AI plant lamp spectrum adjusting method for plant photosynthesis optimization

Abstract

The invention relates to the technical field of spectrum adjustment, and discloses an AI plant lamp spectrum adjustment method for plant photosynthesis optimization, and the method comprises the steps: collecting original plant physiological data through an intelligent spectrum sensor network, and carrying out the preprocessing, and obtaining a plant spectrum interaction feature data set; constructing a three-dimensional spectral response mathematical model by using the plant spectral interaction feature data set; taking the three-dimensional spectral response mathematical model as a teacher network, and training a student network through knowledge distillation to obtain a plant spectrum matching network; processing the real-time plant physiological data through a plant spectrum matching network, executing plant growth stage evaluation, and generating a plant growth stage identifier, a personalized spectrum demand parameter and an illumination time sequence demand parameter; according to the method, real-time correction can be carried out according to the actual spectrum deviation and the plant photosynthetic efficiency deviation, and the accuracy and stability of spectrum adjustment are guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of spectrum regulation technology, and in particular to an AI plant lamp spectrum regulation method for optimizing plant photosynthesis. Background Art

[0002] Plant photosynthesis is the core process of plant growth, and its efficiency directly affects crop yield and quality. Traditional plant spectrum adjustment systems usually use fixed spectrum formulas, which cannot be dynamically adjusted according to the plant growth stage and real-time physiological state, resulting in low efficiency in light energy utilization and difficulty in meeting the specific spectrum requirements of different plant varieties at different growth stages.

[0003] Existing plant spectral regulation technologies face multiple challenges. Due to the complex light scattering, reflection, and absorption characteristics within plant tissues, different plant species have significant differences in their responses to the spectrum, and existing systems find it difficult to accurately capture such differences and make corresponding adjustments. Secondly, plant photosynthesis is highly sensitive to environmental conditions, and spectral requirements vary with growth stages, circadian rhythms, and external environmental changes. Existing technologies lack the ability to adapt to such dynamic changes. In addition, the spectral regulation process involves multi-objective optimization problems. How to strike a balance between photosynthetic efficiency, energy consumption, and economic costs has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The present invention provides an AI plant lamp spectrum adjustment method for optimizing plant photosynthesis. The present invention can perform real-time correction according to actual spectrum deviation and plant photosynthetic efficiency deviation to ensure the accuracy and stability of spectrum adjustment.

[0005] In a first aspect, the present invention provides an AI plant lamp spectrum adjustment method for optimizing plant photosynthesis, the AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis comprising: The original plant physiological data is collected and preprocessed through the intelligent spectral sensor network to obtain the plant spectral interaction feature dataset; Using the plant spectral interaction feature data set to construct a three-dimensional spectral response mathematical model; The three-dimensional spectral response mathematical model is used as a teacher network, and a student network is trained through knowledge distillation to obtain a plant spectrum matching network; Processing real-time plant physiological data through the plant spectrum matching network, performing plant growth stage assessment, and generating plant growth stage identifiers, personalized spectrum requirement parameters, and light timing requirement parameters; A differential co-evolution algorithm is executed according to the plant growth stage identifier, the personalized spectrum requirement parameter and the light timing requirement parameter to generate an LED dimming solution.

[0006] In the technical solution provided by the present invention, the complex three-dimensional spectral response mathematical model is compressed into a lightweight plant spectral matching network through the knowledge distillation technology, so that the number of student network parameters is reduced, the computing resource requirements are significantly reduced, and the model accuracy is maintained, so that the system can operate efficiently in a resource-constrained environment. A plant growth stage evaluator with a three-unit parallel structure is used to comprehensively analyze morphological, physiological and metabolite indicators, and the timing characteristics are processed by feature fusion and gated recurrent unit network to achieve accurate identification of typical plant growth stages. A complete optimization model is constructed through four objective functions (spectral matching, photosynthetic efficiency, energy consumption, and economic cost). Combined with the differential co-evolution algorithm and adaptive mutation operation, energy consumption and economic cost are taken into account while ensuring photosynthetic efficiency, and multi-dimensional balanced optimization is achieved. The hierarchical spectral adjustment controller combines real-time spectral monitoring and plant response monitoring to form a closed-loop feedback control system, which can perform real-time correction according to the actual spectral deviation and plant photosynthetic efficiency deviation to ensure the accuracy and stability of spectral adjustment. When a light source failure, environmental abnormality or plant stress condition is detected, the system automatically switches to a preset safe spectral mode to avoid irreversible damage to plants caused by improper spectral, thereby improving system reliability and safety. The personalized adjustment coefficient is calculated according to the individual characteristics of the plants, and the spectral requirements of different plant individuals are customized to improve the pertinence and accuracy of spectral adjustment and meet the specific spectral requirements of different plant varieties and individuals.

[0007] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0008] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 This is a schematic diagram of an embodiment of an AI plant lamp spectrum adjustment method for optimizing plant photosynthesis in an embodiment of the present invention. DETAILED DESCRIPTION

[0010] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0011] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device end including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or device ends.

[0012] To facilitate understanding of this embodiment, firstly, a method for adjusting the spectrum of an AI plant lamp for optimizing plant photosynthesis disclosed in an embodiment of the present invention is described in detail. Figure 1 As shown, the method comprises the following steps: 101. Collect original plant physiological data through an intelligent spectral sensor network and perform preprocessing to obtain a plant spectral interaction feature data set; It is understandable that the execution subject of the present invention may be an AI plant light spectrum adjustment device for optimizing plant photosynthesis, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0013] Specifically, spectral analysis sensors, plant physiological parameter sensors and growth parameter sensors are configured in the plant growth environment to form an intelligent spectral sensor network. The function of the spectral analysis sensor is to monitor the spectral conditions in the plant's surrounding environment in real time, record the light intensity of different wavelengths, and the impact of these light conditions on the plant. The plant physiological parameter sensor is used to detect the real-time physiological state of the plant, such as physiological indicators such as photosynthesis rate and chlorophyll content. These data help to understand how plants respond to different light conditions. The growth parameter sensor is used to monitor the growth of plants, including parameters directly related to plant growth such as plant height, number of leaves, and leaf surface area. Through the collaborative work of these sensors, the growth and physiological response data of plants under different light conditions are comprehensively collected to form an intelligent spectral sensor network. Through the intelligent spectral sensor network, the original plant physiological data containing spectral conditions and corresponding plant physiological responses are collected. These data include the output of multiple sensors, such as the light intensity of different wavelengths at a specific time, the physiological response of plants to these spectral conditions (such as photosynthesis rate, plant health status, etc.) and the growth status. The original plant physiological data is detected and corrected for outliers to identify and eliminate data points that do not conform to the normal growth law of plants. Outlier detection is achieved through statistical methods such as standard deviation detection, Z-score detection or machine learning-based algorithms. The detected outliers are caused by external environmental interference or sensor errors. They are repaired by linear interpolation or other correction algorithms to ensure the smoothness and consistency of the data and obtain the cleaned plant physiological data. Multi-scale time regularization is performed on the cleaned plant physiological data. The regularization algorithm is used to process the fluctuations of different time scales in the data, making the data more stable in time. The growth and physiological response of plants is a long-term process, but it will be affected by short-term environmental fluctuations (such as instantaneous changes in light intensity). Time regularization helps to smooth out these short-term fluctuations, making the data more consistent with the long-term growth law of plants. Through multi-scale time regularization, the information of different time scales in the data is effectively balanced to obtain standardized time series data. The characteristic points of the spectral response curve, photosynthetic efficiency index and growth phenotypic characteristics are extracted from the standardized time series data. These characteristic points represent the response of plants under specific light conditions. The characteristic points of the spectral response curve include the reaction intensity under different light wavelengths, which can reflect the photosynthetic efficiency and absorption characteristics of plants under different wavelengths of light. The photosynthetic efficiency index is used to quantitatively measure the photosynthetic efficiency of plants under different lighting conditions, while the growth phenotypic characteristics include the morphological indicators of plants, such as plant height, number of leaves, leaf area, etc. Through these extracted features, the physiological performance of plants under different lighting conditions is fully characterized. The plant physiological feature vectors are paired with the corresponding spectral condition data in time series to form a plant spectral interaction feature dataset.The plant physiological feature vector contains information about the photosynthetic efficiency and growth phenotype of plants under different light conditions, while the spectral condition data contains information about the actual light conditions in the environment. By pairing the two in time series, the plant spectral interaction feature dataset formed comprehensively describes the response of plants to spectral conditions.

[0014] 102. Construct a three-dimensional spectral response mathematical model using plant spectral interaction feature dataset; Specifically, the spectral parameters are discretized. The complex light conditions are converted into processable discrete data. The wavelength range of the spectrum is usually continuous, and it is discretized for calculation and modeling. Assuming that the wavelength range is from 300nm to 800nm, the range is divided into A bands, where each band represents a certain range of light wavelengths. The range of light intensity is discretized, and the light intensity is divided into B levels to describe light of different intensities. The light time dimension is also an important parameter. The demand of plants for light time is time-varying, and the time is divided into C time periods. The entire light condition space is described as a three-dimensional discrete parameter space, in which each point consists of light wavelength, light intensity and light time. Based on the plant spectral interaction feature dataset, the corresponding photosynthetic efficiency index is calculated for each three-dimensional coordinate point. Photosynthetic efficiency is the ability of plants to photosynthesize under specific light conditions, which is a key indicator of spectral response. The initial spectral response data is obtained by calculating the photosynthetic efficiency index for each three-dimensional coordinate point (i.e., a specific wavelength, light intensity and time combination) in the spectral interaction feature dataset. These photosynthetic efficiency indices can accurately reflect the photosynthetic effect of plants under different light conditions. The initial spectral response data is interpolated with three-dimensional thin plate splines, and the response data is smoothed and completed by the interpolation method. Three-dimensional thin plate spline interpolation is a multidimensional data smoothing method that generates a smooth response surface based on discrete response data points. This interpolation process generates a continuous and smooth response surface by considering the influence of adjacent data points, allowing the system to smoothly transition between light conditions. In this way, a complete response surface is obtained to describe the changes in photosynthesis of plants under different light conditions. In order to improve the accuracy of the model, a growth stage correction factor is introduced. Different plants have different requirements for light at different growth stages, so the introduction of the growth stage correction factor helps the model adjust the spectral response according to the current growth stage of the plant. This allows the response surface to not only consider the light conditions, but also reflect the growth and development of the plant, resulting in a four-dimensional extended model. The four-dimensional extended model includes the dimension of the plant's development stage, which can dynamically adjust the photosynthetic efficiency of the plant while considering the light conditions. The four-dimensional extended model can more accurately describe the response of plants to light conditions at different growth stages. In order to reduce the complexity of the model and improve the computational efficiency, principal component analysis is performed on the four-dimensional extended model. Principal component analysis is a dimensionality reduction method that converts the original high-dimensional data into low-dimensional data through linear transformation while retaining the important features in the data as much as possible. Through principal component analysis, the dimensions in the four-dimensional extended model are reduced to lower dimensions while retaining the main features of the spectral response. After principal component analysis, the spectral response characteristics after dimension reduction are obtained. The deep autoencoder network is used to perform nonlinear mapping on the spectral response characteristics after dimension reduction to obtain a three-dimensional spectral response mathematical model.An autoencoder is an unsupervised learning algorithm that compresses input data into a low-dimensional space through an encoder and then restores it through a decoder. A deep autoencoder is constructed through a multi-layer neural network, which enables it to learn nonlinear relationships in the data. By training the deep autoencoder network, the system automatically extracts the complex response pattern of plants to light conditions from the spectral response features after dimensionality reduction and constructs a three-dimensional spectral response mathematical model.

[0015] 103. The three-dimensional spectral response mathematical model is used as the teacher network, and the student network is trained through knowledge distillation to obtain the plant spectrum matching network; Specifically, based on the three-dimensional spectral response mathematical model, a teacher network and a student network are constructed. The teacher network consists of 12 fully connected layers, each of which is a computing unit for feature learning and transformation of input data. This deep network structure can effectively extract deep features from complex spectral response data and adapt to complex light regulation tasks. The student network adopts a 5-layer fully connected structure, which is simpler than the teacher network in structure, so that the student network can perform calculations and reasoning more efficiently. Especially in practical applications, the low computational complexity of the student network can improve the operating efficiency of the system. The plant physiological characteristic data and the spectral response model parameters are parallelly extracted to obtain plant physiological feature vectors and spectral response feature vectors. The plant physiological characteristic data contains the physiological responses of plants under different light conditions, such as photosynthesis rate, chlorophyll content, plant height, etc., while the spectral response model parameters describe the response characteristics of plants under different light conditions. The two feature sets represent the relationship between the physiological state of the plant and the external light conditions, respectively. Through parallel feature extraction, the different information of the two feature sets is extracted. The alternating attention mechanism is realized by calculating the similarity matrix between the plant physiological feature vector and the spectral response feature vector, and the aligned feature representation is obtained. The calculation of the similarity matrix can reflect the correlation between the physiological state of the plant and the spectral response, and reveal which spectral conditions are more suitable for the current physiological needs of the plant. The alternating attention mechanism is an optimization algorithm that iteratively adjusts the relationship between two feature vectors to achieve feature alignment. This process can help the system find the optimal match between the physiological characteristics of the plant and the spectral response characteristics, so that the spectral regulation can more accurately meet the growth needs of the plant. The direct splicing method and the cross-modal self-attention mechanism are used to perform multimodal fusion on the aligned feature representations. The cross-modal self-attention mechanism weights the features of different modalities through the self-attention mechanism to learn the correlation between different modalities. The self-attention mechanism enables the model to automatically focus on the features that are most important for spectral regulation, so that the effect of feature fusion is more accurate. Through the above fusion method, the fusion features obtained contain the joint representation of plant physiological information and spectral response information. Based on the fusion features, the teacher network is trained to generate high-precision spectral regulation parameters, and the student network is trained to generate regulation parameters based on the compressed features. The teacher network learns the optimal regulation strategy of plants under different light conditions through a deep network structure and outputs high-precision spectral regulation parameters. These parameters are used as the knowledge output of the teacher network to guide the training of the student network. The student network generates relatively simplified spectral adjustment parameters by learning the compression features. These parameters effectively reduce the computational burden and are suitable for practical application scenarios. During the training process, the soft target knowledge distillation method is used to construct the loss function, so that the student network can learn to generate spectral adjustment parameters by imitating the behavior of the teacher network.Soft target knowledge distillation passes the output of the teacher network as a soft label to the student network, thereby guiding the student network to learn the knowledge of the teacher network at a lower computational complexity. The loss function optimizes the parameters of the student network by calculating the difference between the output of the student network and the output of the teacher network, making it as close to the output of the teacher network as possible. The student network parameters are optimized and adjusted to obtain the final plant spectrum matching network. Through optimization and adjustment, the student network gradually approaches the performance of the teacher network while maintaining computational efficiency. The obtained plant spectrum matching network can dynamically adjust the spectral output according to the physiological characteristics and light conditions of the plant to achieve the best optimization of plant photosynthesis.

[0016] 104. Processing real-time plant physiological data through a plant spectrum matching network, and performing plant growth stage evaluation to generate plant growth stage identifiers, personalized spectrum requirement parameters, and light timing requirement parameters; Specifically, a plant growth stage evaluator is constructed, which includes a composite module of three parallel units: a morphological evaluation unit, a physiological evaluation unit, and a metabolite evaluation unit. The morphological evaluation unit evaluates the growth stage of the plant by collecting morphological data of the plant, such as the height, leaf area, and number of branches of the plant. The physiological evaluation unit evaluates the health status and growth progress of the plant by monitoring the physiological parameters of the plant, such as the photosynthesis rate and chlorophyll content. The metabolite evaluation unit evaluates the metabolic state and energy reserve of the plant by detecting metabolites in the plant, such as sugars and amino acids. Through these three parallel evaluation units, the growth stage of the plant is accurately evaluated by integrating various information. The collected real-time plant physiological data is processed using a plant spectral matching network to obtain multi-stage feature vectors. These feature vectors represent the physiological state and light requirements of the plant at different growth stages and under different environmental conditions. By performing feature fusion on these feature vectors, the feature information of each stage is merged into a comprehensive plant feature representation. Through the plant growth stage evaluator, the obtained comprehensive plant feature representation is mapped to the probability distribution of D typical plant growth stages. This mapping process uses the model of the plant growth stage evaluator, takes the comprehensive plant feature representation as input, and calculates the probability distribution of the plant belonging to each growth stage. By selecting the growth stage corresponding to the maximum probability value, the growth stage of the plant is determined. This process is based on the comprehensive evaluation of the plant's physiological characteristics, morphological indicators and metabolite data, so it has high accuracy and can reflect the actual growth status of the plant in real time. After determining the plant's growth stage identifier, the initial spectral demand parameters are extracted from the preset stage-spectral demand mapping table based on the identifier. The stage-spectral demand mapping table is established through a large amount of experimental data and theoretical research on plant photosynthesis. The table lists the parameters such as the proportion of plant requirements for different wavelength spectra, the required light intensity, and the light cycle at different growth stages. According to the plant's growth stage identifier, the initial spectral demand parameters corresponding to the plant's current growth stage are extracted from the mapping table. Using the plant spectral matching network, the initial spectral demand parameters are adjusted according to the current individual characteristics of the plant to generate personalized spectral demand parameters and light timing demand parameters. The light demand of plants is not only affected by the growth stage, but also closely related to their individual characteristics, such as plant species, health status, leaf area and other factors. Therefore, through the plant spectrum matching network, combined with the individual characteristic data of the plant, the initial spectrum requirement parameters are adjusted to generate personalized spectrum requirement parameters and light timing requirement parameters. The personalized spectrum requirement parameters include the specific demand ratio of plants for light of different wavelengths, while the light timing requirement parameters include the light intensity and light cycle required by the plants. These parameters help the system accurately adjust the spectral output of the light source to ensure that the plants obtain the best lighting conditions at different growth stages.

[0017] 105. Execute a differential co-evolution algorithm according to the plant growth stage identifier, the personalized spectrum requirement parameter and the light timing requirement parameter to generate an LED dimming solution.

[0018] Specifically, a set of optimization variables is defined, and the basic variables for spectrum optimization are obtained. The basic variables for spectrum optimization include parameters such as light intensity, spectrum ratio, and light cycle. These variables are the core elements for adjusting the spectral output of LED lamps. The selection of optimization variables directly affects the optimization effect of the dimming scheme. Therefore, it is necessary to ensure that these variables cover all the key factors required for spectrum regulation. A multi-objective optimization model is constructed based on plant growth stage identifiers, personalized spectrum demand parameters, and light timing demand parameters. The model includes multiple objective functions, such as spectrum matching objective function, photosynthetic efficiency objective function, energy consumption objective function, and economic cost objective function. These four objective functions describe the optimization requirements of different aspects of the light regulation process. The spectrum matching objective function is used to measure the degree of match between the actual light conditions and the plant demand spectrum to ensure that the plants can obtain the most suitable light conditions; the photosynthetic efficiency objective function evaluates the impact of light regulation on plant photosynthesis, with the goal of maximizing photosynthesis efficiency; the energy consumption objective function ensures the energy saving of the dimming scheme, and reduces unnecessary energy waste by controlling the light intensity and cycle; the economic cost objective function focuses on the economy of the dimming scheme, ensuring that the overall operating cost is reduced while meeting the needs of plants. Through the design of these four objective functions, the lighting requirements, energy consumption and economic costs of plants are optimized at the same time. Constraints are set for the multi-objective optimization model to obtain a complete optimization problem. Constraints include power limits of LED light sources, maximum and minimum values ​​of light intensity, and restrictions on photoperiods. These restrictions ensure that the optimization process is feasible in practical applications and does not exceed physical and economic limits while meeting plant needs. According to the basic variables of spectral optimization, the differential coevolution algorithm is performed on the complete optimization problem. The differential coevolution algorithm is an optimization method based on population evolution, which simulates the evolution process of populations in nature to find the optimal solution. According to the basic variables of spectral optimization, the differential coevolution algorithm is performed on the complete optimization problem to generate an initial evolution population. The initial evolution population consists of a group of randomly generated individuals (i.e., possible spectral adjustment schemes), which will undergo subsequent evolution operations. The fitness of each individual is measured by its performance in terms of spectral matching, photosynthetic efficiency, energy consumption and economic cost. Evolution operations are performed on the initial evolution population. Evolution operations include steps such as selection, crossover, and mutation. In this process, an elite retention strategy based on non-dominated sorting is adopted to ensure that the best individuals in each generation can be passed to the next generation. This strategy selects those individuals that are relatively superior in each goal and retains them by comparing the performance of each individual on multiple goals. The introduction of adaptive mutation operation is to enhance the diversity of the search process and prevent the algorithm from falling into the local optimal solution. Adaptive mutation can automatically adjust the mutation rate according to the current state of the population, making the exploration in the search space more uniform and avoiding premature convergence to a certain solution.As evolution proceeds, the differential coevolution algorithm gradually generates a target population, which are individuals that perform best in spectrum adjustment after multiple generations of evolution. Select a non-dominated solution set from the target population. A non-dominated solution set means that among all objective functions, no solution is better than other solutions in all objectives. By selecting these non-dominated solutions, the system finds the best dimming solution from multiple valid solutions. The non-dominated solution set obtained by analysing the fuzzy comprehensive evaluation method is used to generate the final LED dimming solution. The fuzzy comprehensive evaluation method is a multi-criteria decision-making method that can handle the complex relationship between multiple objectives and integrate the results of multiple objective functions into a final decision result by weighted averaging and other methods. Fuzzy evaluation can tolerate uncertainty and ambiguity to a certain extent, so that the final dimming solution can not only balance multiple objectives such as spectral matching, photosynthetic efficiency, energy consumption and economic cost, but also adapt to the changing needs in practical applications.

[0019] The design of the base layer realizes LED drive control through pulse width modulation (PWM) technology. PWM technology adjusts the brightness of LED by controlling the duty cycle of current switching. The light intensity of each LED channel is determined by calculating its PWM duty cycle. The light intensity basic control parameters generated by this process provide preliminary brightness adjustment for each LED channel. These basic control parameters are determined by the light timing demand parameters (such as sunlight cycle, light intensity change, etc.) and the plant demand model. The base layer is mainly responsible for precise control of LED driving to ensure that the brightness and light timing of each LED channel meet the growth needs of plants. In the adaptation layer, adaptive adjustment is performed on the light intensity basic control parameters to cope with the dynamic changes between light intensity and plant needs. Under different plant growth stages and environmental conditions, the light requirements of plants are constantly changing, so the basic control parameters are adaptively adjusted. The adjustment of the adaptation layer is not only based on the basic control parameters of light intensity, but also needs to be combined with the light timing demand parameters for daily spectrum adjustment, which means that the light requirements of plants are different during the day and night and in different time periods. The adaptation layer ensures that plants always get suitable lighting conditions by adjusting lighting parameters in real time. These adjustments reflect the changes in plant growth requirements. In the optimization layer, the adjustment parameters of the adaptation layer are optimized in the long term. The adjustment scheme is optimized according to the long-term growth data of the plant, environmental changes, and the long-term trend of light requirements. The optimization layer mainly focuses on long-term adjustment based on multiple goals such as the overall photosynthesis efficiency of the plant, spectral matching, and energy consumption. The parameters of the adaptation layer are continuously optimized through the accumulated historical data, so that the final spectral adjustment is more in line with the light requirements and growth laws of the plant, and ensures that the plant can maintain efficient photosynthesis under different climatic conditions. In order to verify whether these adjustment parameters are effective, the actual output spectrum is measured with the help of the spectrum monitoring unit according to the spectral requirement parameters adjusted by the optimization layer. The spectrum monitoring unit captures the spectral output of the LED lamp in real time and compares the deviation between the actual output spectrum and the target spectrum. This process can ensure that the adjusted spectrum meets the photosynthesis requirements of the plant and provide data support for subsequent adjustments. At the same time, the plant response monitoring unit is used to collect real-time response data of chlorophyll fluorescence and photosynthetic rate, which reflect the physiological response of the plant under the current light conditions. Chlorophyll fluorescence is the fluorescent response of plants to light stimulation, and its intensity is closely related to photosynthesis efficiency; while photosynthetic rate indicates the photosynthesis efficiency of plants under specific light conditions. By collecting these plant response data in real time, the deviation between the actual spectrum and the target spectrum, as well as the deviation between the actual photosynthetic efficiency of the plant and the expected efficiency are calculated. The deviation value provides a feedback signal to the system to form a closed-loop control deviation signal. Based on these deviation signals, the system generates correction coefficients to correct the adaptation layer adjustment parameters. The correction coefficients can be adjusted according to real-time feedback to ensure the accuracy and stability of the system in practical applications.When the system detects a light source failure, an abnormal environment, or a plant under stress, it can automatically switch to the preset safe spectrum mode. The safe spectrum mode is the basic lighting condition provided by the system to protect the health of plants under abnormal circumstances, to prevent plants from continuing to grow in an unsuitable lighting environment, thereby preventing plant damage. Through this series of control processes, the system adjusts the spectral output in real time to ensure that plants can obtain appropriate light at all growth stages. The hierarchical control structure of the basic layer, adaptation layer, and optimization layer can effectively combine the real-time needs of plants and environmental changes, making plant light regulation more refined and personalized. Through closed-loop control and real-time feedback, continuous optimization of light regulation is achieved to ensure that plants can achieve optimal photosynthesis efficiency in any environment, promoting healthy growth and efficient production of plants.

[0020] In the embodiment of the present invention, the complex three-dimensional spectral response mathematical model is compressed into a lightweight plant spectral matching network through knowledge distillation technology, so that the number of student network parameters is reduced, the computing resource requirements are significantly reduced, and the model accuracy is maintained, so that the system can run efficiently in a resource-constrained environment. A plant growth stage evaluator with a three-unit parallel structure is used to comprehensively analyze morphological, physiological and metabolite indicators, and the timing characteristics are processed by feature fusion and gated recurrent unit network to achieve accurate identification of typical plant growth stages. A complete optimization model is constructed through four objective functions (spectral matching, photosynthetic efficiency, energy consumption, and economic cost). The differential co-evolution algorithm and adaptive mutation operation are combined to ensure photosynthetic efficiency while taking into account energy consumption and economic costs, and realize multi-dimensional balanced optimization. The hierarchical spectral adjustment controller combines real-time spectral monitoring and plant response monitoring to form a closed-loop feedback control system, which can perform real-time correction according to the actual spectral deviation and plant photosynthetic efficiency deviation to ensure the accuracy and stability of spectral adjustment.

[0021] In a specific embodiment, the process of executing step 101 may specifically include the following steps: Configure spectral analysis sensors, plant physiological parameter sensors and growth parameter sensors in the plant growth environment to form an intelligent spectral sensor network; Collect raw plant physiological data including spectral conditions and corresponding plant physiological responses through a network of intelligent spectral sensors; Perform outlier detection and outlier correction on the original plant physiological data to obtain cleaned plant physiological data, and perform multi-scale time regularization processing on the cleaned plant physiological data to obtain standardized time series data; Extract the characteristic points of spectral response curve, photosynthetic efficiency index and growth phenotype characteristics from the standardized time series data to obtain the plant physiological characteristic vector; The plant physiological feature vectors are paired with the corresponding spectral condition data in time series to form a plant spectral interaction feature dataset.

[0022] Specifically, multiple key factors of plant growth are taken into account, including light conditions, plant physiological state, and plant growth parameters. Spectral analysis sensors can monitor the light conditions in the plant growth environment in real time, especially the light intensity of different wavelengths. Plant physiological parameter sensors are used to measure the physiological responses of plants, such as photosynthesis rate, chlorophyll content, etc., which can reflect the growth status of plants under specific light conditions. Growth parameter sensors are used to monitor the morphological characteristics of plants, such as plant height, number of leaves, and leaf area, which can help determine the health status and growth stage of plants. Through the collaborative work of these sensors, a comprehensive intelligent spectral sensor network is formed. The raw data collected by the intelligent spectral sensor network are used to collect the relationship data between spectral conditions and plant physiological responses. These raw data include multiple aspects, such as wavelength distribution of light, light intensity, photosynthetic efficiency of plants under different light conditions, and growth parameters of plants. Outlier detection and outlier correction are performed on the raw data. Outliers refer to extreme data points caused by equipment failure or abnormal fluctuations in the external environment, which will have an adverse effect on subsequent analysis. Outlier detection uses statistical methods, such as the Z-score method and the box plot method, to identify values ​​that deviate from the normal range. These outliers are repaired by interpolation or other correction techniques (such as median filtering) to ensure smooth and accurate data. For example, if the chlorophyll content data measured at a certain time point is abnormally low and differs greatly from the surrounding data points, the average value or interpolation result of the surrounding time points is used to replace the abnormal data, thereby eliminating the impact caused by the error. The cleaned plant physiological data is processed by multi-scale time regularization. The growth and physiological response of plants is a long-term dynamic process, but in the actual collected data, short-term fluctuations will occur due to environmental changes (such as fluctuations in temperature and humidity) or inconsistent system acquisition frequencies. Multi-scale time regularization smoothes the data at different time scales, so that the data remains stable for a long time and eliminates short-term noise. This process involves filtering the original data, such as using a low-pass filter or wavelet transform. In this way, standardized time series data is obtained. Based on the standardized time series data, the characteristic points of the spectral response curve, the photosynthetic efficiency index and the growth phenotypic characteristics are extracted to obtain the plant physiological characteristic vector. The characteristic points of the spectral response curve refer to the response data of the plant under different lighting conditions. These characteristic points help the system understand the optimal response of the plant under different spectral conditions. For example, light of specific wavelengths (such as blue light and red light) has a significant effect on the efficiency of photosynthesis. Therefore, the characteristic points of photosynthetic efficiency in these bands are extracted from the data set to help adjust the light source. The photosynthetic efficiency index is a measure of the ability of a plant to photosynthesize under specific lighting conditions. It is obtained by measuring the carbon dioxide absorption rate or oxygen release rate of the plant.Growth phenotypic characteristics include morphological indicators such as plant height, leaf area, and root length, which reflect the growth and health status of plants. By extracting these important physiological characteristics from standardized time series data, the growth and photosynthesis efficiency of plants under different lighting conditions are fully characterized. The plant physiological feature vectors are paired with the corresponding spectral condition data in time series to form a plant spectral interactive feature dataset. Each plant physiological feature vector contains the physiological response information of the plant at a specific moment, while the corresponding spectral condition data reflects the ambient light conditions at that moment. By pairing the two, an interactive dataset containing plant physiological characteristics and light conditions is obtained.

[0023] In a specific embodiment, the process of executing step 102 may specifically include the following steps: The spectral parameters are discretized, the wavelength range is discretized into A bands, the light intensity range is discretized into B levels, and the illumination time dimension is discretized into C time periods, and a three-dimensional discrete parameter space is obtained; Based on the plant spectral interactive feature dataset, the corresponding photosynthetic efficiency index is calculated for each three-dimensional coordinate point to obtain the initial spectral response data; The initial spectral response data were interpolated with three-dimensional thin plate spline to obtain a complete response surface, which was then extended by introducing a growth stage correction factor to obtain a four-dimensional extended model that included the plant development stage dimension. The principal component analysis of the four-dimensional extended model was performed to obtain the spectral response characteristics after dimension reduction; A deep autoencoder network is used to perform nonlinear mapping on the spectral response characteristics after dimension reduction to obtain a three-dimensional spectral response mathematical model.

[0024] Specifically, the core of the spectral parameter discretization process is to convert the three dimensions of light conditions: wavelength, light intensity and illumination time, into discrete values. These three dimensions are the key influencing factors of plant photosynthesis. Their discretization simplifies data processing and facilitates calculation and analysis. The wavelength range is usually between 400-730nm, among which the light wavelengths required for plant growth are mainly concentrated in the blue and red light regions. Therefore, the wavelength range is discretized into 33 bands, each with a width of 10nm. Considering the plant's demand for light intensity, the light intensity range from 0 to 1000μmol·m⁻²·s⁻¹ is discretized into 20 levels to accurately represent different light intensities. Considering the impact of the light cycle on plant growth, the illumination time is discretized into 8 time periods from 0 to 24 hours. Through these discretization processes, a three-dimensional discrete parameter space is obtained, which contains information in three dimensions: wavelength, light intensity and illumination time. Based on the plant spectral interaction feature dataset, for each three-dimensional coordinate point ( Calculate the corresponding photosynthetic efficiency index (PEI). The photosynthetic efficiency index is an indicator obtained by weighted fusion of photosynthetic rate (Pn), electron transfer rate (ETR) and photochemical quenching coefficient (qP). The calculation formula of PEI is:

[0025] Each indicator is normalized to ensure that indicators of different units can be reasonably combined. The initial spectral response data obtained can reflect the photosynthetic efficiency of plants under different light conditions. The three-dimensional thin plate spline interpolation method is used to interpolate the initial spectral response data, fill the sparse areas in the data, and obtain a complete response surface. The thin plate spline interpolation method is an effective smooth interpolation technique that can ensure the smoothness and continuity of the data. Through this interpolation method, a more complete spectral response model is obtained. However, the demand for light in different growth stages of plants is different. Therefore, the growth stage correction factor G(s) is introduced to expand the response surface. This correction factor reflects the difference in the demand for light in different growth stages (such as germination, vegetative growth, and reproductive growth). By introducing the growth stage dimension, the model can describe the photosynthetic efficiency under different light conditions, and can also adapt to the development process of plants, and obtain a four-dimensional extended model including the plant development stage. For the obtained four-dimensional extended model, principal component analysis is used for dimensionality reduction. The dimension of the data is reduced by extracting the main components of the data while retaining the important information in the data. After dimensionality reduction, the main features of the spectral response are obtained, and redundant information is eliminated, which is helpful for the training and application of subsequent models. Through principal component analysis, the spectral response characteristics obtained can explain more than 90% of the data variation, ensuring the accuracy and effectiveness of the model. The deep autoencoder network is used to perform nonlinear mapping on the spectral response characteristics after dimensionality reduction. An autoencoder is a neural network model that can learn the intrinsic representation of data, map the input data to a low-dimensional space, and reconstruct the original data. By mapping the features after dimensionality reduction through a deep autoencoder, more complex nonlinear relationships are learned, and a more accurate three-dimensional spectral response mathematical model is generated. This three-dimensional model can characterize the photosynthetic response of plants under different light conditions.

[0026] In a specific embodiment, the process of executing step 103 may specifically include the following steps: The teacher network and student network are constructed based on the three-dimensional spectral response mathematical model. The teacher network consists of 12 fully connected layers, and the student network adopts a 5-layer fully connected structure. Performing parallel feature extraction on plant physiological characteristic data and spectral response model parameters to obtain plant physiological characteristic vectors and spectral response characteristic vectors; The alternating attention mechanism is implemented by calculating the similarity matrix between the plant physiological feature vector and the spectral response feature vector to obtain the aligned feature representation; The direct concatenation method and cross-modal self-attention mechanism are used to perform multimodal fusion on the aligned feature representations to obtain fused features; Based on the fusion features, the teacher network is trained to generate high-precision spectral adjustment parameters, and the student network is trained to generate adjustment parameters based on the compression features. The loss function is constructed through the soft target knowledge distillation method to obtain the student network parameters. The parameters of the student network are optimized and adjusted to obtain the plant spectrum matching network.

[0027] Specifically, the architecture of the teacher network and the student network is defined. The teacher network consists of 12 fully connected layers, each containing 256 neurons, and the activation function uses GELU. This structure can effectively capture the complex spectral-physiological feature relationship. The structure of the student network is relatively simplified, using 5 fully connected layers, each containing 64 neurons, and the activation function is ReLU. This simplified structure enables the student network to reduce the consumption of computing resources and is suitable for application in resource-constrained environments. In the process of building the network, parallel feature extraction is performed on plant physiological feature data and spectral response model parameters. These data include physiological response data such as plant photosynthesis efficiency, chlorophyll fluorescence, photosynthetic rate, and spectral characteristics under light conditions. These data are processed separately by two independent feature extraction modules to generate plant physiological feature vectors P and spectral response feature vectors S. The plant physiological feature vector P is extracted by a convolutional neural network, while the spectral response feature vector S is extracted by a one-dimensional convolution and frequency domain analysis layer. Through this parallel processing, features are extracted from both physiological and spectral aspects. The similarity matrix between the plant physiological feature vector and the spectral response feature vector is calculated to implement the alternating attention mechanism and obtain the aligned feature representation. The similarity matrix M is calculated as:

[0028] in, and are plant physiological characteristic vector and spectral response characteristic vector, is the dimension of the feature. By calculating the similarity matrix , quantify the similarity between the two feature vectors, and perform bidirectional feature updates based on this similarity. After three iterations, the feature representations are fully aligned, thereby ensuring the correlation between physiological features and spectral features. Direct concatenation and cross-modal self-attention mechanisms are used to perform multimodal fusion on the aligned features. The direct concatenation method converts the plant physiological feature vectors into and the spectral response eigenvector The two methods are directly concatenated into a large vector and then fused through a multi-layer perceptron; while the cross-modal self-attention mechanism calculates the interaction between different modalities, so that the model can automatically identify which features are most critical for spectral regulation. Through these two fusion methods, the fused features are obtained. , these features can comprehensively consider the physiological needs and light conditions of plants. Based on the fusion features, the teacher network is trained to generate high-precision spectral adjustment parameters , and train the student network to generate adjustment parameters through knowledge distillation method The output of the teacher network is a complete high-precision spectral adjustment parameter, while the student network generates the corresponding adjustment parameters from the compressed features. In order to enable the student network to effectively learn the knowledge of the teacher network, the loss function is constructed by the soft target knowledge distillation method. The loss function of soft target knowledge distillation is expressed as:

[0029] in, is the cross entropy loss, which represents the difference between the output of the student network and the teacher network, is the Kullback-Leibler divergence, which measures the difference in the distribution of the two. is the temperature coefficient, which controls the degree of softening. Through this loss function, the student network can gradually approach the accuracy of the teacher network while maintaining a low computational complexity. The student network parameters are optimized and adjusted, and finally a plant spectral matching network with high computational efficiency and high accuracy is obtained. This optimization process uses standard gradient descent methods, such as the Adam optimizer, to adjust the parameters of the student network by evaluating the network performance on the validation data set, reducing the computational complexity during inference, and making the student network suitable for real-time spectral adjustment tasks. The trained plant spectral matching network can generate matching spectral adjustment parameters based on real-time plant physiological data, thereby optimizing the photosynthesis efficiency of plants.

[0030] In a specific embodiment, the process of executing step 104 may specifically include the following steps: Construct a plant growth stage evaluator, which includes three parallel units: morphological evaluation unit, physiological evaluation unit and metabolite evaluation unit; The plant spectral matching network is used to process the collected real-time plant physiological data to obtain multi-stage feature vectors, and the multi-stage feature vectors are fused to obtain a comprehensive plant feature representation; The comprehensive plant feature representation is mapped to the probability distribution of D typical plant growth stages through the plant growth stage evaluator, and the plant growth stage identifier is determined by the maximum probability value; Based on the plant growth stage identifier, initial spectrum requirement parameters are extracted from a preset stage-spectrum requirement mapping table, and the initial spectrum requirement parameters include spectrum ratio requirements of different wavelengths, light intensity and light cycle; The plant spectral matching network is used to adjust the initial spectral requirement parameters according to the current individual characteristics of the plants to generate personalized spectral requirement parameters and light timing requirement parameters.

[0031] Specifically, the growth stage of a plant is affected by many factors, including morphological characteristics, physiological state, and metabolic activity. The evaluator extracts features from these three dimensions and conducts a comprehensive analysis. The morphological evaluation unit analyzes growth indicators such as plant height growth rate, leaf area expansion rate, and internode length. The gated recurrent unit (GRU) processes the time series features to obtain a stage feature vector reflecting the morphological development of the plant. The physiological evaluation unit analyzes key physiological indicators such as chlorophyll content, photosynthetic rate change trend, and water use efficiency. These indicators can reveal the photosynthetic state and physiological health of plants and generate stage feature vectors. The metabolite evaluation unit generates metabolism-related stage feature vectors by analyzing the changes in the content of metabolites such as sugars, proteins, and secondary metabolites. These feature vectors provide data support for the evaluation of plants at different stages. Through parallel processing, each evaluation unit uses a 128-dimensional GRU hidden layer to extract temporal features, and then the three vectors are fused through a feature fusion. The fusion process is achieved by calculating the weighted sum ,in is the corresponding weight matrix, and the optimization objective is to adjust these weight matrices through the back propagation algorithm. The fused feature vector The probability distribution is mapped to seven typical plant growth stages (such as seed germination, seedling stage, early vegetative growth stage, late vegetative growth stage, flower bud differentiation stage, flowering and fruiting stage, and maturity stage), and the current growth stage identifier of the plant is determined by the maximum probability value. , extract the initial spectral demand parameters from the preset stage-spectral demand mapping table. The mapping table contains the spectral demand parameters corresponding to each growth stage, such as the ratio requirements of blue light (400-500nm), green light (500-600nm), red light (600-700nm) and far-red light (700-730nm), as well as the corresponding light intensity, light cycle and light rhythm. The parameterization of spectral demand provides spectral intensity ratio and timing information for each stage. Using the plant spectral matching network, these initial spectral demand parameters are adjusted according to the physiological characteristics of the current plant individual. The adjustment of personalized spectral demand parameters is achieved by calculating the adjustment coefficient To achieve this, the formula is:

[0032] in, is the individual feature vector, and is the parameter to adjust the network. It is a Sigmoid function that generates personalized spectral requirement parameters and lighting timing requirement parameters These parameters describe the spectral intensity change requirements and light intensity change rules of plants at different growth stages. Through this process, real-time plant physiological data can be processed, and the growth stage of plants can be accurately evaluated to generate personalized spectral demand parameters and light timing demand parameters.

[0033] In a specific embodiment, the process of executing step 105 may specifically include the following steps: Define the optimization variable set and obtain the basic variables for spectrum optimization; Based on the plant growth stage identifier, personalized spectrum requirement parameters and light timing requirement parameters, spectrum matching objective function, photosynthetic efficiency objective function, energy consumption objective function and economic cost objective function were constructed to obtain a multi-objective optimization model. Set constraints on the multi-objective optimization model to obtain the complete optimization problem; According to the basic variables of spectral optimization, the differential co-evolution algorithm is performed on the complete optimization problem to obtain the initial evolution population; Perform evolution operations on the initial evolution population, adopt an elite retention strategy based on non-dominated sorting, and introduce adaptive mutation operations to obtain the target population; A non-dominated solution set is selected from the target population, and the LED dimming solution is obtained by using the fuzzy comprehensive evaluation method.

[0034] Specifically, define the optimization variable set , where each variable Representative The dimming level of the LED light source is in the range of [0,1], which means from completely off to maximum brightness. For example, the LED types in the system , including deep blue light (440nm), blue light (460nm), green light (520nm), yellow light (590nm), red light (660nm) and far red light (730nm). These variables determine the output spectral intensity of each LED and are the basic variables for constructing spectral optimization. Based on the plant growth stage identifier, personalized spectral requirement parameters and light timing requirement parameters, multiple objective functions are constructed. Spectral matching objective function It is used to measure the similarity between the actual output spectrum and the target spectrum, and its form is:

[0035] in, represents the actual output spectrum, It is the target spectrum generated according to the needs of the plant growth stage. This function measures the deviation between the actual spectrum and the target spectrum. The smaller the deviation, the higher the spectrum matching degree. Photosynthetic efficiency objective function The photosynthetic efficiency index (PEl) is calculated based on the three-dimensional spectral response model and its form is:

[0036] in, is the photosynthetic efficiency index, is the plant growth stage identifier, and the function aims to maximize the photosynthesis efficiency. Energy consumption objective function Calculate the total energy consumption of the system:

[0037] in, For the The power consumption of the LEDs, is the dimming level of the LED. The objective of this function is to minimize energy consumption. Economic cost objective function Used to calculate the economic cost of the system:

[0038] in, It is The unit time usage cost of LEDs, is the dimming level of the LED. The goal is to reduce the operating cost of the system. Through the above objective function, a multi-objective optimization model is constructed to optimize the spectral matching, photosynthetic efficiency, energy consumption and economic cost at the same time. In the multi-objective optimization model, constraints are set, such as the total power constraint:

[0039] And the light intensity constraint:

[0040] These constraints ensure that the system meets the power and safety requirements while providing the required spectrum and light intensity. The differential coevolution algorithm is used to solve the optimization problem. The differential coevolution algorithm is an effective multi-objective optimization algorithm that uses population evolution to find the Pareto frontier solution. Define the initial population , the population size is , each individual represents a set of LED dimming solutions. Based on the principle of differential co-evolution algorithm, the evolution operation is performed. Each individual is updated according to the gradient of the objective function, and the evolution of the population evaluates the quality of the solution by calculating the fitness function. In each generation of evolution, the elite retention strategy based on non-dominated sorting is adopted to ensure that the optimal solution of each generation is retained to the next generation. In order to improve the convergence speed of the algorithm, an adaptive mutation operation is introduced, and the mutation probability The initial value of 0.1 decreases linearly to 0.01 with the increase of iteration number. After multiple generations of evolution, the target population is obtained, and the non-dominated solution set is selected from it by the improved Pareto frontier identification method. These non-dominated solution sets represent the trade-off balance points between different objectives. The optimal solution is selected from these non-dominated solutions by the fuzzy comprehensive evaluation method. , the solution achieves the best balance among the four objective functions of spectral matching, photosynthetic efficiency, energy consumption and economic cost.

[0041] In a specific embodiment, executing the AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis also includes the following steps: A hierarchical spectrum regulation controller is designed, which includes a base layer, an adaptation layer, and an optimization layer. The base layer realizes LED drive control through pulse width modulation technology, calculates the PWM duty cycle of each LED channel, and obtains the basic control parameters of light intensity. Adaptive adjustment is performed on the basic control parameters of light intensity, and the daily spectrum adjustment is realized in combination with the light timing demand parameters to obtain the adaptation layer adjustment parameters; Performing long-term optimization on the adjustment parameters of the adaptation layer to obtain the spectrum demand parameters after the optimization layer is adjusted; The actual output spectrum is measured by the spectrum monitoring unit according to the spectrum demand parameters adjusted by the optimization layer, and the real-time response data of chlorophyll fluorescence and photosynthetic rate are collected by the plant response monitoring unit; The deviation between the actual spectrum and the target spectrum and the deviation between the actual photosynthetic efficiency of the plant and the expected efficiency are calculated according to the real-time response data to obtain a closed-loop control deviation signal; The correction coefficient is generated based on the closed-loop control deviation signal, and the adaptation layer adjustment parameters are corrected. When a light source failure, environmental abnormality or plant stress condition is detected, the system automatically switches to the preset safe spectrum mode to obtain the final real-time adjustment parameters.

[0042] Specifically, the hierarchical spectrum regulation controller consists of three key layers: the base layer, the adaptation layer, and the optimization layer. The base layer implements LED drive control through pulse width modulation (PWM) technology. The working principle of PWM is to control the output light intensity by adjusting the duty cycle of the LED. The frequency of PWM is set to 1200Hz, and the dimming resolution is set to 12 bits, that is, the dimming level of each LED channel is determined by the value of the duty cycle, which ranges from 0 to 4095. For each LED channel , its PWM duty cycle The calculation formula is:

[0043] in, For the The formula is based on the optimized spectrum requirement parameters. , the output light intensity of the LED is controlled by PWM signal. The function of the adaptation layer is to dynamically adjust the spectrum according to the real-time plant physiological data and light timing requirements. Every 30 minutes, the adaptation layer will make adaptive adjustments to the basic control parameters of the light intensity, and calculate the adaptive adjustment coefficient based on the plant physiological data. , the adjustment formula is:

[0044] in, The value range is [-0.15, 0.15], which reflects the changing trend of plant photosynthetic efficiency. Adaptation layer combined with light timing requirement parameters The daily spectrum adjustment is realized, which simulates the characteristics of natural light changing over time, so as to optimize and adjust the different time periods of plant growth. The optimization layer is responsible for long-term spectrum adjustment optimization, and the execution cycle is 24 hours. The optimization layer optimizes the spectrum parameters by comprehensively analyzing the plant physiological data and photosynthetic efficiency data of the past 24 hours, and adopts the gradient descent method. The update formula of the spectrum demand parameter is:

[0045] in, is the learning rate, set to 0.05, It is the gradient of photosynthetic efficiency with respect to the spectral requirement parameter. The gradient is calculated by solving the influence of each spectral intensity on the change of photosynthetic efficiency, thereby optimizing the spectral requirement parameter and obtaining the optimized spectral configuration. The actual output spectrum is measured by the spectral monitoring unit according to the spectral requirement parameter adjusted by the optimization layer The spectrum monitoring unit is equipped with a micro-spectrometer with a sampling frequency of once every 5 minutes to monitor the changes in the output spectrum in real time. The plant response monitoring unit collects real-time physiological data of plants, including chlorophyll fluorescence, photosynthetic rate, etc. and the difference between the actual photosynthetic efficiency of plants and the expected efficiency Calculations are performed to obtain closed-loop control deviation signals. These deviation signals reflect the deviation between spectral regulation and plant response, providing feedback information for subsequent correction and optimization. Correction coefficients are generated based on the closed-loop control deviation signals. , this coefficient is used to correct the adaptation layer adjustment parameters. The correction formula is:

[0046] Corrected spectral parameters is sent to the system to achieve more accurate spectral regulation. When the system detects light source failure, environmental abnormalities or plant stress, it automatically switches to the preset safe spectrum mode. This mechanism ensures that plants still get enough light in harsh environments while avoiding plant damage caused by over-regulation. Through the above-mentioned hierarchical control structure and feedback mechanism, the system can continuously adjust and optimize the spectral regulation parameters based on real-time spectral monitoring data and plant response feedback, thereby achieving a dynamic optimal photosynthetic environment, ensuring that plants can obtain the best lighting conditions at each growth stage and improving their photosynthesis efficiency.

[0047] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0048] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0049] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI plant lamp spectrum adjustment method for optimizing plant photosynthesis, characterized in that: The method comprises: The original plant physiological data is collected and preprocessed through the intelligent spectral sensor network to obtain the plant spectral interaction feature dataset; Using the plant spectral interaction feature data set to construct a three-dimensional spectral response mathematical model; The three-dimensional spectral response mathematical model is used as a teacher network, and a student network is trained through knowledge distillation to obtain a plant spectrum matching network; Processing real-time plant physiological data through the plant spectrum matching network, performing plant growth stage assessment, and generating plant growth stage identifiers, personalized spectrum requirement parameters, and light timing requirement parameters; A differential co-evolution algorithm is executed according to the plant growth stage identifier, the personalized spectrum requirement parameter and the light timing requirement parameter to generate an LED dimming solution.

2. The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis according to claim 1, characterized in that: The raw plant physiological data is collected and preprocessed through the intelligent spectral sensor network to obtain a plant spectral interaction feature data set, including: Configure spectral analysis sensors, plant physiological parameter sensors and growth parameter sensors in the plant growth environment to form an intelligent spectral sensor network; collecting raw plant physiological data including spectral conditions and corresponding plant physiological responses through the intelligent spectral sensor network; Performing outlier detection and outlier correction on the original plant physiological data to obtain cleaned plant physiological data, and performing multi-scale time regularization processing on the cleaned plant physiological data to obtain standardized time series data; Extracting spectral response curve characteristic points, photosynthetic efficiency index and growth phenotype characteristics from the standardized time series data to obtain plant physiological characteristic vectors; The plant physiological feature vector is paired with the corresponding spectral condition data in time series to form a plant spectral interaction feature data set.

3. The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis according to claim 1, characterized in that: The method of constructing a three-dimensional spectral response mathematical model using the plant spectral interaction feature data set includes: The spectral parameters are discretized, the wavelength range is discretized into A bands, the light intensity range is discretized into B levels, and the illumination time dimension is discretized into C time periods, and a three-dimensional discrete parameter space is obtained; Based on the plant spectral interaction feature data set, the corresponding photosynthetic efficiency index is calculated for each three-dimensional coordinate point to obtain initial spectral response data; Performing three-dimensional thin plate spline interpolation on the initial spectral response data to obtain a complete response surface, and extending the complete response surface by introducing a growth stage correction factor to obtain a four-dimensional extended model including plant development stage dimensions; Performing principal component analysis on the four-dimensional expansion model to obtain spectral response characteristics after dimension reduction; A deep autoencoder network is used to perform nonlinear mapping on the spectral response characteristics after dimension reduction to obtain a three-dimensional spectral response mathematical model.

4. The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis according to claim 1, characterized in that: The three-dimensional spectral response mathematical model is used as a teacher network, and a student network is trained through knowledge distillation to obtain a plant spectrum matching network, including: Constructing a teacher network and a student network based on the three-dimensional spectral response mathematical model, wherein the teacher network is composed of 12 fully connected layers, and the student network adopts a 5-layer fully connected structure; Performing parallel feature extraction on plant physiological characteristic data and spectral response model parameters to obtain plant physiological characteristic vectors and spectral response characteristic vectors; An alternating attention mechanism is implemented by calculating a similarity matrix between the plant physiological feature vector and the spectral response feature vector to obtain an aligned feature representation; A direct concatenation method and a cross-modal self-attention mechanism are used to perform multimodal fusion on the aligned feature representations to obtain fused features; Based on the fusion features, a teacher network is trained to generate high-precision spectral adjustment parameters, and a student network is trained to generate adjustment parameters based on the compression features, and a loss function is constructed through a soft target knowledge distillation method to obtain student network parameters; The student network parameters are optimized and adjusted to obtain a plant spectrum matching network.

5. The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis according to claim 1, characterized in that: The real-time plant physiological data is processed by the plant spectrum matching network, and plant growth stage evaluation is performed to generate plant growth stage identifiers, personalized spectrum requirement parameters and light timing requirement parameters, including: Constructing a plant growth stage evaluator, wherein the plant growth stage evaluator comprises three parallel units: a morphological evaluation unit, a physiological evaluation unit and a metabolite evaluation unit; The plant spectrum matching network is used to process the collected real-time plant physiological data to obtain a multi-stage feature vector, and the multi-stage feature vector is subjected to feature fusion to obtain a comprehensive plant feature representation; By means of the plant growth stage evaluator, the comprehensive plant feature representation is mapped to a probability distribution of D typical plant growth stages, and a plant growth stage identifier is determined by a maximum probability value; Based on the plant growth stage identifier, extracting initial spectrum requirement parameters from a preset stage-spectrum requirement mapping table, the initial spectrum requirement parameters including spectrum ratio requirements of different wavelengths, light intensity and light cycle; The plant spectrum matching network is used to adjust the initial spectrum requirement parameters according to the current individual characteristics of the plants, so as to generate personalized spectrum requirement parameters and light timing requirement parameters.

6. The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis according to claim 1, characterized in that: The step of executing a differential co-evolution algorithm according to the plant growth stage identifier, the personalized spectrum requirement parameter and the light timing requirement parameter to generate an LED dimming solution includes: Define the optimization variable set and obtain the basic variables for spectrum optimization; Based on the plant growth stage identifier, the personalized spectrum requirement parameter and the light timing requirement parameter, a spectrum matching objective function, a photosynthetic efficiency objective function, an energy consumption objective function and an economic cost objective function are constructed to obtain a multi-objective optimization model; Setting constraints on the multi-objective optimization model to obtain a complete optimization problem; According to the basic variables of the spectrum optimization, a differential co-evolution algorithm is executed on the complete optimization problem to obtain an initial evolution population; Performing an evolution operation on the initial evolution population, adopting an elite retention strategy based on non-dominated sorting, and introducing an adaptive mutation operation to obtain a target population; A non-dominated solution set is selected from the target population, and a fuzzy comprehensive evaluation method is used to analyze and obtain an LED dimming solution.

7. The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis according to claim 1, characterized in that: The AI ​​plant lamp spectrum adjustment method for optimizing plant photosynthesis also includes: Design a hierarchical spectrum adjustment controller including a base layer, an adaptation layer and an optimization layer. The base layer realizes LED drive control through pulse width modulation technology, calculates the PWM duty cycle of each LED channel, and obtains the basic control parameters of light intensity. Adaptively adjusting the light intensity basic control parameter, and implementing daily spectrum adjustment in combination with the light timing demand parameter to obtain an adaptation layer adjustment parameter; Performing long-term optimization on the adaptation layer adjustment parameters to obtain the spectrum demand parameters after the optimization layer adjustment; The actual output spectrum is measured by a spectrum monitoring unit according to the spectrum requirement parameters adjusted by the optimization layer, and the real-time response data of chlorophyll fluorescence and photosynthetic rate are collected by a plant response monitoring unit; Calculating the deviation between the actual spectrum and the target spectrum and the deviation between the actual photosynthetic efficiency of the plant and the expected efficiency according to the real-time response data to obtain a closed-loop control deviation signal; A correction coefficient is generated based on the closed-loop control deviation signal, and a correction operation is performed on the adaptation layer adjustment parameters. When a light source failure, an abnormal environment or a plant stress condition is detected, the system automatically switches to a preset safe spectrum mode to obtain the final real-time adjustment parameters.

Images