A method and system for LED light emitting diode light color detection analysis

Abstract

The application discloses a kind of LED light-emitting diode light color detection analysis method and system, it is related to LED detection technical field, including: light color acquisition module, for obtaining the spectral data generated in the lighting process of target LED;Timing modeling module, for in the continuous preset time period after target LED starts to emit light, multiple spectral data points are collected at fixed time intervals, main wavelength, color coordinates and fitting color temperature value are calculated to obtain, form multiple time light color parameter sequence, and construct time series model based on the sequence, output time series model and corresponding original parameter sequence.The application realizes the automatic judgment of the stable state of LED light emission by constructing the change curve of light intensity and color coordinates and introducing convergence trend analysis mechanism, effectively avoids the false determination caused by early data instability, and improves the representativeness and accuracy of detection data.

Description

Technical Field

[0001] This invention relates to the field of LED testing technology, specifically to a method and system for detecting and analyzing the color of LED light-emitting diodes. Background Technology

[0002] As a high-efficiency, long-life light-emitting device, the consistency and stability of LED color are key factors in quality control. With the widespread application of LEDs in lighting, displays, automotive lighting, and other fields, different application scenarios have placed higher demands on the accuracy and representativeness of parameters such as dominant wavelength, CIE color coordinates, and fitted color temperature values. Therefore, establishing standardized and quantifiable color testing processes in LED production and delivery has become an important direction for the industry.

[0003] A search revealed a Chinese patent (Publication No.: CN112200200B) that discloses a method for detecting the color of LED lights. This patent includes a fixed image acquisition device, a designated LED light acquisition position, and LEDs of various colors placed at the acquisition position. The image acquisition device acquires color digital images of the LED lights, which are then processed by a processor to calculate the range of color feature parameters for each color, forming a database of color feature parameter ranges. The method involves placing the LED to be detected at the acquisition position and turning it on. The image acquisition device then acquires a color digital image of the LED to be detected, which is processed by the processor to calculate the color feature parameter values. These values ​​are then matched with the color feature parameter range database to determine the color of the LED to be detected.

[0004] In practical applications, due to the thermal response process in the initial stage of LED lighting, its luminous intensity and color coordinates often fluctuate within a short period of time. Furthermore, LEDs exhibit spatial differences in light distribution angles, which can lead to inaccurate color difference determination when sampling from a single angle. Traditional static sampling methods typically do not consider the temporal evolution trend of light and color parameters and lack a spatial fusion mechanism for acquiring information from multiple angles, potentially making it difficult to cover LED devices with rapidly changing luminous characteristics or asymmetrical structures. Therefore, this invention proposes a method and system for detecting and analyzing the light and color of LED light-emitting diodes. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for detecting and analyzing the color of LED light-emitting diodes, so as to solve the problems mentioned in the background art.

[0006] This invention can be achieved through the following technical solution: a light color detection and analysis system for LED light-emitting diodes, the system comprising: a light color acquisition module, a time series modeling module, a light color recognition module, and a processing module;

[0007] The light and color acquisition module is used to acquire the spectral data generated by the target LED during the lighting process. The spectral data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value, and light intensity value.

[0008] The time-series modeling module is used to collect multiple spectral data points at fixed time intervals within a continuous preset time period after the target LED starts to emit light, and calculate the dominant wavelength, color coordinates and fitted color temperature value based on each spectral data point to form a multi-time light color parameter sequence, that is, a light color dataset arranged in time.

[0009] The time series modeling module constructs a time series model based on the above light and color parameter sequence to describe the trend of light and color parameters such as dominant wavelength, color coordinates, and color temperature changing over time, and outputs the time series model and the corresponding original parameter sequence.

[0010] The light and color recognition module is used to receive the time series model and light and color parameter sequence output by the time series modeling module;

[0011] Based on preset judgment criteria, identify whether the LED has entered a stable light-emitting state;

[0012] The processing module is used to receive the light color data of the target LED after the light color recognition module confirms the stable state, and extract the average value of the light color data of all sampling points within the stable state, and project it into the established multidimensional light color feature model as the input data for classification judgment.

[0013] The multidimensional light color feature model is a training sample set with labeled light color levels. Each sample contains four indicators: dominant wavelength, color coordinates (x, y), fitted color temperature value, and dynamic change trend direction quantity, which constitute a four-dimensional feature vector. Among them, the dynamic change trend direction quantity is represented by +1, 0, and -1 to indicate rising, stable, or falling, respectively.

[0014] Each vector in the sample library is bound to a light and color level label, which is used for similarity matching and light and color level classification of input vectors;

[0015] The processing module classifies and judges the light and color data using the minimum distance matching algorithm or the K-nearest neighbor method, and outputs the light and color level and color difference evaluation index corresponding to the light and color data.

[0016] A further technical improvement of the present invention is that the method for the light and color acquisition module to acquire spectral data includes:

[0017] A1. The system provides a constant driving current to the target LED, causing it to enter the light-emitting state, and obtains the light intensity of the LED in real time through the light intensity acquisition channel to obtain the light intensity value.

[0018] In this embodiment, the constant driving current is automatically set based on the LED's rated operating current or selected by looking up a table to ensure that the light emission process conforms to typical light emission characteristics.

[0019] If the light intensity values ​​collected at three consecutive sampling times are all higher than the set light intensity threshold, and the rate of change of light intensity is less than the preset threshold, the system confirms that the LED has entered the sampleable state and proceeds to step A2.

[0020] A2. Within each sampling time, the spectral acquisition module continuously acquires no less than three frames of spectral data, with a frame interval of no more than 5ms;

[0021] Furthermore, the spectral acquisition module compares three consecutive frames of spectral data to extract the dominant wavelength and CIE chromaticity coordinates (x,y);

[0022] If the difference in the dominant wavelength between frames is less than 1 nm and the difference in the chromatic coordinates is less than 0.003, the spectral data at that moment is considered valid.

[0023] If the frame is determined to be invalid, the sampling is delayed and retried until a valid data frame is obtained;

[0024] A3. Standardize the spectral data that is determined to be valid to obtain standard spectral data, including the following steps:

[0025] The wavelength sampling range is unified to 380nm to 780nm, with a sampling step size of 1nm, for a total of 401 band points;

[0026] The spectral intensity data is normalized to the maximum value, so that the main peak intensity value is set to 1.0;

[0027] By applying a sliding median filter, local spectral outliers are eliminated, improving data smoothness and anti-interference capabilities.

[0028] A4. Based on the standard spectral data processed in step A3, calculate the following optical color parameters:

[0029] Dominant wavelength: obtained by projecting spectral data onto isochromatic lines in the CIE chromaticity diagram;

[0030] CIE color coordinates (x, y): Two-dimensional color coordinates calculated based on the CIE1931 color space, used to locate the position of light color on the color diagram;

[0031] Fitted color temperature value: The fitted color temperature obtained based on the minimum distance fitting method between CIE color coordinates and blackbody trajectory;

[0032] The light and color acquisition module packages the above light and color parameters and their sampling timestamps into structured single-frame light and color data for use by subsequent recognition or classification modules. Specifically, the structured single-frame light and color data includes: dominant wavelength, CIE color coordinates (x,y), fitted color temperature value, sampling timestamp, sampling frame number, and LED number.

[0033] A further technical improvement of the present invention is that the step of the light color recognition module identifying whether the LED has entered a stable luminous state includes:

[0034] Z1. The light and color acquisition module acquires the light intensity value and light color parameters of the target LED at fixed time intervals, and forms a light intensity change curve and a color coordinate change curve within a preset sampling period.

[0035] Z2 and the light and color recognition module analyze multiple consecutive sampling points corresponding to the light intensity change curve and color coordinate change curve in Z1, and determine whether the following convergence conditions are met respectively:

[0036] Z21, Light Intensity Convergence: The rate of change of light intensity at multiple consecutive sampling points is less than a set threshold for the rate of change of light intensity.

[0037] Z22, Color coordinate convergence: The change range of color coordinates at multiple consecutive sampling points is less than the set change range threshold;

[0038] When both conditions Z21 and Z22 are met simultaneously, the LED is determined to have reached a stable light-emitting state.

[0039] A further technical improvement of the present invention is that: the system sets a delayed inspection time based on the time required for multiple LEDs in the same batch to reach a stable light-emitting state, including:

[0040] In the initial stage of batch testing, several LED samples were selected, and the luminous stability judgment process was performed on each LED. The time required for each LED to reach a stable state was recorded.

[0041] Based on the sample stabilization time, the minimum stabilization time Tmin and the maximum stabilization time Tmax are determined, which constitute the dynamic range of the delayed inspection time [Tmin, Tmax].

[0042] The system initially uses Tmin as the delay check time, and as the number of detected LEDs increases, the actual delay check time is dynamically adjusted from Tmin to Tmax.

[0043] When the number of tests reaches the set threshold or the delay time approaches Tmax, the current delay inspection time is fixed as the waiting time for subsequent LEDs in this batch.

[0044] A further technical improvement of the present invention is that: when the maximum value Tmax and the minimum value Tmin in the dynamic range interval [Tmin, Tmax] change, the system dynamically adjusts the delay check time, including:

[0045] K1. The system obtains a set of stable times from sample LEDs and extracts the original minimum stable time. and maximum settling time Construct the initial delay time interval ;

[0046] K2. When detecting the k-th LED, the system calculates the delay check time using the following formula. ;

[0047] ;

[0048] K3. As more LEDs complete stability testing, the system continuously updates Tmin and Tmax, denoted as follows: , And calculate its relative percentage change:

[0049] ;

[0050] ;

[0051] K4. The system checks the current delay time based on the magnitudes of Δmin and Δmax. Make minor adjustments, and adjust the strategy as follows:

[0052] ;

[0053] α and β are adjustable coefficients (e.g., 0.3 to 0.5) used to control the correction ratio.

[0054] A further technical improvement of the present invention is that the light color acquisition module further includes multiple optical acquisition channels, which are respectively set at multiple angular positions in the light emission direction of the target LED;

[0055] At each sampling moment, the system synchronously acquires spectral data from different angles through the multiple channels, forming a multi-angle light and color data set;

[0056] Multi-angle light and color data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value and light intensity value in multiple directions, with each direction corresponding to a sampling channel and number;

[0057] To eliminate color shift caused by different light distribution angles of LEDs, the system assigns weights to the collected values ​​from each direction based on a preset light distribution model or empirical distribution, and adopts the following weighted fusion strategy:

[0058] For any light color parameter P, the fused value at sampling time ti is:

[0059] , where j represents the direction number, As weight, Let J be the light color parameter of the j-th direction at time i;

[0060] Among them, weight It can be set according to the target LED light distribution curve, angle response function, or empirical statistical model to make the fusion value closer to the subjective color perception of the human eye;

[0061] The fused light and color parameters will serve as representative input data for the current sampling moment, and will be used by subsequent time series modeling, steady-state judgment, and light and color level evaluation modules.

[0062] A further technical improvement of the present invention is that the processing module classifies and judges light and color data, including:

[0063] M1. When the target LED reaches a stable state, the system receives all valid sampling point data of its stable state from the light color recognition module, extracts the light color parameters of each frame, and calculates the average value of all valid frames as the light color feature vector of the target LED.

[0064] M2, the preset multidimensional light and color feature model consists of m training samples, each sample vector is: j=1,2,...,m; each sample vector corresponds to a label. This indicates its color temperature rating;

[0065] M3. The light color level is calculated using minimum Euclidean distance matching or K-nearest neighbor classification.

[0066] pass Calculate color difference evaluation index .

[0067] A method for detecting and analyzing the light color of LED light-emitting diodes, the method comprising the following steps:

[0068] Step 1: Acquisition of light and color data:

[0069] Acquire the spectral data generated by the target LED during the lighting process. The spectral data includes the dominant wavelength value, CIE color coordinates (x, y), fitted color temperature value, and light intensity value.

[0070] Step 2: Temporal Sampling and Modeling

[0071] During a continuous preset time period after the target LED starts emitting light, multiple spectral data points are collected at fixed time intervals, and the dominant wavelength, color coordinates and fitted color temperature value are calculated based on each spectral data point to form a sequence of light and color parameters arranged in time.

[0072] A time series model is constructed based on the light and color parameter sequence to describe the changing trends of the dominant wavelength, color coordinates and color temperature parameters over time.

[0073] Step 3: Stability assessment:

[0074] Based on the constructed time series model and parameter sequence, it is determined whether the target LED has reached a stable light emission state. The determination criteria include preset light color change amplitude and trend threshold judgment standards.

[0075] Step 4: Extraction of average light color vector:

[0076] After the target LED is identified as having reached a stable state, the light color parameters of all valid sampling points within the stable time period are extracted, their average value is calculated, and a stable light color feature vector is formed, which serves as the input for subsequent classification judgment.

[0077] Step 5: Judging the color grade and evaluating the color difference:

[0078] The average light color feature vector is projected onto a pre-constructed multi-dimensional light color feature model, and the light color level corresponding to the target LED is determined by the least Euclidean distance matching algorithm or the K-nearest neighbor classification method, and the color difference evaluation index between it and the reference sample is output.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] This invention constructs a curve showing the change between light intensity and color coordinates and introduces a convergence trend analysis mechanism to automatically determine the stable state of LED light emission, effectively avoiding erroneous judgments caused by unstable early data and improving the representativeness and accuracy of the detection data.

[0081] Furthermore, by recording the time required for multiple LEDs in the same batch to reach a stable state, this invention constructs a dynamic delay inspection time adjustment mechanism and gradually corrects the waiting time according to the number of test samples, effectively reducing waiting redundancy time, while improving the system's automatic adjustment capability and enhancing the intelligence level of batch inspection.

[0082] On the other hand, this invention introduces a multi-angle sampling and weighted fusion mechanism, which combines the dominant wavelength, color coordinates, color temperature and their changing trends to construct a four-dimensional light color feature vector. The vector is then classified and judged using KNN or minimum distance algorithms, which can effectively address local color differences caused by light distribution angles and improve the reliability and consistency of light color level evaluation results. Attached Figure Description

[0083] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0084] Figure 1 This is a system block diagram of the present invention;

[0085] Figure 2 This is a functional schematic diagram of the light and color acquisition module in this invention;

[0086] Figure 3 This is a functional diagram of the timing modeling module in this invention;

[0087] Figure 4 This is a functional diagram of the light and color recognition module in this invention;

[0088] Figure 5 This is a functional diagram of the processing module in this invention. Detailed Implementation

[0089] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0090] Example 1

[0091] Please see Figure 1-5 As shown, the present invention provides a light color detection and analysis system for LED light-emitting diodes, the system comprising: a light color acquisition module, a time series modeling module, a light color recognition module, and a processing module;

[0092] The light and color acquisition module is used to acquire the spectral data generated by the target LED during the lighting process. The spectral data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value, and light intensity value.

[0093] Furthermore, in this embodiment, the light and color acquisition module includes at least one optical acquisition channel and a spectral detection device connected thereto. The wavelength detection range of the spectral detection device is 380nm to 780nm, and the sampling resolution is no greater than 1nm.

[0094] Methods for acquiring spectral data using a light and color acquisition module include:

[0095] A1. The system provides a constant driving current to the target LED, causing it to enter the light-emitting state, and obtains the light intensity of the LED in real time through the light intensity acquisition channel to obtain the light intensity value.

[0096] In this embodiment, the constant driving current is automatically set based on the LED's rated operating current or selected by looking up a table to ensure that the light emission process conforms to typical light emission characteristics.

[0097] The light intensity acquisition channel is a dedicated signal acquisition path for acquiring the light intensity output by the target LED during the light emission process. It includes: optical receiving components, photoelectric sensors, and signal amplification and conversion circuits. It is used to convert the total luminous flux or local illuminance emitted by the target LED into a calculable electrical signal, which is then used as the basis for judging the light emission state in subsequent processing.

[0098] If the light intensity values ​​collected at three consecutive sampling times are all higher than the set light intensity threshold (e.g., 80% of the rated light intensity) and the light intensity change rate is less than the preset threshold (e.g., ±5%), the system confirms that the LED has entered the sampleable state and proceeds to step A2.

[0099] A2. Within each sampling time, the spectral acquisition module continuously acquires no less than three frames of spectral data, with a frame interval of no more than 5ms;

[0100] Furthermore, the spectral acquisition module compares three consecutive frames of spectral data to extract the dominant wavelength and CIE chromaticity coordinates (x,y);

[0101] If the difference in the dominant wavelength between frames is less than 1 nm and the difference in the chromatic coordinates is less than 0.003, the spectral data at that moment is considered valid.

[0102] If the frame is determined to be invalid, the sampling is delayed and retried (i.e., the current sample is discarded, and a complete three-frame sampling process is repeated after a delay period) until a valid data frame is obtained.

[0103] A3. Standardize the spectral data that is determined to be valid to obtain standard spectral data, including the following steps:

[0104] The wavelength sampling range is unified to 380nm to 780nm, with a sampling step size of 1nm, for a total of 401 band points;

[0105] The spectral intensity data is normalized to the maximum value, so that the main peak intensity value is set to 1.0;

[0106] By applying a sliding median filter, local spectral outliers are eliminated, improving data smoothness and anti-interference capabilities.

[0107] A4. Based on the standard spectral data processed in step A3, calculate the following optical color parameters:

[0108] Dominant wavelength: obtained by projecting spectral data onto isochromatic lines in the CIE chromaticity diagram;

[0109] CIE color coordinates (x, y): Two-dimensional color coordinates calculated based on the CIE1931 color space, used to locate the position of light color on the color diagram;

[0110] Fitted color temperature value: The fitted color temperature obtained based on the minimum distance fitting method between CIE color coordinates and blackbody trajectory;

[0111] The light and color acquisition module packages the above light and color parameters and their sampling timestamps into structured single-frame light and color data for use by subsequent recognition or classification modules. Specifically, the structured single-frame light and color data includes: dominant wavelength, CIE color coordinates (x,y), fitted color temperature value, sampling timestamp, sampling frame number, and LED number.

[0112] The temporal modeling module is used to collect multiple spectral data points at fixed time intervals within a continuous preset time period after the target LED starts to emit light, and calculate the dominant wavelength, color coordinates and fitted color temperature value based on each spectral data point to form a multi-time light and color parameter sequence, that is, a light and color dataset arranged in time.

[0113] In this embodiment, the preset time is 1s-10s, the sampling interval is 0.1s-1s, and the number of sampling points is no less than five.

[0114] The time series modeling module constructs a time series model based on the above light and color parameter sequence to describe the trend of light and color parameters such as dominant wavelength, color coordinates, and color temperature changing over time, and outputs the time series model and the corresponding original parameter sequence.

[0115] The light and color recognition module is used to receive the time series model and light and color parameter sequence output by the time series modeling module;

[0116] Based on preset judgment criteria, identify whether the LED has entered a stable light-emitting state;

[0117] The steps for the light color recognition module to identify whether an LED has entered a stable luminous state include:

[0118] Z1. Collection and real-time monitoring of light intensity / color coordinate variation curves:

[0119] The light and color acquisition module starts working and collects the light intensity and light color parameters of the target LED in real time;

[0120] Within a preset sampling period, multiple light intensity values ​​and light color parameters are continuously acquired at fixed time intervals to form light intensity change curves and color coordinate change curves, reflecting the changing trends of light output and light color parameters over time during LED light emission.

[0121] Specifically, the formation of the light intensity variation curve includes:

[0122] Within a preset sampling period T after the target LED is lit, the system continuously samples at fixed time intervals Δt, and obtains the light intensity value Ii of the current light output in each sampling.

[0123] Arrange all sampling points in chronological order to form a two-dimensional data set: {(ti, Ii)}, where, In the formula, This represents the sampling time point corresponding to the i-th sampling point; Indicates the sampling start time (i.e., the time of the first sampling point);

[0124] i represents the index of the sampling point, which is usually counted starting from 0, i.e., i = 0, 1, ..., n-1, where n is the number of sampling points;

[0125] This dataset uses "time ti" as the horizontal axis and "light intensity value Ii" as the vertical axis in a coordinate system. Connecting each point forms a continuous light intensity variation curve, reflecting the trend of LED light intensity change over time during the lighting process.

[0126] The formation of the color coordinate variation curve includes:

[0127] Synchronized with the light intensity acquisition process, each sampling also records the CIE color coordinates (xi, yi) of the target LED at that moment. The data sets of these CIE color coordinates changing over time are represented as {(ti, xi)} and {(ti, yi)}, respectively. With time ti as the horizontal axis and the x-value or y-value as the vertical axis, the following can be formed:

[0128] x-axis variation curve (time-color coordinate x);

[0129] y-axis variation curve (time-color coordinate y);

[0130] These two curves together reflect the changing trend of color output during the LED lighting process;

[0131] Z2. Convergence Trend Analysis and Stability Determination:

[0132] The light and color recognition module analyzes the light intensity and color coordinate data of several consecutive sampling points, and determines whether the variation range meets the stability requirements. Specifically, this includes:

[0133] Light intensity convergence: The rate of change of light intensity at multiple consecutive sampling points is less than a set threshold for the rate of change of light intensity;

[0134] Color coordinate convergence: The variation range of color coordinates at multiple consecutive sampling points is less than a set variation range threshold;

[0135] When both of the above convergence conditions are met simultaneously, the light color recognition module determines that the target LED has reached a stable light-emitting state and allows the system to proceed.

[0136] In this embodiment, the judgment criteria include all of the following conditions:

[0137] (1) The absolute value of the change in the main wavelength at three consecutive sampling points is less than 1 nm;

[0138] (2) The variation range of the corresponding color coordinates (x, y) is less than 0.005;

[0139] (3) The absolute value of the change in the fitted color temperature value is no greater than 50K;

[0140] When all the above judgment conditions are met, the system determines that the LED has reached a stable light-emitting state and allows subsequent light color judgment to be performed;

[0141] The processing module is used to receive the light color data of the target LED after the light color recognition module confirms the stable state, and extract the average value of the light color data of all sampling points within the stable state, which is then projected into the established multidimensional light color feature model as the input data for classification judgment.

[0142] The multidimensional light color feature model is a training sample set with labeled light color levels. Each sample contains four indicators: dominant wavelength, color coordinates (x, y), fitted color temperature value, and dynamic change trend direction quantity, which constitute a four-dimensional feature vector. Among them, the dynamic change trend direction quantity is represented by +1, 0, and -1 to indicate rising, stable, or falling, respectively.

[0143] Each vector in the sample library is bound to a light and color level label, which is used for similarity matching and light and color level classification of input vectors;

[0144] The processing module classifies and judges the light and color data using the minimum distance matching algorithm, and outputs the light and color level and color difference evaluation index corresponding to the light and color data;

[0145] The processing module uses methods to classify and judge light and color data, including:

[0146] M1. When the target LED reaches a stable state, the system receives all valid sampling point data of its stable state from the light color recognition module, extracts the light color parameters of each frame (dominant wavelength, CIE color coordinates (x,y), fitted color temperature value, color temperature drift direction), and calculates the average value of all valid frames as the light color feature vector of the target LED.

[0147] After the target color temperature value of the LED reaches a stable state, a total of n valid time points of light color parameter data are collected. Let the four light color parameters at the i-th time point be: , i=1,2,...,n;

[0148] Then the light and color feature vector in the stable state (i.e., the target vector) The vector is defined as the average value of these n sampling points:

[0149] ;

[0150] In the formula, ;

[0151] ;

[0152] ;

[0153] ;

[0154] M2, the preset multidimensional light and color feature model consists of m training samples, each sample vector is: j=1,2,...,m; each sample vector corresponds to a label. This indicates its color temperature rating;

[0155] M3. The minimum Euclidean distance matching method is used for determination:

[0156] Calculate the Euclidean distance between the target vector and each sample:

[0157] ;

[0158] Take smallest Its corresponding tags This refers to the color temperature rating of the target LED.

[0159] Meanwhile, color difference evaluation index Defined as: ;

[0160] The smaller the value, the closer the target LED light color is to a certain reference level sample, indicating that the light color is stable and acceptable;

[0161] like If the color difference exceeds the preset threshold, it is marked as an abnormal sample or it is tested again.

[0162] Example 2

[0163] A system for detecting and analyzing the light color of LED light-emitting diodes, the system comprising: a light color acquisition module, a time-series modeling module, a light color recognition module, and a processing module;

[0164] The light and color acquisition module is used to acquire the spectral data generated by the target LED during the lighting process. The spectral data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value, and light intensity value.

[0165] The temporal modeling module is used to collect multiple spectral data points at fixed time intervals within a continuous preset time period after the target LED starts to emit light, and calculate the dominant wavelength, color coordinates and fitted color temperature value based on each spectral data point to form a multi-time light and color parameter sequence, that is, a light and color dataset arranged in time.

[0166] The time series modeling module constructs a time series model based on the above light and color parameter sequence to describe the trend of light and color parameters such as dominant wavelength, color coordinates, and color temperature changing over time, and outputs the time series model and the corresponding original parameter sequence.

[0167] The light and color recognition module is used to receive the time series model and light and color parameter sequence output by the time series modeling module;

[0168] Based on preset judgment criteria, identify whether the LED has entered a stable light-emitting state;

[0169] The processing module is used to receive the light color data of the target LED after the light color recognition module confirms the stable state, and extract the average value of the light color data of all sampling points within the stable state, which is then projected into the established multidimensional light color feature model as the input data for classification judgment.

[0170] The processing module classifies and judges the light and color data using the minimum distance matching algorithm or the K-nearest neighbor method, and outputs the light and color level and color difference evaluation index corresponding to the light and color data.

[0171] Compared to Example 1, the system in Example 2 sets a delayed inspection time based on the time required for multiple LEDs in the same batch to reach a stable light-emitting state, including:

[0172] In the initial stage of batch testing, several LED samples were selected, and the luminous stability judgment process was performed on each LED. The time required for each LED to reach a stable state was recorded.

[0173] Based on the sample stabilization time, the minimum stabilization time Tmin and the maximum stabilization time Tmax are determined, which constitute the dynamic range of the delayed inspection time [Tmin, Tmax].

[0174] The system initially uses Tmin as the delay check time, and as the number of detected LEDs increases, the actual delay check time is dynamically adjusted from Tmin to Tmax.

[0175] When the number of tests reaches the set threshold or the delay time approaches Tmax, the current delay inspection time is fixed as the waiting time for subsequent LEDs in this batch.

[0176] Specifically, it includes the following steps:

[0177] B1. Obtaining the sample stabilization time:

[0178] In the initial stage of LED testing, several target LED samples are selected from the current testing batch, and a luminous stability judgment process is performed on each sample. The time required for each LED to reach a stable state is recorded, forming a set of stability time data.

[0179] ;

[0180] B2. Calculation of the stable time range:

[0181] The system extracts the maximum value Tmax and the minimum value Tmin from the above set and constructs the dynamic range interval [Tmin, Tmax] of the delay check time;

[0182] B3. Delayed check time initialization:

[0183] In the initial stage, the system sets the delay check time DCT=Tmin as the initial delay waiting time for the detection of the first few LEDs;

[0184] B4. Dynamic approximation adjustment of delay time:

[0185] As the number of tests increases, the system dynamically approximates the Delayed Inspection Time (DCT) from Tmin to Tmax according to a set strategy, for example, by using... Perform the calculation; where, Let N be the delay check time for the current k-th LED, k be the number of LEDs that have been checked, and N be the upper limit of the number of LEDs that can be adjusted.

[0186] When the maximum value Tmax and the minimum value Tmin in the dynamic range interval [Tmin, Tmax] change, the system dynamically adjusts the delay check time, including:

[0187] K1. The system obtains a set of stable times from sample LEDs and extracts the original minimum stable time. and maximum settling time Construct the initial delay time interval ;

[0188] K2. When detecting the k-th LED, the system calculates the delay check time using the following formula. ;

[0189] ;

[0190] K3. As more LEDs complete stability testing, the system continuously updates Tmin and Tmax, denoted as follows: , And calculate its relative percentage change:

[0191] ;

[0192] ;

[0193] K4. The system checks the current delay time based on the magnitudes of Δmin and Δmax. Make minor adjustments, and adjust the strategy as follows:

[0194] ;

[0195] α and β are adjustable coefficients (e.g., 0.3 to 0.5) used to control the correction ratio;

[0196] Furthermore, in this embodiment, when the cumulative number of detected samples reaches a set proportion (e.g., 30% of the total number of samples), the system will... , Alternative , As a new interval reference, to further improve adaptability.

[0197] Example 3

[0198] A system for detecting and analyzing the light color of LED light-emitting diodes, the system comprising: a light color acquisition module, a time-series modeling module, a light color recognition module, and a processing module;

[0199] The light and color acquisition module is used to acquire the spectral data generated by the target LED during the lighting process. The spectral data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value, and light intensity value.

[0200] The temporal modeling module is used to collect multiple spectral data points at fixed time intervals within a continuous preset time period after the target LED starts to emit light, and calculate the dominant wavelength, color coordinates and fitted color temperature value based on each spectral data point to form a multi-time light and color parameter sequence, that is, a light and color dataset arranged in time.

[0201] The time series modeling module constructs a time series model based on the above light and color parameter sequence to describe the trend of light and color parameters such as dominant wavelength, color coordinates, and color temperature changing over time, and outputs the time series model and the corresponding original parameter sequence.

[0202] The light and color recognition module is used to receive the time series model and light and color parameter sequence output by the time series modeling module;

[0203] Based on preset judgment criteria, identify whether the LED has entered a stable light-emitting state;

[0204] The processing module is used to receive the light color data of the target LED after the light color recognition module confirms the stable state, and extract the average value of the light color data of all sampling points within the stable state, which is then projected into the established multidimensional light color feature model as the input data for classification judgment.

[0205] The processing module classifies and judges the light and color data using the minimum distance matching algorithm or the K-nearest neighbor method, and outputs the light and color level and color difference evaluation index corresponding to the light and color data.

[0206] Compared to Embodiments 1 and 2, the light and color acquisition module in Embodiment 3 further includes multiple optical acquisition channels, which are respectively set at multiple angular positions in the light emission direction of the target LED;

[0207] At each sampling moment, the system synchronously acquires spectral data from different angles through multiple channels, forming a multi-angle light and color data set;

[0208] Multi-angle light and color data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value and light intensity value in multiple directions, with each direction corresponding to a sampling channel and number;

[0209] To eliminate color shift caused by different light distribution angles of LEDs, the system assigns weights to the collected values ​​from each direction based on a preset light distribution model or empirical distribution, and adopts the following weighted fusion strategy:

[0210] For any light color parameter P, the fused value at sampling time ti is:

[0211] , where j represents the direction number, As weight, Let J be the light color parameter of the j-th direction at time i;

[0212] Among them, weight It can be set according to the target LED light distribution curve, angle response function, or empirical statistical model to make the fusion value closer to the subjective color perception of the human eye;

[0213] The fused light and color parameters will serve as representative input data for the current sampling moment, and will be used by subsequent time series modeling, steady-state judgment, and light and color level evaluation modules.

[0214] The processing module uses methods to classify and judge light and color data, including:

[0215] M1. When the target LED reaches a stable state, the system receives all valid sampling point data of its stable state from the light color recognition module, extracts the light color parameters of each frame (dominant wavelength, CIE color coordinates (x,y), fitted color temperature value, color temperature drift direction), and calculates the average value of all valid frames as the light color feature vector of the target LED.

[0216] After the target color temperature value of the LED reaches a stable state, a total of n valid time points of light color parameter data are collected. Let the four light color parameters at the i-th time point be: , i=1,2,...,n;

[0217] Then the light and color feature vector in the stable state (i.e., the target vector) The vector is defined as the average value of these n sampling points:

[0218] ;

[0219] In the formula, ;

[0220] ;

[0221] ;

[0222] ;

[0223] M2, the preset multidimensional light and color feature model consists of m training samples, each sample vector is: j=1,2,...,m; each sample vector corresponds to a label. This indicates its color temperature rating;

[0224] M3. K-Nearest Neighbor (KNN) classification is used for determination:

[0225] Select to make The smallest set of the top k samples is denoted as set. The final label will then be determined by voting.

[0226] In the formula, Light color level;

[0227] Meanwhile, color difference evaluation index Defined as: ;

[0228] The smaller the value, the closer the target LED light color is to a certain reference level sample, indicating that the light color is stable and acceptable;

[0229] like If the color difference exceeds the preset threshold, it is marked as an abnormal sample or it is tested again.

[0230] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0231] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A system for detecting and analyzing the color of light emitted by LED light-emitting diodes, characterized in that, include: The light and color acquisition module is used to acquire the spectral data generated by the target LED during the lighting process; The time series modeling module is used to collect multiple spectral data points at fixed time intervals within a continuous preset time period after the target LED starts emitting light, calculate the dominant wavelength, color coordinates and fitted color temperature values, form a multi-time light and color parameter sequence, and build a time series model based on the sequence, outputting the time series model and the corresponding original parameter sequence. The light color recognition module is used to receive the time series model and light color parameter sequence output by the time series modeling module, and identify whether the LED has entered a stable light-emitting state according to a preset judgment standard. The processing module is used to receive the light color data of the target LED after the light color recognition module confirms the stable state, extract the average value of the light color data of all sampling points within the stable state, project it as the input data for classification judgment into the established multi-dimensional light color feature model, classify and judge the light color data, and output the light color level and color difference evaluation index corresponding to the light color data. The system sets a delayed inspection time based on the time required for multiple LEDs in the same batch to reach a stable lighting state, including: In the initial stage of batch testing, several LED samples were selected, and the luminous stability judgment process was performed on each LED. The time required for each LED to reach a stable state was recorded. Based on the sample stabilization time, the minimum stabilization time Tmin and the maximum stabilization time Tmax are determined, which constitute the dynamic range of the delayed inspection time [Tmin, Tmax]. The system initially uses Tmin as the delay check time, and as the number of detected LEDs increases, the actual delay check time is dynamically adjusted from Tmin to Tmax. When the number of tests reaches the set threshold or the delay time approaches Tmax, the current delay inspection time is fixed as the waiting time for subsequent LEDs in this batch.

2. The LED light-emitting diode light color detection and analysis system according to claim 1, characterized in that, The method for the light and color acquisition module to acquire spectral data includes: A1. The system provides a constant driving current to the target LED, causing it to enter the light-emitting state, and obtains the light intensity of the LED in real time through the light intensity acquisition channel to obtain the light intensity value. If the light intensity values ​​collected at three consecutive sampling times are all higher than the set light intensity threshold, and the rate of change of light intensity is less than the preset threshold, then the LED is confirmed to be in the sampleable state and proceeds to step A2. A2. At each sampling moment, the spectral acquisition module continuously acquires no less than three frames of spectral data; Furthermore, the spectral acquisition module compares three consecutive frames of spectral data to extract the dominant wavelength and CIE chromaticity coordinates (x,y); If the difference in the dominant wavelength between frames is less than 1 nm and the difference in the chromatic coordinates is less than 0.003, the spectral data at that moment is considered valid. If the frame is determined to be invalid, the sampling is delayed and retried until a valid data frame is obtained; A3. Standardize the spectral data that is determined to be valid to obtain standard spectral data; A4. Based on the standard spectral data processed in step A3, calculate its color parameters, including: dominant wavelength, CIE color coordinates (x,y) and fitted color temperature value. The light and color acquisition module packages the above light and color parameters and their sampling timestamps into structured single-frame light and color data for use by subsequent recognition or classification modules.

3. The LED light-emitting diode light color detection and analysis system according to claim 1, characterized in that, The step of the light color recognition module identifying whether the LED has entered a stable luminous state includes: Z1. The light and color acquisition module acquires the light intensity value and light color parameters of the target LED at fixed time intervals, and forms a light intensity change curve and a color coordinate change curve within a preset sampling period. Z2 and the light and color recognition module analyze multiple consecutive sampling points corresponding to the light intensity change curve and color coordinate change curve in Z1, and determine whether the following convergence conditions are met respectively: Z21, Light Intensity Convergence: The rate of change of light intensity at multiple consecutive sampling points is less than a set threshold for the rate of change of light intensity. Z22, Color coordinate convergence: The change range of color coordinates at multiple consecutive sampling points is less than the set change range threshold; When both conditions Z21 and Z22 are met simultaneously, the LED is determined to have reached a stable light-emitting state.

4. The LED light-emitting diode light color detection and analysis system according to claim 1, characterized in that, When the maximum value Tmax and the minimum value Tmin in the dynamic range interval [Tmin, Tmax] change, the system dynamically adjusts the delay check time, including: K1. The system obtains a set of stable times through sample LEDs, extracts the original minimum stable time Tmin0 and the original maximum stable time Tmax0, and constructs the initial delay time interval [Tmin0, Tmax0]. K2. During the detection of the k-th LED, the system calculates the Delayed Check Time (DCT) using the following formula. k ; DCT k =Tmin+(Tmax-Tmin)×(k / N); N is the maximum number of adjustments allowed; K3. As more LEDs complete stability testing, the system continuously updates Tmin and Tmax, denoted as Tmin. n Tmax n And calculate its relative percentage change: Δmin=(Tmin n -Tmin0) / Tmin0; Δmax=(Tmax n -Tmax0) / Tmax0; K4. The system uses the magnitudes of Δmin and Δmax to determine the current delay check time (DCT). k Adjustments were made to obtain the corrected Delayed Check Time (DCT). k ′: DCT k ′=DCT k ×[1+α·Δmin+β·Δmax]; α and β are adjustable coefficients, ranging from 0.3 to 0.5, used to control the correction ratio.

5. The LED light-emitting diode light color detection and analysis system according to claim 1, characterized in that, The light and color acquisition module further includes multiple optical acquisition channels, which are respectively set at multiple angular positions in the light emission direction of the target LED; At each sampling moment, the system synchronously acquires spectral data from different angles through the multiple channels, forming a multi-angle light and color data set; Multi-angle light and color data includes the dominant wavelength value, CIE color coordinates (x,y), fitted color temperature value and light intensity value in multiple directions, with each direction corresponding to a sampling channel and number; The system assigns weights to the collected values ​​from each direction and performs weighted fusion of the light and color parameters from each direction; The fused light and color parameters will serve as representative input data for the current sampling moment, and will be used by subsequent time series modeling, steady-state judgment, and light and color level evaluation modules.

6. The LED light-emitting diode light color detection and analysis system according to claim 1, characterized in that, The processing module uses methods to classify and judge light and color data, including: M1. When the target LED reaches a stable state, the system receives all valid sampling point data of its stable state from the light color recognition module, extracts the light color parameters of each frame, and calculates the average value of all valid frames as the light color feature vector of the target LED. M2, the preset multidimensional light and color feature model consists of m training samples, each sample vector is: j=1,2,...,m; each sample vector corresponds to a label. This indicates its color temperature rating; M3. The light color level is calculated using minimum Euclidean distance matching or K-nearest neighbor classification. pass Calculate color difference evaluation index .

7. A method for detecting and analyzing the color of light emitted by LED light-emitting diodes, characterized in that, The method employs the analysis system of any one of claims 1-6, comprising: Step 1: Collect spectral data of the target LED during the lighting process, including dominant wavelength, CIE color coordinates (x, y), fitted color temperature value, and light intensity value; Step 2: Collect multiple spectral data points at fixed time intervals within a continuous preset time period to form a sequence of light and color parameters, and construct a time series model to describe the changing trend of light and color parameters over time; Step 3: Based on the time series model and parameter series, determine whether the LED has reached a stable light-emitting state according to the preset light color change amplitude and trend threshold; Step 4: After identifying a stable state, extract all valid light and color parameters within that time period and calculate the average value as the stable light and color feature vector; Step 5: Project the average light color vector into the multidimensional light color feature model, determine the light color level through minimum distance matching or K-nearest neighbor classification algorithm, and output the color difference evaluation index.

Images