Fluorescent-powder-free multi-primary-color LED testing method and system

Abstract

The invention relates to the field of LED testing, in particular to a fluorescent powder-free multi-primary-color LED testing method and a fluorescent powder-free multi-primary-color LED testing system. A fluorescent-powder-free multi-primary-color LED test system comprises an LED single-primary-color test frequency acquisition module, a sampling frequency analysis module and a fluorescent-powder-free multi-primary-color LED test module. According to the method, the LED single-primary-color test frequency for testing the fluorescent-powder-free multi-primary-color LED is sent to the sampling frequency analysis model to be analyzed, so that the LED single-primary-color acquisition frequency for acquiring the electrical variables of the fluorescent-powder-free multi-primary-color LED is output, and the electrical variables of the fluorescent-powder-free multi-primary-color LED during testing are acquired on the basis of the LED single-primary-color acquisition frequency. Therefore, the test is carried out through the collected electrical variables, the collection frequency of the electrical variables does not need to be manually adjusted, the accuracy of electrical variable collection can be improved, and the accuracy of the fluorescent powder-free multi-primary-color LED test is further improved.

Description

Technical Field

[0001] The invention relates to the field of LED testing, and in particular to a method and system for testing a phosphor-free multi-primary color LED. Background Art

[0002] In LED lighting and display technology, multi-primary color LED systems (such as red, green, blue, yellow, etc.) are widely used in display, intelligent lighting, stage lighting and other fields due to their rich colors, high energy efficiency and fast response speed. Traditional LED lighting systems rely on phosphor conversion technology to produce white light, but due to the problems of phosphor light decay, heat generation and unstable color temperature, phosphor-free multi-primary color LED technology has gradually received attention in recent years. This technology directly combines multiple primary color LEDs (such as red, green, blue, etc.) to emit light, achieving efficient color output and flexible color temperature adjustment.

[0003] However, in the test process of the phosphor-free multi-primary color LED system, the electrical characteristics of different primary color LEDs vary significantly, especially when collecting electrical variables such as voltage, current and power, the setting of the collection frequency is particularly critical. Since the operating frequency, forward voltage and response speed of the primary color LEDs such as red, green, blue and yellow are different, if the collection frequency of each primary color LED is not properly set during the test, the collection frequency mismatch problem will occur.

[0004] Mismatched acquisition frequencies can lead to a series of adverse effects: First, too high or too low acquisition frequencies can cause the signal details of certain primary colors to be lost or cannot be accurately captured, resulting in data distortion or signal aliasing. Second, inconsistent acquisition frequencies may also cause synchronization problems between multiple primary colors, resulting in the inability to accurately reflect the actual luminous state when evaluating the output color or brightness. In addition, frequency mismatch may also lead to uneven allocation of system resources, increase test errors, and thus affect the accuracy and stability of test results. Summary of the invention

[0005] The present invention sends an LED single-primary-color test frequency for testing a multi-primary-color LED without phosphors into a sampling frequency analysis model for analysis, so as to output an LED single-primary-color collection frequency for collecting electrical variables for the multi-primary-color LED without phosphors, and collects electrical variables during the test of the multi-primary-color LED without phosphors based on the LED single-primary-color collection frequency, and then performs testing through the collected electrical variables, without the need to manually adjust the collection frequency of the electrical variables, and can improve the accuracy of the electrical variable collection, thereby improving the accuracy of the test of the multi-primary-color LED without phosphors.

[0006] The present invention provides a method for testing a phosphor-free multi-primary color LED, comprising: Obtain N LED single-primary color test frequencies, form the N LED single-primary color test frequencies into a sampling frequency analysis vector, then send the sampling frequency analysis vector to a sampling frequency analysis model for processing, and output N LED single-primary color acquisition frequencies; Based on the N LED single-primary-color acquisition frequencies, the electric variable of each primary-color lamp in the phosphor-free multi-primary-color LED is acquired to obtain an LED electric variable data set, and then the test operation of the phosphor-free multi-primary-color LED is performed based on the LED electric variable data set; The sampling frequency analysis model is established based on the BP neural network model, including an input layer, a feature enhancement layer, a hidden layer and an output layer, wherein the input layer is used to receive a sampling frequency analysis vector; the feature enhancement layer is used to perform feature enhancement on the sampling frequency analysis vector to construct a sampling frequency analysis enhancement vector; the hidden layer is used to perform feature learning and feature analysis on the sampling frequency analysis enhancement vector to output N LED single-primary color acquisition frequencies; the output layer is used to output N LED single-primary color acquisition frequencies; and the sampling frequency analysis model is trained by sampling frequency analysis training samples obtained through a population optimization algorithm.

[0007] As a preferred aspect of the present invention, in the feature enhancement layer, feature enhancement is performed on the sampling frequency analysis vector to construct the sampling frequency analysis enhancement vector, which specifically includes the following steps: Calculate H attention weight vectors ATT h =W h F T F, where h = 1, 2, 3, ..., H, W h is the hth attention weight matrix, F is the sampling frequency analysis vector, and T is the transposition operation; H attention weight vectors ATT h After the addition, a normalization operation is performed to obtain a comprehensive attention weight matrix, and then the sampling frequency analysis vector is multiplied by the comprehensive attention weight matrix to construct a sampling frequency analysis enhancement vector.

[0008] As a preferred aspect of the present invention, training is performed on a sampling frequency analysis model, specifically comprising the following steps: A number of sampling frequency analysis training samples are obtained, the sampling frequency analysis training samples include a sampling frequency analysis vector and its corresponding N LED single-primary color acquisition frequencies, and the sampling frequency analysis training samples are obtained by simulation calculation of a swarm optimization algorithm, all sampling frequency analysis training samples are combined into a sampling frequency analysis training set, and the sampling frequency analysis model is trained through the sampling frequency analysis training set. During the training, the sampling frequency analysis vector is used as input, and the corresponding N LED single-primary color acquisition frequencies are used as target outputs to determine whether the training conditions are met. If the training conditions are met, the trained sampling frequency analysis model is output; otherwise, the sampling frequency analysis model continues to be trained through the sampling frequency analysis training set.

[0009] As a preferred aspect of the present invention, the swarm optimization algorithm simulates and calculates to obtain the sampling frequency analysis training samples, which specifically includes the following steps: Step S1: randomly setting N LED single-primary-color test frequencies, and forming a sampling frequency analysis vector from the N LED single-primary-color test frequencies; Step S2: randomly setting a number of acquisition frequency simulation individuals, wherein the acquisition frequency simulation individuals include N LED single primary color acquisition frequencies, forming a population set with all acquisition frequency simulation individuals, and setting a maximum number of iterations; Step S3: Calculate the fitness corresponding to the individual simulated by the acquisition frequency; Step S4: based on the fitness corresponding to the individual simulated by the acquisition frequency, the population set is iteratively updated by the population optimization algorithm; Step S5: until the number of iterations reaches the maximum number of iterations, output the acquisition frequency simulation individual with the largest fitness as the N LED single-primary color acquisition frequencies corresponding to the sampling frequency analysis vector, and construct a sampling frequency analysis training sample with the sampling frequency analysis vector and the corresponding N LED single-primary color acquisition frequencies.

[0010] As a preferred aspect of the present invention, the fitness corresponding to the individual of the acquisition frequency simulation is calculated. The specific calculation method is: according to the N LED single-primary color test frequencies, the multi-primary color LED without phosphor is driven and controlled, and based on the acquisition frequency simulation individual corresponding to the N LED single-primary color acquisition frequencies to collect electrical variables within a preset time period, to obtain the experimental electrical variable R i , i=1, 2, 3, ..., I, I is the total number of experimental electrical variables, and obtain the experimental electrical variable R i The corresponding theoretical electric variable E i , calculate the acquisition accuracy , and then calculate the fitness of the individual corresponding to the acquisition frequency simulation δ=α1μ+α2I -1 , α1+α2 are the first weight coefficient and the second weight coefficient respectively, and satisfy α1+α2=1.

[0011] As a preferred aspect of the present invention, the hidden layer includes M hidden units, and the M value is obtained by simulation calculation through a swarm optimization algorithm as a hyperparameter.

[0012] As a preferred aspect of the present invention, the swarm optimization algorithm adopts the sparrow search algorithm.

[0013] The present invention also provides a phosphor-free multi-primary color LED testing system, comprising: LED single-primary-color test frequency acquisition module, used to acquire N LED single-primary-color test frequencies; The sampling frequency analysis module is used to form a sampling frequency analysis vector from N LED single-primary color test frequencies, and then send the sampling frequency analysis vector to the sampling frequency analysis model for processing, and output N LED single-primary color acquisition frequencies; The phosphor-free multi-primary color LED test module is used to collect electrical variables for each primary color lamp in the phosphor-free multi-primary color LED based on the N LED single primary color collection frequencies to obtain an LED electrical variable data set, and then perform a test operation of the phosphor-free multi-primary color LED based on the LED electrical variable data set; The sampling frequency analysis model is established based on the BP neural network model, including an input layer, a feature enhancement layer, a hidden layer and an output layer, wherein the input layer is used to receive a sampling frequency analysis vector; the feature enhancement layer is used to perform feature enhancement on the sampling frequency analysis vector to construct a sampling frequency analysis enhancement vector; the hidden layer is used to perform feature learning and feature analysis on the sampling frequency analysis enhancement vector to output N LED single-primary color acquisition frequencies; the output layer is used to output N LED single-primary color acquisition frequencies; and the sampling frequency analysis model is trained by sampling frequency analysis training samples obtained through a population optimization algorithm.

[0014] The present invention has the following advantages: The present invention sends an LED single-primary-color test frequency for testing a multi-primary-color LED without phosphors into a sampling frequency analysis model for analysis, so as to output an LED single-primary-color collection frequency for collecting electrical variables for the multi-primary-color LED without phosphors, and collects electrical variables during the test of the multi-primary-color LED without phosphors based on the LED single-primary-color collection frequency, and then performs testing through the collected electrical variables, without the need to manually adjust the collection frequency of the electrical variables, and can improve the accuracy of the electrical variable collection, thereby improving the accuracy of the test of the multi-primary-color LED without phosphors. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the structure of a phosphor-free multi-primary color LED testing system used in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0017] Embodiment 1, a method for testing a phosphor-free multi-primary color LED, comprising: Get N LED single-primary color test frequencies. When testing the non-phosphor multi-primary color LED, it is generally necessary to control the non-phosphor multi-primary color LED through PWM dimming, and then test the optical performance of the non-phosphor multi-primary color LED. The LED single-primary color test frequency is the switching frequency corresponding to each primary color lamp in the non-phosphor multi-primary color LED. Therefore, when testing the non-phosphor multi-primary color LED, the tester will set the PWM signal of each primary color lamp in the non-phosphor multi-primary color LED through the driving circuit, and then obtain the corresponding LED single-primary color test frequency. N is the total number of primary color lamps in the non-phosphor multi-primary color LED. It is generally 4, indicating four primary colors, namely red, blue, green and amber. The N LED single-primary color test frequencies are composed of a sampling frequency analysis vector, and then the sampling frequency analysis vector is sent to the sampling frequency analysis model for processing, and N LED single-primary color acquisition frequencies are output. The LED single-primary color acquisition frequency is the frequency of the corresponding electrical variable acquisition of each primary color lamp in the multi-primary color LED without phosphor. The electrical variables of each primary color lamp in the multi-primary color LED without phosphor are collected through the LED single-primary color acquisition frequency, which can take into account the effects of improving the acquisition accuracy and the reasonable use of storage space. Once the acquisition frequency is set unreasonably, it is easy to cause aliasing effect, affecting the acquisition accuracy of the electrical variables; Based on the N LED single-primary-color acquisition frequencies, the electric variables of each primary-color lamp in the phosphor-free multi-primary-color LED are collected to obtain the LED electric variable data set. The electric variables include voltage and current, etc. Then, the phosphor-free multi-primary-color LED is tested based on the LED electric variable data set. The specific test method is to compare the electric variables in the LED electric variable data set with the corresponding standard test conditions. The standard test conditions are set by the test personnel to specify whether the electric variables meet the industry or quality standards. The sampling frequency analysis model is established based on the BP neural network model, including an input layer, a feature enhancement layer, a hidden layer and an output layer, wherein the input layer is used to receive the sampling frequency analysis vector; the feature enhancement layer is used to perform feature enhancement on the sampling frequency analysis vector to construct the sampling frequency analysis enhancement vector. It should be noted that under the influence of current coupling, etc., the electrical variable acquisition between different primary color lamps will affect each other, so it is necessary to learn these influences through feature enhancement operations, and enhance the sampling frequency analysis vector through these influence relationships; the hidden layer includes M hidden units, and the M value is generally 3, which can also be used as a hyperparameter to be simulated and calculated by the group optimization algorithm. The hidden layer is used to perform feature learning and feature analysis on the sampling frequency analysis enhancement vector to output N LED single primary color acquisition frequencies; the output layer is used to output N LED single primary color acquisition frequencies; and the sampling frequency analysis model is trained by the sampling frequency analysis training samples obtained by the group optimization algorithm; The present application sends the LED single-primary color test frequency for testing the multi-primary color LED without phosphor into the sampling frequency analysis model for analysis, so as to output the LED single-primary color collection frequency for collecting electrical variables for the multi-primary color LED without phosphor, and collects the electrical variables during the test of the multi-primary color LED without phosphor based on the LED single-primary color collection frequency, and then performs testing through the collected electrical variables, without the need to manually adjust the collection frequency of the electrical variables, and can improve the accuracy of the electrical variable collection, thereby improving the accuracy of the test of the multi-primary color LED without phosphor.

[0018] In the feature enhancement layer, feature enhancement is performed on the sampling frequency analysis vector to construct a sampling frequency analysis enhancement vector, which specifically includes the following steps: Calculate H attention weight vectors ATT h =W h F T F, where h = 1, 2, 3, ..., H, W h is the hth attention weight matrix, F is the sampling frequency analysis vector, T is the transposition operation, and the correlation between different single-primary color test frequencies is calculated by inner product calculation, and then the influence of electrical variable collection between different primary color lights is analyzed; H attention weight vectors ATT h After the addition, a normalization operation is performed to obtain a comprehensive attention weight matrix, and then the sampling frequency analysis vector is multiplied by the comprehensive attention weight matrix to construct a sampling frequency analysis enhancement vector.

[0019] Training the sampling frequency analysis model includes the following steps: A number of sampling frequency analysis training samples are obtained, the sampling frequency analysis training samples include a sampling frequency analysis vector and its corresponding N LED single-primary color acquisition frequencies, and the sampling frequency analysis training samples are obtained by simulation calculation of a swarm optimization algorithm, all sampling frequency analysis training samples are combined into a sampling frequency analysis training set, and the sampling frequency analysis model is trained through the sampling frequency analysis training set. During the training, the sampling frequency analysis vector is used as input, and the corresponding N LED single-primary color acquisition frequencies are used as target outputs to determine whether the training conditions are met. The training conditions may be that they are met a certain number of times, or that the accuracy of the sampling frequency analysis model reaches a certain value. If the training conditions are met, the trained sampling frequency analysis model is output; otherwise, the sampling frequency analysis model continues to be trained through the sampling frequency analysis training set.

[0020] The swarm optimization algorithm simulates and calculates the sampling frequency analysis training samples, which specifically includes the following steps: Step S1: randomly set N LED single-primary color test frequencies, and form the N LED single-primary color test frequencies into a sampling frequency analysis vector. The developer can perform a priori setting based on the model of the non-phosphor multi-primary color LED; Step S2: Randomly set a number of acquisition frequency simulation individuals, including N LED single-primary color acquisition frequencies. It should be noted that the N LED single-primary color acquisition frequencies are random values ​​set by the developer according to the corresponding range. All acquisition frequency simulation individuals are composed of a population set, and the maximum number of iterations is set. Step S3: Calculate the fitness corresponding to the individual simulated by the acquisition frequency. The specific calculation method is: according to the N LED single-primary-color test frequencies, the multi-primary-color LED without phosphor is driven and controlled, and the multi-primary-color LED without phosphor here is a multi-primary-color LED without phosphor without problem, and based on the N LED single-primary-color acquisition frequencies corresponding to the acquisition frequency simulation individual, the electrical variable is collected within a preset time period. The preset time period is determined by the developer and is generally 2 minutes, and the experimental electrical variable R is obtained. i , i=1, 2, 3, ..., I, I is the total number of experimental electrical variables, such as the first voltage collected by the red primary color lamp, the second voltage collected by the red primary color lamp, the first current collected by the green primary color lamp, and the experimental electrical variable R i The corresponding theoretical electric variable E i The theoretical electrical variables are calculated by developers using simulation software or manually for N LED single-primary color test frequencies. The calculation accuracy is , and then calculate the fitness of the individual corresponding to the acquisition frequency simulation δ=α1μ+α2I -1, α1+α2 are the first weight coefficient and the second weight coefficient, which are set by the developer and satisfy α1+α2=1. In the process of fitness calculation, the direction of group optimization is guided to improve the collection accuracy and reduce the number of collected experimental electrical variables, improve the accuracy of electrical variable collection and use storage space reasonably; Step S4: based on the fitness corresponding to the individual simulated by the acquisition frequency, the population set is iteratively updated by the sparrow search algorithm; Step S5: until the number of iterations reaches the maximum number of iterations, output the acquisition frequency simulation individual with the largest fitness as the N LED single-primary color acquisition frequencies corresponding to the sampling frequency analysis vector, and construct a sampling frequency analysis training sample with the sampling frequency analysis vector and the corresponding N LED single-primary color acquisition frequencies; Embodiment 2, a phosphor-free multi-primary color LED testing system, such as Figure 1 As shown, including: LED single-primary color test frequency acquisition module, used to obtain N LED single-primary color test frequencies. When testing a phosphor-free multi-primary color LED, it is generally necessary to control the phosphor-free multi-primary color LED through PWM dimming, and then test the optical performance of the phosphor-free multi-primary color LED. The LED single-primary color test frequency is the switching frequency corresponding to each primary color lamp in the phosphor-free multi-primary color LED. Therefore, when testing the phosphor-free multi-primary color LED, the tester will set the PWM signal of each primary color lamp in the phosphor-free multi-primary color LED through the driving circuit, and then obtain the corresponding LED single-primary color test frequency. N is the total number of primary color lamps in the phosphor-free multi-primary color LED, which is generally 4, indicating four primary colors, namely red, blue, green and amber; The sampling frequency analysis module is used to form a sampling frequency analysis vector from N LED single-primary color test frequencies, and then send the sampling frequency analysis vector to the sampling frequency analysis model for processing, and output N LED single-primary color acquisition frequencies. The LED single-primary color acquisition frequency is the frequency of the corresponding acquisition of electrical variables for each primary color lamp in the multi-primary color LED without phosphor. The electrical variables of each primary color lamp in the multi-primary color LED without phosphor are collected through the LED single-primary color acquisition frequency, which can take into account the effects of improving the acquisition accuracy and reasonably using the storage space. Once the acquisition frequency is set unreasonably, it is easy to cause aliasing effect, affecting the acquisition accuracy of the electrical variables; The phosphor-free multi-primary color LED test module is used to collect electrical variables for each primary color lamp in the phosphor-free multi-primary color LED based on the N LED single primary color collection frequencies to obtain the LED electrical variable data set. The electrical variables include voltage and current, etc., and then the phosphor-free multi-primary color LED test operation is performed based on the LED electrical variable data set. The specific test method can be to compare the electrical variables in the LED electrical variable data set with the corresponding standard test conditions. The standard test conditions are set by the test personnel to specify whether the electrical variables meet the industry or quality standards; The sampling frequency analysis model is established based on the BP neural network model, including an input layer, a feature enhancement layer, a hidden layer and an output layer, wherein the input layer is used to receive the sampling frequency analysis vector; the feature enhancement layer is used to perform feature enhancement on the sampling frequency analysis vector to construct the sampling frequency analysis enhancement vector. It should be noted that under the influence of current coupling, etc., the electrical variable acquisition between different primary color lamps will affect each other, so it is necessary to learn these influences through feature enhancement operations, and enhance the sampling frequency analysis vector through these influence relationships; the hidden layer includes M hidden units, and the M value is generally 3, which can also be obtained as a hyperparameter through simulation and calculation of the group optimization algorithm. The hidden layer is used to perform feature learning and feature analysis on the sampling frequency analysis enhancement vector to output N LED single primary color acquisition frequencies; the output layer is used to output N LED single primary color acquisition frequencies; and the sampling frequency analysis model is trained by the sampling frequency analysis training samples obtained through the group optimization algorithm.

[0021] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A method for testing a phosphor-free multi-primary color LED, characterized in that: include: Obtain N LED single-primary color test frequencies, form the N LED single-primary color test frequencies into a sampling frequency analysis vector, then send the sampling frequency analysis vector to a sampling frequency analysis model for processing, and output N LED single-primary color acquisition frequencies; Based on the N LED single-primary-color acquisition frequencies, the electric variable of each primary-color lamp in the phosphor-free multi-primary-color LED is acquired to obtain an LED electric variable data set, and then the test operation of the phosphor-free multi-primary-color LED is performed based on the LED electric variable data set; The sampling frequency analysis model is established based on the BP neural network model, including an input layer, a feature enhancement layer, a hidden layer and an output layer, wherein the input layer is used to receive a sampling frequency analysis vector; the feature enhancement layer is used to perform feature enhancement on the sampling frequency analysis vector to construct a sampling frequency analysis enhancement vector; the hidden layer is used to perform feature learning and feature analysis on the sampling frequency analysis enhancement vector to output N LED single-primary color acquisition frequencies; the output layer is used to output N LED single-primary color acquisition frequencies; and the sampling frequency analysis model is trained by sampling frequency analysis training samples obtained through a population optimization algorithm.

2. A method for testing a phosphor-free multi-primary color LED according to claim 1, characterized in that: In the feature enhancement layer, feature enhancement is performed on the sampling frequency analysis vector to construct a sampling frequency analysis enhancement vector, which specifically includes the following steps: Calculate H attention weight vectors ATT h =W h F T F, where h = 1, 2, 3, ..., H, W h is the hth attention weight matrix, F is the sampling frequency analysis vector, and T is the transposition operation; H attention weight vectors ATT h After the addition, a normalization operation is performed to obtain a comprehensive attention weight matrix, and then the sampling frequency analysis vector is multiplied by the comprehensive attention weight matrix to construct a sampling frequency analysis enhancement vector.

3. A method for testing a phosphor-free multi-primary color LED according to claim 2, characterized in that: Training the sampling frequency analysis model includes the following steps: A number of sampling frequency analysis training samples are obtained, the sampling frequency analysis training samples include a sampling frequency analysis vector and its corresponding N LED single-primary color acquisition frequencies, and the sampling frequency analysis training samples are obtained by simulation calculation of a swarm optimization algorithm, all sampling frequency analysis training samples are combined into a sampling frequency analysis training set, and the sampling frequency analysis model is trained through the sampling frequency analysis training set. During the training, the sampling frequency analysis vector is used as input, and the corresponding N LED single-primary color acquisition frequencies are used as target outputs to determine whether the training conditions are met. If the training conditions are met, the trained sampling frequency analysis model is output; otherwise, the sampling frequency analysis model continues to be trained through the sampling frequency analysis training set.

4. A method for testing a phosphor-free multi-primary color LED according to claim 3, characterized in that: The swarm optimization algorithm simulates and calculates the sampling frequency analysis training samples, which specifically includes the following steps: Step S1: randomly setting N LED single-primary-color test frequencies, and forming a sampling frequency analysis vector from the N LED single-primary-color test frequencies; Step S2: randomly setting a number of acquisition frequency simulation individuals, wherein the acquisition frequency simulation individuals include N LED single primary color acquisition frequencies, forming a population set with all acquisition frequency simulation individuals, and setting a maximum number of iterations; Step S3: Calculate the fitness corresponding to the individual simulated by the acquisition frequency; Step S4: based on the fitness corresponding to the individual simulated by the acquisition frequency, the population set is iteratively updated by the population optimization algorithm; Step S5: until the number of iterations reaches the maximum number of iterations, output the acquisition frequency simulation individual with the largest fitness as the N LED single-primary color acquisition frequencies corresponding to the sampling frequency analysis vector, and construct a sampling frequency analysis training sample with the sampling frequency analysis vector and the corresponding N LED single-primary color acquisition frequencies.

5. A method for testing a phosphor-free multi-primary color LED according to claim 4, characterized in that: Calculate the fitness corresponding to the individual simulated by the acquisition frequency. The specific calculation method is: drive and control the multi-primary LED without phosphor according to the N LED single-primary test frequencies, and collect the electrical variables within the preset time period based on the N LED single-primary acquisition frequencies corresponding to the individual simulated by the acquisition frequency to obtain the experimental electrical variable R i , i=1, 2, 3, ..., I, I is the total number of experimental electrical variables, and obtain the experimental electrical variable R i The corresponding theoretical electric variable E i , calculate the acquisition accuracy , and then calculate the fitness of the individual corresponding to the acquisition frequency simulation δ=α1μ+α2I -1 , α1+α2 are the first weight coefficient and the second weight coefficient respectively, and satisfy α1+α2=1.

6. A method for testing a phosphor-free multi-primary color LED according to claim 5, characterized in that: The hidden layer includes M hidden units, and the M value is calculated as a hyperparameter through simulation of the swarm optimization algorithm.

7. A method for testing a phosphor-free multi-primary color LED according to claim 6, characterized in that: The swarm optimization algorithm uses the sparrow search algorithm.

8. A phosphor-free multi-primary color LED testing system, characterized in that: The system applies a phosphor-free multi-primary color LED testing method as described in any one of claims 1 to 7, including: LED single-primary-color test frequency acquisition module, used to acquire N LED single-primary-color test frequencies; The sampling frequency analysis module is used to form a sampling frequency analysis vector from N LED single-primary color test frequencies, and then send the sampling frequency analysis vector to the sampling frequency analysis model for processing, and output N LED single-primary color acquisition frequencies; The phosphor-free multi-primary color LED test module is used to collect electrical variables for each primary color lamp in the phosphor-free multi-primary color LED based on the N LED single primary color collection frequencies to obtain an LED electrical variable data set, and then perform a test operation of the phosphor-free multi-primary color LED based on the LED electrical variable data set; The sampling frequency analysis model is established based on the BP neural network model, including an input layer, a feature enhancement layer, a hidden layer and an output layer, wherein the input layer is used to receive a sampling frequency analysis vector; the feature enhancement layer is used to perform feature enhancement on the sampling frequency analysis vector to construct a sampling frequency analysis enhancement vector; the hidden layer is used to perform feature learning and feature analysis on the sampling frequency analysis enhancement vector to output N LED single-primary color acquisition frequencies; the output layer is used to output N LED single-primary color acquisition frequencies; and the sampling frequency analysis model is trained by sampling frequency analysis training samples obtained through a population optimization algorithm.

Images