LED lamp bead sorting method and system based on spectral characteristics
By extracting the spectral curve feature parameters and similarity of LED beads and combining them with clustering algorithms for grading, the problem of insufficient spectral sorting accuracy in existing technologies is solved, and high-precision spectral consistency sorting is achieved.
Patent Information
- Application Number
- CN202511959978.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-24
- Publication Date
- 2026-03-20
AI Technical Summary
Existing LED bead sorting methods lack effective characterization and quantitative comparison of spectral shape, resulting in insufficient sorting accuracy and failing to meet the spectral consistency requirements of high-end applications.
By acquiring the spectral curves of LED beads, extracting shape feature parameters and fitting parameters, calculating the similarity between the spectral curves and the reference spectrum, and using a clustering algorithm for grading, high-precision spectral feature sorting is achieved.
It significantly improves the spectral consistency of LED beads, meets the stringent requirements of high-end applications for spectral quality, and improves sorting accuracy.
Smart Images
Figure CN121696141A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of semiconductor technology, and in particular to an LED lamp bead sorting method and system based on spectral characteristics. BACKGROUND
[0002] LED lamp beads are widely used in lighting, display (such as Mini / Micro LED backlight, direct display), medical treatment, plant growth and other fields. In these application fields, the consistency of spectral distribution will affect the color difference, brightness and color rendering of the product, and ultimately affect the visual effect and quality of the product, so the consistency of spectral distribution is crucial.
[0003] The existing LED lamp bead sorting methods mainly include: 1. According to the main wavelength and luminous flux / brightness to sort: this sorting method ignores the overall shape characteristics of the spectrum, even if the main wavelength and luminous flux / brightness are the same, the slight difference of the spectral curve in a certain waveband (such as the peak shift of the blue light chip, the different red / green light ratios caused by the difference in the excitation efficiency of the fluorescent powder) will also cause obvious visual inconsistency in mixed light or display.
[0004] 2. Simple color coordinate sorting: this sorting method increases the color coordinates (such as CIEx,y or u',v') as the basis for sorting, but the color coordinates are the integral results of the spectrum, and different spectral shapes may also calculate the same color coordinates (metamerism phenomenon), so the two kinds of lamp beads may perform completely differently in certain applications (such as narrowband applications, applications requiring specific spectral power distribution).
[0005] Therefore, the existing LED lamp bead sorting methods lack effective characterization, quantitative comparison and intelligent grouping of "spectrum shape", and cannot meet the actual application. SUMMARY
[0006] The technical problem to be solved by the present application is to provide an LED lamp bead sorting method and system based on spectral characteristics, which can effectively improve the sorting accuracy.
[0007] In order to solve the above technical problems, the present application provides an LED lamp bead sorting method based on spectral characteristics, comprising: acquiring the spectral curve of each LED lamp bead in a target batch respectively; extracting spectral feature information according to the spectral curve respectively; performing clustering processing on the spectral feature information of all LED lamp beads in the target batch to generate a clustering result; and performing sorting processing on the LED lamp beads according to the clustering result.
[0008] As an improvement to the above scheme, the step of extracting the spectral feature information from the spectral curve includes: extracting shape feature parameters from the spectral curve, wherein the shape feature parameters include any one or more of the following: dominant wavelength, peak wavelength, half-maximum width, center wavelength, integral intensity ratio of a specific band, average intensity of a specific band, and intensity value of a characteristic wavelength point; and using the shape feature parameters as the spectral feature information.
[0009] As an improvement to the above scheme, the specific band includes any one or more of the blue light de-band, green light band, and red light band.
[0010] As an improvement to the above scheme, the step of extracting the spectral feature information based on the spectral curve includes: fitting the spectral curve to generate a fitting function; extracting the fitting parameters of the fitting function; and using the fitting parameters as the spectral feature information.
[0011] As an improvement to the above scheme, the step of extracting the spectral feature information based on the spectral curve includes: calculating the similarity between the spectral curve and the reference spectral curve; and using the similarity as the spectral feature information.
[0012] As an improvement to the above scheme, the method for calculating the similarity between the spectral curve and the reference spectral curve includes: calculating the Pearson correlation coefficient or cosine similarity between the spectral curve and the reference spectral curve at discrete wavelength points according to the correlation coefficient method; or calculating the angle between the spectral curve and the reference spectral curve according to the spectral angle mapping method; or calculating the minimum cumulative distance between the spectral curve and the reference spectral curve considering local wavelength stretching / compression according to the dynamic time warping method.
[0013] As an improvement to the above scheme, the reference spectral curve includes a preset reference spectral curve or the average spectral curve of all LED beads in the target batch.
[0014] As an improvement to the above scheme, the LED bead sorting method based on spectral characteristics further includes: calculating the standard deviation of the spectral curve of the LED beads in each grade; comparing the standard deviation with a preset threshold to determine whether the standard deviation exceeds the preset threshold; and if so, generating an alarm message and / or re-sorting the LED beads.
[0015] Accordingly, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described LED bead sorting method based on spectral characteristics.
[0016] Accordingly, the present invention also provides an LED bead sorting system based on spectral characteristics, comprising: a testing device for performing spectral tests on each LED bead in the target batch to obtain spectral curves; and the aforementioned computer device.
[0017] The LED chip sorting method and system based on spectral characteristics of this invention can be applied to the LED chip production process. Implementing this invention has the following beneficial effects: This invention extracts the shape features of the entire spectral curve of LED beads and combines them with clustering to achieve high-precision sorting of LED beads from the perspective of shape features. Under the premise of reasonably controlling the number of categories, it can significantly improve the spectral consistency of LED beads in the same group and meet the stringent requirements of high-end applications for spectral quality. Furthermore, this invention employs methods such as shape features, curve fitting, and similarity calculation to extract spectral feature information, thereby greatly improving the accuracy of sorting. Attached Figure Description
[0018] Figure 1 This is a flowchart of the first embodiment of the LED bead sorting method based on spectral characteristics of the present invention; Figure 2 This is a flowchart of the second embodiment of the LED bead sorting method based on spectral characteristics of the present invention; Figure 3 This is a schematic diagram of an embodiment of the LED bead sorting system based on spectral characteristics of the present invention. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. It is hereby declared that the directional terms such as up, down, left, right, front, back, inside, and outside used in this text are based solely on the accompanying drawings and are not intended to specifically limit the invention.
[0020] See Figure 1 , Figure 1 The flowchart of the first embodiment of the LED bead sorting method based on spectral characteristics of the present invention is shown, which includes: S101, Obtain the spectral curves of each LED bead in the target batch respectively; Each LED in the target batch is placed on the production line and its spectrum is measured separately under specific operating conditions (specific driving current and temperature) to collect the spectral data emitted by each LED.
[0021] In this embodiment, the spectral curve is preferably a spectral power distribution curve (SPD), whose wavelength range can cover the emission band of LED beads (such as 380nm~780nm visible light), but is not limited thereto.
[0022] S102, extract spectral feature information based on the spectral curves respectively; In existing technologies, sorting is mainly based on one or a few parameters such as dominant wavelength, luminous flux, brightness, and color coordinates. Unlike existing technologies, this invention extracts features from the shape characteristics of the entire spectral curve of LED beads to achieve high-precision sorting of LED beads from another perspective.
[0023] In some embodiments, the step of extracting spectral feature information from the spectral curve includes: (1) Extract shape feature parameters from the spectral curve; Shape characteristic parameters include any one or more of the following: dominant wavelength, peak wavelength, full width at half maximum (FWHM), center wavelength, integral intensity ratio of a specific band, average intensity of a specific band, and intensity value of a characteristic wavelength point. Among them, the specific bands include any one or more of the blue light band, the green light band, and the red light band.
[0024] (2) Use shape feature parameters as spectral feature information.
[0025] Therefore, this embodiment can achieve the extraction and quantification of spectral feature information, which is highly targeted.
[0026] In other embodiments, the step of extracting spectral feature information from the spectral curve includes: (1) Fit the spectral curve to generate a fitting function; In practical applications, preset function models (such as Gaussian function, Lorentz function, or superposition of multiple Gaussian / Lorentz functions) can be used to fit the spectral curve.
[0027] (2) Extract the fitting parameters of the fitting function; Accordingly, fitting parameters (such as the center wavelength, amplitude, and standard deviation of the Gaussian function) can be used as eigenvectors characterizing the shape of the spectral curve.
[0028] (3) Use the fitted parameters as spectral feature information.
[0029] In other embodiments, the step of extracting spectral feature information from the spectral curve includes: (1) Calculate the similarity between the spectral curve and the reference spectral curve; The reference spectrum curve includes a preset reference spectrum curve or the average spectrum curve of all LEDs in the target batch.
[0030] Furthermore, the method for calculating the similarity between the spectral curve and the reference spectral curve includes any one of the following methods: (1.1) Calculate the Pearson correlation coefficient or cosine similarity of the spectral curve and the reference spectral curve at discrete wavelength points according to the correlation coefficient method; (1.2) Calculate the angle between the spectral curve and the reference spectral curve using the Spectral Angle Mapper (SAM) method; It should be noted that in this embodiment, the spectral curve and the reference spectral curve are treated as vectors in a high-dimensional space. The smaller the angle between the spectral curve and the reference spectral curve, the more similar the shapes of the spectral curve and the reference spectral curve are.
[0031] (1.3) Calculate the minimum cumulative distance of the spectral curve and the reference spectral curve under the consideration of local wavelength stretching / compression according to the Dynamic Time Warping (DTW) method.
[0032] It should be noted that the dynamic time warping method is particularly suitable for cases where the spectral peaks shift slightly.
[0033] (2) Similarity is used as spectral feature information.
[0034] S103, perform clustering processing on the spectral feature information of all LED beads in the target batch to generate clustering results; In practical applications, unsupervised clustering algorithms (such as K-Means, K-Medoids, hierarchical clustering, DBSCAN, fuzzy C-means clustering, etc.) can be used to cluster the spectral feature information of all LED beads in the target batch.
[0035] Therefore, the LED bead sorting method based on spectral characteristics of this invention no longer uses the traditional "grid sorting" based on fixed one-dimensional / multi-dimensional thresholds, but instead uses intelligent clustering sorting, which has higher accuracy. S104, The LED beads are sorted according to the clustering results.
[0036] Based on the clustering results, each LED bead is sorted to the collection location corresponding to its cluster, thereby automatically grouping LED beads with highly similar spectral shapes into the same category (level).
[0037] As can be seen from the above, the LED bead sorting method based on spectral characteristics of the present invention extracts the shape features of the entire spectral curve of the LED bead and achieves high-precision sorting of the LED bead from the perspective of shape features. Under the premise of reasonably controlling the number of grades, it can significantly improve the spectral consistency of LED beads in the same group and meet the stringent requirements of high-end applications for spectral quality.
[0038] See Figure 2 , Figure 2 The flowchart of a second embodiment of the LED bead sorting method based on spectral characteristics of the present invention is shown, which includes: S201, Obtain the spectral curves of each LED bead in the target batch; S202, extract spectral feature information based on the spectral curves respectively; S203, perform clustering processing on the spectral feature information of all LED beads in the target batch to generate clustering results; S204, LED beads are sorted according to clustering results.
[0039] S205, calculate the standard deviation of the spectral curve of the LED beads in each gear; S206, compare the standard deviation with a preset threshold to determine whether the standard deviation exceeds the preset threshold; S207, if the determination is yes, generate alarm information and / or re-classify the LED beads.
[0040] and Figure 1 Unlike the first embodiment shown, this embodiment can perform consistency verification and feedback on the grading results.
[0041] In practical applications, steps S205 to S207 can be used to conduct random checks on the spectral consistency of LED beads in each grade after grading. If the standard deviation of the spectral curve of LED beads in a group exceeds the limit, an alarm will be triggered, and the group will be re-clustered or subdivided.
[0042] Accordingly, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described LED bead sorting method based on spectral characteristics.
[0043] See Figure 3 , Figure 3 The specific structure of the LED bead sorting system 100 based on spectral characteristics of the present invention is shown, which includes a testing device 1 and a computer device 2, specifically: Test equipment 1 is used to perform spectral tests on each LED bead in the target batch to collect spectral curves; Computer device 2 includes a memory and a processor. The memory stores a computer program, wherein the processor executes the computer program to implement the steps of the LED bead sorting method based on spectral characteristics described above.
[0044] Therefore, the LED bead sorting system based on spectral characteristics of the present invention can be applied to the LED bead production process. It takes the entire emission spectrum curve of the LED bead as the core basis for sorting, focuses on its morphological characteristics to perform high-precision sorting, so as to achieve highly consistent spectral sorting within the same group.
[0045] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.
Claims
1. A method for sorting LED beads based on spectral characteristics, characterized in that, include: Obtain the spectral curves of each LED bead in the target batch; Spectral feature information is extracted from the spectral curves respectively; Clustering is performed on the spectral characteristic information of all LED beads in the target batch to generate clustering results; The LED beads are categorized based on the clustering results.
2. The LED bead sorting method based on spectral characteristics as described in claim 1, characterized in that, The steps for extracting the spectral feature information based on the spectral curve include: Shape feature parameters are extracted from the spectral curve. The shape feature parameters include any one or more of the following: dominant wavelength, peak wavelength, full width at half maximum (FWHM), center wavelength, integral intensity ratio of a specific band, average intensity of a specific band, and intensity value of a characteristic wavelength point. The shape feature parameters are used as the spectral feature information.
3. The LED bead sorting method based on spectral characteristics as described in claim 2, characterized in that, The specific bands include any one or more of the blue light band, the green light band, and the red light band.
4. The LED bead sorting method based on spectral characteristics as described in claim 1, characterized in that, The steps for extracting the spectral feature information based on the spectral curve include: The spectral curve is fitted to generate a fitting function; Extract the fitting parameters of the fitting function; The fitting parameters are used as the spectral feature information.
5. The LED bead sorting method based on spectral characteristics as described in claim 1, characterized in that, The steps for extracting the spectral feature information based on the spectral curve include: Calculate the similarity between the spectral curve and the reference spectral curve; The similarity is used as the spectral feature information.
6. The LED bead sorting method based on spectral characteristics as described in claim 5, characterized in that, The method for calculating the similarity between the spectral curve and the reference spectral curve includes: Calculate the Pearson correlation coefficient or cosine similarity of the spectral curve and the reference spectral curve at discrete wavelength points using the correlation coefficient method; or The angle between the spectral curve and the reference spectral curve is calculated using the spectral angle mapping method; or The minimum cumulative distance between the spectral curve and the reference spectral curve, considering local wavelength stretching / compression, is calculated using the dynamic time warping method.
7. The LED bead sorting method based on spectral characteristics as described in claim 5, characterized in that, The reference spectral curve includes a preset reference spectral curve or the average spectral curve of all LED beads in the target batch.
8. The LED bead sorting method based on spectral characteristics as described in claim 1, characterized in that, Also includes: Calculate the standard deviation of the spectral curves of the LED beads in each gear setting; The standard deviation is compared with a preset threshold to determine whether the standard deviation exceeds the preset threshold. If the determination is yes, generate an alarm message and / or re-classify the LED beads.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the LED bead sorting method based on spectral characteristics as described in any one of claims 1 to 8.
10. An LED bead sorting system based on spectral characteristics, characterized in that, include: The testing equipment is used to perform spectral tests on each LED bead in the target batch to obtain spectral curves. The computer device as claimed in claim 9.