A multi-wavelength spectrum tunable method and system for LED light source

Abstract

The present invention relates to the technical field of light source adjustment. The present invention discloses a multi-wavelength spectral tunable method and system for LED light sources. The method includes first processing real-time spectral data according to an adaptive filtering algorithm to obtain precise spectral characteristics, then performing independent component analysis on the precise spectral characteristics to obtain a single spectral characteristic corresponding to each light source, determining a contribution coefficient corresponding to the single spectral characteristic by a least squares method, and finally comparing different contribution coefficients according to a minimum value method to determine an overlap integral, and adjusting the real-time spectral data according to the overlap integral. The present invention processes the real-time spectral data through an adaptive filtering algorithm, thereby improving the accuracy and stability of the spectral data, avoiding interference caused by spectral overlap, adjusting the real-time spectral data according to the overlap integral, optimizing the spectral output, ensuring more accurate color mixing and uniform lighting effects, and improving the adaptability and accuracy of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of light source adjustment, and more particularly to a multi-wavelength spectrum adjustable method and system for an LED light source. Background Art

[0002] With the rapid development of modern lighting technology, LEDs (light-emitting diodes) have been widely used in various lighting scenarios due to their high efficiency, low energy consumption, long life, and environmental friendliness. However, traditional LED lighting systems typically use LED modules with fixed wavelengths or a limited combination of wavelengths. Their fixed spectral output has poor adaptability to different environmental conditions and cannot dynamically adjust the light color according to the actual usage environment.

[0003] For example, the patent application with publication number CN106793326A provides a high-color rendering multi-spectrum LED light source, exhibition lighting for museum exhibits, and exhibition and lighting methods. This technical solution includes a variety of different LED light source chips, each of which has different spectral wavelengths and spectral bandwidths. Through spectral independent driving technology and multi-spectral light source testing and automatic control systems, it can achieve precise control of the light intensity of different bands in the lighting source to meet the needs of different cultural relics.

[0004] In existing technologies, although some high-end LED systems achieve a certain degree of color temperature and color variation by adjusting the driving current of LEDs of different colors, these systems often rely on a preset spectral characteristic database and lack real-time environmental perception and adaptive adjustment capabilities. This reliance on a preset database not only limits the flexibility of spectral output, but may also cause spectral overlap and interference areas in multi-wavelength LED combinations, thereby affecting the stability and uniformity of the overall spectrum. In addition, spectral overlap will lead to inaccurate color mixing, reducing the accuracy of the lighting effect.

[0005] In view of this, the present invention proposes a multi-wavelength spectrum tunable method and system for LED light sources to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides a multi-wavelength spectrum tunable method and system for an LED light source.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] In a first aspect, a multi-wavelength spectrum tunable method for an LED light source is provided, comprising:

[0009] Acquire real-time spectral data and process it according to the adaptive filtering algorithm to obtain accurate spectral features;

[0010] Perform independent component analysis on the precise spectral features to obtain a single spectral feature corresponding to each light source, and determine the contribution coefficient corresponding to the single spectral feature using the least squares method. The contribution coefficient represents the weight of each light source in the current spectral output;

[0011] Within a preset wavelength range, different contribution coefficients are compared according to a minimum method to determine an overlap integral, and the real-time spectral data is adjusted according to the overlap integral, where the overlap integral characterizes the degree of overlap of different spectra within the preset wavelength range.

[0012] In some embodiments, a method for processing real-time spectral data according to an adaptive filtering algorithm to obtain accurate spectral features includes:

[0013] Acquire real-time environmental data, process the real-time spectral data according to the adaptive filtering algorithm to obtain initial spectral features, determine the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data, correct the initial spectral features according to the filter estimation parameters, and obtain accurate spectral features.

[0014] In some embodiments, a method for determining a filter estimation parameter in an adaptive filtering algorithm based on real-time environmental data includes:

[0015] The real-time environmental data is input into the pre-trained neural network model to obtain the filter estimation parameters, which include signal power spectral density and noise power spectral density.

[0016] In some embodiments, the method of correcting the initial spectral features according to the filter estimation parameters to obtain accurate spectral features includes:

[0017] The frequency response coefficient is calculated according to the filter estimation parameters, and the frequency response coefficient is subjected to inverse Fourier transform to obtain the impulse response function. The initial spectral features are convolved according to the impulse response function to obtain the accurate spectral features.

[0018] In some embodiments, the method of calculating the frequency response coefficient according to the filter estimation parameter includes:

[0019] The sum of the signal power spectrum density and the noise power spectrum density is calculated, and the ratio of the signal power spectrum density to the sum of the signal power spectrum density and the noise power spectrum density is calculated, and the ratio is used as the frequency response coefficient.

[0020] In some embodiments, a method of performing independent component analysis on the precise spectral signature to obtain a single spectral signature corresponding to each light source includes:

[0021] The precise spectral features are centralized and decomposed using the ICA algorithm to obtain a single spectral feature corresponding to each light source.

[0022] In some embodiments, the method of determining the contribution coefficient corresponding to a single spectral feature by the least squares method includes:

[0023] A spectral feature model is constructed based on the precise spectral features and the single spectral features corresponding to each light source. A first feature matrix is ​​constructed based on the single spectral features, and a second eigenvector is constructed based on the precise spectral features. A linear equation system is constructed according to the first feature matrix and the second eigenvector. The linear equation system is solved by the least squares method to obtain the contribution coefficient corresponding to the single spectral feature.

[0024] In some embodiments, a method for comparing different contribution coefficients according to a minimum method to determine an overlap integral includes:

[0025] Determine each wavelength point within a preset wavelength range, traverse each wavelength point, determine the minimum contribution coefficient of all light sources at the wavelength point by the minimum value method, traverse the wavelength range, integrate the minimum contribution coefficient of each wavelength point, and obtain the overlapping integral value within the wavelength range.

[0026] In some embodiments, the method of determining the minimum contribution coefficient of all light sources at the wavelength point by the minimum value method includes:

[0027] For each wavelength point, the contribution coefficients of all light sources at that wavelength point are compared, the minimum value is selected, and the corresponding minimum contribution coefficient is determined.

[0028] In a second aspect, a multi-wavelength spectrum tunable system for an LED light source is provided, which is used to implement the above-mentioned multi-wavelength spectrum tunable method for an LED light source, including:

[0029] Filtering module: used to obtain real-time spectral data, process the real-time spectral data according to the adaptive filtering algorithm to obtain accurate spectral characteristics;

[0030] Data processing module: used to perform independent component analysis on the precise spectral characteristics to obtain the single spectral characteristics corresponding to each light source, and determine the contribution coefficient corresponding to the single spectral characteristics through the least squares method. The contribution coefficient represents the weight of each light source in the current spectral output;

[0031] Overlap analysis module: used to compare different contribution coefficients within a preset wavelength range according to the minimum method, determine the overlap integral, and adjust the real-time spectral data according to the overlap integral. The overlap integral represents the degree of overlap of different spectra within the preset wavelength range.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention first processes real-time spectral data according to an adaptive filtering algorithm to obtain precise spectral features, then performs independent component analysis on the precise spectral features to obtain a single spectral feature corresponding to each light source, determines the contribution coefficient corresponding to the single spectral feature by the least squares method, and finally compares different contribution coefficients according to the minimum method to determine the overlap integral, and adjusts the real-time spectral data according to the overlap integral. The present invention processes real-time spectral data through an adaptive filtering algorithm, thereby improving the accuracy and stability of the spectral data and avoiding interference caused by spectral overlap. The single spectral feature of each light source is extracted through independent component analysis, and the contribution coefficient is determined by the least squares method, which solves the problem of poor spectral output flexibility in traditional systems and realizes dynamic adjustment. Finally, the real-time spectral data is adjusted according to the overlap integral, the spectral output is optimized, more accurate color mixing and uniform lighting effects are ensured, and the adaptability and accuracy of the system are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Schematic diagram of a process for a multi-wavelength spectrum tunable method for an LED light source in the present invention;

[0035] Figure 2 This is a schematic structural diagram of a multi-wavelength spectrum tunable system for an LED light source in the present invention;

[0036] Figure 3 Schematic diagram of the process of the method for processing real-time spectral data according to the adaptive filtering algorithm to obtain accurate spectral features in the present invention;

[0037] Figure 4 The figure is a flow chart of the method for correcting the initial spectral characteristics according to the filter estimation parameters to obtain the accurate spectral characteristics in the present invention. DETAILED DESCRIPTION

[0038] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings. In the following detailed description, many specific details are set forth to provide a thorough understanding of the described exemplary embodiments. However, it is obvious to those skilled in the art that the described embodiments can be practiced without some or all of these specific details. In other exemplary embodiments, well-known structures are not described in detail to avoid unnecessarily obscuring the concepts of the present disclosure. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. At the same time, the various aspects described in the embodiments can be arbitrarily combined without conflict.

[0039] Example 1

[0040] See also Figure 1As shown, this embodiment discloses a multi-wavelength spectrum tunable method for an LED light source, comprising:

[0041] S10: Acquire real-time spectral data, process the real-time spectral data according to an adaptive filtering algorithm, and obtain accurate spectral features;

[0042] In this embodiment, the real-time spectral data includes at least a spectral intensity value, a spectral wavelength value, and a spectral change rate value. The spectral change rate value refers to the rate at which the spectral intensity changes with the wavelength. Specifically, the spectral change rate value quantifies the changing trend of the spectral intensity with the wavelength, and represents the slope or curvature of the spectral curve.

[0043] like Figure 3 As shown, the method of processing real-time spectral data according to the adaptive filtering algorithm to obtain accurate spectral features includes:

[0044] Acquire real-time environmental data, process the real-time spectral data according to the adaptive filtering algorithm to obtain initial spectral features, determine the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data, correct the initial spectral features according to the filter estimation parameters, and obtain accurate spectral features.

[0045] It can be understood that the initial spectral characteristics include but are not limited to spectral peak position, spectral peak intensity, spectral width value and spectral contrast. The spectral peak position refers to the position of the maximum absorption or emission wavelength in the spectrum, the spectral peak intensity refers to the light intensity at a wavelength point in the spectrum, the spectral width value represents the light intensity distribution within a specific wavelength range, and the spectral contrast represents the degree of difference between the spectral signal and the background or noise, specifically measuring the significance or clarity of the signal.

[0046] It should be added that the adaptive filtering algorithm is an algorithm that can dynamically adjust its filtering parameters according to the characteristics of the input signal. Unlike the traditional fixed filtering method, the adaptive filtering algorithm can adjust the filter parameters in real time when the signal changes or the noise environment changes, thereby optimizing the signal extraction effect. In this embodiment, the purpose of processing real-time spectral data according to the adaptive filtering algorithm is to optimize the extraction of spectral features, reduce noise interference, and improve signal quality, thereby improving the accuracy and reliability of subsequent spectral analysis.

[0047] The method for determining the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data includes:

[0048] The real-time environmental data is input into the pre-trained neural network model to obtain the filter estimation parameters, which include signal power spectral density and noise power spectral density.

[0049] It can be understood that the signal power spectral density represents the power distribution of the signal at different frequency components, and the noise power spectral density represents the power distribution of the noise signal at different frequencies. The power spectral density is usually obtained by converting the time domain signal into the frequency domain signal through Fourier transform. It can show the energy distribution of the signal at each frequency component. The training method of the neural network model can be based on the fully connected neural network as the basic model. The input layer of the fully connected neural network receives historical environmental data, and the output layer of the fully connected neural network outputs historical filtering estimation parameters. When training the fully connected neural network, the cross entropy loss function is selected as the loss function, and the loss function is minimized by the gradient descent method. The weight parameters of the fully connected neural network are updated, and the neural network model is obtained through iterative training.

[0050] It should be noted that real-time environmental data includes but is not limited to real-time environmental temperature and real-time environmental humidity. Taking real-time environmental temperature and real-time environmental humidity as an example, as the temperature rises, the electronic noise of the sensor will generally increase. This is because the increase in temperature increases the thermal motion of the sensor elements, thereby increasing the noise level. Therefore, in a high temperature environment, the filter estimation parameters should be adjusted higher. Increased humidity will cause the electrical characteristics of the sensor to change, thereby affecting the signal quality and noise level. When the humidity is too high, water vapor will adhere to the sensor surface to form a thin film, which will cause additional interference to the signal. In a high humidity environment, the filter estimation parameters need to be adjusted higher. It should be added that in the adaptive filtering algorithm, the filter estimation parameters (such as signal power spectral density and noise power spectral density) are adjusted according to real-time environmental data in order to optimize the signal extraction effect, reduce noise interference and improve signal quality.

[0051] like Figure 4 As shown, the method of correcting the initial spectral features according to the filter estimation parameters to obtain accurate spectral features includes:

[0052] The frequency response coefficient is calculated according to the filter estimation parameters, and the frequency response coefficient is subjected to inverse Fourier transform to obtain the impulse response function. The initial spectral features are convolved according to the impulse response function to obtain the accurate spectral features.

[0053] Methods for calculating frequency response coefficients based on filter estimation parameters include:

[0054] The sum of the signal power spectrum density and the noise power spectrum density is calculated, and the ratio of the signal power spectrum density to the sum of the signal power spectrum density and the noise power spectrum density is calculated, and the ratio is used as the frequency response coefficient.

[0055] The specific methods for calculating the frequency response coefficient include:

[0056]

[0057] Where H(ω) is the frequency response coefficient, S xx (ω) is the signal power spectrum density, S nn (ω) is the noise power spectral density, where ω represents the frequency.

[0058] In this embodiment, the frequency response coefficient represents the response of the filter to signals of different frequencies. Performing an inverse Fourier transform on the frequency response coefficient refers to converting the frequency response coefficient in the frequency domain back into an impulse response function in the time domain. The impulse response function represents the time domain signal. The impulse response function describes the response of the filter to a unit pulse signal and is one of the most basic characteristics of the filter. The transformation between the frequency response coefficient and the impulse response function is an existing technology and will not be elaborated on in this embodiment.

[0059] Methods for convolving the initial spectral features according to the impulse response function to obtain accurate spectral features include:

[0060]

[0061] Where RSF(t0) represents the precise spectral feature, ISF(t) represents the initial spectral feature, H(·) represents the impulse response function, and τ is the time delay variable.

[0062] It should be added that the time delay variable refers to the time variable used to shift the impulse response function in the convolution operation. Specifically, it represents the time offset of the impulse response function relative to the initial signal when the signal is convolved. In this embodiment, the filter estimation parameters are adjusted according to the real-time environmental data. This process helps the system better respond to environmental changes and optimizes the signal extraction process. Through these dynamic adjustments, the system can more accurately extract spectral features, reduce noise interference, and ensure the accuracy of subsequent analysis.

[0063] S20: Performing independent component analysis on the precise spectral features to obtain a single spectral feature corresponding to each light source, and determining a contribution coefficient corresponding to the single spectral feature by a least squares method, wherein the contribution coefficient represents the weight of each light source in the current spectral output;

[0064] Methods for performing independent component analysis on precise spectral features to obtain a single spectral feature corresponding to each light source include:

[0065] The precise spectral features are centralized and decomposed using the ICA algorithm to obtain a single spectral feature corresponding to each light source.

[0066] In this embodiment, the precise spectral features may be centralized by calculating the mean of the precise spectral features and subtracting it from each data point, thereby ensuring that the mean is zero before performing independent component analysis (ICA), which helps the ICA algorithm to more effectively extract the single spectral features of each light source. The ICA algorithm is a technique used in signal processing and statistics, which aims to separate independent components from multiple signal sources. The light sources mentioned above may be LED lights of different wavelengths, each LED light source emits light in a different wavelength range, and the single spectral features of these light sources can be separated by the ICA algorithm.

[0067] Methods for determining the contribution coefficient corresponding to a single spectral feature by the least squares method include:

[0068] A spectral feature model is constructed based on the precise spectral features and the single spectral features corresponding to each light source. A first feature matrix is ​​constructed based on the single spectral features, and a second eigenvector is constructed based on the precise spectral features. A linear equation system is constructed according to the first feature matrix and the second eigenvector. The linear equation system is solved by the least squares method to obtain the contribution coefficient corresponding to the single spectral feature.

[0069] Methods for constructing a spectral feature model based on precise spectral features and a single spectral feature corresponding to each light source include:

[0070]

[0071] Where Y(λ) represents the precise spectral characteristics, X j (λ) is represented by the single spectral feature corresponding to the j-th light source, a j It is represented as the contribution coefficient corresponding to the j-th light source, M is the total number of light sources, and ∈(λ) is represented as the error term.

[0072] It should be noted that the error term refers to the difference between the weighted sum of each single spectral feature and the exact spectral feature, which is caused by multiple factors. This difference usually reflects the fitting error of the model, including the influence of measurement noise, model incompleteness and other factors.

[0073] It should be noted that the first feature matrix is ​​formed by combining the single spectral features corresponding to each light source to form a matrix representing the spectrum of the light source. The second feature vector is formed by combining the precise spectral features to form a vector. The spectral peak intensity in the precise spectral features is used as an example to illustrate:

[0074]

[0075] Among them, X(λ) represents the first characteristic matrix.

[0076] The method of constructing a linear equation system based on the first characteristic matrix and the second characteristic vector includes:

[0077] a=[X(λ) T X(λ)] -1 X(λ) T Y(λ);

[0078] Among them, a represents the contribution coefficient, X(λ) T is the transposed matrix of the first characteristic matrix, [X(λ) T X(λ)] -1 is X(λ) T The inverse matrix of X(λ).

[0079] It can be understood that the least squares method is used to solve a set of linear equations in order to find the optimal solution to a set of linear equations, thereby minimizing the error term. Specifically, the least squares method solves the coefficients of a set of unknown variables by minimizing the objective function, ensuring that the difference between the weighted sum of each single spectral feature and the exact spectral feature is minimized, thereby achieving the best fit. For example, the a obtained by the solution is shown below:

[0080]

[0081] Among them, a1 corresponds to X1(λ), ​​and a2 corresponds to X2(λ).

[0082] This embodiment first uses an adaptive filtering algorithm to adjust filtering parameters, optimize spectral data processing, and reduce the impact of environmental noise, thereby obtaining a more accurate initial spectral signature. This precise spectral signature is then centralized and the ICA algorithm is used to extract the single spectral signature of each light source. This effectively separates the contribution of each light source, eliminating interference between different light sources. By combining the least squares method to further solve for the contribution coefficient, the weight of each light source in the current spectral output can be accurately calculated, further optimizing the regulation and control of the light sources.

[0083] S30: Within a preset wavelength range, different contribution coefficients are compared according to a minimum method to determine an overlap integral, and the real-time spectral data is adjusted according to the overlap integral, wherein the overlap integral represents the degree of overlap of different spectra within the preset wavelength range;

[0084] In this embodiment, the wavelength range refers to a specific wavelength interval selected during the spectral data analysis process. The spectral data will be analyzed and processed within this interval to determine the contribution coefficient of the light source and its weight in different spectral outputs. The purpose of presetting the wavelength range is to reduce the computational complexity by selecting an appropriate wavelength interval in actual operation. Limiting the range of analysis wavelengths helps to reduce the amount of calculation, thereby improving analysis efficiency.

[0085] Methods for determining overlap integrals by comparing different contribution coefficients using the minimum method include:

[0086] Determine each wavelength point within a preset wavelength range, traverse each wavelength point, determine the minimum contribution coefficient of all light sources at the wavelength point by the minimum value method, traverse the wavelength range, integrate the minimum contribution coefficient of each wavelength point, and obtain the overlapping integral value within the wavelength range.

[0087] In this embodiment, the minimum method refers to calculating the minimum value of the contribution coefficients of different light sources at each wavelength point. This means that for each wavelength point, the contribution coefficients of all light sources at that wavelength are selected, and the minimum value among them is taken to represent the degree of overlap of different light sources at that wavelength. Then, the entire preset wavelength range is traversed, and this process is repeated. The minimum contribution coefficient of each wavelength point is integrated to obtain the overlap integral within the wavelength range.

[0088] Methods for determining the minimum contribution coefficient of all light sources at this wavelength point by the minimum value method include:

[0089] For each wavelength point, the contribution coefficients of all light sources at that wavelength point are compared, the minimum value is selected, and the corresponding minimum contribution coefficient is determined.

[0090] It should be noted that the overlap integral represents the degree of overlap between the spectral outputs of different light sources within a preset wavelength range. If the overlap integral is high (i.e., the degree of overlap is large), the contribution coefficients of certain light sources need to be corrected or adjusted to reduce unnecessary overlap and optimize the output effect of the light source. The adjustment method can be weighted adjustment, filtering, or optimizing the overall spectral data by adjusting the output intensity of the light source. For example, the output intensity of different light sources can be adjusted to optimize the contribution of different light sources within a specific wavelength range. If the contribution coefficient of a certain light source is high, resulting in an increase in the overlap integral, the contribution of the light source can be reduced by adjusting the output intensity of the light source, thereby reducing overlap and making the contribution of each light source more balanced.

[0091] This embodiment first processes the real-time spectral data according to an adaptive filtering algorithm to obtain precise spectral features, then performs independent component analysis on the precise spectral features to obtain a single spectral feature corresponding to each light source, determines the contribution coefficient corresponding to the single spectral feature by the least squares method, and finally compares different contribution coefficients according to the minimum method to determine the overlap integral, and adjusts the real-time spectral data according to the overlap integral. This embodiment processes the real-time spectral data through an adaptive filtering algorithm, thereby improving the accuracy and stability of the spectral data and avoiding interference caused by spectral overlap. The single spectral feature of each light source is extracted through independent component analysis, and the contribution coefficient is determined using the least squares method, which solves the problem of poor spectral output flexibility in traditional systems and realizes dynamic adjustment. Finally, the real-time spectral data is adjusted according to the overlap integral, the spectral output is optimized, more accurate color mixing and uniform lighting effects are ensured, and the adaptability and accuracy of the system are improved.

[0092] Example 2

[0093] See also Figure 2 As shown, based on the same inventive concept, this embodiment discloses a multi-wavelength spectrum tunable system for an LED light source. For details not provided in this embodiment, please refer to the description of the relevant parts in Example 1. The system includes:

[0094] Filtering module: used to obtain real-time spectral data, process the real-time spectral data according to the adaptive filtering algorithm to obtain accurate spectral characteristics;

[0095] In this embodiment, the real-time spectral data includes at least a spectral intensity value, a spectral wavelength value, and a spectral change rate value. The spectral change rate value refers to the rate at which the spectral intensity changes with the wavelength. Specifically, the spectral change rate value quantifies the changing trend of the spectral intensity with the wavelength, and represents the slope or curvature of the spectral curve.

[0096] Methods for processing real-time spectral data using adaptive filtering algorithms to obtain accurate spectral features include:

[0097] Acquire real-time environmental data, process the real-time spectral data according to the adaptive filtering algorithm to obtain initial spectral features, determine the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data, correct the initial spectral features according to the filter estimation parameters, and obtain accurate spectral features.

[0098] The method for determining the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data includes:

[0099] The real-time environmental data is input into the pre-trained neural network model to obtain the filter estimation parameters, which include signal power spectral density and noise power spectral density.

[0100] Methods for correcting the initial spectral features based on the filter estimation parameters to obtain accurate spectral features include:

[0101] The frequency response coefficient is calculated according to the filter estimation parameters, and the frequency response coefficient is subjected to inverse Fourier transform to obtain the impulse response function. The initial spectral features are convolved according to the impulse response function to obtain the accurate spectral features.

[0102] Data processing module: used to perform independent component analysis on the precise spectral characteristics to obtain the single spectral characteristics corresponding to each light source, and determine the contribution coefficient corresponding to the single spectral characteristics through the least squares method. The contribution coefficient represents the weight of each light source in the current spectral output;

[0103] Methods for performing independent component analysis on precise spectral features to obtain a single spectral feature corresponding to each light source include:

[0104] The precise spectral features are centralized and decomposed using the ICA algorithm to obtain a single spectral feature corresponding to each light source.

[0105] Methods for determining the contribution coefficient corresponding to a single spectral feature by the least squares method include:

[0106] A spectral feature model is constructed based on the precise spectral features and the single spectral features corresponding to each light source. A first feature matrix is ​​constructed based on the single spectral features, and a second eigenvector is constructed based on the precise spectral features. A linear equation system is constructed according to the first feature matrix and the second eigenvector. The linear equation system is solved by the least squares method to obtain the contribution coefficient corresponding to the single spectral feature.

[0107] Methods for constructing a spectral feature model based on precise spectral features and a single spectral feature corresponding to each light source include:

[0108]

[0109] Where Y(λ) represents the precise spectral characteristics, X j (λ) is represented by the single spectral feature corresponding to the j-th light source, a j It is represented as the contribution coefficient corresponding to the j-th light source, M is the total number of light sources, and ∈(λ) is represented as the error term.

[0110] Overlap analysis module: used to compare different contribution coefficients within a preset wavelength range according to the minimum method, determine the overlap integral, and adjust the real-time spectral data based on the overlap integral. The overlap integral represents the degree of overlap of different spectra within the preset wavelength range;

[0111] In this embodiment, the wavelength range refers to a specific wavelength interval selected during the spectral data analysis process. The spectral data will be analyzed and processed within this interval to determine the contribution coefficient of the light source and its weight in different spectral outputs. The purpose of presetting the wavelength range is to reduce the computational complexity by selecting an appropriate wavelength interval in actual operation. Limiting the range of analysis wavelengths helps to reduce the amount of calculation, thereby improving analysis efficiency.

[0112] Methods for determining overlap integrals by comparing different contribution coefficients using the minimum method include:

[0113] Determine each wavelength point within a preset wavelength range, traverse each wavelength point, determine the minimum contribution coefficient of all light sources at the wavelength point by the minimum value method, traverse the wavelength range, integrate the minimum contribution coefficient of each wavelength point, and obtain the overlapping integral value within the wavelength range.

[0114] Methods for determining the minimum contribution coefficient of all light sources at this wavelength point by the minimum value method include:

[0115] For each wavelength point, the contribution coefficients of all light sources at that wavelength point are compared, the minimum value is selected, and the corresponding minimum contribution coefficient is determined.

[0116] The detailed description set forth above in conjunction with the accompanying drawings describes examples and does not represent all examples that can be implemented or fall within the scope of the claims. The terms "example" and "exemplary" when used in this specification mean "used as an example, instance or illustration" and do not mean "better than or better than other examples."

[0117] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Therefore, use of these phrases may refer to more than just one embodiment, and further, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0118] It should also be noted that these embodiments may be described as a process depicted as a flowchart, structure diagram, or block diagram, and that although the flowchart may describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently, and the order of the operations may be rearranged.

Claims

1. A multi-wavelength spectrum tunable method for LED light source, characterized in that: include: Acquire real-time spectral data and process it according to the adaptive filtering algorithm to obtain accurate spectral features; Perform independent component analysis on the precise spectral features to obtain a single spectral feature corresponding to each light source, and determine the contribution coefficient corresponding to the single spectral feature using the least squares method. The contribution coefficient represents the weight of each light source in the current spectral output; Within a preset wavelength range, different contribution coefficients are compared according to the minimum method to determine the overlap integral, and the real-time spectral data is adjusted according to the overlap integral, wherein the overlap integral represents the degree of overlap of different spectra within the preset wavelength range; The method of comparing different contribution coefficients according to the minimum value method to determine the overlap integral includes: Determine each wavelength point within a preset wavelength range, traverse each wavelength point, determine the minimum contribution coefficient of all light sources at the wavelength point by the minimum value method, traverse the wavelength range, integrate the minimum contribution coefficient of each wavelength point, and obtain the overlapping integral value within the wavelength range.

2. A multi-wavelength spectrum tunable method for LED light source according to claim 1, characterized in that: The method of processing real-time spectral data according to the adaptive filtering algorithm to obtain accurate spectral features includes: Acquire real-time environmental data, process the real-time spectral data according to the adaptive filtering algorithm to obtain initial spectral features, determine the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data, correct the initial spectral features according to the filter estimation parameters, and obtain accurate spectral features.

3. The multi-wavelength spectrum tunable method for LED light source according to claim 2, characterized in that: The method for determining the filter estimation parameters in the adaptive filtering algorithm according to the real-time environmental data includes: The real-time environmental data is input into the pre-trained neural network model to obtain the filter estimation parameters, which include signal power spectral density and noise power spectral density.

4. A multi-wavelength spectrum tunable method for LED light source according to claim 3, characterized in that: The method of correcting the initial spectral characteristics according to the filter estimation parameters to obtain accurate spectral characteristics includes: The frequency response coefficient is calculated according to the filter estimation parameters, and the frequency response coefficient is subjected to inverse Fourier transform to obtain the impulse response function. The initial spectral features are convolved according to the impulse response function to obtain the accurate spectral features.

5. The multi-wavelength spectrum tunable method for LED light source according to claim 4, characterized in that: The method for calculating the frequency response coefficient according to the filter estimation parameter includes: The sum of the signal power spectrum density and the noise power spectrum density is calculated, and the ratio of the signal power spectrum density to the sum of the signal power spectrum density and the noise power spectrum density is calculated, and the ratio is used as the frequency response coefficient.

6. The multi-wavelength spectrum tunable method for LED light source according to claim 4, characterized in that: The method of performing independent component analysis on the precise spectral features to obtain a single spectral feature corresponding to each light source includes: The precise spectral features are centralized and decomposed using the ICA algorithm to obtain a single spectral feature corresponding to each light source.

7. A multi-wavelength spectrum tunable method for LED light source according to claim 6, characterized in that: The method for determining the contribution coefficient corresponding to a single spectral feature by the least squares method includes: A spectral feature model is constructed based on the precise spectral features and the single spectral features corresponding to each light source. A first feature matrix is ​​constructed based on the single spectral features, and a second eigenvector is constructed based on the precise spectral features. A linear equation system is constructed according to the first feature matrix and the second eigenvector. The linear equation system is solved by the least squares method to obtain the contribution coefficient corresponding to the single spectral feature.

8. The multi-wavelength spectrum tunable method for LED light source according to claim 7, characterized in that: The method for determining the minimum contribution coefficient of all light sources at the wavelength point by the minimum value method includes: For each wavelength point, the contribution coefficients of all light sources at that wavelength point are compared, the minimum value is selected, and the corresponding minimum contribution coefficient is determined.

9. A multi-wavelength spectrum tunable system for an LED light source, which is used to implement a multi-wavelength spectrum tunable method for an LED light source according to any one of claims 1 to 8, characterized in that: include: Filtering module: used to obtain real-time spectral data, process the real-time spectral data according to the adaptive filtering algorithm to obtain accurate spectral characteristics; Data processing module: used to perform independent component analysis on the precise spectral characteristics to obtain the single spectral characteristics corresponding to each light source, and determine the contribution coefficient corresponding to the single spectral characteristics through the least squares method. The contribution coefficient represents the weight of each light source in the current spectral output; Overlap analysis module: used to compare different contribution coefficients within a preset wavelength range according to the minimum method, determine the overlap integral, and adjust the real-time spectral data according to the overlap integral. The overlap integral represents the degree of overlap of different spectra within the preset wavelength range.

Images