximation coefficients and detail coefficients; performing soft threshold processing on the detail coefficients and then reconstructing the signal, and its calculation expression is:

[0018] Y1(t) = WaveletDenoise(X(t));

[0019] Among them, Y(t) is the data sequence after wavelet denoising, X(t) is the original data sequence, WaveletDenoise() is the wavelet denoising function, specifically:

[0020] WaveletDenoise(X) = IDWT(DWT(X, db4, 3), λ);

[0021] Among them, DWT is the discrete wavelet transform, IDWT is the inverse discrete wavelet transform, db4 is the Deaubechies-4 wavelet basis function, λ is the soft threshold, σ is the noise standard deviation, n is the number of data points collected, that is, the sampling times, t is the sampling moment, representing the specific time point of each data collection;

[0022] The neural network prediction model includes: constructing a deep learning model containing 2 LSTM layers and 1 fully connected layer, with the input being the data in the historical period and the output being the predicted value at the current moment, and its calculation expression is:

[0023] Y2(t) = LSTM(X(t - z:t));

[0024] Wherein, X(t - z:t) represents the input data, and t - z:t represents the data interval from the current time step t pushed back z time steps, that is, it includes the data points of the past historical cycle of z time steps.

[0025] As a further optimization scheme of the present invention, the ensemble learning model is used to process the denoised data sequence value and the predicted value at the current time, and output the initial weighted fusion data. Its calculation expression is:

[0026] Y(t) = α·Y1(t) + (1 - α)·Y2(t);

[0027] Where α is the fusion weight;

[0028] By performing normalization processing on the initial weighted fusion data, the fusion environment data is obtained. Its calculation expression is:

[0029]

[0030] Where μ is the average value of historical data and σ is the standard deviation.

[0031] As a further optimization scheme of the present invention, based on the threshold judgment result, the environmental regulation strategy is analyzed to obtain the target environmental regulation mechanism, including: by comparing and analyzing the environmental parameter data corresponding to a single environmental factor with the corresponding threshold judgment result, when the analysis result is that the environmental parameter data exceeds the threshold judgment result, the output target environmental regulation mechanism is to correspondingly single - regulate the environmental parameter data within the threshold judgment result.

[0032] As a further optimization scheme of the present invention, based on the threshold judgment result, the environmental regulation strategy is analyzed to obtain the target environmental regulation mechanism, including:

[0033] Based on multiple environmental parameter data, a state vector space composed of multi - parameter state vectors is constructed. The expression of the multi - parameter state vector is:

[0034] P(t) = [L(t), S(t), T(t), H(t)];

[0035] The state vector space is divided. The state vector space includes a normal region A, an over - limit region B, an insufficient region C, and a mixed region D. Among them, the normal region A means that all parameters are within the appropriate range, the over - limit region B means that at least one parameter exceeds the upper limit, the insufficient region C means that at least one parameter is lower than the lower limit, and the mixed region D means that there are both over - limit and insufficient parameters;

[0036] According to the divided regions of the state vector space, a joint environmental regulation strategy is formulated correspondingly as the target environmental regulation mechanism.

[0037] As a further optimization scheme of the present invention, the steps for formulating the joint environmental regulation strategy include:

[0038] When the multi-parameter state vector is in the over-limit region B, that is, when the multi-environment parameter data exceeds the upper limit, the corresponding environmental parameter data is regulated to within the threshold judgment result according to the following priority: air temperature > air humidity > light intensity > soil humidity;

[0039] When the multi-parameter state vector is in the insufficient region C, that is, when the multi-environment parameter data is lower than the lower limit, the corresponding environmental parameter data is regulated to within the threshold judgment result according to the following priority: air temperature > soil humidity > air humidity > light intensity;

[0040] When the multi-parameter state vector is in the mixed region D, a weighted decision function is used to formulate the joint environmental regulation strategy, where the weighted decision function is:

[0041] D(t) = w1·δ L + w2·δ S + w3·δ T + w4·δ H ;

[0042] where, w i (i = 1, 2, 3, 4) are the weights of each parameter, and δ i is the parameter deviation value, and its calculation formula is:

[0043]

[0044] As a further optimization scheme of the present invention, the threshold judgment result is processed through an early warning regulation network, the early warning level is divided, and an early warning regulation strategy is obtained based on the early warning level, including:

[0045] The early warning regulation network includes an environmental parameter early warning evaluation module for processing environmental parameter data and outputting an early warning level, and its calculation expression is:

[0046]

[0047] where, x is the percentage of the parameter deviating from the median value of the appropriate range, and the expression is: α is the mild early warning threshold, and β is the severe early warning threshold;

[0048] Based on the early warning level data corresponding to multiple environmental parameters, a comprehensive early warning index is calculated, and its expression is:

[0049] WI = max{W(δ L ), W(δS ),W(δ T ),W(δ H )};

[0050] Based on the comprehensive warning index, a warning control strategy is formulated. Among them, WI = 0 indicates normal operation, WI = 1 indicates the triggering of a yellow warning, the buzzer starts intermittent alarm, and WI = 2 indicates the triggering of a red warning, the buzzer starts continuous alarm and sends a text message notification.

[0051] In a second aspect, the present invention also provides a plant growth environment monitoring and intelligent control system for implementing the above-mentioned plant growth environment monitoring and intelligent control method, including:

[0052] A data acquisition module for respectively acquiring environmental parameter data of multiple environmental factors in the plant growth area through multiple sensors;

[0053] A data fusion module for fusing the environmental parameter data to obtain fused environmental data;

[0054] A threshold judgment module for judging parameter thresholds based on the fused environmental data and according to the suitable target environmental parameter range for plant growth to obtain threshold judgment results corresponding to multiple environmental factors;

[0055] An intelligent control module for analyzing the environmental control strategy based on the threshold judgment results to obtain a target environmental control mechanism;

[0056] An intelligent warning module for processing the threshold judgment results through a warning control network, classifying the warning levels, and obtaining a warning control strategy based on the warning levels.

[0057] The beneficial effects of the present invention are as follows: A method and system for monitoring and intelligent regulation of a plant growth environment provided by the present invention integrates multiple sensors such as light intensity, soil humidity, air humidity, and air temperature to real-time monitor multiple key factors of the Dendrobium growth environment and provide accurate data feedback. These data can be used to timely understand the growth status of Dendrobium and avoid poor growth of Dendrobium caused by environmental changes. Through an automatic control system, combined with sensor data, automatic regulation of the irrigation system, temperature and humidity regulation equipment, lighting system, etc. is achieved to solve the problem of certain lag and error existing in traditional manual operations. Through intelligent regulation, the irrigation time and water volume are accurately controlled, combined with soil humidity and climate information, to reduce water resource waste. At the same time, the greenhouse environmental conditions are automatically adjusted to avoid excessive energy consumption and achieve water and energy conservation in agricultural production. The agricultural environment monitoring and intelligent regulation method based on light intensity, soil humidity, air temperature, and humidity solves a series of technical problems such as environmental perception, data analysis, intelligent decision-making, and automatic regulation in the cultivation of Huoshan Dendrobium, improves the cultivation efficiency of Huoshan Dendrobium, reduces resource consumption, and at the same time helps to promote the sustainable development of the Huoshan Dendrobium industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is a flowchart of the steps of a method for monitoring and intelligent regulation of a plant growth environment provided by an embodiment of the present invention;

[0059] Figure 2 is a schematic structural diagram of a system for monitoring and intelligent regulation of a plant growth environment provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] Now, the subject matter described herein will be discussed with reference to example embodiments. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. The functions and arrangements of the elements discussed can be changed without departing from the scope of protection of the content of this specification. Each example can omit, substitute, or add various processes or components as needed. Additionally, the features described in some examples can be combined in other examples.

[0061] Traditional monitoring methods often rely on manual observation or simple sensors, lacking real-time and comprehensiveness, and unable to accurately reflect the changes of multiple environmental factors such as soil, air, and light. Through the above embodiments, by integrating multiple sensors such as light intensity, soil humidity, and air temperature and humidity, multiple key factors of the Dendrobium growth environment are monitored in real time, providing accurate data feedback. These data can promptly understand the growth status of Dendrobium and avoid poor growth of Dendrobium caused by environmental changes. And through the automatic control system, combined with sensor data, automatic regulation of the irrigation system, temperature and humidity adjustment equipment, light system, etc. is realized, solving the certain lag and error existing in traditional manual operation.

[0062] It should be noted that the Chinese herbal medicine mentioned in this application is Dendrobium huoshanense, and the plant growth environment monitoring and intelligent regulation method of this application can also be applied to the cultivation of other Chinese herbal medicines, such as Bletilla striata, Panax ginseng, Gastrodia elata, Anoectochilus roxburghii, Coptis chinensis, etc., which are not limited here.

[0063] Embodiment 1

[0064] As Figure 1 shown, the present invention provides a plant growth environment monitoring and intelligent regulation method, including the following steps:

[0065] Step 1, data collection: Collect environmental parameter data of multiple environmental factors in the plant growth area through multiple sensors;

[0066] The environmental parameter data includes environmental light intensity value L(t), soil water content value S(t), environmental temperature value T(t), and environmental humidity value H(t), where t is the sampling time;

[0067] The step of collecting environmental parameter data of multiple environmental factors in the plant growth area includes: Collecting the environmental light intensity value L(t) through a light intensity sensor, with the unit of lux; Collecting the soil water content S(t) through a soil humidity sensor, expressed as volume water content (%). Collecting the environmental temperature T(t) through an air temperature sensor, with the unit of °C; Collecting the environmental humidity H(t) through an air humidity sensor, with the unit of %RH.

[0068]

[0069] Step 2, data preprocessing: Perform data fusion on the environmental parameter data to obtain fused environmental data;

[0070] The step of performing data fusion on the environmental parameter data to obtain fused environmental data includes:

[0071] Sub-step 201: Denoise the collected environmental parameter data through a wavelet transform model to obtain a denoised data sequence;

[0072] Sub-step 202: Process the environmental parameter data within the historical period through a neural network prediction model to obtain the predicted value at the current moment;

[0073] Sub-step 203: Fuse and process the denoised data sequence value and the predicted value at the current moment through an ensemble learning model to obtain the fused environmental data.

[0074] Among them, the wavelet transform model includes: decomposing the original data into 3 layers using the db4 wavelet basis function to obtain the approximation coefficient and the detail coefficient; reconstructing the signal after performing soft threshold processing on the detail coefficient, and its calculation expression is:

[0075] Y1(t) = WaveletDenoise(X(t));

[0076] Among them, Y(t) is the data sequence after wavelet denoising, X(t) is the original data sequence, and WaveletDenoise() is the wavelet denoising function, specifically:

[0077] WaveletDenoise(X) = IDWT(DWT(X, db4, 3), λ);

[0078] Among them, DWT is the discrete wavelet transform, IDWT is the inverse discrete wavelet transform, db4 is the Deaubechies - 4 wavelet basis function, λ is the soft threshold, σ is the noise standard deviation, n is the number of data points collected, that is, the sampling times, and t is the sampling moment, representing the specific time point of each data collection;

[0079] It should be noted that the approximation coefficient represents the low - frequency part of the signal, reflecting the overall trend of the signal; the detail coefficient represents the high - frequency part of the signal, reflecting the detailed changes of the signal; performing soft threshold processing on the detail coefficient means that when the absolute value of the detail coefficient is greater than the threshold, subtract the threshold from it; when the absolute value of the detail coefficient is less than the threshold, set it to 0.

[0080] The neural network prediction model includes: constructing a deep learning model containing 2 LSTM layers and 1 fully - connected layer, with the input being the data within the historical period and the output being the predicted value at the current moment, and its calculation expression is:

[0081] Y2(t) = LSTM(X(t - z:t));

[0082] Among them, X(t - z:t) represents the input data, and t - z:t represents the data interval from the current time step t backward by z time steps, that is, it includes the data points at z time moments in the past historical period;

[0083] It should be noted that the historical period is a period of time in the past of the t sampling moment. For example, if the historical period z is 24 hours, that is, z takes 24, then t - 24:t represents the data interval from the current time step t backward by 24 time steps. In other embodiments, the historical period z can be 20 hours, 30 hours, 48 hours, etc., which is not limited here.

[0084] The ensemble learning model is used to process the denoised data sequence values and the predicted values at the current time step, and output the initial weighted fusion data. Its calculation expression is:

[0085] Y(t) = α·Y1(t) + (1 - α)·Y2(t);

[0086] Among them, α is the fusion weight; through this fusion method, combining the high - frequency noise suppression ability of wavelet denoising and the time - series prediction advantage of LSTM, the reliability and prediction accuracy of the data are improved, the deviation and error that may be brought by a single method are reduced, and the adaptability and robustness of the system to environmental parameter changes are enhanced. Robustness refers to the ability of the system to still operate stably when environmental parameters fluctuate.

[0087] Among them, for different environmental parameters, Y(t) is the processed data, and the units are respectively: for the light intensity Y L (t), the unit is lux; for the soil humidity Y S (t), the unit is %; for the air temperature Y T (t), the unit is °C; for the air humidity Y H (t), the unit is %RH.

[0088] By performing standardization processing on the initial weighted fusion data, the fused environmental data is obtained. Its calculation expression is:

[0089]

[0090] Among them, μ is the average value of historical data, and σ is the standard deviation. The standardized data follows a standard normal distribution, with a mean of 0 and a standard deviation of 1, which is beneficial for subsequent parameter analysis and regulation.

[0091] Step 3: Parameter threshold judgment: Based on the fused environmental data and according to the suitable target environmental parameter range for plant growth, parameter threshold judgment is performed to obtain the threshold judgment results corresponding to multiple environmental factors;

[0092] In this step, the appropriate target environmental parameter range for plant growth is preset according to the appropriate environmental parameter range for plant growth. For example, the appropriate parameter ranges for each environment of the growth environment of Dendrobium officinale are as follows: the appropriate range of light intensity is 2000 - 20000 lux, the appropriate range of soil humidity is 60% - 80%, the appropriate range of air temperature is 18 - 28 °C, and the appropriate range of air humidity is 70% - 85%. Then, based on these preset appropriate parameter range values, corresponding threshold judgments are made on the processed data value Y(t) above, and a threshold judgment result is obtained. For example, whether the environmental parameter data Y L (t) exceeds 20000 lux, is between 2000 - 20000 lux, or is less than 2000 lux. Another example is whether the environmental parameter data Y H (t) exceeds 85%, is between 70% - 85%, or is less than 70%, etc.

[0093] Step 4, Intelligent Regulation: Analyze the environmental regulation strategy based on the threshold judgment result to obtain the target environmental regulation mechanism;

[0094] Specifically, analyze the environmental regulation strategy based on the threshold judgment result to obtain the target environmental regulation mechanism, including: by comparing and analyzing the environmental parameter data corresponding to a single environmental factor with the corresponding threshold judgment result, when the analysis result shows that the environmental parameter data exceeds the threshold judgment result, the output target environmental regulation mechanism is to correspondingly regulate the environmental parameter data to within the threshold judgment result for the single regulation.

[0095] Exemplarily, when a single environmental parameter data exceeds the appropriate parameter range value, for example, when Y L (t) > 20000 lux, turn off the artificial light source; when Y S (t) > 80%, turn off the irrigation system; when Y T (t) > 28 °C, start the cooling system; when Y H (t) > 85%, turn off the humidification system;

[0096] Exemplarily, when a single environmental parameter data is lower than the appropriate parameter range value, for example, when Y L (t) < 2000 lux, turn on the artificial light source; when Y S (t) < 60%, turn on the irrigation system; when Y T (t) < 18 °C, turn on the heating system; when Y H (t) < 70%, turn on the humidification system.

[0097] Step 5, Intelligent Early Warning: Process the threshold judgment result through the early warning regulation network, divide the early warning level, and obtain the early warning regulation strategy based on the early warning level.

[0098] Specifically: the judgment result of the threshold is processed by the early warning control network, the early warning levels are divided, and the early warning control strategies are obtained based on the early warning levels, including:

[0099] The early warning control network includes an environmental parameter early warning evaluation module for processing environmental parameter data and outputting early warning levels, and its calculation expression is:

[0100]

[0101] Where x is the percentage of the parameter deviating from the median of the suitable range, and the expression is: α is the mild early warning threshold, and β is the severe early warning threshold;

[0102] The comprehensive early warning index is calculated based on the early warning level data corresponding to multiple environmental parameters, and its expression is:

[0103] WI = max{W(δ L ), W(δ S ), W(δ T ), W(δ H )};

[0104] The early warning control strategy is formulated based on the comprehensive early warning index. Among them, WI = 0 indicates normal operation, WI = 1 indicates the trigger of a yellow early warning, the buzzer starts intermittent alarm, WI = 2 indicates the trigger of a red early warning, the buzzer starts continuous alarm and sends a text message notification to the staff.

[0105] It should be noted that α is the mild early warning threshold, which is set to 15% according to the growth characteristics of Dendrobium and expert experience, indicating that when the environmental parameter deviates from the median of the suitable range by more than 15%, a mild early warning is triggered; β is the severe early warning threshold, which is set to 30% according to a large amount of experimental data and the critical growth conditions of Dendrobium, indicating that when the environmental parameter deviates from the median of the suitable range by more than 30%, a severe early warning is triggered, and at this time, the growth of Dendrobium may be seriously affected. In other embodiments, the values of α and β can be set according to the growth characteristics of Dendrobium or other indoor plants.

[0106] Based on the intelligent early warning steps in the above embodiments, the early warning control strategy is obtained, and intelligent early warning is carried out. It can monitor the deviation of each parameter in real time through a mathematical model, and can detect and give early warning in time before the problem expands; moreover, a hierarchical response is set, and a multi-level early warning mechanism (normal, mild, severe) is adopted, and corresponding early warning measures are taken for different severities to avoid false alarms and missed alarms; in addition, automatic decision-making is carried out based on the comprehensive early warning index WI, and the corresponding early warning response can be triggered without manual intervention; and multiple methods such as sound and light alarm and text message notification are combined to ensure that the management personnel can timely learn about abnormal situations; the early warning thresholds α and β can be flexibly adjusted according to actual needs to adapt to the growth requirements of different types of Dendrobium.

[0107] Example 2

[0108] Based on the content of the above embodiments, in step 4, the environmental regulation strategy is analyzed based on the threshold judgment result to obtain the target environmental regulation mechanism, including:

[0109] Construct a state vector space composed of multi-parameter state vectors based on multiple environmental parameter data. The expression of the multi-parameter state vector is:

[0110] P(t) = [L(t), S(t), T(t), H(t)];

[0111] Divide the state vector space. The state vector space includes a normal area A, an over-limit area B, a shortage area C, and a mixed area D. Among them, the normal area A means that all parameters are within the appropriate range, the over-limit area B means that at least one parameter exceeds the upper limit, the shortage area C means that at least one parameter is lower than the lower limit, and the mixed area D means that there are both over-limit and shortage parameters;

[0112] According to the divided areas of the state vector space, formulate a combined environmental regulation strategy as the target environmental regulation mechanism correspondingly.

[0113] The steps for formulating the combined environmental regulation strategy include:

[0114] When the multi-parameter state vector is in the over-limit area B, that is, when the multi-environment parameter data exceeds the upper limit, adjust the corresponding environmental parameter data to within the threshold judgment result according to the following priority: air temperature > air humidity > light intensity > soil humidity;

[0115] It should be noted that this priority setting is based on the following considerations: Temperature is the primary regulation parameter because it has the greatest direct impact on the growth of Dendrobium, and temperature changes will cause chain reactions in other parameters; air humidity ranks second because it is closely related to temperature and has an important impact on the photosynthesis and transpiration of Dendrobium; light ranks third because Dendrobium has a relatively wide adaptation range to light and a relatively fast adjustment response time; the regulation priority of soil humidity is the lowest because its change is relatively slow and Dendrobium has a certain drought tolerance.

[0116] When the multi-parameter state vector is in the shortage area C, that is, when the multi-environment parameter data is lower than the lower limit, adjust the corresponding environmental parameter data to within the threshold judgment result according to the following priority: air temperature > soil humidity > air humidity > light intensity;

[0117] It should be noted that the priority setting is based on the following considerations: Temperature remains the primary control parameter, soil humidity is promoted to the second place because insufficient moisture will directly affect the growth and survival of Dendrobium; air humidity is in the third place as it can be partially improved by increasing soil humidity; light intensity is the last because when the parameters are insufficient, the basic survival conditions should be ensured first.

[0118] When the multi-parameter state vector is in the mixed region D, a weighted decision function is used to formulate a joint environmental control strategy. Among them, the weighted decision function is:

[0119] D(t) = w1·δ L + w2·δ S + w3·δ T + w4·δ H ;

[0120] Among them, w i (i = 1, 2, 3, 4) are the weights of each parameter, and δ i is the parameter deviation value, and its calculation formula is:

[0121]

[0122] Among them, w i is the weight of each parameter, which is adjusted according to actual needs. For example, during the growth period, the weights are: w1 = 0.2, w2 = 0.3, w3 = 0.3, w4 = 0.2; during the flowering period, the weights are: w1 = 0.3, w2 = 0.2, w3 = 0.3, w4 = 0.2; during the dormant period, the weights are: w1 = 0.1, w2 = 0.2, w3 = 0.4, w4 = 0.3.

[0123] The joint environmental control strategy formulated by the weighted decision function takes into account the mutual influence between parameters, reflects the importance of different parameters and their interaction relationships through the weight coefficients, adjusts the weights according to different growth periods of Dendrobium to achieve more precise environmental control, and avoids the frequent on-off regulation caused by parameter fluctuations in the traditional method. In addition, by comprehensively considering the deviation degrees of multiple parameters, the coordinated optimization of the overall environment is realized.

[0124] Specifically, when D(t) > 0, it means that the overall is on the high side, and even within the appropriate range, fine-tuning is required: then a single judgment and regulation are carried out on the parameters corresponding to the following multiple environmental parameter data. For example: if δ T > 0, start the cooling system; if δ H > 0, turn off the humidification system; if δ L > 0, turn off the artificial light source; if δ S > 0, turn off the irrigation system.

[0125] When D(t) < 0, it indicates that the overall value is on the low side, and fine-tuning is required even within the appropriate range. Then, a single judgment and regulation are performed on the parameters corresponding to the following multiple environmental parameter data. For example, if δ T < 0, turn on the heating system; if δ H < 0, turn on the humidifying system; if δ L < 0, turn on the artificial light source; if δ S < 0, turn on the irrigation system.

[0126] When D(t) = 0, the system is in a stable state, and the current control strategy is maintained.

[0127] Embodiment 3

[0128] Continue to refer to Figure 2 , the present invention also provides a plant growth environment monitoring and intelligent regulation system for implementing the above-mentioned plant growth environment monitoring and intelligent regulation method, including:

[0129] A data acquisition module for collecting environmental parameter data of multiple environmental factors in the plant growth area through multiple sensors;

[0130] A data fusion module for fusing the environmental parameter data to obtain fused environmental data;

[0131] A threshold judgment module for performing parameter threshold judgment based on the fused environmental data and according to the appropriate target environmental parameter range for plant growth to obtain threshold judgment results corresponding to multiple environmental factors;

[0132] An intelligent regulation module for analyzing the environmental regulation strategy based on the threshold judgment results to obtain a target environmental regulation mechanism;

[0133] An intelligent early warning module for processing the threshold judgment results through an early warning regulation network, classifying the early warning levels, and obtaining an early warning regulation strategy based on the early warning levels.

[0134] This system simultaneously collects multi-dimensional environmental parameters such as light intensity, soil humidity, air temperature, and air humidity through multiple sensors, realizing the comprehensive monitoring of the plant growth environment. Moreover, it adopts a data fusion method combining wavelet transform denoising and LSTM neural network prediction, effectively improving the reliability and prediction accuracy of environmental data. And based on the threshold judgment results, the system can automatically analyze and generate the optimal environmental regulation strategy, realizing the intelligent adjustment of environmental parameters. In addition, through a multi-level early warning mechanism, the system can timely detect environmental anomalies and trigger corresponding-level early warning responses, effectively preventing the deterioration of the plant growth environment. And the system design has good scalability and customizability, and can flexibly adjust parameter thresholds and control strategies according to the growth requirements of different plants.

[0135] The embodiments of the present invention have been described above, but these embodiments are not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative rather than restrictive. Under the inspiration of this embodiment, those of ordinary skill in the art can also make more equivalent embodiments in various forms, all of which fall within the protection scope of this embodiment.

Claims

1. A method for monitoring and intelligent regulation of plant growth environment, characterized in that, Including: Data acquisition: Collecting environmental parameter data of multiple environmental factors in the plant growth area through multiple sensors; Data preprocessing: Performing data fusion on the environmental parameter data to obtain fused environmental data; Parameter threshold judgment: Based on the fused environmental data and according to the suitable target environmental parameter range for plant growth, performing parameter threshold judgment to obtain threshold judgment results corresponding to multiple environmental factors; Intelligent regulation: Analyzing the environmental regulation strategy based on the threshold judgment results to obtain the target environmental regulation mechanism; Intelligent warning: Processing the threshold judgment results through the warning regulation network, dividing the warning levels, and obtaining the warning regulation strategy based on the warning levels.

2. The method for monitoring and intelligently regulating the plant growth environment according to claim 1, characterized in that, The environmental parameter data includes the environmental light intensity value L(t), the soil water content value S(t), the environmental temperature value T(t), and the environmental humidity value H(t), where t is the sampling time; The steps of collecting environmental parameter data of multiple environmental factors in the plant growth area include: Collecting the environmental light intensity value L(t) through a light intensity sensor, with the unit of lux; Collecting the soil water content S(t) through a soil humidity sensor, expressed in volume moisture content (%); Collecting the environmental temperature T(t) through an air temperature sensor, with the unit of °C; Collecting the environmental humidity H(t) through an air humidity sensor, with the unit of %RH.

3. The method for monitoring and intelligently regulating a plant growth environment according to claim 2, characterized in that, The steps of performing data fusion on the environmental parameter data to obtain fused environmental data include: Performing denoising processing on the collected environmental parameter data through a wavelet transform model to obtain a denoised data sequence; Processing the environmental parameter data within the historical period through a neural network prediction model to obtain the predicted value at the current moment; Fusing and processing the denoised data sequence value and the predicted value at the current moment through an ensemble learning model to obtain the fused environmental data.

4. A method for monitoring and intelligent regulation of a plant growth environment according to claim 3, characterized in that, The wavelet transform model includes: Performing 3-layer decomposition on the original data using the db4 wavelet basis function to obtain the approximation coefficient and the detail coefficient; Reconstructing the signal after performing soft threshold processing on the detail coefficient, and its calculation expression is: Y1(t) = WaveletDenoise(X(t)); Where, Y(t) is the data sequence after wavelet denoising, X(t) is the original data sequence, WaveletDenoise() is the wavelet denoising function, specifically: WaveletDenoise(X) = IDWT(DWT(X, db4, 3), λ); Among them, DWT is the discrete wavelet transform, IDWT is the inverse discrete wavelet transform, db4 is the Deaubechies-4 wavelet basis function, λ is the soft threshold, σ is the standard deviation of the noise, n is the number of data points collected, that is, the number of sampling times, and t is the sampling time, representing the specific time point of each data collection; The neural network prediction model includes: Constructing a deep learning model containing 2 LSTM layers and 1 fully connected layer, with the input being the data within the historical period and the output being the predicted value at the current moment, and its calculation expression is: Y2(t) = LSTM(X(t - z:t)); Where, X(t - z:t) represents the input data, and t - z:t represents the data interval from z time steps before the current moment t, that is, it contains the data points of the past historical period of z moments.

5. The method for monitoring and intelligent regulation of a plant growth environment according to claim 4, wherein The ensemble learning model is used to process the denoised data sequence value and the predicted value at the current moment, and output the initial weighted fusion data, and its calculation expression is: Y(t) = α·Y1(t) + (1 - α)·Y2(t); where α is the fusion weight; By normalizing the initial weighted fusion data, the fused environmental data is obtained, and its calculation expression is: where μ is the average value of historical data and σ is the standard deviation.

6. The method for monitoring and intelligently regulating a plant growth environment according to claim 5, characterized in that, Based on the threshold judgment result, the environmental regulation strategy is analyzed to obtain the target environmental regulation mechanism, including: by comparing and analyzing the environmental parameter data corresponding to a single environmental factor with the corresponding threshold judgment result, when the analysis result shows that the environmental parameter data exceeds the threshold judgment result, the output target environmental regulation mechanism is to regulate the corresponding environmental parameter data to within the threshold judgment result singly.

7. A method for monitoring and intelligent regulation of a plant growth environment according to claim 5, characterized in that, Based on the threshold judgment result, the environmental regulation strategy is analyzed to obtain the target environmental regulation mechanism, including: Based on multiple environmental parameter data, a state vector space composed of multi-parameter state vectors is constructed, and the expression of the multi-parameter state vector is: P(t) = [L(t), S(t), T(t), H(t)]; The state vector space is divided. The state vector space includes a normal region A, an over-limit region B, an under-limit region C, and a mixed region D. Among them, the normal region A means that all parameters are within the appropriate range, the over-limit region B means that at least one parameter exceeds the upper limit, the under-limit region C means that at least one parameter is lower than the lower limit, and the mixed region D means that there are both over-limit and under-limit parameters; According to the divided regions of the state vector space, a joint environmental regulation strategy is formulated correspondingly as the target environmental regulation mechanism.

8. A method for monitoring and intelligent regulation of a plant growth environment according to claim 7, characterized in that, The steps for formulating the joint environmental regulation strategy include: When the multi-parameter state vector is in the over-limit region B, that is, when the multi-environmental parameter data exceeds the upper limit, the corresponding environmental parameter data is regulated to within the threshold judgment result according to the following priority: air temperature > air humidity > light intensity > soil humidity; When the multi-parameter state vector is in the under-limit region C, that is, when the multi-environmental parameter data is lower than the lower limit, the corresponding environmental parameter data is regulated to within the threshold judgment result according to the following priority: air temperature > soil humidity > air humidity > light intensity; When the multi-parameter state vector is in the mixed region D, a weighted decision function is used to formulate a joint environmental regulation strategy, where the weighted decision function is: D(t) = w1·δ L + w2·δ S + w3·δ T + w4·δ H ; Among them, w i (i = 1, 2, 3, 4) are the weights of each parameter, and δ i is the parameter deviation value, and its calculation formula is:

9. A method for monitoring and intelligent regulation of a plant growth environment according to claim 6 or 8, characterized in that, The threshold judgment result is processed by the warning regulation network to divide the warning levels, and a warning regulation strategy is obtained based on the warning levels, including: The warning regulation network includes an environmental parameter warning evaluation module for processing environmental parameter data and outputting warning levels, and its calculation expression is: where x is the percentage of the parameter deviating from the median of the appropriate range, and the expression is: α is the mild warning threshold, and β is the severe warning threshold; Based on the warning level data corresponding to multiple environmental parameters, a comprehensive warning index is calculated, and its expression is: WI = max{W(δ L ), W(δ S ), W(δ T ), W(δ H )}; Based on the comprehensive warning index, a warning regulation strategy is formulated. Among them, WI = 0 means normal operation, WI = 1 means triggering a yellow warning, starting intermittent alarm of the buzzer, and WI = 2 means triggering a red warning, starting continuous alarm of the buzzer and sending a text message notification.

10. A plant growth environment monitoring and intelligent regulation system, characterized in that, For implementing a method for monitoring and intelligent regulation of a plant growth environment according to any one of claims 1-9, including: A data acquisition module for collecting environmental parameter data of multiple environmental factors in a plant growth area through multiple sensors; A data fusion module for fusing the environmental parameter data to obtain fused environmental data; A threshold judgment module, which is used to perform parameter threshold judgment based on the fused environmental data and according to the suitable target environmental parameter range for plant growth, so as to obtain threshold judgment results corresponding to multiple environmental factors; An intelligent regulation module, which is used to analyze the environmental regulation strategy based on the threshold judgment result to obtain a target environmental regulation mechanism; An intelligent warning module, which is used to process the threshold judgment result through a warning regulation network, divide the warning level, and obtain a warning regulation strategy based on the warning level.

Images