Dynamic spectrum optimization method and system of full-spectrum LED light source system

Abstract

The invention relates to the technical field of LED illumination, and discloses a dynamic spectrum optimization method and system for a full-spectrum LED light source system, and the method comprises the steps: obtaining initial light intensity distribution data to obtain a basic data set, and obtaining light intensity distribution offset; grouping according to the basic data set to obtain a spectrum grouping scheme; performing scene optimization according to the light intensity distribution offset and the scheme to obtain spectrum grouping configuration of scene optimization; and finally, performing fine adjustment and parameter mapping on the configuration according to real-time illumination feedback data to obtain final spectrum management output. The method can improve the precision and environmental adaptability of spectrum management.

Description

Technical Field

[0001] This invention relates to the field of LED lighting technology, and in particular to a dynamic spectral optimization method and system for a full-spectrum LED light source system. Background Technology

[0002] Currently, in the field of lighting technology, full-spectrum LED light source systems, by simulating the natural spectrum, have become a core support for achieving healthy and comfortable lighting, and are widely used in scenarios such as medical care, education, and agriculture. However, current spectral management methods often ignore the dynamic changes in actual applications, making it difficult for the light source output to adapt to complex environments.

[0003] In existing technologies, spectral management methods primarily rely on fixed divisions of wavelength ranges and preset patterns for light intensity distribution. This rigid configuration cannot flexibly address the specific requirements of different scenarios. These shortcomings stem from the lack of effective quantification of the interaction relationships between different wavelength groups within the spectral combination. Specifically, due to the lack of correlation modeling for factors such as light intensity distribution and color temperature characteristics, the system cannot coordinate the output ratios of each group in real time when facing changes in ambient lighting. This leads to the color temperature of the light source deviating from the required range or causing energy waste, resulting in overcompensation or underresponse in the system output.

[0004] In summary, existing technologies suffer from the technical problem that full-spectrum LED light source systems cannot respond accurately to dynamic environments and user needs. Summary of the Invention

[0005] This invention provides a dynamic spectral optimization method and system for a full-spectrum LED light source system to solve the technical problem that a full-spectrum LED light source system cannot respond accurately to dynamic environments and user needs.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a dynamic spectral optimization method for a full-spectrum LED light source system, comprising: Acquire initial light intensity distribution data and perform data preprocessing to obtain the basic dataset; Feature extraction and grouping are performed on the basic dataset to obtain a spectral grouping scheme; Obtain the light intensity distribution offset, and based on the light intensity distribution offset, perform scene optimization and boundary adjustment on the spectral grouping scheme to obtain the labeling data to be optimized; The labeled data to be optimized is then optimized for the scene to obtain a scene-optimized spectral grouping configuration. Real-time acquisition of illumination feedback data for the scene-optimized spectral grouping configuration; fine-tuning of the scene-optimized spectral grouping configuration based on the illumination feedback data to obtain wavelength range adjustment parameters; Based on the wavelength range, parameters are adjusted to perform environmental response adjustments and parameter mapping, resulting in the final spectral management output.

[0007] In one optional implementation, the step of acquiring initial light intensity distribution data and performing data preprocessing to obtain a basic dataset includes: Real-time acquisition of initial light intensity distribution data; The initial light intensity distribution data is subjected to color temperature correction and noise reduction processing to obtain the basic dataset.

[0008] In one optional implementation, the step of extracting features and grouping the base dataset to obtain a spectral grouping scheme includes: For the basic dataset, feature extraction and spectral grouping boundary extraction are performed to obtain the first grouping scheme; Interference correction and drift optimization are performed on the first grouping scheme to obtain the spectral grouping scheme.

[0009] In one optional implementation, the step of obtaining the light intensity distribution shift, and optimizing the spectral grouping scheme based on the light intensity distribution shift to obtain the labeling data to be optimized, includes: Obtain user scenario datasets; The light intensity distribution data of the spectral grouping scheme is collected in real time, and the difference between the data and the initial light intensity distribution data is calculated to obtain the light intensity distribution shift. Based on the user scenario dataset, scene matching is performed to obtain the light intensity distribution offset threshold of the current scene; If the light intensity distribution offset exceeds the light intensity distribution offset threshold, anomaly marking is performed to obtain the marking data to be optimized.

[0010] In one optional implementation, the step of performing scene optimization on the labeled data to be optimized to obtain a scene-optimized spectral grouping configuration includes: For the labeled data to be optimized, scene optimization target matching is performed to obtain the scheme optimization target; Based on the optimization objective, the light intensity distribution of the spectral grouping scheme is optimized to obtain the first optimized configuration. For the first optimized configuration, interference analysis and fine-tuning of spectral grouping boundaries are performed to obtain a scene-optimized spectral grouping configuration.

[0011] In one optional implementation, the real-time acquisition of illumination feedback data for the scene-optimized spectral grouping configuration, and the fine-tuning of the scene-optimized spectral grouping configuration based on the illumination feedback data to obtain wavelength range adjustment parameters, includes: Real-time acquisition of illumination feedback data from the optimized spectral grouping configuration of the scene, followed by fluctuation analysis to obtain color temperature fluctuation characteristic data; The color temperature fluctuation feature data is compared with the scene data to obtain the scene color temperature deviation value; Based on the scene color temperature deviation value, the optimized spectral grouping configuration for the scene is optimized by gradient descent and parameter transformation is performed to obtain wavelength range adjustment parameters.

[0012] In one optional implementation, the step of adjusting parameters according to the wavelength range, performing environmental response adjustment and parameter mapping to obtain the final spectral management output includes: Frequency component analysis was performed on the illumination feedback data to obtain spectral grouping data; Cluster analysis and quantization analysis are performed on the spectral grouping data to obtain a preliminary distribution of the output intensity; The initial distribution of the output intensity is mapped using data mapping, and the final spectral management output is obtained by adjusting the parameters based on the wavelength range.

[0013] Secondly, the present invention provides a dynamic spectral optimization system for a full-spectrum LED light source system, comprising: The data processing module is used to acquire initial light intensity distribution data and perform data preprocessing to obtain the basic dataset; The spectral grouping module is used to extract features and group the basic dataset to obtain a spectral grouping scheme. The boundary adjustment module is used to obtain the light intensity distribution offset and perform scene optimization and boundary adjustment on the spectral grouping scheme according to the light intensity distribution offset to obtain the labeling data to be optimized. The grouping configuration module is used to perform scene optimization on the labeled data to be optimized, and obtain a scene-optimized spectral grouping configuration; The real-time fine-tuning module is used to collect illumination feedback data of the scene-optimized spectral grouping configuration in real time, and to fine-tune the scene-optimized spectral grouping configuration based on the illumination feedback data to obtain wavelength range adjustment parameters. The spectral output module is used to adjust parameters according to the wavelength range, perform environmental response adjustment and parameter mapping, and obtain the final spectral management output.

[0014] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention constructs a basic dataset reflecting real-time dynamics by acquiring initial light intensity distribution data and performing preprocessing; then, it uses a spectral analysis model to extract features and group the dataset to determine a dynamic spectral grouping scheme. This method overcomes the limitations of fixed wavelength range division in existing technologies, making the basis for spectral grouping more closely match the complexity and variability of the actual lighting environment, and laying a high-quality data foundation for subsequent accurate optimization.

[0015] (2) After detecting a shift in light intensity distribution, this invention optimizes the correlation mapping between intensity distribution and color temperature characteristics, and classifies dynamic changes using a clustering algorithm. This enables scenario optimization and boundary adjustment based on the specific needs of different application scenarios. This method allows spectral configuration to intelligently match the refined requirements of specific scenarios (such as medical and agricultural applications), realizing a shift from "passive adaptation" to "active optimization," and significantly improving the targeting and intelligence level of spectral management.

[0016] (3) This invention forms a closed-loop feedback control system by collecting real-time illumination feedback data and continuously fine-tuning the spectral grouping configuration optimized for the scene using algorithms such as gradient descent. This mechanism can accurately adjust the environmental response and parameter mapping according to the real-time illumination effect, and finally obtain and output a stable spectral management scheme, ensuring the accuracy, stability and environmental adaptability of the full-spectrum LED light source system in dynamically changing environments. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a dynamic spectral optimization method for a full-spectrum LED light source system provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of the dynamic spectral optimization system structure of a full-spectrum LED light source system provided in the second embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Reference Figure 1 The first embodiment of the present invention provides a dynamic spectral optimization method for a full-spectrum LED light source system, comprising the following steps: S11: Obtain initial light intensity distribution data and perform data preprocessing to obtain the basic dataset; S12, perform feature extraction and grouping on the basic dataset to obtain a spectral grouping scheme; S13, obtain the light intensity distribution offset, and perform scene optimization and boundary adjustment on the spectral grouping scheme according to the light intensity distribution offset to obtain the labeling data to be optimized; S14, perform scene optimization on the labeled data to be optimized to obtain a scene-optimized spectral grouping configuration; S15, collect the illumination feedback data of the scene-optimized spectral grouping configuration in real time, and fine-tune the scene-optimized spectral grouping configuration according to the illumination feedback data to obtain wavelength range adjustment parameters; S16, adjust parameters according to the wavelength range, perform environmental response adjustment and parameter mapping, and obtain the final spectral management output.

[0020] In step S11, initial light intensity distribution data is acquired and preprocessed to obtain a basic dataset, including: Real-time acquisition of initial light intensity distribution data; The initial light intensity distribution data is subjected to color temperature correction and noise reduction processing to obtain the basic dataset.

[0021] It is important to note that real-time acquisition of initial light intensity distribution data is fundamental to the entire dynamic spectral optimization process, aiming to accurately obtain the real-time illumination status of the light source system itself and the external environment. This step uses high-precision sensors, such as spectrophotometers, to capture the current output spectral data of the full-spectrum LED light source system and ambient light illumination information. This information includes light intensity values ​​accurate to specific wavelengths. To standardize the acquired raw data and facilitate subsequent processing, the system employs a spectral resampling algorithm to adjust the wavelength range. Especially when the wavelength range of the original data exceeds a wavelength range threshold, resampling generates initial light intensity distribution data with a uniform format and continuous data points, providing high-quality data input for subsequent precise analysis and adjustments.

[0022] It is worth noting that the wavelength range threshold (400nm) is primarily based on common knowledge and industry standards in the lighting field. The core of full-spectrum lighting lies in simulating the visible spectrum, which has the most significant impact on human vision and circadian rhythms. This spectral range is typically defined between 380nm (violet) and 780nm (red), with a total span of exactly 400nm (780nm - 380nm). Therefore, setting the threshold to 400nm serves as a data validity checkpoint, ensuring that the data collected by the sensor completely covers the entire visible light band effective for the human eye. If the collection range exceeds this value, it may mean that the data contains invalid information from non-visible light bands (such as ultraviolet or infrared); if it is much smaller than this value, it may indicate incomplete data collection. In both cases, initiating a resampling algorithm not only standardizes the data step size but also provides an opportunity to normalize the data to the standard 380nm to 780nm visible light range, ensuring the validity of subsequent analysis.

[0023] It is worth noting that the spectral resampling algorithm uses linear interpolation, assuming that the change in light intensity is linear between two adjacent original data points. Based on the input target wavelength range and target wavelength step size, a standardized, equally spaced target wavelength sequence λ' is first generated. Then, each target wavelength point generated in the previous step is traversed... Calculate the corresponding light intensity value. For any target wavelength point In the raw spectral data, find the... The two nearest original wavelength points and , making ≤ ≤ These two points correspond to light intensities respectively. and The light intensity I'_j at λ'_j is calculated using a linear interpolation formula. This formula is derived based on the two-point linear equation: Shin Kong Strong equal to the light intensity at the starting point Add an increment. This increment is based on λ'_j in [ , The relative position within the interval, calculated proportionally to the change in light intensity. All calculated ( , The data pairs are combined to form the final resampled spectral data.

[0024] For example, the visible spectrum from 380 nm to 780 nm can be measured using a spectrophotometer, with the light intensity value of each point recorded at 10 nm intervals. For instance, the system might detect an internal light intensity of 50 mW / nm at a wavelength of 450 nm at a certain moment, while ambient light contributes an additional 20 mW / nm at a wavelength of 550 nm. If the wavelength range of the acquired spectral data exceeds the 400 nm wavelength range threshold, the system will perform spectral resampling, unifying the originally unevenly spaced wavelength points to a step size of 5 nm, generating a standardized initial distribution curve. During this process, due to errors introduced by interpolation calculations, the total integrated light intensity may change from the original 1000 mW to approximately 950 mW. Subsequently, the system will normalize the total luminous flux back to the target value to ensure stable illumination intensity. This processing not only unifies the data standard but also effectively reduces redundancy in subsequent calculations, improving the system's response speed.

[0025] It should be noted that after obtaining the initial light intensity distribution data, a series of preprocessing operations are required to improve data quality, mainly including color temperature correction and noise reduction. Color temperature correction refers to the system extracting color temperature characteristic values ​​from the initial light intensity distribution data and calculating a correction value using a blackbody radiation model to adjust the light intensity distribution to better suit natural light or specific scene requirements, thereby improving visual comfort. Specifically, the calculation process is as follows: First, based on a preset target color temperature value (e.g., 3200K), the system generates a standard, ideal spectral distribution curve using Planck's blackbody radiation law. This curve represents the most natural spectral form at that color temperature. Then, the system compares the spectral distribution curve actually measured by the sensor with this ideal curve band by band. The "correction value" is a correction function or a set of scaling factors calculated through this comparison; it quantifies the difference or ratio between the actual light intensity and the ideal light intensity at each wavelength. For example, to increase the color temperature from 2800K to 3200K, model comparisons show that the current spectrum is insufficient in the blue region compared to the ideal spectrum, while being relatively excessive in the red region. Therefore, the calculated correction value guides the system to increase the output power of the blue LED channel while potentially reducing the power of the red LED channel, thus making the final synthesized spectrum approximate the ideal 3200K blackbody radiation spectrum. Next, when integrating the corrected light intensity distribution data, the system uses a Fourier transform filtering algorithm to denoise the data to eliminate environmental or electrical noise that may be introduced during sensor acquisition, resulting in smooth and reliable filtered light intensity distribution data. Finally, a basic dataset is constructed based on this high-quality filtered data, and a label verification process ensures the dataset's integrity and accuracy, providing support for subsequent spectral grouping and model training.

[0026] It is worth noting that the preset target color temperature value is determined based on well-known standards in the lighting field and the needs of specific application scenarios. Different color temperature values ​​correspond to different lighting atmospheres and functional uses, and are key parameters for improving visual comfort and matching environmental requirements. For example, low color temperature values ​​(such as 2700K-3200K) are often used to create a warm and relaxing home environment, simulating the sunlight of early morning or dusk; neutral color temperatures (such as 4000K-5000K) are suitable for offices or commercial spaces, providing clear and efficient work lighting; while high color temperature values ​​(such as 6500K) simulate midday sunlight and are often used in professional fields that require high color rendering or accurate color judgment. Therefore, setting the target color temperature to a specific value (such as 3200K) aims to adjust the spectrum output by the system to best meet the lighting quality requirements of the target application scenario (such as a home environment).

[0027] For example, suppose the system sets a preset threshold of 3000K for color temperature characteristics, while the actual color temperature value extracted from the initial light intensity distribution data is 2800K. In this case, the system will activate the blackbody radiation model for correction. By comparing the deviation between the actual spectrum and the ideal blackbody radiation spectrum, the color temperature will be increased from 2800K to the target value of 3200K. This process may correspondingly increase the light intensity in the blue light region by 15% (this correction ratio is derived by calculating the root mean square error between the ideal and actual spectra in each band; 15% is an example adjustment). In the denoising process, assuming that the original data experienced a 5% numerical fluctuation due to high-frequency interference when integrating the total luminous flux, the Fourier transform filtering algorithm can effectively filter out these noise components, stabilizing the integral value at 900lm. Finally, a basic dataset is constructed based on the processed data. This dataset may contain 1000 spectral samples with precise light intensity and color temperature labels. After confirming that its label coverage reaches 95%, the dataset is finalized and can be used for subsequent machine learning training, thereby supporting intelligent spectral adaptive optimization.

[0028] In step S12, feature extraction and grouping are performed on the basic dataset to obtain a spectral grouping scheme, including: For the basic dataset, feature extraction and spectral grouping boundary extraction are performed to obtain the first grouping scheme; Interference correction and drift optimization are performed on the first grouping scheme to obtain the spectral grouping scheme.

[0029] It should be noted that the core of this step lies in the in-depth analysis of the preprocessed basic dataset to scientifically classify the spectrum. First, spectral reconstruction is performed on the interaction between wavelength range and light intensity distribution in the basic dataset. This process aims to extract key feature values ​​that reflect the dynamic trends of spectral changes. When the analyzed feature values, such as the slope of the light intensity peak change (a slope that quantifies the overall trend of recent light intensity changes by performing linear regression analysis on the spectral intensity peaks over a continuous time period, for example, linear regression analysis on the spectral intensity peaks over the last 5 seconds), exceed the peak change threshold, it indicates a significant change pattern in the spectral distribution, at which point the system triggers the next step. Next, the spectral data with significant features are processed using a non-negative matrix factorization method to accurately identify and extract the boundaries between different spectral bands, thus forming a preliminary, objective spectral grouping scheme that reflects the current illumination characteristics—the first grouping scheme. The specific operation of the non-negative matrix factorization method is as follows: the system first constructs a non-negative matrix V from the continuously collected, preprocessed spectral data over a period of time. In this matrix, each row corresponds to a specific wavelength, and each column corresponds to a spectral measurement sample at a given time. The element values ​​in the matrix are the light intensity at the corresponding wavelength. Matrix factorization (V≈W×H): The goal of the NMF algorithm is to decompose the original data matrix V into the product of two low-rank non-negative matrices: a basis matrix W and a coefficient matrix H. Basis matrix W: Its column vectors are called "basic spectra." These basic spectra represent the most fundamental and core spectral components that constitute all original spectra. In a full-spectrum LED system, these basic spectra typically correspond to the characteristic spectra of the main color channels in the system (such as blue, green, and red LEDs). Before decomposition, the number of basic spectra k needs to be preset (e.g., k=3 for an RGB system). Coefficient matrix H: Its column vectors contain weight information, indicating how much weight each basic spectrum in W needs to be multiplied by before linearly superimposed to reconstruct the spectrum at a given time in the original spectrum. After decomposition, the system focuses on analyzing the basis matrix W. Each column of W is a pure basic spectral curve. The system analyzes the shape of these k spectral curves, such as finding the peak wavelength position of each curve. The peaks of these fundamental spectra clearly indicate the center wavelengths of different core color channels. The "boundaries" of spectral groups are determined at key locations where these adjacent fundamental spectral curves transition or overlap, such as the intersection of two fundamental spectral curves or a midpoint between two peak wavelengths. Through this data-driven approach, NMF can adaptively and accurately find the group boundaries that best represent the intrinsic structure of the spectrum.

[0030] It is worth noting that the peak change threshold (0.5 mW / nm / s) is primarily based on dynamic algorithm calculation, aiming to find an optimal balance between system response speed and computational efficiency for the spectral grouping trigger mechanism. If the threshold is too low, the system may frequently perform unnecessary spectral grouping recalculations due to minor noise interference, increasing energy consumption; if the threshold is too high, the system may be slow to respond to real changes in ambient light. To determine the optimal threshold, a grid search or similar optimization algorithm can be performed on a historical spectral dataset containing a large amount of labeled data (distinguishing between "stable states" and "significantly changing states"). The system will test a series of candidate thresholds (e.g., 0.1, 0.3, 0.5, 0.7 mW / nm, etc.) and evaluate the accuracy and recall of the system in correctly identifying "significantly changing" events at each threshold. Finally, the candidate value (e.g., 0.5 mW / nm / s) that maximizes the overall performance index (e.g., F1 score) is selected as the fixed preset threshold.

[0031] For example, when processing the basic dataset, the system analyzes the mapping relationship between wavelength and light intensity in the spectral data to reconstruct the spectrum and capture the light intensity variation patterns within a specific wavelength range. If the calculated peak slope of light intensity in a certain segment exceeds the peak variation threshold of 0.5 mW / nm / s, the system determines that grouping is necessary. Subsequently, a non-negative matrix factorization method is initiated to divide the entire visible spectrum from 380 nm to 780 nm into three core sub-regions: blue, green, and red, with their boundaries located at 450 nm and 600 nm, respectively. This initial grouping scheme lays the foundation for subsequent differentiated and refined adjustments for different spectral segments.

[0032] It should be noted that after obtaining the first grouping scheme, interference correction and drift optimization are required to ensure its adaptability and stability in complex environments. Interference correction aims to eliminate the interference of external ambient light on the grouping boundaries. The system extracts the peak position of light intensity and the ambient light interference factor from the initial scheme. If the interference factor exceeds the interference factor threshold, a convolution filter is used to smooth the spectral curve and fine-tune the grouping boundaries to obtain the corrected spectral grouping scheme. Subsequently, drift optimization is performed based on the corrected scheme. Correlation calculations are performed to analyze the magnitude of dynamic spectral shift and color temperature drift. If the drift magnitude is less than the drift magnitude threshold, vector matching is performed to fuse the gradient information of light intensity distribution to adjust the spectral characteristics of each group, ultimately obtaining an optimized spectral grouping scheme that can adapt to dynamic changes and has a stable color temperature. The specific implementation process of the vector matching is as follows: First, the spectral distribution is defined as a mathematical "vector". The "target vector" is the standard spectral power distribution (SPD) curve corresponding to the preset ideal color temperature (e.g., 3500K before drift). The "current vector" is the spectral power distribution curve that has drifted and is measured by the sensor in real time. An "error vector" is obtained by subtracting the "target vector" from the "current vector." This error vector precisely reveals whether the current light intensity is excessive or insufficient at each wavelength. For example, when the color temperature shifts from 3500K to a higher 3600K, the error vector will typically show a negative value in the blue light band (indicating insufficient intensity) and a positive value in the red and yellow light bands (indicating excessive intensity). The core of this method is "integrating the gradient information of the light intensity distribution." The system internally stores the mathematical relationship between the power change of each independent LED color channel (such as red, green, and blue) and its impact on the overall spectrum and color temperature, i.e., the "gradient." Upon receiving the error vector, it uses this gradient information to calculate an optimal correction scheme. It doesn't simply correct point by point, but intelligently determines which color channel(s) should be adjusted and to what extent to most efficiently counteract the error vector (especially to compensate for deficiencies in the blue light region). This is essentially an optimization problem: finding a set of minimum adjustments to each color channel that minimizes the "distance" or "difference" between the adjusted composite spectral vector and the target vector. After the matching calculation is complete, a set of specific adjustment parameters is output, such as "increase the driving current of the blue channel by 15%" or "decrease the driving current of the red channel by 5%." These parameters are sent to the LED driver for execution, directly adjusting the spectral characteristics of each group.

[0033] It is worth noting that the interference factor threshold (5mW / nm) is primarily based on experimental statistics, aiming to define an acceptable minimum signal-to-interference ratio for the system. This threshold is typically determined through a series of tests in a controlled experimental environment: engineers introduce interference light sources of varying intensities, simulating common ambient light, and continuously monitor the system's key performance indicators (KPIs), such as color temperature stability and color rendering index (CRI). The threshold (e.g., 5mW / nm) is set at a critical point: when the interference factor is below this value, the fluctuations in various system performance indicators are within an acceptable small range; however, when the interference factor exceeds this value, the performance indicators begin to deteriorate significantly. This value is similar to an empirical lower limit for signal quality, ensuring that the main light source signal dominates the data analysis. The principle is similar to selecting principal components that retain more than 85% of the information in principal component analysis to ensure data representativeness.

[0034] It's worth noting that the threshold for the drift amplitude is primarily based on well-known facts in the lighting field and human visual perception standards in color science. Research shows that the human eye has a "Just-Noticeable Difference" (JND) threshold for perceiving color temperature changes. When the color temperature change is less than this threshold, the human eye usually cannot distinguish the difference from the original color temperature. The 3% threshold is a well-founded value set based on these perception models, aiming to ensure that any color temperature drift automatically corrected by the system occurs at a critical point that is about to be or has already been detected by the user. This ensures both the long-term stability of the lighting environment, avoiding unnecessary frequent adjustments to minor, imperceptible normal fluctuations due to oversensitivity, and timely intervention and correction when real, perceptible changes occur in light and color quality.

[0035] For example, the system detects from the first grouping scheme that the peak intensity of blue light is 60 mW / nm, while the ambient light interference factor at 550 nm reaches 10 mW / nm, exceeding the interference factor threshold of 5 mW / nm. At this point, the convolutional filter automatically intervenes, smoothing the effect caused by the abrupt change in the spectral curve and fine-tuning the grouping boundary originally set at 450 nm to 455 nm to obtain a more accurate correction scheme. Next, based on this correction scheme, the system detects that the color temperature has drifted from 3500 K to 3600 K, with a drift amplitude of approximately 2.86%, which meets the drift amplitude threshold of less than 3%. Therefore, vector matching is performed to counteract the color temperature drift by enhancing the light intensity distribution in the blue light region, thereby generating the final optimized spectral grouping scheme and improving the adaptability and stability of the spectral output.

[0036] In step S13, the light intensity distribution shift is obtained. Based on the light intensity distribution shift, the spectral grouping scheme is optimized in terms of scene and boundary adjustment to obtain the labeling data to be optimized, including: Obtain user scenario datasets; The light intensity distribution data of the spectral grouping scheme is collected in real time, and the difference between the data and the initial light intensity distribution data is calculated to obtain the light intensity distribution shift. Based on the user scenario dataset, scene matching is performed to obtain the light intensity distribution offset threshold of the current scene; If the light intensity distribution offset exceeds the light intensity distribution offset threshold, anomaly marking is performed to obtain the marking data to be optimized.

[0037] It's important to note that the user scenario dataset is a pre-built database used to store ideal spectral configuration parameters for different application scenarios (such as offices, medical operating rooms, and agricultural greenhouses). This dataset was generated through cluster analysis of a large amount of spectral data, specifically divided according to the color temperature characteristics required in different scenarios. Standard spectral grouping boundaries, light intensity distribution models, and dynamic adjustment rules were defined for each specific scenario. Obtaining this dataset is a prerequisite for achieving scene adaptive optimization, providing a data foundation for subsequent scene matching and threshold determination.

[0038] For example, when constructing a user scenario dataset for "indoor lighting," cluster analysis is performed to subdivide the color temperature requirements of this scenario into three categories: 6000K cool light (suitable for focused work), 3000K warm light (suitable for relaxation), and 4500K neutral light (suitable for daily activities). The dataset stores the spectral grouping boundaries corresponding to these three requirements; for example, the boundaries can be set to 430nm and 580nm respectively, thereby ensuring that the system can accurately match and call the most suitable lighting scheme according to specific user needs or time periods.

[0039] It should be noted that this step is used to quantify the deviation between the current lighting environment and the ideal state. The system uses a light sensor to continuously collect real-time light intensity distribution data under the current spectral grouping scheme. This real-time data is then compared with the initial light intensity distribution data obtained in step S11 as a reference, with differences calculated point-by-point or band-by-band. Through this subtraction operation, a precise quantified "light intensity distribution offset" value can be obtained, which intuitively reflects the degree of deviation between the actual lighting and the initial setting due to environmental changes, equipment aging, or other factors.

[0040] For example, suppose that initially, the system's light intensity at a wavelength of 520 nm is 50 mW / nm. After the system has been running for a period of time, the real-time light intensity data at that point becomes 60 mW / nm. By calculating the difference (60 mW / nm - 50 mW / nm), the system obtains a light intensity distribution offset of +10 mW / nm at 520 nm. This 10 mW / nm offset will be used for subsequent judgment to determine whether an optimization adjustment procedure needs to be initiated.

[0041] It's important to note that, to achieve intelligent and refined management, the system does not employ a fixed offset threshold. Instead, it dynamically matches and retrieves a scene-specific light intensity distribution offset threshold from the user scenario dataset based on the current application scenario. The system first identifies current environmental characteristics, such as time and external light intensity, and then matches them with scenes in the user scenario dataset. Once a match is successful, the system extracts the preset offset tolerance for that scene, i.e., the light intensity distribution offset threshold. This approach makes threshold setting more flexible and scientific, accommodating the varying light stability requirements of different scenarios.

[0042] For example, the system identifies the current application scenario as "conference room lighting." By querying the user scenario dataset, the system matches this scenario and reads the pre-set light intensity distribution offset threshold of 8mW / nm. This threshold means that in a conference room scenario, as long as the real-time fluctuation of light intensity does not exceed 8mW / nm, the system considers the lighting environment to be stable and requires no adjustment. In other scenarios, such as medical scenarios requiring higher precision, this threshold may be set even lower.

[0043] It's important to note that this is a decision-making and triggering step. The system compares the real-time calculated light intensity distribution offset (from the previous step) with the current scene light intensity distribution offset threshold obtained from the scene dataset. If the absolute value of the calculated offset exceeds the scene's allowed threshold, the system determines that the current spectral output has a significant deviation, exceeding the acceptable range. At this point, the system marks that time point or data segment as "abnormal" or "needs optimization," thereby generating data to be optimized.

[0044] For example, the system compares the calculated light intensity distribution shift of 10 mW / nm at 520 nm with the shift threshold of 8 mW / nm obtained from the "conference room lighting" scene. Since 10 mW / nm exceeds the 8 mW / nm threshold, the system determines that the current illumination has shifted abnormally. Therefore, the system marks the current spectral data containing this shift value, generates a "marked data to be optimized," and passes it to the next processing module. In step S14, scene optimization is performed on the labeled data to be optimized to obtain a scene-optimized spectral grouping configuration, including: For the labeled data to be optimized, scene optimization target matching is performed to obtain the scheme optimization target; Based on the optimization objective, the light intensity distribution of the spectral grouping scheme is optimized to obtain the first optimized configuration. For the first optimized configuration, interference analysis and fine-tuning of spectral grouping boundaries are performed to obtain a scene-optimized spectral grouping configuration.

[0045] It's important to note that this step is the first step in responding to the labeled data to be optimized and initiating targeted scene optimization. The system will process and match the pre-established scene dataset using a clustering algorithm based on the current user input or environmental state. The clustering algorithm analyzes core features such as light intensity, color temperature, and wavelength distribution in the dataset, dividing the complex scene data into multiple sets with distinct characteristics. By matching the data to be optimized with these categorized scene feature sets, the system can accurately determine the most suitable scene features. These matched features (such as the target color temperature range and light intensity range) collectively constitute the "optimization target" for this optimization.

[0046] For example, when addressing user needs for a single indoor lighting scenario, clustering algorithms can automatically categorize light intensity and color temperature data from different time periods within the scenario dataset into three scene features: cool light requirements in the morning, neutral light requirements at midday, and warm light requirements at night. If the optimization target of the currently matched solution is "cool light requirements in the morning," this target will be specifically defined as: a target light intensity range of 500-600 lx and a target color temperature range of 5500-6000 K. This clear target provides precise guidance for subsequent adjustments to light intensity and spectrum.

[0047] It should be noted that after determining the optimization target, the system will correct for deviations in the current light intensity distribution. This process uses threshold comparison logic to process the environmental response data, comparing the currently acquired real-time light intensity distribution with the target light intensity defined in the optimization target. By analyzing the deviation between the two, the system can calculate specific corrected intensity distribution parameters. These parameters aim to adjust the current light intensity to the target range, thereby eliminating deviations and ensuring the stability and compliance of the illumination. The spectral grouping scheme after applying these corrected parameters constitutes the first optimized configuration.

[0048] For example, suppose the light intensity deviation threshold set in the optimization target is 7mW / nm, but the system detects an actual deviation of 9mW / nm at the 510nm wavelength during real-time monitoring, exceeding the threshold. In this case, the threshold comparison logic will flag this deviation and generate a correction parameter, instructing the system to reduce the light intensity at that wavelength by 3mW / nm to bring it back within the target range. After this adjustment, the light intensity distribution of the entire system reaches the first optimized configuration, preparing for subsequent fine-tuning.

[0049] It should be noted that, to enable the spectral configuration to adapt to more complex dynamic changes, the system further performs interference analysis and boundary fine-tuning based on the initial optimized configuration. First, the system extracts dynamic intensity variation data from scene features and analyzes the frequency components of this data using methods such as Fourier transform to identify the presence of periodic external light source interference. Then, the system combines the interference analysis results, the corrected intensity distribution parameters obtained in the previous step, and the scene adaptation boundary condition data (such as color temperature range limitations) to perform a final fine-tuning of the proportions and boundaries of the spectral groups using data mapping methods. This results in a final scene-optimized spectral grouping configuration that can adapt to dynamic changes and user needs.

[0050] For example, in the interference analysis, the system detected a periodic light intensity fluctuation with a frequency of 0.15 Hz and an amplitude of 5 mW / nm near the 510 nm wavelength using Fourier transform. This indicates potential interference caused by flickering from other indoor lights. Taking this interference into account, and combining it with scene boundary condition data (e.g., specifying that the indoor lighting color temperature must be between 3000-5500K), the data mapping method transforms the corrected light intensity parameters, ultimately adjusting the spectral grouping configuration to an output ratio of 40% blue light, 35% green light, and 25% red light. This configuration satisfies both the overall requirements for light intensity and color temperature, while also considering dynamic interference, representing the final result of scene optimization.

[0051] In step S15, illumination feedback data of the scene-optimized spectral grouping configuration is collected in real time. Based on the illumination feedback data, the scene-optimized spectral grouping configuration is fine-tuned to obtain wavelength range adjustment parameters, including: Real-time acquisition of illumination feedback data from the optimized spectral grouping configuration of the scene, followed by fluctuation analysis to obtain color temperature fluctuation characteristic data; The color temperature fluctuation feature data is compared with the scene data to obtain the scene color temperature deviation value; Based on the scene color temperature deviation value, the optimized spectral grouping configuration for the scene is optimized by gradient descent and parameter transformation is performed to obtain wavelength range adjustment parameters.

[0052] It should be noted that this step is the starting point for achieving closed-loop fine-tuning control, aiming to obtain the true effect of system operation. The system continuously collects real-time ambient light data, including light intensity and color temperature, under the current spectral grouping configuration through a light sensor, forming time-series feedback data. To gain a deeper understanding of the dynamic characteristics of illumination, the system uses Fourier transform to process this time-series data, analyzing its frequency components to reveal the periodic variation and distribution of color temperature. This process can identify fluctuations caused by external light source interference or equipment instability, ultimately generating color temperature fluctuation characteristic data that can quantitatively describe the dynamic changes in color temperature.

[0053] For example, in a single indoor lighting scenario, the ambient light data collected by the light sensor at a certain moment is 520 lx light intensity and 5700 K color temperature. After performing Fourier transform analysis on the continuously collected light feedback data, it was found that the color temperature fluctuates periodically at a frequency of 0.2 Hz within the range of 5600 K to 5800 K. This fluctuation may indicate the presence of imperceptible light flicker or the influence of external light sources. This dynamic information, including frequency, amplitude, and range, collectively constitutes the color temperature fluctuation characteristic data, providing a basis for subsequent precise adjustments.

[0054] It's important to note that after obtaining the color temperature fluctuation characteristic data, the system matches and compares it with a pre-established dataset of scene requirements. This dataset stores the ideal color temperature range or threshold for each specific application scenario (such as an office in the morning). Through comparison logic, the system determines whether the actual distribution of the current color temperature matches the scene requirements. If there is a mismatch, for example, if the current color temperature is below the lower limit of the scene requirements, the system will accurately calculate the deviation between the two and determine the adjustment direction (such as increasing or decreasing the color temperature), thus obtaining a scene color temperature deviation value that provides a clear target for subsequent optimization steps.

[0055] For example, the current scenario is "morning indoor lighting," whose ideal color temperature range specified in the scenario requirement dataset is 5500K to 6000K. Based on the color temperature fluctuation characteristic data obtained in the previous step, the system calculates the average color temperature for the current period to be 5400K. Since 5400K is lower than the threshold of 5500K, the comparison logic determines that there is a deviation and calculates the field color temperature deviation value as -100K, while determining the adjustment direction as "increasing the color temperature." This quantified deviation value ensures the targetedness and accuracy of subsequent optimization adjustments.

[0056] It's important to note that this step is the core execution stage for fine-tuning. The system uses the scene color temperature deviation value and adjustment direction obtained in the previous step as input to the gradient descent algorithm. The gradient descent algorithm systematically adjusts the light intensity distribution data (e.g., the output ratio of each spectral group) through multiple iterations. Each step aims to reduce the color temperature deviation value until the deviation converges to an acceptable range, thus obtaining optimized light intensity distribution parameters. Subsequently, to apply the algorithm-level parameters to the physical device, the system uses a data mapping method to convert the optimized light intensity distribution parameters into specific, executable wavelength range adjustment parameters, completing the final precise wavelength range configuration.

[0057] It's worth noting that the specific implementation process of the gradient descent algorithm is as follows: First, define the cost function. The core of the algorithm is to minimize a "cost" or "loss," which in this case is the deviation between the actual color temperature and the target color temperature. The cost function J can be defined as the square of the difference between the two, i.e., J = (CCT_current - CCT_target) 2 Here, J represents the cost, CCT_current refers to the actual color temperature of the LED light source as measured in real time by the system's light sensor, and CCT_target refers to the system's preset ideal color temperature value. The algorithm's goal is to adjust the parameters to bring the value of J close to zero. The adjustable parameters of the algorithm are variables that control the output intensity of each spectral group (color channel), such as the power of the blue channel (P_blue), the power of the green channel (P_green), and the power of the red channel (P_red). The "gradient" refers to the rate of change (i.e., partial derivative) of the cost function J with respect to each adjustable parameter; that is, (∂J / ∂P_blue, ∂J / ∂P_green, ...) is the gradient. This gradient indicates the direction in which the cost function decreases the fastest. This gradient can be calculated using an established spectral-color temperature physical model or empirically determined through small perturbation tests during actual operation. The gradient descent algorithm updates each parameter using the following iterative formula. In each iteration, each parameter moves a small step along its negative gradient direction (the step size is controlled by the learning rate α = 0.01), thereby reducing the total cost J. This process is repeated until the maximum number of iterations is reached (100) or the cost function J is less than 100K. 2 The iteration stops when the power of each channel (P_blue, P_green, P_red) is reached, and the optimized light intensity distribution parameters are then obtained. Subsequently, in order to apply the algorithm-level parameters to the physical device, the system uses a data mapping method to convert the optimized light intensity distribution parameters into specific, executable wavelength range adjustment parameters, thus completing the final precise wavelength range configuration.

[0058] For example, based on a scene color temperature deviation of -100K and the adjustment direction of "increasing color temperature", the gradient descent algorithm begins to iteratively optimize the initial light intensity distribution of "50% blue light, 30% green light, and 20% red light". The algorithm gradually increases the proportion of blue light, and after several iterations, the optimized light intensity distribution is obtained as "55% blue light, 28% green light, and 17% red light". Next, the data mapping method translates this new proportion configuration into hardware instructions. For example, the increase in the proportion of blue light corresponds to enhancing the output in the 450-480nm wavelength range. The final generated wavelength range adjustment parameter may be "increase the proportion of blue light wavelength output to 42%", thereby accurately meeting the scene's requirement for a high color temperature.

[0059] In step S16, environmental response adjustment and parameter mapping are performed according to the wavelength range adjustment parameters to obtain the final spectral management output, including: Frequency component analysis was performed on the illumination feedback data to obtain spectral grouping data; Cluster analysis and quantization analysis are performed on the spectral grouping data to obtain a preliminary distribution of the output intensity; The initial distribution of the output intensity is mapped using data mapping, and the final spectral management output is obtained by adjusting the parameters based on the wavelength range.

[0060] It should be noted that this step aims to scientifically establish the boundaries of spectral groupings by analyzing the dynamic characteristics of real-time illumination feedback data. The system employs the Fast Fourier Transform (FFT) method to process the time-series feedback data continuously acquired by the illumination sensor. The Fast Fourier Transform can decompose complex time-series signals into sinusoidal components of different frequencies, thereby revealing potential periodic fluctuations in light intensity or color temperature. By identifying these frequency components, the system can determine whether interference caused by factors such as power instability or external light sources exists, and accordingly determine or optimize the boundary distribution of spectral groupings, ultimately generating spectral grouping data for subsequent processing.

[0061] It is worth noting that the specific implementation process of the Fast Fourier Transform (FFT) method is as follows: The system first continuously collects feedback data on light intensity or color temperature at a fixed sampling rate within a preset time window, forming a time-series signal containing N sampling points. For example, collecting data at a frequency of 10Hz for 20 seconds will yield a signal buffer containing 200 data points. Before performing the FFT, to reduce spectral leakage errors that may occur due to signal truncation, a window function (such as the Hanning window) is usually applied to the collected time-series data. Windowing causes the signal to smoothly decay to zero at both ends of the buffer, improving the accuracy of the FFT analysis. The preprocessed data is then input into the FFT algorithm. The FFT efficiently converts this time-domain signal of length N into a frequency-domain complex sequence of length N. The algorithm's output is a frequency-domain view; typically, we are interested in its "amplitude spectrum." The horizontal axis of the amplitude spectrum represents frequency (Hz), and the vertical axis represents the intensity or amplitude of that frequency component. A stable DC light source will have most of its energy concentrated at 0Hz. If significant peaks appear at other non-zero frequency points in the spectrum, it indicates the presence of a periodic fluctuation at a corresponding frequency in the original signal. The results of FFT analysis (i.e., the identified frequency components) primarily serve as a diagnostic and triggering mechanism. It does not directly calculate the wavelength boundaries of spectral grouping (e.g., 450nm), but rather determines whether the current spectral state is suitable for grouping or needs to be regrouped by identifying the signal's "instability." When significant low-frequency fluctuations are detected, the system determines that the current lighting environment is unstable, which triggers a deeper spectral structure analysis algorithm (such as the aforementioned non-negative matrix factorization) to ensure that the boundaries of spectral grouping remain valid and optimal under the current (even unstable) lighting conditions.

[0062] For example, after analyzing real-time illumination feedback data, the system found periodic fluctuations in light intensity at a frequency of 0.2 Hz, which may be attributed to slight instability in the light source power supply. This 0.2 Hz frequency peak was obtained by analyzing the amplitude spectrum using the aforementioned FFT process. Based on this frequency analysis result, the system divided the entire visible spectrum into three main bands: 380-450 nm (blue light region), 450-590 nm (green light region), and 590-780 nm (red light region). This clearly defined set of bands and their boundaries together constitute the spectral grouping data guiding the next step of cluster analysis.

[0063] It should be noted that after establishing the spectral grouping data (i.e., the wavelength range of each group), the system uses the K-means clustering algorithm to process the real-time illumination feedback data to quantify the current actual output intensity of each spectral group. The K-means clustering algorithm iteratively assigns the collected spectral data points to the corresponding categories according to preset group boundaries. After clustering, the system performs statistical analysis on the data within each group, calculating the group's average intensity, fluctuation range, and other quantitative characteristics. These quantitative characteristics collectively constitute a preliminary distribution of the output intensity reflecting the current illumination state.

[0064] For example, based on the spectral grouping data obtained in the previous step, the K-means clustering algorithm will group all illumination data points falling within the 450-590nm wavelength range into one category (green light group). After quantitative analysis, the system calculates that the average intensity of this group is 42%, with a fluctuation range of ±4%. This 42% average intensity value is the preliminary distribution of the current green light band output intensity, which will serve as the basis for subsequent comparison with the target value.

[0065] It's important to note that this step is crucial for generating the final control commands. The system first compares the initial distribution of output intensity with the target threshold in a pre-established environmental response dataset. Through comparison logic, it calculates the deviation between the two and determines the adjusted target output intensity accordingly. Subsequently, the system uses a data mapping method to convert this adjusted output intensity, expressed as a percentage, into specific wavelength range adjustment parameters that can be directly executed by the lighting equipment. This process ensures that the output adjustment not only corrects deviations but also adapts to dynamic changes, ultimately generating a final spectral management output that achieves both uniform illumination and environmental adaptability.

[0066] It should be noted that the pre-established environmental response dataset is a database or lookup table storing ideal spectral parameters for various preset scenarios, providing a precise target benchmark for closed-loop feedback control of the spectrum. The core structure of this dataset is the mapping relationship from "scene conditions" to "target spectral thresholds." For example, an entry for a "standard office scene" might correspond to a set of parameters, including total illuminance (500 lx), target color temperature (5800 K), and the ideal light intensity percentage for each major wavelength band (e.g., 48% for the green band). The construction method of this dataset mainly combines theoretical calculations with experimental calibration. First, based on international lighting standards (such as EN 12464-1) or ergonomic studies for specific application scenarios, the ideal illuminance, color temperature, and other macroscopic indicators for each scenario are theoretically determined. Subsequently, in a standardized experimental environment, real-time measurements are performed using precision instruments such as a spectral analyzer, and professionals finely adjust the output power of each color channel in the LED system until the actual output spectrum precisely matches the aforementioned theoretical indicators. The system ultimately records the output intensity ratio of each channel at this time and uses it as the 'target threshold' for this scenario, which is then stored in the dataset for comparison and correction when subsequent control commands are generated.

[0067] It is worth noting that the final spectral management output is used to control the current light source output on the one hand, and its data also serves as the initial data or feedback for the next control cycle, thereby initiating a new round of optimization; the grouping in this step is based on microsecond / millisecond-level dynamic micro-grouping of real-time feedback data, which aims to cope with the instantaneous changes in the lighting environment; while the grouping in step S12 is based on macroscopic grouping of the basic dataset, which aims to establish the overall spectral framework.

[0068] For example, the system compares the initial green light intensity of 42% with the centralized requirement of 48% in the environmental response data set, calculating a deviation of -6%. To compensate for this deviation, the system determines the adjusted target intensity to be 46%. Next, the data mapping method translates the instruction to "increase green light intensity to 46%" into a specific adjustment to the output proportion in the 450-590nm band. After this adjustment, the system achieves a stable final spectral management output, for example, stabilizing the ambient light intensity within the range of 515-525 lx and the color temperature between 5750-5850 K, significantly improving the accuracy and dynamic adaptability of spectral management.

[0069] In summary, this invention constructs a closed-loop spectral management process, from initial data acquisition and preprocessing, dynamic spectral grouping, scene optimization based on light intensity distribution shift, to real-time feedback fine-tuning and final parameter mapping. It combines scene adaptive optimization based on ambient light intensity shift with a closed-loop fine-tuning mechanism based on real-time illumination feedback. This solves the technical problem in existing spectral management schemes that use fixed divisions and preset modes, resulting in an inability to accurately respond to dynamic environments and user needs. It significantly improves the spectral management accuracy, environmental adaptability, and intelligence level of full-spectrum LED light source systems.

[0070] Reference Figure 2 The second embodiment of the present invention provides a dynamic spectral optimization system for a full-spectrum LED light source system, comprising: The data processing module is used to acquire initial light intensity distribution data and perform data preprocessing to obtain the basic dataset; The spectral grouping module is used to extract features and group the basic dataset to obtain a spectral grouping scheme. The boundary adjustment module is used to obtain the light intensity distribution offset and perform scene optimization and boundary adjustment on the spectral grouping scheme according to the light intensity distribution offset to obtain the labeling data to be optimized. The grouping configuration module is used to perform scene optimization on the labeled data to be optimized, and obtain a scene-optimized spectral grouping configuration; The real-time fine-tuning module is used to collect illumination feedback data of the scene-optimized spectral grouping configuration in real time, and to fine-tune the scene-optimized spectral grouping configuration based on the illumination feedback data to obtain wavelength range adjustment parameters. The spectral output module is used to adjust parameters according to the wavelength range, perform environmental response adjustment and parameter mapping, and obtain the final spectral management output.

[0071] It should be noted that the dynamic spectral optimization system for a full-spectrum LED light source system provided in this embodiment of the invention is used to execute all the process steps of the dynamic spectral optimization method for a full-spectrum LED light source system in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0072] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a dynamic spectral optimization program for a full-spectrum LED light source system. When the processor executes the computer program, it implements the steps described in the embodiments of the dynamic spectral optimization method for a full-spectrum LED light source system, for example... Figure 1The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above-described device embodiments, such as the data processing module.

[0073] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0074] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0075] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0076] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0077] Wherein, if the modules / units integrated in the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0078] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0079] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method of dynamic spectral optimization of a full spectrum LED light source system, characterized in that, The method comprises the following steps: acquiring initial light intensity distribution data and performing data preprocessing to obtain a basic data set; performing feature extraction and grouping on the basic data set to obtain a spectral grouping scheme; acquiring a light intensity distribution offset, performing scene optimization and boundary adjustment on the spectral grouping scheme according to the light intensity distribution offset to obtain to-be-optimized marked data; performing scene optimization on the to-be-optimized marked data to obtain a scene-optimized spectral grouping configuration; real-time acquisition of light feedback data of the scene-optimized spectral grouping configuration, fine-tuning of the scene-optimized spectral grouping configuration according to the light feedback data, and obtaining of a wavelength range adjustment parameter; environmental response adjustment and parameter mapping according to the wavelength range adjustment parameter, and obtaining of a final spectral management output.

2. A dynamic spectral optimization method of a full-spectrum LED light source system according to claim 1, characterized in that, The method comprises the following steps: real-time acquisition of initial light intensity distribution data; color temperature correction and denoising processing on the initial light intensity distribution data to obtain a basic data set.

3. A dynamic spectral optimization method of a full-spectrum LED light source system according to claim 1, characterized in that, The method comprises the following steps: feature extraction on the basic data set, and spectral grouping boundary extraction to obtain a first grouping scheme; interference correction and drift optimization on the first grouping scheme to obtain a spectral grouping scheme.

4. The dynamic spectral optimization method of a full-spectrum LED light source system according to claim 1, wherein, The method comprises the following steps: acquiring a user scene data set; real-time acquisition of light intensity distribution data of the spectral grouping scheme, difference calculation between the initial light intensity distribution data and the light intensity distribution data, and obtaining of a light intensity distribution offset; scene matching according to the user scene data set to obtain a light intensity distribution offset threshold value of a current scene; abnormal marking if the light intensity distribution offset exceeds the light intensity distribution offset threshold value to obtain to-be-optimized marked data.

5. The dynamic spectral optimization method of a full-spectrum LED light source system according to claim 1, wherein, The method comprises the following steps: scene optimization target matching on the to-be-optimized marked data to obtain a scheme optimization target; light intensity distribution optimization on the spectral grouping scheme according to the scheme optimization target to obtain a first optimization configuration; interference analysis and spectral grouping boundary fine-tuning on the first optimization configuration to obtain a scene-optimized spectral grouping configuration.

6. A dynamic spectral optimization method of a full-spectrum LED light source system according to claim 1, characterized in that, The method comprises the following steps: real-time acquisition of light feedback data of the scene-optimized spectral grouping configuration, and fluctuation analysis to obtain color temperature fluctuation feature data; scene comparison on the color temperature fluctuation feature data to obtain a scene color temperature deviation value; gradient descent optimization on the scene-optimized spectral grouping configuration according to the scene color temperature deviation value, and parameter conversion to obtain a wavelength range adjustment parameter.

7. A dynamic spectral optimization method of a full-spectrum LED light source system according to claim 6, characterized in that, The method comprises the following steps: The light feedback data is subjected to frequency component analysis to obtain spectral grouping data; The spectral grouping data is subjected to clustering analysis and quantitative analysis to obtain a preliminary distribution of output intensity; The preliminary distribution of output intensity is subjected to data mapping, and combined with the wavelength range adjustment parameter, to obtain a final spectral management output.

8. A dynamic spectral optimization system for a full spectrum LED light source system, characterized by, Comprise: A data processing module for obtaining initial light intensity distribution data and performing data preprocessing to obtain a basic data set; A spectral grouping module for performing feature extraction and grouping on the basic data set to obtain a spectral grouping scheme; A boundary adjustment module for obtaining a light intensity distribution offset and performing scene optimization and boundary adjustment on the spectral grouping scheme according to the light intensity distribution offset to obtain to-be-optimized marked data; A grouping configuration module for performing scene optimization on the to-be-optimized marked data to obtain a scene-optimized spectral grouping configuration; A real-time fine-tuning module for real-time collection of light feedback data of the scene-optimized spectral grouping configuration and fine-tuning of the scene-optimized spectral grouping configuration according to the light feedback data to obtain a wavelength range adjustment parameter; A spectral output module for performing environment response adjustment and parameter mapping according to the wavelength range adjustment parameter to obtain a final spectral management output.

Images