An AI-driven method for supplemental lighting control in fig cultivation

Abstract

This invention relates to the technical field of supplemental lighting regulation, and discloses an artificial intelligence-driven method for supplemental lighting regulation in fig cultivation. The method includes: collecting cultivation environment data and fig physiological response data from the fig cultivation area; extracting spectral shift features and fig physiological characteristics; using a spectral demand intelligent inference model based on the fig's physiological state to output target spectral parameters corresponding to the fig's growth stage; and using a dynamic spectral correction model to output spectral correction parameters to dynamically regulate the light power of multiple spectral bands in the fig cultivation area. This invention, by constructing a physiological demand-driven intelligent spectral regulation mechanism, achieves quantitative calculation of target spectral parameters related to the fig's growth stage, and dynamically corrects the light power of multiple spectral bands by combining spectral shift features. This enables precise supplemental lighting regulation of light power in fig cultivation areas under complex environmental conditions, avoiding physiological stress caused by sudden changes in light intensity.

Description

Technical Field

[0001] This invention relates to the field of supplemental lighting control through big data analysis, specifically an artificial intelligence-driven method for supplemental lighting control in fig cultivation, which is particularly suitable for supplemental lighting control during seed and seedling cultivation. Background Technology

[0002] Fig cultivation, as a high-value specialty fruit tree, has seen rapid development in protected cultivation in recent years. Figs are highly sensitive to light conditions, especially during winter when low temperatures, weak light, or prolonged periods of rain can significantly impact flower bud differentiation, fruit enlargement, and sugar accumulation. Therefore, artificial lighting has become a crucial means to improve the productivity and quality of figs grown in protected environments. However, in actual production environments, LED supplemental lighting operates for extended periods in high-humidity, low-temperature greenhouses, often resulting in condensation or water film formation on the light source surface. This leads to a shift in spectral output characteristics and a decrease in light efficiency, ultimately distorting the light signal received by the plant. Existing supplemental lighting systems are typically based on fixed spectrum and light intensity settings, failing to dynamically adjust according to actual spectral changes, resulting in reduced lighting efficiency, unstable physiological responses, and fluctuations in fruit quality.

[0003] Existing research has yielded supplemental lighting optimization schemes to address the problem of uneven crop light exposure. For example, patent CN119836950A provides a supplemental lighting system and method for densely planted tomato cultivation. By setting up a first and second reflector and adjusting the angle of the reflectors based on the detected light intensity information, external light sources are reflected into the cultivation layer, thereby improving the uniformity of light exposure in the lower cultivation area of ​​tomatoes and increasing light resource utilization. This scheme effectively alleviates the shading problem under dense planting conditions at the structural design level and has certain engineering practical value. However, it only focuses on the light intensity distribution and does not conduct detailed modeling of spectral composition and spectral shift, making it difficult to match the actual photosynthetic needs of crops. Furthermore, the supplemental lighting regulation logic is mainly based on rule-based or structural adjustments, with limited intelligence, making it difficult to cope with complex and rapidly changing microclimate environments.

[0004] To address this issue, this invention proposes an artificial intelligence-driven method for supplemental lighting control in fig cultivation. By monitoring spectral information in real time and automatically correcting the light power in spectral bands, it can not only achieve precise dynamic control of supplemental lighting for figs under complex microclimate conditions, but also provide a new technical path for intelligent light environment management in plant and fruit cultivation areas. Summary of the Invention

[0005] This invention proposes an AI-driven method for regulating supplemental lighting in fig cultivation. Steps S1-S2 continuously collect cultivation environment data and extract spectral shift features, solving the problems of existing supplemental lighting methods being unable to perceive spectral changes in real time and the long-term mismatch between the supplemental lighting spectrum and the actual environmental spectrum. Step S3 extracts physiological characteristics such as photosynthetic efficiency per unit leaf area, coordination rate of change, and physiological response stability from fig physiological response data, and introduces an intelligent inference model for spectral demand, outputting the target dominant wavelength and target light intensity. This overcomes the problem of poor adaptability of fixed-band, fixed-intensity supplemental lighting to different growth stages and solves the problem that traditional empirical supplemental lighting cannot accurately reflect the real photosynthetic needs at different growth stages. Step S4 uses a dynamic spectral correction model to jointly model the target spectral parameters and spectral shift features, solving the problem that multispectral band power allocation depends on manual rules and is difficult to optimize collaboratively.

[0006] To achieve the above objectives, this invention provides an artificial intelligence-driven method for regulating supplemental lighting in fig cultivation, comprising the following steps: S1: Collect cultivation environment data and fig physiological response data from fig cultivation areas; S2: Extract spectral shift features based on the cultivation environment data of the fig cultivation area; S3: Based on the fig physiological response data, extract the fig physiological characteristics, use the intelligent inference model of spectral demand based on the fig physiological state to receive the fig physiological characteristics, and output the target spectral parameters corresponding to the fig growth period. S4: Based on the target spectral parameters and spectral shift characteristics of the fig during its growth period, a dynamic spectral correction model is used to output spectral correction parameters. Based on these spectral correction parameters, the light power of multiple spectral bands in the fig cultivation area is dynamically adjusted to supplement light.

[0007] As a further improvement of the present invention: Furthermore, in step S1, the collection of cultivation environment data and fig physiological response data in the fig cultivation area includes: S11: Deploy environmental sensing devices in the fig cultivation area and periodically collect cultivation environment data of the fig cultivation area using distributed environmental sensing devices. The cultivation environment data includes spectral data, ambient temperature data, relative humidity data, and carbon dioxide concentration data. The spectral data is the intensity of reflected light on the surface of fig leaves at multiple spectral wavelengths collected by the environmental sensing devices. S12: Based on the cultivation environment data of the fig cultivation area, perform inversion calculation on the net photosynthetic rate and leaf area index of fig leaves in the fig cultivation area. S13: The net photosynthetic rate and leaf area index are used as the physiological response data of figs in the fig cultivation area.

[0008] Further, step S12 involves inverting the calculation of the net photosynthetic rate and leaf area index of fig leaves in the fig cultivation area, including: S121: Based on the spectral data, extract the reflected light intensity in the red light band and near-infrared band respectively, and obtain the incident light intensity in the red light band and near-infrared band respectively from the environmental sensing device, calculate the ratio of reflected light intensity to incident light intensity, and use it as the band reflectivity of the red light band and near-infrared band. S122: Based on the band reflectance of the red light band and the near-infrared band, the normalized vegetation index is calculated, and the normalized vegetation index is inverted into the light absorption ratio. S123: Perform integration on the spectral data to generate photosynthetically effective reflectance, and calculate the net photosynthetic rate based on the light absorption ratio and photosynthetically effective reflectance. S124: The leaf area index is obtained by inverting the normalized vegetation index based on the empirical regression coefficient.

[0009] Further, step S2 involves extracting spectral shift features, including: S21: Based on the cultivation environment data of the fig cultivation area, calculate the spectral centroid of the spectral data, and calculate the offset value between the spectral centroid and the standard center wavelength, as the spectral main peak offset value. S22: Based on the ambient temperature data, relative humidity data, and carbon dioxide concentration data, the spectral main peak shift value is modulated by environmental modulation to obtain the environmentally modulated spectral main peak shift value. S23: Extract the reflected light intensity corresponding to the spectral centroid and the standard center wavelength from the spectral data respectively, and calculate the ratio between the reflected light intensity corresponding to the spectral centroid and the reflected light intensity corresponding to the standard center wavelength as the light intensity attenuation ratio. S24: Construct standard reflected light intensities for different spectral wavelengths, perform difference integration on the spectral data, and generate effective spectral offset; S25: The main peak shift value of the environmental modulation spectrum, the light intensity attenuation ratio, and the effective spectral shift are used as spectral shift features.

[0010] Furthermore, the formula for environmental modulation of the spectral main peak shift value based on the ambient temperature data, relative humidity data, and carbon dioxide concentration data in step S22 is as follows: ; ; in, Indicates the spectral peak shift value Environmental modulation results Indicates the ambient temperature modulation parameters. This represents the relative humidity modulation parameter. This represents the carbon dioxide concentration modulation parameter. The modulation weights of the ambient temperature modulation parameter, relative humidity modulation parameter, and carbon dioxide concentration modulation parameter are represented in turn. The data represent ambient temperature, relative humidity, and carbon dioxide concentration, respectively. These represent the reference values ​​for ambient temperature, relative humidity, and carbon dioxide concentration, respectively. These represent the normalized scaling factors for ambient temperature data, relative humidity data, and carbon dioxide concentration data, respectively. Represents the hyperbolic tangent function. This represents the natural logarithm function.

[0011] Further, step S3 involves extracting the physiological characteristics of figs, including: S31: Based on the aforementioned fig physiological response data, calculate the ratio between net photosynthetic rate and leaf area index, which is used as the photosynthetic efficiency per unit leaf area. S32: Calculate the coordinated rate of change between the net photosynthetic rate and the leaf area index; Specifically, the formula for calculating the coordinated change rate between the net photosynthetic rate and the leaf area index is as follows: ; in, This represents the coordinated rate of change between net photosynthetic rate and leaf area index. Indicates net photosynthetic rate, Indicates leaf area index, This represents the net photosynthetic rate collected in the previous cycle. This represents the leaf area index collected in the previous period. This represents the rate of change in net photosynthetic rate. This indicates the rate of change of the leaf area index; S33: Construct a time window, take the currently collected net photosynthetic rate as the end data value of the time window, calculate the mean and standard deviation of all net photosynthetic rates within the time window, and calculate the ratio between the mean and the standard deviation as an indicator of the physiological response stability of the net photosynthetic rate. S34: The photosynthetic efficiency per unit leaf area, the rate of coordinated change, and the physiological response stability index are used as physiological characteristics of figs.

[0012] Furthermore, step S3, which utilizes a spectral demand intelligent inference model based on the physiological state of figs to receive the physiological characteristics of figs and output the target spectral parameters corresponding to the fig's growth stage, also includes: The intelligent inference model for spectral demand includes a target dominant wavelength calculation module and a target light intensity calculation module; S35: The target dominant wavelength calculation module calculates the target dominant wavelength based on the response of the fig's physiological characteristics to the photosynthetic sensitive area; S36: The target light intensity calculation module adjusts the baseline light energy of the fig cultivation area by combining the photosynthetic efficiency per unit leaf area and the physiological response stability index to generate the target light intensity. S37: Use the target dominant wavelength and target light intensity as target spectral parameters.

[0013] Furthermore, in step S4, the spectral correction parameters are output using a dynamic spectral correction model, including: The dynamic spectral correction model includes an input layer, a feature mapping layer, and a power allocation output layer, wherein the feature mapping layer is a hidden layer structure. The process of outputting spectral correction parameters using the dynamic spectral correction model is as follows: S41: The output layer receives the target spectral parameters and spectral shift features of the fig growth period, and splices the target main wavelength and spectral shift features in the target spectral parameters to obtain the fig cultivation spliced ​​feature vector. S42: The feature mapping layer performs multi-layer mapping on the fig cultivation splicing feature vector to obtain a multi-layer state vector; Specifically, the multi-level mapping formula for the fig cultivation splicing feature vector is as follows: ; in, This represents the feature vector of fig cultivation splicing. The corresponding state vector of the h-th layer, This represents the mapping weight matrix of the h-th layer. Let H represent the mapping bias at layer h, where H represents the layer number of the state vector. This indicates the activation function; we set the activation function to the Sigmoid function. S43: The power distribution output layer receives the target light intensity and the multi-layer state vector, and generates the illumination power of each spectral band as a spectral correction parameter.

[0014] Furthermore, step S4, which involves dynamically adjusting the light power across multiple spectral bands in the fig cultivation area based on the spectral correction parameters, also includes: The spectral correction parameters are sent to the multi-channel supplemental lighting control unit of the fig cultivation area. The multi-channel supplemental lighting control unit includes multiple supplemental lighting units, each of which provides supplemental lighting for a specific spectral band. By adjusting the light power of the supplemental lighting units, the light power of each spectral band is dynamically adjusted.

[0015] Compared with existing technologies, this invention proposes an artificial intelligence-driven method for regulating supplemental lighting in fig cultivation, which has the following beneficial effects: First, this invention uses a nonlinear coupling model to model the spectral peak shift value with ambient temperature, relative humidity, and carbon dioxide concentration, achieving adaptive correction of the spectral peak shift value to microclimate changes. Specifically, this invention utilizes a hyperbolic tangent function to smooth and limit deviations in ambient temperature and relative humidity, avoiding excessive amplification of spectral modulation caused by extreme environmental fluctuations. A logarithmic function is introduced to characterize the decreasing marginal effect of carbon dioxide concentration, consistent with the photosynthetic response mechanism. Simultaneously, this invention, through differentiated weight allocation, makes the modulation results more closely match the actual physiological needs of figs, improving the stability and accuracy of spectral peak correction, and facilitating precise and continuous supplemental lighting control under complex environmental conditions.

[0016] Meanwhile, this invention achieves multi-dimensional quantitative characterization of the physiological state of figs by constructing the coordinated change rate of photosynthetic efficiency per unit leaf area, net photosynthetic rate and leaf area index, as well as physiological response stability index. Among them, photosynthetic efficiency per unit leaf area reflects the level of light energy utilization, the coordinated change rate describes the matching relationship between photosynthesis and growth expansion, and the physiological response stability index is used to evaluate the temporal fluctuation characteristics of the photosynthetic process. The constructed fig physiological characteristics are beneficial to improving the sensitivity and robustness of subsequent spectral demand inference to physiological changes. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an artificial intelligence-driven method for adjusting supplemental lighting in fig cultivation, as provided in an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of a data acquisition and supplemental lighting control structure for a fig cultivation area, provided in an embodiment of the present invention. Detailed Implementation

[0019] The realization of the objectives, functional characteristics, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] This invention provides an artificial intelligence-driven method for regulating supplemental lighting in fig cultivation. The executing entity of this AI-driven method includes, but is not limited to, at least one electronic device configured to execute the method provided in this invention, such as a server or a terminal. In other words, the AI-driven method for regulating supplemental lighting in fig cultivation can be executed by software or hardware installed on a terminal device or a server device, and the software may be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0021] Reference Figure 1 as well as Figure 2 Embodiment 1 of the present invention is as follows: An artificial intelligence-driven method for regulating supplemental lighting in fig cultivation, the method comprising: S1: Collect cultivation environment data and fig physiological response data from fig cultivation areas.

[0022] Data on the cultivation environment and physiological responses of figs were collected from the fig cultivation area, including: S11: Deploy environmental sensing devices in the fig cultivation area and periodically collect cultivation environment data of the fig cultivation area using distributed environmental sensing devices. The cultivation environment data includes spectral data, ambient temperature data, relative humidity data, and carbon dioxide concentration data. The spectral data is the intensity of reflected light on the surface of fig leaves at multiple spectral wavelengths collected by the environmental sensing devices. As an embodiment of the present invention, the environmental sensing device includes a spectrometer, a temperature sensor, a humidity sensor, and a carbon dioxide concentration sensor. In the fig cultivation area, the environmental sensing device is deployed at the position of the plant canopy at the same height as the light-receiving surface of the fig leaves. The spectrometer is used to collect the reflected light intensity of the fig leaf surface at multiple spectral wavelengths to form spectral data. The spectrometer can output the reflected light intensity covering multiple spectral bands such as blue light, red light, and far-red light. The spectral wavelength range corresponding to the spectrometer is 380 to 1100 nanometers. The temperature sensor, humidity sensor, and carbon dioxide concentration sensor are used to obtain the environmental temperature, relative humidity, and carbon dioxide concentration that affect the transpiration, photosynthetic efficiency, and stomatal opening and closing state of the fig, respectively. S12: Based on the cultivation environment data of the fig cultivation area, perform inversion calculation on the net photosynthetic rate and leaf area index of fig leaves in the fig cultivation area. S13: The net photosynthetic rate and leaf area index are used as the physiological response data of figs in the fig cultivation area.

[0023] Step S12 involves inverting the calculation of the net photosynthetic rate and leaf area index of fig leaves in the fig cultivation area, including: S121: Based on the spectral data, extract the reflected light intensity in the red light band and near-infrared band respectively, and obtain the incident light intensity in the red light band and near-infrared band respectively from the environmental sensing device, calculate the ratio of reflected light intensity to incident light intensity, and use it as the band reflectivity of the red light band and near-infrared band. Specifically, the wavelength range of the red light band is 620 to 680 nanometers. Based on the spectral data, the average reflected light intensity of the spectral wavelength range of 620 to 680 nanometers is calculated as the reflected light intensity of the red light band. The wavelength range of the near-infrared band is 700 to 1100 nanometers. The average reflected light intensity of the spectral wavelength range of 700 to 1100 nanometers is calculated as the reflected light intensity of the near-infrared band. S122: Based on the band reflectance of the red light band and the near-infrared band, the normalized vegetation index is calculated, and the normalized vegetation index is inverted into the light absorption ratio. Specifically, the formula for calculating the normalized vegetation index is as follows: ; in, Represents the normalized vegetation index. The reflectance of the red band and the near-infrared band are respectively. The formula for reversing the light absorption ratio is: ; in, Indicates the proportion of light absorption. All represent inversion coefficients. Indicates selection The maximum value in, Indicates selection The minimum value in; Optionally, inversion coefficients are set according to the current growth stage of the fig tree, with the inversion coefficient corresponding to the vegetative growth stage being... The inversion coefficients for the flowering and fruiting periods are 0.6 and 0.1 respectively. The inversion coefficients for maturity are 0.5 and 0.3 respectively. They are 0.3 and 0.4 respectively; S123: Perform integration on the spectral data to generate photosynthetically effective reflectance, and calculate the net photosynthetic rate based on the light absorption ratio and photosynthetically effective reflectance. Specifically, the formula for calculating the net photosynthetic rate is as follows: ; in, Indicates net photosynthetic rate, Denotes Avogadro's constant. Represents the speed of light. Denotes Planck's constant. Represents the spectral wavelength in the spectral data The corresponding reflected light intensity, The differential component (integral increment) representing the wavelength of the spectrum. The light energy utilization efficiency parameter is used to indicate the light energy utilization efficiency parameter. Optionally, the light energy utilization efficiency parameter is set according to the current growth stage of the fig, where the light energy utilization efficiency parameter corresponding to the vegetative growth stage is 0.04, the light energy utilization efficiency parameter corresponding to the flowering and fruiting stage is 0.03, and the light energy utilization efficiency parameter corresponding to the ripening stage is 0.02. S124: The leaf area index is obtained by inverting the normalized vegetation index based on the empirical regression coefficient.

[0024] Specifically, the inversion formula for the leaf area index is: ; in, Indicates leaf area index, The empirical regression coefficient is represented by the actual normalized vegetation index and leaf area index measured in a laboratory setting, and the empirical regression coefficient is fitted based on the actual normalized vegetation index and leaf area index.

[0025] It should be noted that the hierarchical inversion link constructed in this invention, consisting of band reflectance, normalized vegetation index, light absorption ratio, and fig physiological response data, conforms to the basic principles of plant spectral remote sensing and crop physiological modeling. Specifically, by utilizing the differential responses of red and near-infrared bands to chlorophyll absorption and canopy structure, the growth status of fig leaves can be stably characterized. Combined with adaptive settings of inversion coefficients and light energy utilization efficiency parameters during the growth period, the calculated results of net photosynthetic rate and leaf area index are made closer to the actual physiological process, thereby ensuring the reliability and engineering feasibility of fig physiological response data acquisition.

[0026] S2: Extract spectral shift features based on the cultivation environment data of the fig cultivation area.

[0027] Extracting spectral shift features, including: S21: Based on the cultivation environment data of the fig cultivation area, calculate the spectral centroid of the spectral data, and calculate the offset value between the spectral centroid and the standard center wavelength, as the spectral main peak offset value. Specifically, based on a comprehensive evaluation of the absorption intensity of photosynthetic pigments in fig leaves at different wavelengths and their corresponding physiological responses, the wavelength with the optimal photosynthetic response is selected as the standard spectral center wavelength. In one specific embodiment, the standard spectral center wavelength is selected in the red light band, preferably 660 nanometers. The formula for calculating the spectral centroid of the spectral data is: ; in, Indicates the spectral barycenter of the spectral data. This indicates the range of spectral wavelengths in the spectral data. This indicates the smallest spectral wavelength (380 nm) in the spectral data. This indicates the maximum spectral wavelength (1100 nm) in the spectral data. Represents the spectral wavelength in the spectral data The corresponding reflected light intensity; The formula for calculating the spectral main peak shift value is as follows: ,in Indicates the standard center wavelength; S22: Based on the ambient temperature data, relative humidity data, and carbon dioxide concentration data, the spectral main peak shift value is modulated by environmental modulation to obtain the environmentally modulated spectral main peak shift value. S23: Extract the reflected light intensity corresponding to the spectral centroid and the standard center wavelength from the spectral data respectively, and calculate the ratio between the reflected light intensity corresponding to the spectral centroid and the reflected light intensity corresponding to the standard center wavelength as the light intensity attenuation ratio. S24: Construct standard reflected light intensities for different spectral wavelengths, perform difference integration on the spectral data, and generate effective spectral offset; As an embodiment of the present invention, during the calibration stage when the fig's growth status is stable and the supplemental lighting conditions are controlled, a spectrometer is used to collect the reflected light intensity of the fig leaves at multiple spectral wavelengths, which is used as the standard reflected light intensity of the spectral wavelength. Specifically, the formula for calculating the effective spectral shift is as follows: ; in, Indicates spectral wavelength The corresponding standard reflected light intensity, Indicates the effective spectral offset; S25: The main peak shift value of the environmental modulation spectrum, the light intensity attenuation ratio, and the effective spectral shift are used as spectral shift features.

[0028] Specifically, the environmental modulation spectral peak shift value, light intensity attenuation ratio, and effective spectral shift are respectively as follows: .

[0029] It should be noted that this invention accurately characterizes the overall drift trend of the current spectral center relative to the optimal photosynthetic response range of figs by calculating the offset value between the spectral centroid and the standard center wavelength. Furthermore, it combines ambient temperature, relative humidity, and carbon dioxide concentration to modulate the spectral peak offset value, so that the spectral offset characteristics can truly reflect the impact of microclimate changes on the photosynthetic spectrum requirements. At the same time, this invention introduces the light intensity attenuation ratio and the effective spectral offset to quantitatively evaluate the spectral offset from two dimensions: energy level and spectral structure. This avoids the local distortion problem caused by adjusting only a single wavelength. The light intensity attenuation ratio represents the energy change ratio of the reflected light intensity at the spectral centroid relative to the reflected light intensity at the standard center wavelength, and is used to characterize the effective energy attenuation or enhancement of the supplementary light spectrum in the key photosynthetic sensitive band. The effective spectral offset characterizes the cumulative deviation of the current overall spectral shape from the ideal spectral template.

[0030] The formula for environmental modulation of the spectral main peak shift value based on the ambient temperature data, relative humidity data, and carbon dioxide concentration data in step S22 is as follows: ; ; in, Indicates the spectral peak shift value Environmental modulation results Indicates the ambient temperature modulation parameters. This represents the relative humidity modulation parameter. This represents the carbon dioxide concentration modulation parameter. The modulation weights of the ambient temperature modulation parameter, relative humidity modulation parameter, and carbon dioxide concentration modulation parameter are represented in sequence, and are set accordingly. The values ​​are 0.4, 0.4, and 0.2 respectively. The data represent ambient temperature, relative humidity, and carbon dioxide concentration, respectively. These represent the reference values ​​for ambient temperature, relative humidity, and carbon dioxide concentration, respectively. These represent the normalized scaling factors for ambient temperature data, relative humidity data, and carbon dioxide concentration data, respectively. Represents the hyperbolic tangent function. Represents the natural logarithm function; Optionally, the reference values ​​for the ambient temperature data, relative humidity data, and carbon dioxide concentration data are set to 25 degrees Celsius, 65%, and 400 ppm, respectively, where ppm represents the unit of carbon dioxide concentration, and the normalization scaling factors for the ambient temperature data, relative humidity data, and carbon dioxide concentration data are set to 4 degrees Celsius, 10%, and 200 ppm, respectively.

[0031] S3: Based on the fig physiological response data, extract the fig physiological characteristics, use the intelligent inference model of spectral demand based on the fig physiological state to receive the fig physiological characteristics, and output the target spectral parameters corresponding to the fig growth period.

[0032] Physiological characteristics of figs were extracted, including: S31: Based on the aforementioned fig physiological response data, calculate the ratio between net photosynthetic rate and leaf area index, which is used as the photosynthetic efficiency per unit leaf area. S32: Calculate the coordinated rate of change between the net photosynthetic rate and the leaf area index; Specifically, the formula for calculating the coordinated change rate between the net photosynthetic rate and the leaf area index is as follows: ; in, This represents the coordinated rate of change between net photosynthetic rate and leaf area index. Indicates net photosynthetic rate, Indicates leaf area index, This represents the net photosynthetic rate collected in the previous cycle. This represents the leaf area index collected in the previous period. This represents the rate of change in net photosynthetic rate. This indicates the rate of change of the leaf area index; S33: Construct a time window, take the currently collected net photosynthetic rate as the end data value of the time window, calculate the mean and standard deviation of all net photosynthetic rates within the time window, and calculate the ratio between the mean and the standard deviation as an indicator of the physiological response stability of the net photosynthetic rate. Specifically, the length of the time window is 8; S34: The photosynthetic efficiency per unit leaf area, the rate of coordinated change, and the physiological response stability index are used as physiological characteristics of figs.

[0033] Specifically, the photosynthetic efficiency per unit leaf area, the rate of coordinated change, and the physiological response stability index are respectively as follows: .

[0034] Step S3 utilizes a spectral demand intelligent inference model based on the physiological state of figs to receive the physiological characteristics of figs and outputs the target spectral parameters corresponding to the fig's growth stage. This also includes: The intelligent inference model for spectral demand includes a target dominant wavelength calculation module and a target light intensity calculation module; S35: The target dominant wavelength calculation module calculates the target dominant wavelength based on the response of the fig's physiological characteristics to the photosynthetic sensitive area; Specifically, the formula for calculating the target dominant wavelength is as follows: ; in, Indicates the target dominant wavelength. Indicates the standard center wavelength; S36: The target light intensity calculation module adjusts the baseline light energy of the fig cultivation area by combining the photosynthetic efficiency per unit leaf area and the physiological response stability index to generate the target light intensity. Specifically, the formula for calculating the target light intensity is: ; in, Indicates the target light intensity. Indicates the reference light energy. Indicates the light adjustment coefficient, set Both are 0.5; S37: Use the target dominant wavelength and target light intensity as target spectral parameters.

[0035] It should be noted that this invention achieves a quantitative correlation between the physiological state of figs and light parameters by constructing an intelligent inference model of spectral demand that includes a target dominant wavelength calculation module and a target light intensity calculation module. Specifically, based on the physiological characteristics of figs and their response to the photosynthetic sensitive zone, this invention dynamically calculates the target dominant wavelength, enabling the main peak of the output spectrum to adaptively adjust with changes in photosynthetic demand, effectively improving the absorption efficiency of fig leaves for effective wavelength light energy. Simultaneously, this invention combines photosynthetic efficiency per unit leaf area and physiological response stability indicators to implement dual-factor regulation of baseline light energy. While promoting enhanced photosynthesis, it suppresses physiological stress caused by excessive light, thereby providing refined and differentiated light configurations for figs at different growth stages, reducing ineffective energy consumption, improving light energy utilization and the stability of cultivation environment regulation, thus contributing to the synergistic improvement of fig growth efficiency and quality.

[0036] S4: Based on the target spectral parameters and spectral shift characteristics of the fig during its growth period, a dynamic spectral correction model is used to output spectral correction parameters. Based on these spectral correction parameters, the light power of multiple spectral bands in the fig cultivation area is dynamically adjusted to supplement light.

[0037] The S4 step utilizes a dynamic spectral correction model to output spectral correction parameters, including: The dynamic spectral correction model includes an input layer, a feature mapping layer, and a power allocation output layer, wherein the feature mapping layer is a hidden layer structure. The process of outputting spectral correction parameters using the dynamic spectral correction model is as follows: S41: The output layer receives the target spectral parameters and spectral shift features of the fig growth period, and splices the target main wavelength and spectral shift features in the target spectral parameters to obtain the fig cultivation spliced ​​feature vector. S42: The feature mapping layer performs multi-layer mapping on the fig cultivation splicing feature vector to obtain a multi-layer state vector; Specifically, the multi-level mapping formula for the fig cultivation splicing feature vector is as follows: ; in, This represents the feature vector of fig cultivation splicing. The corresponding state vector of the h-th layer, This represents the mapping weight matrix of the h-th layer. Let H represent the mapping bias at layer h, where H represents the layer number of the state vector. This indicates the activation function; we set the activation function to the Sigmoid function. S43: The power distribution output layer receives the target light intensity and the multi-layer state vector, and generates the illumination power of each spectral band as a spectral correction parameter.

[0038] As an embodiment of the present invention, the formula for generating the illumination power of each spectral band is as follows: ; ; in, Let M represent the illumination power of the m-th spectral band, and M represent the total number of spectral bands. This represents the power mapping weight matrix of the m-th spectral band to the h-th layer state vector. This represents the bias term for the m-th spectral band. This represents the center of the spectral wavelength range of the m-th spectral band. This represents an exponential function with the natural constant as its base.

[0039] As an embodiment of the present invention, multiple sets of target spectral parameters, spectral shift features, and light power of each spectral band after manual adjustment from different fig cultivation areas are obtained as training datasets. Based on the training dataset, a training loss function is constructed with the objective of minimizing the difference between the light power of each spectral band after manual adjustment and the light power output by the dynamic spectral correction model. The training loss function is optimized using gradient descent algorithm or Adam optimizer to achieve iterative optimization of trainable parameters in the dynamic spectral correction model. The trainable parameters include the mapping weight matrix, the mapping bias, and the power mapping weight matrix.

[0040] Step S4, based on the spectral correction parameters, involves dynamically adjusting the supplemental lighting power across multiple spectral bands in the fig cultivation area, and also includes: The spectral correction parameters are sent to the multi-channel supplemental lighting control unit of the fig cultivation area. The multi-channel supplemental lighting control unit includes multiple supplemental lighting units, each of which provides supplemental lighting for a specific spectral band. By adjusting the light power of the supplemental lighting units, the light power of each spectral band is dynamically adjusted.

[0041] It should be noted that this invention achieves unified modeling of target spectral requirements and actual spectral shift states by constructing a dynamic spectral correction model that includes a feature mapping layer and a power allocation output layer. Specifically, this invention fully characterizes the comprehensive impact of environmental disturbances on spectral distribution by splicing and mapping the target dominant wavelength and multidimensional spectral shift features, and then calculates multi-layer state vectors to obtain nonlinear spectral control relationships. Under the constraint of target light intensity, the power allocation output layer adaptively allocates the illumination power of each band according to the degree of deviation between the center wavelength of each spectral band and the target dominant wavelength, so that the supplementary light energy is concentrated in the highly efficient photosynthetic sensitive area (i.e., the target dominant wavelength region). At the same time, this invention combines a multi-channel supplementary light control unit to achieve continuous and smooth adjustment of multi-band illumination power, which is beneficial to improve the accuracy of supplementary light, reduce ineffective energy consumption, and enhance the stability and consistency of the physiological response of figs.

[0042] For reference Figure 2 The diagram shows a data acquisition and supplemental lighting control structure for a fig cultivation area, including the deployment of environmental sensing devices and a multi-channel supplemental lighting control unit. The diagram includes environmental sensing devices, a multi-channel supplemental lighting control unit, and figs. The environmental sensing device integrates temperature, humidity, carbon dioxide concentration sensors and a spectrometer, which can collect planting environment parameters and spectral data in real time; the multi-channel supplementary lighting control unit has multiple supplementary lighting units built in, which can dynamically adjust the supplementary lighting parameters based on the sensing data.

[0043] During use, the environmental sensing device continuously monitors the cultivation environment data of the fig cultivation area. The multi-channel supplementary lighting control unit calculates the spectral correction parameters based on the monitoring results and precisely controls the light power of each supplementary lighting unit to achieve intelligent adaptation of the light environment for fig cultivation, thereby improving planting efficiency and crop quality.

[0044] It should be noted that the terms "comprising," "including," or any other variations thereof used herein are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0045] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0046] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An artificial intelligence-driven method for regulating supplemental lighting in fig cultivation, characterized in that, The method includes: S1: Collect cultivation environment data and fig physiological response data from fig cultivation areas; S2: Extract spectral shift features based on the cultivation environment data of the fig cultivation area; S3: Based on the fig physiological response data, extract the fig physiological characteristics, use the intelligent inference model of spectral demand based on the fig physiological state to receive the fig physiological characteristics, and output the target spectral parameters corresponding to the fig growth period. S4: Based on the target spectral parameters and spectral shift characteristics of the fig during its growth period, a dynamic spectral correction model is used to output spectral correction parameters. Based on these spectral correction parameters, the light power of multiple spectral bands in the fig cultivation area is dynamically adjusted to supplement light.

2. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 1, characterized in that, Step S1 involves collecting cultivation environment data and fig physiological response data from the fig cultivation area, including: S11: Deploy environmental sensing devices in the fig cultivation area and periodically collect cultivation environment data of the fig cultivation area using distributed environmental sensing devices. The cultivation environment data includes spectral data, ambient temperature data, relative humidity data, and carbon dioxide concentration data. The spectral data is the intensity of reflected light on the surface of fig leaves at multiple spectral wavelengths collected by the environmental sensing devices. S12: Based on the cultivation environment data of the fig cultivation area, perform inversion calculation on the net photosynthetic rate and leaf area index of fig leaves in the fig cultivation area. S13: The net photosynthetic rate and leaf area index are used as the physiological response data of figs in the fig cultivation area.

3. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 2, characterized in that, Step S12 involves inverting the calculation of the net photosynthetic rate and leaf area index of fig leaves in the fig cultivation area, including: S121: Based on the spectral data, extract the reflected light intensity in the red light band and near-infrared band respectively, and obtain the incident light intensity in the red light band and near-infrared band respectively from the environmental sensing device, calculate the ratio of reflected light intensity to incident light intensity, and use it as the band reflectivity of the red light band and near-infrared band. S122: Based on the band reflectance of the red light band and the near-infrared band, the normalized vegetation index is calculated, and the normalized vegetation index is inverted into the light absorption ratio. S123: Perform integration on the spectral data to generate photosynthetically effective reflectance, and calculate the net photosynthetic rate based on the light absorption ratio and photosynthetically effective reflectance. S124: The leaf area index is obtained by inverting the normalized vegetation index based on the empirical regression coefficient.

4. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 1, characterized in that, The extraction of spectral shift features in step S2 includes: S21: Based on the cultivation environment data of the fig cultivation area, calculate the spectral centroid of the spectral data, and calculate the offset value between the spectral centroid and the standard center wavelength, as the spectral main peak offset value. S22: Based on the ambient temperature data, relative humidity data, and carbon dioxide concentration data, the spectral main peak shift value is modulated by environmental modulation to obtain the environmentally modulated spectral main peak shift value. S23: Extract the reflected light intensity corresponding to the spectral centroid and the standard center wavelength from the spectral data respectively, and calculate the ratio between the reflected light intensity corresponding to the spectral centroid and the reflected light intensity corresponding to the standard center wavelength as the light intensity attenuation ratio. S24: Construct standard reflected light intensities for different spectral wavelengths, perform difference integration on the spectral data, and generate effective spectral offset; S25: The main peak shift value of the environmental modulation spectrum, the light intensity attenuation ratio, and the effective spectral shift are used as spectral shift features.

5. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 4, characterized in that, The formula for environmental modulation of the spectral main peak shift value based on the ambient temperature data, relative humidity data, and carbon dioxide concentration data in step S22 is as follows: ； ； in, Indicates the spectral peak shift value Environmental modulation results Indicates the ambient temperature modulation parameters. This represents the relative humidity modulation parameter. This represents the carbon dioxide concentration modulation parameter. The modulation weights of the ambient temperature modulation parameter, relative humidity modulation parameter, and carbon dioxide concentration modulation parameter are represented in turn. The data represent ambient temperature, relative humidity, and carbon dioxide concentration, respectively. These represent the reference values ​​for ambient temperature, relative humidity, and carbon dioxide concentration, respectively. These represent the normalized scaling factors for ambient temperature data, relative humidity data, and carbon dioxide concentration data, respectively. Represents the hyperbolic tangent function. This represents the natural logarithm function.

6. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 1, characterized in that, The step S3 involves extracting physiological characteristics of figs, including: S31: Based on the aforementioned fig physiological response data, calculate the ratio between net photosynthetic rate and leaf area index, which is used as the photosynthetic efficiency per unit leaf area. S32: Calculate the coordinated rate of change between the net photosynthetic rate and the leaf area index; S33: Construct a time window, take the currently collected net photosynthetic rate as the end data value of the time window, calculate the mean and standard deviation of all net photosynthetic rates within the time window, and calculate the ratio between the mean and the standard deviation as an indicator of the physiological response stability of the net photosynthetic rate. S34: The photosynthetic efficiency per unit leaf area, the rate of coordinated change, and the physiological response stability index are used as physiological characteristics of figs.

7. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 6, characterized in that, Step S3 utilizes a spectral demand intelligent inference model based on the physiological state of figs to receive the physiological characteristics of figs and outputs the target spectral parameters corresponding to the fig's growth stage. This also includes: The intelligent inference model for spectral demand includes a target dominant wavelength calculation module and a target light intensity calculation module; S35: The target dominant wavelength calculation module calculates the target dominant wavelength based on the response of the fig's physiological characteristics to the photosynthetic sensitive area; S36: The target light intensity calculation module adjusts the baseline light energy of the fig cultivation area by combining the photosynthetic efficiency per unit leaf area and the physiological response stability index to generate the target light intensity. S37: Use the target dominant wavelength and target light intensity as target spectral parameters.

8. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 1, characterized in that, The S4 step utilizes a dynamic spectral correction model to output spectral correction parameters, including: The dynamic spectral correction model includes an input layer, a feature mapping layer, and a power allocation output layer, wherein the feature mapping layer is a hidden layer structure. The process of outputting spectral correction parameters using the dynamic spectral correction model is as follows: S41: The output layer receives the target spectral parameters and spectral shift features of the fig growth period, and splices the target main wavelength and spectral shift features in the target spectral parameters to obtain the fig cultivation spliced ​​feature vector. S42: The feature mapping layer performs multi-layer mapping on the fig cultivation splicing feature vector to obtain a multi-layer state vector; S43: The power distribution output layer receives the target light intensity and the multi-layer state vector, and generates the illumination power of each spectral band as a spectral correction parameter.

9. The artificial intelligence-driven method for supplemental lighting control in fig cultivation as described in claim 8, characterized in that, Step S4, based on the spectral correction parameters, involves dynamically adjusting the supplemental lighting power across multiple spectral bands in the fig cultivation area, and also includes: The spectral correction parameters are sent to the multi-channel supplemental lighting control unit of the fig cultivation area. The multi-channel supplemental lighting control unit includes multiple supplemental lighting units, each of which provides supplemental lighting for a specific spectral band. By adjusting the light power of the supplemental lighting units, the light power of each spectral band is dynamically adjusted.

Images