A method and device for staged switching control of LED multispectral signal in white chicken farming

Abstract

This invention discloses a method and device for staged switching control of LED multispectral spectrum in white chicken farming. By acquiring spectral sampling data and band configuration data, it extracts rhythmic response coefficients according to diurnal rhythm periods to construct a band adaptation benchmark. It then conducts hierarchical identification of matching degree to determine the growth stages with varying strengths and weaknesses. Combining historical batch decay baselines and batch deviation sequences, it dynamically corrects and determines the threshold, assesses the band decay demand level, and identifies the tiered spectral response zones. Synchronous drift detection and independence verification are performed on the duty cycle offset spectrum to generate band correction coefficients. Combined with coupling interference detection and dispersion verification, it determines the effective spectral boundary. Based on the energy consumption offset rate and efficiency ratio sequence, it outputs the optimal band configuration and extracts the energy-saving buffer space. Combining the control priority sequence and switching trigger conditions, it completes the step-by-step elimination of redundant bands and outputs LED spectral switching control commands, achieving dynamic optimization of spectral configuration and precise control of switching timing at each growth stage.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for livestock and poultry farming, and in particular to a method and device for staged switching control of LED multispectral technology in the breeding of white chickens. Background Technology

[0002] The retinas of broiler chickens contain multiple types of cone cells, and their physiological responses to different wavelengths of light exhibit significant stage-specific differences. The spectral composition and illuminance requirements for each growth stage from hatching to market weight continuously change with age. Currently, most farms use LED lighting systems with fixed spectral parameters, failing to dynamically adjust the wavelength configuration according to the actual needs of each growth stage. This results in a persistent discrepancy between illuminance supply and physiological requirements at certain growth stages, affecting the triggering of feeding behavior and the normal establishment of growth hormone secretion rhythms.

[0003] During continuous operation, LED light sources experience a gradual decrease in actual illuminance across different wavelengths due to device aging. The varying drive intensities across different wavelength channels lead to inconsistent attenuation rates. Existing control systems lack the ability to dynamically track the actual aging state of each wavelength band, and the discrepancy between switching commands and actual light output increases with runtime. Furthermore, existing systems lack differentiated assessment of the physiological response contribution per unit power for each wavelength band when driving multiple bands simultaneously. They fail to identify redundant wavelengths with low physiological response benefits and release corresponding power resources, resulting in overall low power efficiency in the poultry house. Therefore, it is necessary to develop a new switching control method to meet the practical requirements of spectral configuration accuracy and energy consumption optimization for chicks at different growth stages. Summary of the Invention

[0004] This invention discloses a method and device for staged switching control of LED multispectral spectrum in broiler chicken farming. The aim is to establish a band adaptation benchmark and, through hierarchical identification of the strength and weakness of demand during growth stages and the level of band decay demand, determine the tiered spectral response zone, thereby achieving dynamic matching of spectral configuration with circadian rhythm requirements at each growth stage. Effective configuration boundaries for each band are determined through duty cycle drift analysis and dispersion verification. Optimal band configuration is generated by combining energy consumption offset rate and efficiency ratio evaluation, and energy-saving buffer space is extracted. Redundant bands are progressively eliminated and LED spectral switching control commands are output according to the control priority sequence and switching trigger conditions, providing spectral supply that meets the physiological needs of broiler chickens at each growth stage.

[0005] The first aspect of this invention proposes a method for staged switching control of LED multispectral technology in white chicken farming, comprising the following steps: Acquire spectral sampling data and band configuration data of the white chick, and perform band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark; Based on the band adaptation benchmark, the matching degree is hierarchically identified to identify the strong and weak demand growth stages. The band decay demand level is evaluated based on the stage distribution of the strong and weak demand growth stages. The gradient spectral response region is determined based on the band decay demand level. Duty cycle drift analysis is performed on the spectral sampling data to form a duty cycle shift map. Systematic shift features are extracted from the duty cycle shift map to generate band correction coefficients. Based on the duty cycle shift map and the band correction coefficients, a dispersion check is performed to determine the effective spectral boundary. Based on the band decay demand level, the energy consumption offset rate is extracted. Based on the energy consumption offset rate and the band correction coefficient, the optimal band configuration is generated. The non-critical band reduction margin is extracted from the optimal band configuration to generate an energy-saving buffer space. The control priority sequence is formulated by combining the energy-saving buffer space and the stepped spectral response region. The control priority sequence is used to carry out band hierarchical arrangement to form a driving scheduling arrangement. Based on the driving scheduling arrangement and the effective spectral boundary, the switching trigger condition is determined. According to the switching trigger condition, the driving scheduling arrangement is used to eliminate redundant bands step by step and output LED spectral switching control command.

[0006] The second aspect of this invention provides a multi-spectral LED staged switching control device for raising white chickens, comprising: The benchmark construction module is used to acquire the spectral sampling data and band configuration data of the white chicken, and to perform band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark. The decay analysis module is used to identify the strong and weak demand growth stages based on the band adaptation benchmark, evaluate the band decay demand level based on the stage distribution of the strong and weak demand growth stages, and determine the gradient spectral response region based on the band decay demand level. The offset evaluation module is used to perform duty cycle drift analysis on the spectral sampling data to form a duty cycle offset map, extract systematic offset features from the duty cycle offset map to generate band correction coefficients, and perform dispersion verification based on the duty cycle offset map and the band correction coefficients to determine the effective spectral boundary. The configuration planning module is used to extract the energy consumption offset rate based on the band decay demand level, generate the optimal band configuration based on the energy consumption offset rate and the band correction coefficient, extract the non-critical band reduction margin from the optimal band configuration to generate an energy-saving buffer space, and formulate a control priority sequence by combining the energy-saving buffer space with the gradient spectral response region. The switching output module is used to perform band hierarchical arrangement to form a driving scheduling arrangement using the control priority sequence, determine the switching trigger condition based on the driving scheduling arrangement and the effective spectral boundary, and output LED spectrum switching control command by eliminating redundant bands in the driving scheduling arrangement according to the switching trigger condition.

[0007] The beneficial effects of this invention are reflected in the following points: 1. By extracting the rhythm response coefficients and calculating the rhythm matching degree from the spectral sampling data according to the diurnal rhythm time period, a dynamic correspondence between band configuration and the physiological needs of chicks at each growth stage is established; by combining the historical batch band decay baseline and batch deviation sequence to dynamically correct the judgment threshold, adaptive evaluation of the band decay demand level is realized, and the demand matching accuracy of spectral configuration at each growth stage continues to improve with the accumulation of batch data. 2. By conducting multi-band synchronous drift detection and band independence verification on the duty cycle offset spectrum, the global disturbance component and the independent aging drift component of each band are effectively separated; by combining power coupling interference detection and dispersion verification, the effective boundary of each band spectrum is determined on the basis of eliminating false deviations introduced by high-power joint drive, so that subsequent switching control commands are executed within the actual achievable illuminance range of each band, reducing the systematic bias between control commands and actual light output state. 3. By using the efficiency ratio sequence and comprehensive elimination scoring mechanism to identify band combinations that do not contribute enough to the physiological response per unit power, driving resources are concentrated and allocated to high-efficiency bands under the constraint of total power in the breeding house. The switching trigger sequence of each channel is determined by combining the advance trigger compensation amount and the urgency ranking mechanism. LED spectrum switching control commands are generated by eliminating redundant bands step by step. While ensuring that the target illuminance of each growth stage arrives on time, the energy-saving buffer space of non-critical bands is released, thereby improving the overall power efficiency of the breeding house. Attached Figure Description

[0008] Figure 1 This is a flowchart of a multi-spectral LED phased switching control method for raising white chickens according to the present invention.

[0009] Figure 2 This is a structural block diagram of an LED multispectral staged switching control device for raising white chickens according to the present invention. Detailed Implementation

[0010] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0011] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0012] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0013] The technical solutions of the embodiments of this application will be described below.

[0014] like Figure 1 As shown, this embodiment of the invention provides a method for staged switching control of LED multispectral technology in white chicken farming, including the following steps S110-S150: Step S110: Obtain the spectral sampling data and band configuration data of the white chicken, and perform band adaptation analysis on the spectral sampling data and band configuration data to construct a band adaptation benchmark.

[0015] Specifically, spectral sampling data and band configuration data of broiler chickens were acquired. The retina of broiler chickens contains four types of cone cells, which have different photosensitivity characteristics to four bands: ultraviolet (320-400nm), blue light (440-490nm), green light (500-570nm), and red light (620-700nm). Different bands of light stimulation have differential effects on feeding behavior, resting state, and growth hormone secretion at different times of the day and night. Continuous output of high-intensity short-wavelength light during the resting period will inhibit melatonin secretion and lead to group stress. Long-term accumulation of this deviation affects the feed conversion ratio and the slaughter cycle. Spectral sampling data was continuously collected at 15-minute intervals by multi-channel spectral sensors evenly distributed on the top of the breeding shed. The sensors cover the complete photosensitive range of 320nm to 750nm. The spectral sampling data is composed of illuminance values ​​for each band, organized and stored using a dual index of sampling time and band center wavelength. The unit of illuminance values ​​for each band is μmol·m⁻¹. -2 ·s -1The coverage range of the center wavelength of the spectral sampling data corresponds to the center wavelength of each channel in the band configuration data. The band configuration data is generated from the current configuration file of the LED driver controller in the breeding house, and includes the center wavelength, rated power, current duty cycle setting value and corresponding measured illuminance value of each LED channel. The band configuration data is read independently with the channel number as the main index. The rated power of each channel in the band configuration data is used for the calculation of the band adaptation reference power constraint.

[0016] In some embodiments, the step of performing band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark includes: extracting the response intensity distribution of each band from the spectral sampling data according to the diurnal rhythm time period to generate a rhythm response distribution table; identifying rhythm-sensitive bands from the rhythm response distribution table to generate rhythm response coefficients; calculating the rhythm matching degree using the rhythm response coefficients and the band configuration data to generate a rhythm adaptation sequence; and constructing a band adaptation benchmark according to the rhythm adaptation sequence.

[0017] The spectral sampling data was processed according to the diurnal rhythm to extract the response intensity distribution of each band and generate a rhythmic response distribution table. The spectral sampling data was grouped into four time periods: the morning light activation period (04:00 to 06:00), the daytime high-activity period (06:00 to 18:00), the dusk transition period (18:00 to 20:00), and the nighttime rest period (20:00 to 04:00 the next day). Within each time period, the mean and standard deviation of illuminance were calculated for each band. The morning light activation period lasted 120 minutes and corresponded to 8 sampling frames, while the daytime high-activity period lasted 720 minutes and corresponded to 48 sampling frames. The visual sensitivity characteristics of chicks exhibit systematic differences across different time periods. The response intensity to the blue light band was significantly higher during the morning light activation period than during the nighttime rest period. Even slight fluctuations in red light illuminance during the nighttime rest period could cause disturbance to the flock. If a uniform model was used to calculate the mean throughout the day, the differences in the needs of each time period would cancel each other out, resulting in a loss of segmented identification ability. During the daytime high-activity period, the increased activity of the chickens leads to enhanced background scattering from the sensors. For high-variability bands with a standard deviation exceeding 30% of the mean for the same period, the median is used instead of the arithmetic mean to mitigate the outlier effect. Missing sampled frames are filled by linear interpolation of adjacent valid frames with added missing markers; frames with missing markers are included in the statistical calculation with a 0.5x weight. The rhythm response distribution table establishes a dual-index structure based on the band center wavelength and the diurnal rhythm period, with each intersection recording the mean response of that band during that period. The rhythm response distribution table is continuously updated across breeding batches. When new batch data arrives, each unit is refreshed based on the weighted mean of the old and new batches. The weighting coefficients are distributed exponentially in reverse batch number order, ensuring that recent batch sampling characteristics have a higher response weight.

[0018] Rhythm-sensitive bands are identified and rhythm response coefficients are generated from the rhythm response distribution table. In the rhythm response distribution table, the ratio of the maximum difference between the mean values ​​of each band across time periods to the mean value across all time periods is defined as the rhythm sensitivity index of that band. Bands with a rhythm sensitivity index higher than 0.25 are identified as rhythm-sensitive bands. This threshold is calibrated based on the distribution characteristics of historical aquaculture batch spectral sampling data. Under this threshold, the false negative rate for identifying rhythm-sensitive bands is controlled within 8%. Bands in the rhythm response distribution table whose rhythm sensitivity index does not exceed the threshold indicate that their response differences between time periods do not have significant segmentation identification capabilities. Including them in the weight calculation only introduces invalid perturbations. Therefore, non-rhythm-sensitive bands are assigned a rhythm response coefficient of 1.0 in each time period to indicate that no differential weight is applied. The specific values ​​of the rhythm response coefficients are obtained through normalization: using the average response of each band throughout the entire time period as the benchmark, the response values ​​of each time period in the rhythm response distribution table are divided by the benchmark to obtain the dimensionless coefficients. A coefficient greater than 1.0 indicates that the response of that time period is higher than the daily average level, and a coefficient less than 1.0 indicates that it is lower than the daily average level. For the aforementioned high-variability bands, the median is used instead of the arithmetic mean in the normalization calculation, consistent with the processing method in the statistical extraction stage. The rhythm response coefficients are organized by the band center wavelength index. Each band corresponds to four normalized coefficient values ​​for the morning light activation segment, the daytime high-activity segment, the dusk transition segment, and the nighttime rest segment. The time period number of the rhythm response coefficient is used as the lookup key in the rhythm matching degree calculation to match the real-time time period.

[0019] A rhythm matching sequence is generated by calculating the rhythm matching degree using rhythm response coefficients and band configuration data. The center wavelength of each LED channel in the band configuration data is aligned to the main index by performing a nearest neighbor search in the wavelength index of the rhythm response coefficient. Channels with a center wavelength deviation exceeding 10nm are marked as wavelength deviation channels, and the rhythm response coefficient of the nearest neighbor band is substituted into the data, with an additional wavelength deviation correction factor added for compensation. The rhythm matching degree is calculated as M_t=Σ(D_i×R_i,t×α_i) / ΣD_i, where M_t is the rhythm matching degree of time period t, D_i is the duty cycle setting value of the i-th channel, R_i,t is the normalized value of the i-th channel in the rhythm response coefficient at time period t, and α_i is the wavelength deviation correction factor of the i-th channel. For channels with a center wavelength deviation of less than 10nm, α_i is set to 1.0; for channels with a deviation of more than 10nm, α_i=1.0-0.1×(deviation / 10nm), so that the correction factor is set to 1.0 when the deviation is 0 and 0.9 when the deviation reaches 10nm. The boundaries of each time period are dynamically offset with reference to the sunrise and sunset times of the area where the breeding house is located. The offset amount does not exceed ±30 minutes of the nominal boundary of each time period to adapt to the seasonal light cycle drift. The rhythm adaptation sequence covers the 24-hour time domain and is segmented according to the diurnal rhythm period. Within each segment, a matching degree vector and a time period number are recorded. Channel combinations with a matching degree lower than 0.6 are marked as low-fit and are not included in the candidate benchmark set for band adaptation. When the band configuration data is updated and the duty cycle setting of any channel changes by more than 5%, the rhythm adaptation sequence is recalculated to avoid frequent recalculations caused by daily fine-tuning, which could lead to oscillations in the control strategy.

[0020] Band-matching benchmarks are constructed based on the rhythm-matching sequence. The matching degree vectors for each time period of the rhythm-matching sequence are sorted in descending order, and the top three channel combinations constitute the candidate benchmark set for that time period. Band-matching benchmarks are selected from the candidate benchmark set based on power constraints. The power constraint requires that the total power of each channel configuration does not exceed 85% of the upper limit of the rated power of the poultry house. The upper limit of the rated power is obtained by summing the rated power of each channel in the band configuration data. Since low-fitting-mark channel combinations have been excluded from the candidate benchmark set during the rhythm-matching sequence generation stage, the benchmark configurations for each time period of the band-matching benchmark all meet the lower limit requirement of rhythm matching degree. The degree of difference in channel combinations between candidate benchmark sets of adjacent time periods is measured using the Jaccard coefficient. When the Jaccard coefficient is below 0.4, it indicates that the transition at that time period involves a significant channel switch. The band-matching benchmark inserts a transition segment marker at this transition point to guide the switching control to gradually complete the band transition in a gradual manner to prevent abrupt spectral jumps from causing stress in the flock. The band adaptation benchmark is organized according to the diurnal rhythm time period. Each time period records the benchmark channel combination, its benchmark duty cycle setting value, and rhythm matching degree. If the matching degree vector of a certain time period is missing in the rhythm adaptation sequence, the benchmark configuration of the adjacent time period is used to fill the gap and a filling mark is added. The time period with the filling mark is given corresponding weight reduction in the subsequent matching degree stratification and control priority sequence formulation stage.

[0021] Step S120: Based on the band adaptation benchmark, identify the strong and weak demand growth stages by matching degree stratification, assess the band decay demand level according to the stage distribution of the strong and weak demand growth stages, and determine the gradient spectral response zone based on the band decay demand level.

[0022] Specifically, based on the band adaptation benchmark, a stratified identification of strong and weak demand growth periods is conducted. From hatching to market weight, chicks undergo three growth stages: the brooding period (0-7 days old), the early rearing period (8-21 days old), and the late rearing period (22 days old to market weight). Each stage exhibits systematic differences in physiological response requirements to different bands. Insufficient energy supply in high-demand stages will suppress daily weight gain, while excessive band stimulation in low-demand stages will cause flock disturbance and additional energy consumption. The band adaptation benchmark divides the rhythm matching degree of each time period into preset stratification thresholds: below 0.65 is defined as a strong demand period, above 0.80 as a weak demand period, and 0.65 to 0.80 is classified as a transition period. Setting an intermediate range aims to avoid repeated changes in classification labels due to slight fluctuations in matching degree near the threshold. The time periods with added filler marks are grouped based on the majority of the adjacent valid time periods, and are not directly compared to the threshold using the filler matching degree value. During each growth stage, periods with strong demand exceeding 40% are marked as strong-demand-dominated growth periods, and periods with weak demand exceeding 50% are marked as weak-demand-dominated growth periods. When both are triggered simultaneously, strong-demand dominance takes precedence, as insufficient illumination supply has a greater inhibitory effect on daily weight gain than excessive illumination causes disturbance. If neither is triggered, the growth period is classified as balanced and recorded with a neutral label. The stage distribution of strong and weak demand growth periods covers the entire breeding cycle and the four diurnal rhythm periods. The labeling results are used as a grouping index in the stages of decline baseline extraction and band decline demand level determination.

[0023] In some embodiments, assessing the band decline demand level based on the phase distribution of the strong and weak demand growth periods includes: extracting historical batch band demand decline baselines for the strong and weak demand growth periods to generate a decline baseline distribution table; calculating the deviation between the decline baseline distribution table and the strong and weak demand growth periods to generate a batch deviation sequence; dynamically correcting a preset decline level judgment threshold according to the batch deviation sequence to generate a correction threshold set; and fusing the correction threshold set with the phase distribution of the strong and weak demand growth periods to determine the band decline demand level.

[0024] A degradation baseline distribution table is generated by extracting historical batch band demand degradation baselines for the strong and weak demand growth periods. These historical batch band demand degradation baselines are extracted from spectral sampling data of the last three aquaculture batches. The measured illuminance degradation rate for each historical batch corresponding to the same band at each stage of the strong and weak demand growth period is calculated as follows: the difference between the average measured illuminance when the current batch enters this growth stage and the average measured illuminance on the first day of the same batch is divided by the average of the first day, and the absolute value is expressed as a percentage. LED light sources degrade batch by batch due to chip junction temperature accumulation and phosphor aging. The degradation rate is faster during the strong demand-dominated growth period due to the higher drive duty cycle, and slower during the weak demand-dominated growth period due to the lower duty cycle. The differences in degradation rates at each stage of the strong and weak demand growth period must be independently recorded in the degradation baseline distribution table to support the stage-by-stage classification. The degradation baseline values ​​at the same position from three historical batches are aggregated using a weighted method that increases in order of batch age, with weighting coefficients of 0.6, 0.8, and 1.0 respectively. The weighted mean is calculated by multiplying the degradation baseline values ​​of the three batches by their respective weighting coefficients, summing the results, and then dividing by the sum of the weighting coefficients, which is 2.4. The most recent batch has the highest weight to reflect the actual aging state of the current LED devices. The degradation baseline distribution table is indexed by a dual index based on the center wavelength of the band and the growth stage of the growth period of the demand for strong and weak light. At each intersection position of the degradation baseline distribution table, the within-group standard deviation of the weighted mean degradation baseline value and the degradation baseline values ​​of the three batches is recorded. The within-group standard deviation of the degradation baseline distribution table is used for confidence weight conversion during the deviation sequence calculation stage. When there are fewer than three historical batches, the data is filled with typical LED degradation curve data from the industry and an external filling mark is added. The position of the external filling mark is included in the deviation sequence calculation of subsequent batches with a reduced weight of 0.7 to reflect the adaptation deviation between the external data and the actual operating conditions of the breeding shed.

[0025] A batch deviation sequence is generated by calculating the deviation between the degradation baseline distribution table and the growth periods of strong and weak demand. The deviation is defined as the difference between the measured degradation rate of the current batch at each stage of the growth period of strong and weak demand and the weighted average degradation baseline at the corresponding position in the degradation baseline distribution table. A positive difference indicates that the current batch's degradation rate is higher than the historical baseline, i.e., aging is accelerated, while a negative difference indicates that it is lower than the historical baseline. The degradation baseline distribution table provides the expected degradation rate for each band based on historical experience. When the deviation is consistently positive, it indicates that the aging progress of the current batch of LED devices is exceeding expectations, and band compensation adjustment must be initiated in advance. If this deviation is not calculated and historical thresholds are used directly, the degradation level determination will lag behind the actual aging progress, resulting in a delay in the timing of illuminance compensation. The deviations at each stage of the growth period of strong and weak demand are accumulated sequentially in the independent sequences of each band according to the age. The within-group standard deviation of each band in the degradation baseline distribution table is used to calculate the confidence weight of the deviation within the band. The stage grouping index of the growth period of strong and weak demand serves as the basis for organizing the deviation sequence. The batch deviation sequence is organized according to the stage number of the growth period of strong and weak demand. For each band within each stage of the batch deviation sequence, the mean deviation and the confidence weight calculated based on the standard deviation within the group of the decline baseline distribution table are recorded. The deviation at the external padding marker position is also included with a 0.7-fold weight reduction. When the mean deviation of any band exceeds the baseline value by 15% for two consecutive growth stages, an accelerated decline warning marker is added to the batch deviation sequence. This marker will trigger additional threshold tightening during the correction threshold set generation stage. The bands in the batch deviation sequence with the accelerated decline warning marker receive additional threshold tightening in the correction coefficient calculation.

[0026] The preset degradation level judgment thresholds are dynamically corrected according to the batch deviation sequence to generate a set of corrected thresholds. The preset degradation level judgment thresholds are calibrated based on the industry standard LED life curve and include three levels: Level 1 (illuminance degradation rate exceeding 10%), Level 2 (exceeding 20%), and Level 3 (exceeding 35%). These static thresholds fail to reflect the impact of actual operating intensity in different breeding sheds and individual device differences. Dynamic correction tightens or maintains the preset thresholds for each level through a correction coefficient κ. The correction coefficient κ is calculated as κ = 1 + k × Δ, where k is the correction sensitivity coefficient, and Δ is the average deviation of the current growth stage of that band in the batch deviation sequence, substituted in decimal form. The corrected threshold is calculated as T' = T / κ, where T is the original threshold of the corresponding level in the preset degradation level judgment thresholds, and T' is the threshold of that level after dynamic correction. When κ is greater than 1, T' is less than T, indicating that the threshold is tightened, and the degradation level judgment is more easily triggered. During the high-demand-driven growth phase, k is set to 1.2 to increase the threshold tightening, as this stage has high driving intensity and rapid device aging. During the low-demand-driven growth phase, k is set to 0.8 to reduce threshold perturbation and avoid overly sensitive judgments during the gradual aging phase. For bands in the batch deviation sequence with added accelerated degradation warning markers, κ is further increased by 5% based on the above calculations to further tighten the judgment threshold. The three threshold levels for each band in the modified threshold set must maintain a monotonically increasing constraint. If the modified secondary threshold is lower than the primary threshold, it is reset by adding a fixed interval of 2% to the primary correction value, and the same applies to the third level. The modified threshold set is updated synchronously after each growth stage of the batch deviation sequence is completed, ensuring that the level judgment boundary continuously tracks the actual aging state of the device.

[0027] The band decay demand level is determined by integrating the modified threshold set with the stage distribution of strong and weak demand growth periods. The measured illuminance decay rate of each band at each growth stage of the strong and weak demand growth periods is compared step-by-step with the three threshold levels corresponding to the modified threshold set: below the first-level modified threshold is determined as level zero; above the first level but below the second level is determined as level one decay; above the second level but below the third level is determined as level two decay; and above the third level is determined as level three decay. During the strong demand-dominated growth period, the tightening threshold corresponding to k=1.2 in the modified threshold set is used for judgment; during the weak demand-dominated growth period, the lenient threshold corresponding to k=0.8 in the modified threshold set is used. The judgment thresholds for strong and weak demand growth periods are strictly distinguished and not mixed to avoid applying the tightening judgment standard applicable to high-intensity stages to low-intensity operation stages, resulting in misjudgment. The band decay demand level is organized by a dual index of the band center wavelength and the growth stage sequence number. Each position of the band decay demand level records the decay level label (integer from level 0 to 3) and the corresponding measured illuminance decay rate. When the number of bands with degradation levels of level 2 or above exceeds 30% of the total number of bands during the strong demand-dominated growth phase in the growth stage, a group degradation warning flag is added to the band degradation demand level, indicating that the overall aging of LED luminaires in this growth stage has reached a level requiring centralized treatment. For growth stages where the correction threshold set has not been corrected due to insufficient batch deviation sequence data, the band degradation demand level determination is reverted to the preset threshold and an uncorrected flag is added. This flag is used to reduce the confidence rating of the corresponding partition in the gradient spectral response zone delineation.

[0028] The tiered spectral response zones are determined based on the band decay demand level. In the growth stage dimension, band decay demand levels are defined as continuous zones with consecutive segments of the same level value, and each band independently identifies the boundaries of its continuous zone. In the cross-band dimension, using the band decay demand level vectors of each growth stage as clustering features, a hierarchical clustering method is employed to group bands according to the similarity of their level vectors. The clustering truncation threshold is set to 0.6 times the mean of the pairwise Euclidean distances of all band level vectors. A gradient partitioning structure for decay degree is simultaneously established in both the spectral and growth stage dimensions. The tiered spectral response zones use a two-layer identification system consisting of cluster numbers and continuous zone numbers. Clusters of bands in the growth stage where the group decay alarm marker is located within the band decay demand level are preferentially assigned to high-priority tiered spectral response zones. Each tiered spectral response zone records the list of center wavelengths of the bands it contains, the corresponding growth stage interval, the highest decay level within the zone, and the high-priority marker. Adjacent partitions with a level difference of less than 1 are merged to control the total number of partitions to not exceed 12. In the band decay demand level, the confidence level of the gradient spectral response zone corresponding to the uncorrected marker stage is set to 0.7. During the regulation priority sequence formulation stage, this type of low confidence zone is ranked after the high confidence zone of the same level to prevent zones with insufficient data reliability from obtaining excessive regulation priority.

[0029] Step S130: Perform duty cycle drift analysis on the spectral sampling data to form a duty cycle shift map, extract systematic shift features from the duty cycle shift map to generate band correction coefficients, and perform dispersion verification based on the duty cycle shift map and band correction coefficients to determine the effective spectral boundary.

[0030] Specifically, duty cycle drift analysis is performed on spectral sampling data to generate a duty cycle offset map. During the brooding period, red light illuminance must be maintained above 90% of the rated value to ensure stable triggering of phototaxis and feeding behavior in chicks. In the later stages of growth, if the actual deviation of green light illuminance exceeds 10% of the rated value, it will cause disturbances in the group's movement rhythm. The high precision requirements of illuminance at each growth stage necessitate accurate identification of drive deviations. However, due to factors such as crystal oscillator frequency deviation, thermal drift, and power supply ripple, the actual duty cycle of each channel in the LED driver circuit continuously drifts relative to the set value. If the drift component in the spectral sampling data is not separated, the adjustment of band configuration parameters will be based on observations containing drift deviations, resulting in a continuous systematic bias between control commands and the actual light output state. The duty cycle drift is extracted by subtracting the illuminance value of each band in the spectral sampling data from the theoretical illuminance calculated based on the rated efficiency coefficient. The difference is the duty cycle drift for that frame. The drift values ​​of each band in the spectral sampling data are arranged in order of sampling time and expanded into a two-dimensional matrix, the duty cycle shift map, using the band center wavelength as the row index and the sampling time number as the column index. Each cell in the duty cycle shift map stores the drift value of the corresponding band at the corresponding time. Missing sampling frames are filled by linear interpolation of adjacent valid frames and marked with missing markers. Cells with missing markers participate in subsequent calculations with reduced weights. The temporal coverage of the duty cycle shift map is strictly aligned with the running time of the current aquaculture batch to ensure a complete temporal record of drift characteristics at each stage. This complete temporal alignment of the duty cycle shift map ensures temporal consistency for subsequent segmented fitting and dispersion checks.

[0031] In some embodiments, the step of extracting systematic migration features from the duty cycle migration map to generate band correction coefficients includes: performing multi-band synchronous drift detection on the duty cycle migration map to generate a synchronous drift segment set; performing band independence verification on the synchronous drift segment set to identify asynchronous drift segments and form a decoupled segment set; calibrating the cumulative migration position of the decoupled segment set to generate a cumulative migration marker; and using the cumulative migration marker to perform segmented fitting on the duty cycle migration map to generate band correction coefficients.

[0032] Multi-band synchronous drift detection is performed on the duty cycle offset spectrum to generate a set of synchronous drift segments. The total band drift of each column of the duty cycle offset spectrum constitutes the drift cross-section vector at that moment. The difference between the drift cross-section vectors of two adjacent columns of the duty cycle offset spectrum is defined as the cross-band drift change vector. When all components in the cross-band drift change vector have the same sign and the ratio of the amplitudes of any two components falls within the range of 0.8 to 1.2, synchronous drift is determined to exist in that time period. Continuous time periods that meet the conditions are merged into a synchronous drift segment. The overall fluctuation of power supply voltage or the sudden change in the temperature of the drive controller motherboard during the concentrated power consumption period in the breeding shed are typical sources of synchronous drift in each band. If this type of global disturbance component is mixed with the drift component caused by the independent aging of each band, the subsequent correction coefficient will produce a compensation residual in the opposite direction during the next global disturbance. Synchronization drift segment sets are organized by segment number. Each segment in the synchronization drift segment set records the start and end time numbers and a list of band numbers participating in synchronization. This information is used to define segment boundaries and perform grouping judgments in subsequent band independence verification. For columns containing missing marker units, synchronization drift judgment is only performed if the number of available bands is not less than 60% of the total number of bands. Columns that do not meet the condition are marked as empty and excluded from the segment merging judgment of adjacent valid columns.

[0033] Band independence verification is performed on the synchronous drift segment set to identify asynchronous drift segments and form a decoupled segment set. Band independence verification uses the cross-correlation coefficient of the drift sequence of each band within each segment of the synchronous drift segment set as an indicator. Segments with a mean cross-correlation coefficient exceeding 0.80 indicate highly synchronized drift behavior and are classified as purely synchronous segments, directly marked as synchronous elimination segments. Segments with a mean cross-correlation coefficient below 0.80 typically belong to mixed drift segments. For these mixed segments, principal component analysis is used to decompose the drift of each band into global common components and band-independent components. The band-independent components are calculated by subtracting the projected reconstruction value of the global common components from the original drift and then retaining them. Directly eliminating mixed segments from the start and end boundaries of the synchronous drift segment set would lose effective information about the true aging characteristics of each band; therefore, the separation and extraction of independent components is crucial for preserving the drift characteristics of each band. The decoupled segment set consists of the original valid time periods not covered by the synchronous drift segment set in the duty cycle offset map, as well as independent components extracted from the mixed segments. Each segment in the decoupled segment set is labeled with its source type to distinguish it from the original valid segments and the decomposed segments of the mixed segments. The latter participates in the intensity classification calculation with a slightly lower weight in the offset accumulation position calibration. When the temporal coverage of the decoupled segment set is less than 50%, a low coverage warning mark is added. The low coverage warning mark of the decoupled segment set is passed as a low confidence mark during the offset accumulation mark generation stage, indicating that the confidence of the drift feature extraction in the corresponding time period is insufficient.

[0034] For example, the step of calibrating the offset accumulation position of the decoupled segment set to generate offset accumulation markers includes: extracting the offset direction of the same band within each period according to the illumination period for the decoupled segment set to generate a periodic offset direction set; performing multi-period consistency verification on the periodic offset direction set to identify stable offset direction segments to generate a steady-state offset segment set; using the steady-state offset segment set to perform offset intensity classification to generate an offset intensity classification table; and calibrating the offset accumulation position according to the offset intensity classification table to generate offset accumulation markers.

[0035] For the decoupled segment set, the offset direction of the same band within each cycle is extracted and grouped according to the illumination cycle to generate a periodic offset direction set. After determining the start and end times of each illumination cycle according to the illumination control log, the mode of the difference sign of the drift amount sequence within each cycle segment of the decoupled segment set is used as the offset direction label for that cycle: a positive value of more than 60% of the difference sign is considered positive, a negative value of more than 60% is considered negative, otherwise it is considered that there is no dominant direction. The breeding house implements a strict 24-hour illumination system. During the illumination period, the drive circuit continuously heats up, causing the drift amount of each band to accumulate unidirectionally. After the illumination is turned off, the circuit cools down naturally, causing the drift amount to partially recover. This periodic heating-cooling process makes the unidirectional accumulation characteristic within each cycle stably reflected in the mode of the difference sign. The periodic offset direction set establishes a dual index based on the center wavelength of the band and the illumination cycle number. The offset direction label and the percentage value of the difference sign are recorded at each position of the periodic offset direction set. The percentage value is used as a reference weight for flip correction in subsequent multi-cycle consistency verification. A low integrity marker is added to the location where the effective sampled frames in the decoupled segment are less than 50% of the total number of frames in the cycle. The direction label of the marker location is included with low weight in the multi-cycle consistency verification.

[0036] Multi-period consistency verification is performed on the periodic offset direction set to identify stable segments and generate a steady-state offset segment set. The multi-period consistency verification scans the direction label sequence of each band in the periodic offset direction set using a sliding window of three light cycles. If the three direction labels within the window are completely consistent, the direction is considered stable. If one flip occurs, the flip position is corrected using the multiple values ​​of the other two direction labels, and the stability determination is still maintained. If two or more inconsistencies occur, the direction is considered unstable. During the brooding period, frequent temperature control in the rearing house can cause occasional local heat dissipation disturbances, leading to isolated direction flips in the periodic offset direction set. These flips are presented as single inconsistencies in the sliding window and are not included in the unstable segment after correction. In contrast, the drift characteristics accumulated over multiple consecutive cycles in the same direction during the later rearing stage form a stable determination during verification. The steady-state drift characteristics of these two growth stages are clearly distinguished in the steady-state offset segment set. When the number of consecutive illumination cycles for directional stability determination exceeds three, they are merged into a single steady-state offset segment. The steady-state offset segment set is grouped by wavelength and indexed primarily by segment number. Each segment in the steady-state offset segment set records the center wavelength of the band, the start and end cycle numbers, and the stable direction label. If a low-integrity marker position in the periodic offset direction set is included in a steady-state offset segment, the corresponding segment is marked with a low-integrity periodic marker.

[0037] A migration intensity classification table is generated using a set of steady-state migration segments. For each segment in the steady-state migration segment set, the absolute value of the difference between the drift at the end of the segment and the drift at the beginning of the segment is taken as the monotonic cumulative amplitude of that segment. The cumulative amplitude of all segments and all bands in the steady-state migration segment set is classified according to quantiles: below the 25th quantile is considered a weak migration, between the 25th and 75th quantiles is considered a medium migration, and above the 75th quantile is considered a strong migration. In the later stages of growth, the cumulative amplitudes of red and green light channels in the steady-state migration segment set usually fall into the strong migration category, while the cumulative amplitudes of the same bands in the brooding period are mostly in the weak migration category. The quantile classification solidifies the difference in illuminance compensation priority between the two growth stages in the form of a grade. The offset intensity classification table establishes a dual index based on the center wavelength of the band and the steady-state offset segment number. Each position in the offset intensity classification table records the offset intensity level and the corresponding cumulative amplitude value. For segments with low integrity periodic markers, a low reliability marker is added to the corresponding position in the offset intensity classification table. The offset intensity level at the marked position is downgraded by one level to compensate for the amplitude underestimation that may be caused by insufficient sampling integrity.

[0038] The offset accumulation positions are calibrated according to the offset intensity grading table to generate offset accumulation markers. The starting time of strong offset level segments in the offset intensity grading table is independently calibrated as offset accumulation position candidates for each band. When the cumulative amplitude of two consecutive offset level segments in the offset intensity grading table exceeds the strong offset threshold, they are merged and upgraded to offset accumulation position candidates. Weak offset level segments do not generate offset accumulation position candidates. The resulting distribution of offset accumulation positions in the time domain exhibits a characteristic that naturally aligns with the sensitivity of illumination requirements at each growth stage: strong offset positions are concentrated in the red and green light channels in the later stages of growth, precisely corresponding to the interval with the highest illumination compensation accuracy requirements during the muscle gain stage. The offset accumulation markers are organized using a dual index of the band center wavelength and the offset accumulation position time. Each offset accumulation marker records the offset direction, offset intensity level, and cumulative amplitude value. The offset direction inherits the stable direction label of the corresponding segment in the steady-state offset segment set. Offset accumulation markers corresponding to low reliability marker positions in the offset intensity grading table are appended with low-reliability markers. The offset accumulation markers within each band are arranged in descending order of offset intensity level to support the priority processing of segment nodes during subsequent segmented fitting.

[0039] The duty cycle shift map was segmented using offset cumulative markers to generate band correction coefficients. The drift sequence of each band in the duty cycle shift map was divided into segments using the time index of the offset cumulative markers as the dividing node. The slope of the drift within each segment was calculated using least-squares linear fitting, and the slope characterizes the drift rate of that band within that segment. In the later stages of growth, the red and green light channels had higher distribution density of offset cumulative markers, corresponding to shorter segments with larger fitting slopes; in the brooding stage, the corresponding segments were longer with smaller fitting slopes. This segmented structure allowed for independent expression of the drift rate differences between the two growth stages. A segment was considered to have insufficient fitting quality if the root mean square error of the fitting residual exceeded 1.5 times the standard deviation of the drift in that segment. Segments with insufficient fitting quality were replaced by the weighted average of the slopes of two adjacent segments, with weights inversely proportional to the ratio of the fitting residuals. Missing marker units in the duty cycle shift map were not included in the least-squares fitting of that segment. The band correction coefficients are organized by the band center wavelength index. Each band's correction coefficient records the drift rate value, start and end time sequence number, and fitting quality marker for each segment. The theoretical illuminance correction at any given time is calculated by multiplying the drift rate of the segment to which it belongs by the cumulative duration within that segment. The segments corresponding to the low-confidence marker positions in the offset accumulation markers have a low-confidence marker added to the band correction coefficients. This marker guides the corresponding conservative processing in the subsequent determination of the effective spectral boundary.

[0040] In some embodiments, determining the effective spectral boundary based on the discreteness check of the duty cycle offset spectrum and the band correction coefficient includes: performing power coupling interference detection on segments in the duty cycle offset spectrum where adjacent bands are simultaneously driven at high power to generate a set of coupling interference intensity; identifying band combinations whose coupling interference exceeds a preset limit for the set of coupling interference intensity to generate a set of coupling limit exceedance markers; using the set of coupling limit exceedance markers to perform discreteness elimination on the corresponding band configurations in the duty cycle offset spectrum to generate a set of effective configuration intervals; and determining the effective spectral boundary based on the set of effective configuration intervals and the band correction coefficient.

[0041] Power coupling interference detection is performed on segments in the duty cycle offset spectrum where adjacent bands are simultaneously driven at high power to generate a set of coupling interference intensity. In each column of the duty cycle offset spectrum, band pairs with a center wavelength spacing of no more than 50 nm between adjacent bands are defined as adjacent band pairs. When the duty cycle settings of both bands at the same time exceed 75% of the rated duty cycle, it is determined to be a state of simultaneous high power drive. The drift residual of the interfered band is characterized by the difference after removing the piecewise fitted prediction value of the band correction coefficient. Since the piecewise fitting of the band correction coefficient uses the offset accumulation mark as the dividing node and the least squares method is used to approximate the trend of the whole segment within the fitting segment, the local additional deviation caused by coupling interference is still retained in the residual within the segment as an over-limit amplitude and is not absorbed by the trend fitting. When the absolute value of the residual exceeds 1.5 times the standard deviation of the full-time domain drift of the band, it is determined that coupling interference exists. During the later stages of growth, when high demand dominates, multiple band channels need to be driven simultaneously with high duty cycles. The PCB traces and heat dissipation structures of adjacent channels are spatially close, and power coupling caused by electromagnetic induction and heat conduction adds an additional deviation to the drift residual. This deviation manifests as amplitude exceeding limits in the residual distribution during high-power joint drive periods. The coupling interference intensity set is organized using a dual-index system of adjacent band pair numbers and sampling time sequence numbers. Interference threshold judgment markers are recorded at each position in the coupling interference intensity set. When generating subsequent coupling limit exceeding marker sets, the judgment markers of each unit in the coupling interference intensity set are used as the basic elements for window statistics. Coupling interference detection at times with missing marker units is skipped, and the corresponding positions are filled with empty markers.

[0042] A coupling interference exceedance marker set is generated by identifying band combinations whose coupling interference exceeds a preset limit based on the coupling interference intensity set. Adjacent band pairs within the coupling interference intensity set are statistically analyzed using 10 sampling frames as a decision window. When the proportion of interference exceeding the limit in the coupling interference intensity set window exceeds 40%, it is determined to be a continuous coupling interference band combination. The start and end times are marked at the beginning and end frames of the window where the proportion of exceeding the limit first exceeds 40%, respectively. The reason for using a window-based statistical mechanism instead of frame-by-frame judgment is that occasional ground vibrations caused by chickens running around in the poultry house occasionally trigger abnormal sensor readings. Such isolated exceedances usually account for less than 40% within the 10-frame window and do not trigger continuous coupling interference judgment. Window statistics effectively distinguish between occasional disturbances and continuous coupling interference caused by multi-band combined high-power drive in the later stages of rearing. The coupling limit violation marker set is organized by adjacent band pairs. Each band pair of the coupling limit violation marker set records the start and end time sequence of the continuous coupling interference period, the number of the main interference band, and the peak value of the excess ratio. For coupling limit violation marker sets with missing marker time periods, incomplete data markers are added to the corresponding positions. The coupling limit violation determination of the marked time period is extrapolated by the excess ratio of the adjacent valid time periods.

[0043] The effective configuration interval set is generated by using a coupling over-limit marker set to perform discrete elimination on the corresponding band configurations in the duty cycle offset map. The drift sequence of each band in the duty cycle offset map is eliminated using the start and end times of the coupling over-limit marker set as the elimination boundary. The cell markers corresponding to the time periods within the elimination boundary are considered coupled and are not included in the discrete calculation for that band. The discreteness of the drift sequence of each band within the non-coupled elimination time periods of the duty cycle offset map is calculated using a sliding window of 20 sampling frames. Time periods with a drift standard deviation exceeding twice the full-time standard deviation within the window are identified as high-discretion segments and excluded. In the later stages of development, the apparent discreteness of the drift during the multi-band high-power joint drive period is relatively large without coupling interference elimination. After elimination using the coupling over-limit marker set, the discreteness of this period returns to a normal level, ensuring that the effective configuration interval set retains sufficient usable time period width during the high-demand stage of the later development phase. The effective configuration interval set is organized by the center wavelength index of the band. Each band in the effective configuration interval set records the start and end time sequence number of the continuous available time period, the standard deviation of the drift amount within the corresponding time period, and the effective coverage rate. The effective configuration interval set unit corresponding to the incomplete data period in the coupled over-limit mark set is marked with a low integrity mark.

[0044] The effective spectral boundary is determined based on the effective configuration interval set and band correction coefficients. Within the available time period of each band in the effective configuration interval set, the corrected residual is calculated using the corresponding piecewise correction function of the band correction coefficient. A moment when the absolute value of the corrected residual exceeds 8% of the rated illuminance is considered insufficiently corrected. When the low-confidence marker segment of the band correction coefficient overlaps with the available time period of the effective configuration interval set, the insufficient correction threshold for this overlapping portion of the effective configuration interval set is tightened to 5% of the rated illuminance to compensate for the decrease in correction accuracy. Moments where the corrected residual exceeds the limit correspond to intervals where the actual illuminance will deviate from the rated range after the issuance of spectral configuration commands for each growth stage; these must be excluded from the effective spectral boundary. The effective spectral boundary is based on the upper and lower bounds of the available duty cycle for each band. The upper bound is the highest duty cycle setting value in the effective configuration interval set where the corrected residual does not exceed the limit, and the lower bound is the lowest duty cycle setting value where the corrected residual does not exceed the limit. Both define the range of duty cycles that can be safely adjusted for each band under the current aging state and drift characteristics. The effective spectral boundaries are organized by the center wavelength index of the bands. Each band of the effective spectral boundary records the upper and lower limits of the duty cycle and the corresponding effective coverage. For bands with low integrity markers added to the effective configuration interval set, the effective spectral boundaries are executed with conservative boundaries when the switching control command is generated, i.e., the upper limit is lowered by 5% and the lower limit is raised by 5%, to leave a safety margin to resist the evaluation deviation caused by incomplete sampling. When the same band has both a low confidence marker for the band correction coefficient and a low integrity marker for the effective configuration interval set, the two conservative treatments are executed in combination. That is, the insufficient correction judgment threshold is tightened to 5% of the rated illuminance, while the upper limit of the effective spectral boundary is lowered by 5% and the lower limit is raised by 5%, to simultaneously compensate for the dual deficiencies in correction accuracy and sampling integrity.

[0045] Step S140: Extract the energy consumption offset rate based on the band decay demand level, generate the optimal band configuration based on the energy consumption offset rate and the band correction coefficient, extract the non-critical band reduction margin from the optimal band configuration to generate the energy-saving buffer space, and formulate the control priority sequence by combining the energy-saving buffer space and the stepped spectral response region.

[0046] Specifically, energy consumption offset rate is extracted based on the band decay demand level. The energy consumption offset rate is calculated by dividing the difference between the actual driving power of each band channel at the current growth stage and the rated driving power at the aging level corresponding to the band decay demand level by the rated power. A positive value indicates that the actual energy consumption is higher than the rated level, and a negative value indicates that it is lower than the rated level. The continuous accumulation of junction temperature of LED light chips in the breeding house accelerates the aging of phosphors. In the later stage of rearing, the red and green light channels have a higher band decay demand level due to the long-term high duty cycle drive. In order to maintain the target illuminance, the duty cycle of these channels must be continuously increased. Their energy consumption offset rate is often significantly higher than that of low-intensity operating channels such as blue light during the brooding period. The energy consumption offset rate shows a gradient structure in the two-dimensional space formed by the band and the growth stage, which is highly corresponding to the distribution of the band decay demand level. For channels with a band decay demand level of zero, the rated drive power is used directly from the factory calibration value. For channels of levels one to three, the factory calibration value is multiplied by the corresponding power compensation factor. The compensation factor is obtained from the preset decay level-power compensation mapping table according to the band center wavelength and the band decay demand level. The compensation factor at each position in the mapping table is calibrated based on historical measured data. The energy consumption offset rate is organized by dual indexing of the band center wavelength and the growth stage number. For stages in the band decay demand level with an uncorrected mark, the energy consumption offset rate is estimated by the average offset rate of adjacent stages and an estimation mark is added. The position of the estimation mark participates in the benefit ratio calculation with a low weight when generating the optimal band configuration in the future, so as to avoid the estimation value of the incomplete data stage from having too much impact on the configuration optimization result.

[0047] In some embodiments, generating the optimal band configuration based on the energy consumption offset rate and the band correction coefficient includes: performing gradient tracking analysis on the energy consumption offset rate to locate configurable regions and generate candidate band groups; calculating the benefit ratio according to the physiological response intensity per unit power of each band combination in the candidate band group to generate a benefit ratio sequence; identifying band combinations with benefit ratios lower than a preset threshold in the benefit ratio sequence to generate an inefficient band marker set; and using the inefficient band marker set in combination with the band correction coefficient to perform comprehensive sorting to generate the optimal band configuration.

[0048] Gradient tracking analysis of energy consumption offset rate is used to locate configurable regions and generate candidate band groups. The first-order difference of energy consumption offset rate in the growth stage dimension is calculated independently for each band. Continuous growth stage segments with an absolute difference value lower than 0.5 times the full-time standard deviation of energy consumption offset rate are identified as low-gradient segments. Within low-gradient segments, the duty cycle configuration of each band has a stable adjustment margin. In the late growth stage, the energy consumption offset rate of red and green light channels tends to stabilize during the high duty cycle continuous driving stage, and the absolute value of the difference usually meets the low-gradient judgment condition and is identified as a configurable segment. In the brooding stage, the driving intensity of blue light channel is relatively low, and the energy consumption offset rate fluctuates less with age, also forming a low-gradient segment, but the direction of the corresponding duty cycle adjustment margin is opposite to that of the late growth stage channel. The two types of segments must be distinguished based on the physiological response intensity during the benefit ratio calculation stage. Low-gradient segments are grouped into candidate band groups based on the merging condition that the center wavelength spacing between adjacent bands does not exceed 50 nm. When the same low-gradient segment spans multiple consecutive growth stages, the entire segment is included in the candidate band group based on the continuity of the stage sequence number. Adjacent low-gradient segments that do not meet the continuity requirement form their own candidate band group entries independently. Candidate band groups are indexed by group number. Each candidate band group records the list of center wavelengths of the bands, the start and end numbers of the configurable growth stages, and the average energy consumption offset rate within the group. Configurable units at the energy consumption offset rate estimation marker positions participate in the calculation of the group's average with a low confidence weight.

[0049] The benefit ratio sequence is generated by calculating the physiological response intensity per unit power for each band combination in the candidate band group. The benefit ratio measures the average physiological response intensity per unit driving power for each group within the configurable growth phase, and its calculation formula is: Where Qg is the benefit ratio of the g-th group (unit: μmol·m -2 ·s -1 ·W -1 ), where g is the candidate band group number, Tg is the set of times corresponding to the configurable growth stage of the g-th group, |Tg| is the total number of times in the set, i∈g represents the band channel belonging to the g-th group, and Ei,t is the measured illuminance of the i-th channel at time t (unit: μmol·m). -2 ·s -1Ri,t is the normalized value of the rhythm response coefficient at time t corresponding to the diurnal rhythm period, and Pg,t is the sum of the actual driving power (in W) of all channels in group g at time t. The combined excitation intensity of the candidate bands covering green and red light in the late rearing stage of chicks is higher than that of combinations covering only a single band at the same driving power, and its Qg value is ranked higher in the benefit ratio sequence. The combination of blue and ultraviolet bands in the brooding stage also has a higher physiological response intensity per unit power than the combination of non-sensitive bands during the visual development stage of chicks. The benefit ratio distribution of different growth stages is therefore highly correlated with the demand of diurnal rhythm periods. The benefit ratio sequence is arranged in descending order of Qg value. Ei,t corresponding to the energy consumption offset rate estimation mark period is included in the Qg mean calculation with a weight of 0.6. The descending order of the benefit ratio sequence makes the efficient combination covering the key bands in the late rearing stage ranked first and the single band low response combination ranked last. The benefit ratio sequence provides a clear priority basis for the subsequent identification and elimination of inefficient bands.

[0050] For example, the step of identifying band combinations with benefit ratios lower than a preset threshold and generating an inefficient band label set for the benefit ratio sequence includes: extracting differential features from the benefit ratio sequence to obtain a differential label sequence; using the differential label sequence to adaptively segment the benefit ratio sequence to generate candidate boundary positions; performing boundary stability analysis on the candidate boundary positions to screen effective boundaries; and extracting inefficient segments from the benefit ratio sequence based on the effective boundaries to generate an inefficient band label set.

[0051] A grade difference feature extraction process is performed on the benefit ratio sequence to obtain a grade difference marker sequence. After the benefit ratio sequence is arranged in descending order, the difference in benefit ratio between adjacent groups is calculated as an absolute value to form a grade difference sequence. Positions with a grade difference value exceeding 1.5 times the overall mean are marked with a higher grade difference marker. When multiple bands are used in conjunction with the aquaculture shed, there is a significant discontinuity in the physiological excitation efficiency per unit power between the high-response band combination covering red and green light in the later stages of rearing and the low-response combination covering only secondary bands. This discontinuity causes a sharp jump in the benefit ratio of adjacent groups at the corresponding positions in the benefit ratio sequence, with the grade difference value far exceeding the mean, thus forming a higher grade difference marker. At other positions, the grade difference values ​​in the gradually decreasing range of benefit ratio are generally lower than the mean, and the numerical distribution patterns of the two types of positions are clearly distinguishable in the grade difference marker sequence. The grade difference marker sequence is arranged in descending order of group number, and the number of higher grade difference markers in the grade difference marker sequence is usually proportional to the number of effective band groups covered by the aquaculture shed lights. If the group number of the higher grade difference marker position shifts by more than two positions between aquaculture batches, the integrity of the measured illuminance data of each channel must be re-verified in the benefit ratio sequence verification step. The high-level differential marker positions in the differential marker sequence serve as the main basis for determining subsequent candidate boundary positions, while the non-high-level differential marker positions do not participate in the generation of candidate boundary positions because the difference in the combined benefits of adjacent bands is relatively small.

[0052] Candidate boundary positions are generated by adaptively segmenting the benefit ratio sequence using a grade difference marker sequence. Each high grade difference marker position in the grade difference marker sequence is used as the midpoint index of the corresponding two adjacent group numbers in the benefit ratio sequence as the candidate boundary position. Each candidate boundary position is labeled with the corresponding grade difference value, the mean benefit ratio on the left, and the mean benefit ratio on the right. The absolute value of the difference between the means on the left and right sides is used as the boundary strength index for that candidate boundary position. Generating candidate boundary positions using the high grade difference marker positions in the grade difference marker sequence can accurately divide the benefit ratio sequence into high-efficiency and low-efficiency segments along the benefit fault between the key and secondary bands in the late growth stage. If a uniform fixed-interval segmentation method is used, the non-uniform distribution of the actual benefit differences between band combinations at each growth stage will be ignored, resulting in miscutting at the boundary between high-response and low-response bands in the late growth stage, causing some red light channel combinations to be mistakenly classified into the low-efficiency segment. If the difference between the mean of the left and right benefit ratios of a candidate boundary position is less than 20% of the mean of the entire benefit ratio sequence, it is judged as a weak boundary candidate position and a weak boundary mark is added. The position of the weak boundary mark participates in the verification with low weight in the boundary stability analysis. The higher the boundary intensity index of the candidate boundary position, the more obvious the benefit fault of the band combination on both sides of the candidate position.

[0053] A stability analysis was conducted on candidate boundary points to screen for effective boundaries. Boundary stability was measured by the ratio of the boundary intensity index of each candidate boundary point to the boundary intensity index of its adjacent candidate boundary points. A ratio exceeding 1.5 was considered a high-stability boundary, while a ratio below 1.0 was considered a low-stability boundary. The 1.5 threshold was taken from the trough of the bimodal distribution of the boundary intensity ratio of the two types of candidate boundary points in multiple batches of aquaculture data. When there was only one candidate boundary point and no adjacent candidate boundary points, the ratio of the boundary intensity index of that candidate boundary point to the mean of the difference between all levels of the benefit ratio sequence was used instead of the stability judgment. A ratio exceeding 1.5 was also considered a high-stability boundary. High-stability boundaries are concentrated at the fault locations where the difference in benefits between key and secondary band combinations is most significant in the later stages of rearing. These locations appear stably in the benefit ratio sequences of different breeding batches. Low-stability boundaries, on the other hand, are concentrated at secondary disturbance locations introduced by density deviations in individual batches or fluctuations in sensor readings. These latter locations drift with batch data differences and lack reproducibility stability in multi-batch comparisons. Locations with added weak boundary markers among candidate boundaries are not included in the high-stability boundary candidate list, regardless of whether the boundary intensity ratio reaches 1.5 times the threshold, to prevent weak benefit fault locations from being misjudged as valid boundaries due to noise-induced ratio increases. Valid boundaries are selected from candidate boundaries with the highest boundary intensity and high stability. The number of valid boundaries is set to 1 to 2. When multiple high-stability candidate boundaries are parallel, the one with the largest difference in the mean benefit ratios on both sides is selected first. The final valid boundary location is recorded by its index number in the benefit ratio sequence.

[0054] An inefficient band marker set is generated by extracting inefficient segments from the benefit ratio sequence based on the effective boundary. The effective boundary index number is used to locate the segmentation node in the benefit ratio sequence. All candidate band groups to the right of the segmentation node constitute entries in the inefficient band marker set. The inefficient segments to the right of the benefit ratio sequence concentrate secondary band combinations that contribute limited to the physiological responses of chicks at various growth stages. The efficient segments to the left retain key band combinations covered by combined red and green light in the later stages of rearing. The average benefit ratio of the efficient segments to the left of the effective boundary serves as the benchmark for calculating the difference in the magnitude of each entry in the inefficient segment. Entries with a difference exceeding 30% of this average must be assigned the highest elimination priority in the comprehensive ranking. Each entry in the inefficient band marker set records its benefit ratio, the magnitude of the difference below the average benefit ratio to the left of the effective boundary, its original benefit ratio sequence position number, and the configurable growth stage interval of its candidate band group. The magnitude of the difference serves as the weighting basis for elimination priority in the comprehensive ranking, and the configurable growth stage interval is used to limit the time domain range of the replacement effect of each inefficient entry. When the number of inefficient band marker entries exceeds 50% of the total number of candidate band groups, an early warning marker indicating that the proportion of inefficiency is too high is added, indicating that the overall band configuration benefit ratio structure of the current aquaculture batch needs to be re-examined.

[0055] The optimal band configuration is generated by comprehensively ranking inefficient band marker sets combined with band correction coefficients. The comprehensive ranking uses the normalized difference magnitude of each entry in the inefficient band marker set as the base weight, plus a penalty term for the correction uncertainty of bands within the group corresponding to low-confidence markers of the band correction coefficient. The comprehensive elimination score is calculated using the following formula: Where Sg is the overall elimination score for the g-th item. The normalized value of the difference magnitude of the g-th item is (normalized to the interval [0,1] using the maximum difference magnitude of all items in the inefficient band marker set as the denominator). ρg is the ratio of the number of low-confidence marker bands in the band correction coefficient group to the total number of bands in the group, with a value range of [0,1]. The penalty coefficient of 0.3 is calibrated based on the regression analysis of the impact of the proportion of low-confidence markers on the accuracy of illuminance compensation in historical batch data. Band combinations with large Δ'g and high ρg in the inefficient band marker set mean that the combination not only has an efficiency ratio lower than the effective boundary but also insufficient correction confidence. The combination of these two factors causes Sg to be significantly high, resulting in it being processed earlier in the configuration replacement sequence. The optimal band configuration is constructed with the candidate band group combination with the lowest Sg as the core. The baseline duty cycle setting value of each channel in the optimal band configuration is based on the configuration of the band adaptation baseline for the corresponding time period, and the achievable illuminance range constraint of each channel after illuminance correction is adjusted in conjunction with the band correction coefficient. The optimal band configuration is organized by a dual index of the band center wavelength and the growth stage number. The recommended duty cycle setting and the corresponding expected illuminance value are recorded at each position of the optimal band configuration.

[0056] Energy-saving buffer space is generated by extracting reduction margins for non-critical bands from the optimal band configuration. Non-critical bands are determined based on the benefit ratio ranking of each band combination in the benefit ratio sequence; channels within the bottom 30% of band combinations are marked as non-critical band channels. The reduction margin for non-critical band channels in the optimal band configuration is defined as the difference between the recommended duty cycle setting and the minimum duty cycle required to meet the target illuminance lower limit for the current growth stage. The minimum duty cycle is calculated by dividing the target illuminance lower limit by the band correction factor to obtain the unit duty cycle illuminance. Channels with recommended duty cycle settings close to the minimum duty cycle in the optimal band configuration have near-zero reduction margins and contribute little to the energy-saving buffer space. The recommended duty cycles for the brooding stage UV channels and the early growth stage blue light channels in the optimal band configuration are typically 10% to 20% higher than the minimum duty cycle corresponding to the target illuminance lower limit; their reduction margins constitute the main source of the energy-saving buffer space. The energy-saving buffer space is constituted by the weighted sum of the reduction margins of all non-critical band channels. The weights are allocated according to the proportion of each channel's rated power in the total rated power of the non-critical band channels. The reduction margin of channels with higher rated power contributes more to the energy-saving buffer space. The energy-saving buffer space is organized according to the growth stage sequence, and the total energy-saving buffer amount is recorded for each stage. The reduction margin derived from the estimated marker is included in the total energy-saving buffer amount calculation with a weight reduction of 0.7 times when summarizing.

[0057] A control priority sequence was established by combining energy-saving buffer space and tiered spectral response zone. The control priority sequence uses the tiered spectral response zone partition number as the primary index. The control priority score for each partition within the tiered spectral response zone is calculated by weighting four indicators: the highest decay level within the partition, the high-priority marker, the total amount of energy-saving buffer space at the corresponding growth stage, and the partition confidence level. The highest decay level, after normalization, has a weight of 0.4; the high-priority marker has a binary value (0.2 if marked, 0.0 if unmarked); the total amount of energy-saving buffer space, after normalization, has a weight of 0.3; and the partition confidence level, after inverse mapping, has a weight of 0.1. The sum of these four factors constitutes the priority score, with a maximum score of 1.0. In the later stages of growth, when the decay levels of red and green light channels are high and the confidence level of the tiered spectral response zone is sufficient, the power margin released by the energy-saving buffer space is concentrated to compensate for the illuminance loss of these channels. The four indicators for the corresponding tiered spectral response zone are all at a high level, and the priority score naturally ranks before the secondary band partitions during the brooding period in the control priority sequence. In the priority sequence of regulation, partitions with the same priority score are ranked secondary based on the strength of demand as the primary criterion. Partitions in the growth period dominated by strong demand are prioritized over those in the growth period dominated by weak demand. The buffer amount corresponding to the estimated marker position in the energy-saving buffer space is included with a 0.7-fold weight reduction when calculating the score, ensuring that partitions with insufficient data reliability do not receive excessively high regulation priority due to estimation errors in buffer amount. The priority sequence of regulation is indexed by tiered spectral response and organized in descending order of priority score. Each partition records the regulation priority score, the secondary ranking marker, and the values ​​of the four indicators used in the priority score calculation.

[0058] Step S150: The bands are arranged in a hierarchical manner using a priority sequence to form a drive scheduling arrangement. The switching triggering conditions are determined based on the drive scheduling arrangement and the effective spectral boundary. The redundant bands of the drive scheduling arrangement are eliminated step by step according to the switching triggering conditions, and the LED spectral switching control command is output.

[0059] Specifically, a priority sequence for regulation is used to create a hierarchical arrangement of bands to drive the scheduling. The priority sequence is indexed by spectral response levels and organized in descending order of priority score. The priority score recorded in each zone directly serves as the weighting for the driving scheduling arrangement. The zone with the highest priority score corresponds to an area where the red and green light channels have a high decay rate in the later stages of development and ample energy-saving buffer space. Within this zone, each band channel is prioritized in the driving scheduling arrangement, receiving priority power allocation and illuminance compensation resources. In the priority sequence, the secondary sorting markers recorded in each zone determine the arrangement order within the same score range based on the rule that strong demand takes precedence over weak demand. The four indicators used to calculate the priority score of the priority sequence are traceable during the driving scheduling generation stage to verify the integrity of the arrangement basis for each zone. The drive scheduling is indexed by channel number. Each channel records its corresponding spectral response zone number, control priority score, arrangement sequence number, growth stage number, and baseline duty cycle setting. The baseline duty cycle setting is written with the recommended duty cycle for the channel corresponding to the optimal band configuration. The arrangement sequence number is determined by the channel's position in the global sequence after the control priority sequence is expanded in descending order. Channels with adjacent arrangement sequences are triggered sequentially with the minimum time interval when the switching control command is executed. Channels corresponding to low-confidence zones in the control priority sequence are marked with a low-confidence flag in the drive scheduling and are arranged after high-confidence channels with the same priority score. The total number of channels in the drive scheduling is capped at the number of valid channels in the current batch; invalid channels are not included in the arrangement.

[0060] In some embodiments, determining the switching triggering conditions based on the drive scheduling arrangement and the effective spectral boundary includes: extracting the duty cycle adjustment response delay time for each band in the drive scheduling arrangement to generate a delay time distribution table; using the delay time distribution table to estimate the lead time of the target power arrival time for each band to generate a lead triggering compensation amount; performing a margin mapping between the lead triggering compensation amount and the effective spectral boundary to generate a candidate triggering condition set; and determining the switching triggering conditions based on the urgency ranking of the candidate triggering condition set.

[0061] A delay time distribution table is generated by extracting the duty cycle adjustment response delay time for each band in the drive scheduling arrangement. The arrangement sequence number of each channel in the drive scheduling arrangement and the growth stage sequence number jointly determine the scope of the delay time extraction. The delay time of each channel in the drive scheduling arrangement is extracted independently according to the channel number and growth stage sequence number. The thermal state differences of the drive circuit in different growth stages must be calibrated segment by segment. The duty cycle adjustment response delay time refers to the time elapsed from the change of the duty cycle set value of each channel to the actual illuminance stabilizing within the range of ±3% of the target value. Due to the long-term high temperature operation of the drive circuit in the red light channel in the later stage of growth, its delay time is usually 15% to 25% higher than that of the blue light channel in the brooding period. Therefore, each channel in the drive scheduling arrangement is calibrated independently instead of being uniformly substituted with the global mean. The delay time distribution table is organized by dual indexing of channel number and growth stage sequence number. The delay time mean and standard deviation are recorded at each position in the delay time distribution table. For channels with fewer than 5 historical samples, the mean of adjacent channels in the same band is used for estimation and a fill mark is added. The standard deviation of the fill mark channel is conservatively estimated at 1.3 times the mean of the same band. The delay time distribution table is updated on a rolling basis after each breeding batch ends. The update weight is allocated according to the most recent batch 0.6, the second most recent batch 0.3, and the earliest batch 0.1. The continuous update of the delay time distribution table enables dynamic tracking of the aging changes of the response characteristics of each channel.

[0062] The advance trigger compensation amount is generated by estimating the arrival time of the target power for each band using a delay time distribution table. The formula for calculating the advance trigger compensation amount is Ci = μi + 1.5 × σi, where Ci is the advance trigger compensation amount for the i-th channel, μi is the mean delay time (in ms) of the corresponding position in the delay time distribution table for that channel, and σi is the standard deviation of the delay time (in ms) of the corresponding position in the delay time distribution table for that channel. The delay time is recorded with millisecond precision by the built-in timing module of the drive controller. The safety margin coefficient of 1.5 is calibrated based on the frequency of events in historical batches where insufficient advance amount leads to a delay in the arrival of the target illuminance value. Under this coefficient, the occurrence rate of delay events is controlled within 5%. The standard deviation of the delay time of the red light channel in the later stage of growth is usually larger, and its Ci value is significantly higher than that of the blue light channel in the brooding period due to the amplification effect of the safety margin term. The corresponding switching command must be triggered earlier to ensure the timely arrival of the target illuminance in this growth stage. For channels with added fill marks in the delay time distribution table, the advance trigger compensation amount is calculated using a conservative standard deviation of 1.3. The calculation results are then marked with a low-confidence flag. The advance trigger compensation amount for channels with fill marks in the delay time distribution table is conservatively set to allow for additional timing margins to compensate for insufficient correction accuracy. The advance trigger compensation amount is organized using a dual index of channel number and growth stage number. Each position of the advance trigger compensation amount records the Ci value and the source flag of the corresponding delay time mean and standard deviation. The source flag of the advance trigger compensation amount is used in the margin mapping stage to identify low-confidence channels and trigger joint low-confidence flags.

[0063] Candidate triggering condition sets are generated by mapping the advance trigger compensation amount to the effective spectral boundary. The upper and lower bounds of the duty cycle for each band within the effective spectral boundary define the duty cycle adjustment range within which each channel can safely perform switching during the current growth stage. The duty cycle adjustment margin is calculated by dividing the difference between the upper bound of the effective spectral boundary and the current duty cycle setting by the duty cycle adjustment rate per unit time. The duty cycle adjustment rate per unit time is provided by the rated adjustment rate parameter of each channel's drive controller (in % / ms) and read from the corresponding channel entry in the band configuration data. The calculated time margin is in ms and is consistent with the unit of the advance trigger compensation amount Ci. The advance trigger compensation amount is compared with this time margin to determine the margin of each channel: channels with a time margin greater than the advance trigger compensation amount Ci are considered to have sufficient margin, while channels with a time margin less than Ci are considered to have insufficient margin and are marked as having insufficient margin. In the later stages of growth, the red channel's effective spectral boundary has been lowered by 5% due to the influence of the low-confidence marker of the band correction coefficient. This results in a relatively narrower margin for duty cycle adjustment. Combined with the channel's relatively high advance trigger compensation, the triggering frequency of the insufficient margin marker is significantly higher in the later stages of growth than in other stages, reflecting the urgency of the switching control at this stage. The candidate triggering condition set is organized using a dual index of channel number and growth stage number. Each position in the candidate triggering condition set records the sufficient or insufficient margin judgment marker and the corresponding upper and lower bounds of the effective spectral boundary. Channels with low-confidence advance trigger compensation are additionally marked with a joint low-confidence marker in the candidate triggering condition set. Candidate entries at the joint low-confidence marker positions in the candidate triggering condition set are sorted with a conservative priority during urgency ranking to prevent channels with insufficient correction accuracy from receiving excessive triggering priority.

[0064] The switching trigger conditions are determined by ranking the urgency of the candidate trigger condition set. The urgency of each channel in the candidate trigger condition set is calculated by weighting three indicators: insufficient margin flag weight, control priority sequence priority score, and advance trigger compensation amount Ci. The urgency score is calculated as F = 0.5 × N + 0.3 × P' + 0.2 × C', where N is the insufficient margin flag value (1.0 when the margin is insufficient, and 0.0 when the margin is sufficient), P' is the control priority sequence priority score normalized to [0,1] with the maximum priority score of all channels in the candidate trigger condition set as the denominator, and C' is the advance trigger compensation amount Ci normalized to [0,1] with the maximum Ci of all channels in the candidate trigger condition set as the denominator. After normalization, the three items have the same dimensions, and the sum of the weight coefficients is 1.0. The channel with a higher urgency score indicates that its duty cycle adjustment margin is more strained and its control priority is higher in the current growth stage, and it should be executed first in the switching trigger conditions. In the later stages of candidate triggering conditions, insufficient red and green light channels resulted in high triggering frequencies and a high priority score in the regulation priority sequence. These two factors combined led to the high urgency score among all channels. Therefore, the triggering sequence of the red and green channels was prioritized for switching triggering conditions at this stage to ensure timely illumination compensation during the critical muscle growth phase. Switching triggering conditions were indexed by channel number and sorted in descending order of urgency score. Each channel recorded the trigger time advance, urgency score, and remaining capacity judgment marker. For candidate triggering conditions, the trigger time advance of the low-confidence marked channels in the switching triggering conditions was conservatively set, i.e., Ci increased by 10%. The ranking results in the switching triggering conditions served as the basis for the execution of the redundant band elimination and command issuance in the third part.

[0065] The LED spectrum switching control command is output to eliminate redundant bands in the drive scheduling arrangement according to the switching trigger conditions. The channel with the highest urgency score in the switching trigger conditions corresponds to the band with the most urgent need for illuminance compensation. The channel with the later arrangement number in the drive scheduling arrangement and whose urgency score is less than 0.5 times the average urgency score of all channels (i.e., less than half the average) is identified as a redundant band. The redundancy band identification uses the last 30% of the drive scheduling arrangement as the initial candidate range. Within the candidate range, channels are eliminated step by step from low to high urgency score. After eliminating each channel, the total power of the remaining channels is checked to see if it still exceeds the rated power limit of 85%. The total power of the remaining channels is calculated as the sum of the current drive power of the non-eliminated channels and the power increment required for illuminance compensation of the key band. If the limit is exceeded, elimination continues; otherwise, the process terminates. This step-by-step elimination mechanism relies on the urgency ranking result of the switching trigger conditions to concentrate the power margin to the key bands in the later stage of development. LED spectrum switching control commands are indexed by channel number. Each channel records the target duty cycle setting, trigger time advance amount, and redundancy elimination flag. The target duty cycle setting of the redundant elimination flag channel is set to zero to shut down the output. Non-eliminated channels are written with the reference duty cycle setting of the corresponding channel in the drive scheduling arrangement. LED spectrum switching control commands are issued to each channel drive controller in descending order of urgency score in the switching trigger conditions. The channel with the highest urgency completes the duty cycle switching first, ensuring that the band with the most urgent illuminance compensation needs reaches the target illuminance level in the shortest time.

[0066] To implement the LED multispectral staged switching control method for raising white chickens corresponding to the above method embodiments, in order to achieve the corresponding functions and technical effects. See also Figure 2 , Figure 2 This paper shows a structural block diagram of an LED multispectral staged switching control device 200 for raising white chickens, according to an embodiment of this application, comprising: The benchmark construction module 201 is used to acquire the spectral sampling data and band configuration data of the white chicken, and to perform band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark. The decay analysis module 202 is used to perform matching degree hierarchical identification of strong and weak demand growth stages based on the band adaptation benchmark, evaluate the band decay demand level based on the stage distribution of the strong and weak demand growth stages, and determine the gradient spectral response region based on the band decay demand level. The offset evaluation module 203 is used to perform duty cycle drift analysis on the spectral sampling data to form a duty cycle offset map, extract systematic offset features from the duty cycle offset map to generate band correction coefficients, and perform dispersion verification based on the duty cycle offset map and the band correction coefficients to determine the effective spectral boundary. The configuration planning module 204 is used to extract the energy consumption offset rate based on the band decay demand level, generate the optimal band configuration based on the energy consumption offset rate and the band correction coefficient, extract the non-critical band reduction margin from the optimal band configuration to generate an energy-saving buffer space, and formulate a control priority sequence by combining the energy-saving buffer space and the stepped spectral response region. The switching output module 205 is used to perform band hierarchical arrangement to form a driving scheduling arrangement using the control priority sequence, determine the switching trigger condition based on the driving scheduling arrangement and the effective spectral boundary, and output LED spectrum switching control command by eliminating redundant bands in the driving scheduling arrangement according to the switching trigger condition.

[0067] The aforementioned LED multispectral staged switching control device 200 for raising white chickens can implement the LED multispectral staged switching control method for raising white chickens described in the above method embodiments. The options in the above method embodiments are also applicable to this embodiment and will not be detailed here. The remaining contents of this application embodiment can be referred to the contents of the above method embodiments, and will not be repeated in this embodiment.

[0068] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for staged switching control of LED multispectral signals in white chicken farming, characterized in that, include: Acquire spectral sampling data and band configuration data of the white chick, and perform band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark; Based on the band adaptation benchmark, the matching degree is hierarchically identified to identify the strong and weak demand growth stages. The band decay demand level is evaluated based on the stage distribution of the strong and weak demand growth stages. The gradient spectral response region is determined based on the band decay demand level. Duty cycle drift analysis is performed on the spectral sampling data to form a duty cycle shift map. Systematic shift features are extracted from the duty cycle shift map to generate band correction coefficients. Based on the duty cycle shift map and the band correction coefficients, a dispersion check is performed to determine the effective spectral boundary. Based on the band decay demand level, the energy consumption offset rate is extracted. Based on the energy consumption offset rate and the band correction coefficient, the optimal band configuration is generated. The non-critical band reduction margin is extracted from the optimal band configuration to generate an energy-saving buffer space. The control priority sequence is formulated by combining the energy-saving buffer space and the stepped spectral response region. The control priority sequence is used to carry out band hierarchical arrangement to form a driving scheduling arrangement. Based on the driving scheduling arrangement and the effective spectral boundary, the switching trigger condition is determined. According to the switching trigger condition, the driving scheduling arrangement is used to eliminate redundant bands step by step and output LED spectral switching control command.

2. The method according to claim 1, characterized in that, The step of performing band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark includes: The spectral sampling data is processed according to the diurnal rhythm time period to extract the response intensity distribution of each band and generate a rhythm response distribution table; The rhythm response distribution table is used to identify rhythm-sensitive bands and generate rhythm response coefficients. The rhythm response coefficients and the band configuration data are used to calculate the rhythm matching degree and generate a rhythm adaptation sequence. Construct a band adaptation benchmark according to the rhythm adaptation sequence.

3. The method according to claim 1, characterized in that, The assessment of the band decline demand level based on the phase distribution of the growth period of the strong and weak demand includes: For the aforementioned periods of strong and weak demand growth, historical batch band demand decline baselines are extracted to generate a decline baseline distribution table. The deviation between the recession baseline distribution table and the strong and weak demand growth periods is calculated to generate a batch deviation sequence. The preset degradation level determination threshold is dynamically corrected according to the batch deviation sequence to generate a correction threshold set; The band decline demand level is determined by integrating the modified threshold set with the stage distribution of the strong and weak demand growth periods.

4. The method according to claim 1, characterized in that, The step of extracting systematic migration features from the duty cycle shift map to generate band correction coefficients includes: Multi-band synchronous drift detection is performed on the duty cycle offset map to generate a set of synchronous drift segments; The synchronous drift segment set is subjected to band independence verification to identify asynchronous drift segments and form a decoupled segment set; The offset accumulation position is calibrated and offset accumulation markers are generated for the decoupled segment set; The duty cycle offset spectrum is segmented and fitted using the offset accumulation marker to generate band correction coefficients.

5. The method according to claim 1, characterized in that, The step of determining the effective spectral boundary by performing a dispersion check based on the duty cycle offset spectrum and the band correction coefficient includes: Power coupling interference detection is performed on the segments in the duty cycle offset spectrum where adjacent bands are simultaneously driven by high power to generate a set of coupling interference intensity; For the set of coupling interference strength, identify band combinations where the coupling interference exceeds a preset limit and generate a set of coupling limit exceedance markers; The set of coupling over-limit markers is used to perform discrete elimination on the corresponding band configuration in the duty cycle offset map to generate an effective configuration interval set; The effective spectral boundary is determined based on the set of effective configuration intervals and the band correction coefficients.

6. The method according to claim 1, characterized in that, The process of generating the optimal band configuration based on the energy consumption offset rate and the band correction coefficient includes: Gradient tracking analysis is performed on the energy consumption offset rate to locate configurable regions and generate candidate band groups; The benefit ratio is calculated based on the physiological response intensity per unit power of each band combination in the candidate band group to generate a benefit ratio sequence. For the benefit ratio sequence, identify band combinations with benefit ratios lower than a preset threshold and generate an inefficient band label set; The optimal band configuration is generated by combining the inefficient band label set with the band correction coefficients for comprehensive sorting.

7. The method according to claim 1, characterized in that, The determination of the switching triggering conditions based on the drive scheduling arrangement and the effective spectral boundary includes: The duty cycle adjustment response delay time of each band in the drive scheduling arrangement is extracted to generate a delay time distribution table; The delay time distribution table is used to estimate the lead time of target power arrival time for each band and generate the lead trigger compensation amount. A set of candidate triggering conditions is generated by performing margin mapping between the advance triggering compensation amount and the effective spectral boundary. The switching trigger condition is determined by sorting the candidate trigger condition set according to their urgency.

8. The method according to claim 4, characterized in that, The step of generating offset accumulation markers by offset accumulation position calibration of the decoupled segment set includes: For the decoupled segment set, the offset direction of the same band within each period is extracted according to the illumination period to generate a period offset direction set; Perform multi-cycle consistency verification on the set of periodic offset directions to identify stable segments of the offset direction and generate a set of steady-state offset segments; The offset intensity classification table is generated by classifying the offset intensity using the set of steady-state offset segments. An offset accumulation marker is generated by calibrating the offset accumulation position according to the offset intensity grading table.

9. The method according to claim 6, characterized in that, The step of generating an inefficient band marker set for band combinations whose benefit ratio is lower than a preset threshold based on the benefit ratio sequence includes: Perform differential feature extraction on the benefit ratio sequence to obtain a differential label sequence; The benefit ratio sequence is adaptively segmented using the grade difference marker sequence to generate candidate boundary positions; Perform boundary stability analysis on the candidate boundary positions to screen for effective boundaries; Based on the effective boundary, inefficient segments in the efficiency ratio sequence are extracted to generate an inefficient band label set.

10. A multi-spectral LED staged switching control device for raising white chickens, characterized in that, include: The benchmark construction module is used to acquire the spectral sampling data and band configuration data of the white chicken, and to perform band adaptation analysis on the spectral sampling data and the band configuration data to construct a band adaptation benchmark. The decay analysis module is used to identify the strong and weak demand growth stages based on the band adaptation benchmark, evaluate the band decay demand level based on the stage distribution of the strong and weak demand growth stages, and determine the gradient spectral response region based on the band decay demand level. The offset evaluation module is used to perform duty cycle drift analysis on the spectral sampling data to form a duty cycle offset map, extract systematic offset features from the duty cycle offset map to generate band correction coefficients, and perform dispersion verification based on the duty cycle offset map and the band correction coefficients to determine the effective spectral boundary. The configuration planning module is used to extract the energy consumption offset rate based on the band decay demand level, generate the optimal band configuration based on the energy consumption offset rate and the band correction coefficient, extract the non-critical band reduction margin from the optimal band configuration to generate an energy-saving buffer space, and formulate a control priority sequence by combining the energy-saving buffer space with the gradient spectral response region. The switching output module is used to perform band hierarchical arrangement to form a driving scheduling arrangement using the control priority sequence, determine the switching trigger condition based on the driving scheduling arrangement and the effective spectral boundary, and output LED spectrum switching control command by eliminating redundant bands in the driving scheduling arrangement according to the switching trigger condition.

Images