CN-120594046-B - Light color detection and analysis method and system for LED

CN120594046BCN 120594046 BCN120594046 BCN 120594046BCN-120594046-B

Abstract

The invention discloses a light color detection analysis method and system for an LED, which relate to the technical field of LED detection and comprise a light color acquisition module and a time sequence modeling module, wherein the light color acquisition module is used for acquiring spectrum data generated in the process of lighting a target LED, the time sequence modeling module is used for acquiring a plurality of spectrum data points at fixed time intervals in a continuous preset time period after the target LED starts to emit light, calculating and obtaining a dominant wavelength, a color coordinate and a fitting color temperature value to form a multi-time light color parameter sequence, constructing a time sequence model based on the sequence, and outputting the time sequence model and a corresponding original parameter sequence. According to the invention, by constructing the change curve of the light intensity and the color coordinates and introducing a convergence trend analysis mechanism, the automatic judgment of the LED luminous steady state is realized, the erroneous judgment caused by the instability of early data is effectively avoided, and the representativeness and the accuracy of the detection data are improved.

Inventors

YANG JINQUAN
GUI LIPING
CHEN JIAN
FENG MEI

Assignees

深圳市鸿屹光科技有限公司

Dates

Publication Date: 20260508
Application Date: 20250627

Claims (7)

1. A system for detecting and analyzing light color of an LED, comprising: the light color acquisition module is used for acquiring spectrum data generated in the process of lighting the target LED; The time sequence modeling module is used for collecting a plurality of spectrum data points at fixed time intervals in a continuous preset time period after the target LED starts to emit light, calculating to obtain a dominant wavelength, a color coordinate and a fitting color temperature value, forming a multi-time light color parameter sequence, constructing a time sequence model based on the sequence, and outputting the time sequence model and a corresponding original parameter sequence; The light color identification module is used for receiving the time sequence model and the light color parameter sequence output by the time sequence modeling module and identifying whether the LED enters a light-emitting stable state according to a preset judgment standard; The processing module is used for receiving the light color data of the target LED after the light color identification module confirms the stable state, extracting the average value of the light color data of all sampling points in the stable state, projecting the average value as input data for classification judgment into an established multidimensional light color characteristic model, classifying and judging the light color data, and outputting a light color grade and a color difference evaluation index corresponding to the light color data; The system sets a delay inspection time according to the time required for a plurality of LEDs in the same batch to reach a stable light emitting state, and comprises: in the initial stage of batch detection, selecting a plurality of LED samples, respectively executing a luminescence stability judging process, and recording the time required by each LED to reach a stable state; Determining a minimum stabilization time Tmin and a maximum stabilization time Tmax based on the sample stabilization time, and forming a dynamic value range [ Tmin, tmax ] of the delay checking time; The system initially adopts Tmin as delay checking time, and dynamically adjusts the delay checking time actually adopted from Tmin to Tmax along with the increase of the number of detected LEDs; When the detected number reaches a set threshold or the delay time has approached Tmax, the current delay check time is fixed as the detection latency of the subsequent LEDs of the batch.
2. The system of claim 1, wherein the method for obtaining spectral data by the light color collection module comprises: A1, the system provides constant driving current for a target LED to enable the target LED to enter a luminous state, and obtains the luminous intensity of the LED in real time through a luminous intensity acquisition channel to obtain a luminous intensity value; if the light intensity values acquired at three continuous sampling moments are higher than the set light intensity threshold and the light intensity change rate is smaller than the preset threshold, confirming that the LED enters a collectable state and entering the step A2; a2, continuously collecting at least three frames of spectrum data by the spectrum collecting module in each sampling time; The spectrum acquisition module compares three frames of spectrum data acquired continuously and extracts a dominant wavelength and CIE color coordinates (x, y); if the difference value of the main wavelengths among frames is smaller than 1nm and the difference value of the color coordinates is smaller than 0.003, judging that the spectrum data at the moment is valid; If the frame is judged to be invalid, delaying the sampling retry until the valid data frame is obtained; a3, carrying out standardized operation on the spectrum data judged to be effective so as to obtain standard spectrum data; A4, calculating light color parameters based on the standard spectrum data processed in the step A3, wherein the light color parameters comprise a main wavelength, CIE color coordinates (x, y) and fitting color temperature values; The light color acquisition module packages the light color parameters and the sampling time stamps thereof into structured single-frame light color data for the subsequent recognition or classification module.
3. The system of claim 1, wherein the light color recognition module recognizes whether the LED is in a steady state, comprising: Z1, a light color acquisition module acquires a light intensity value and a light color parameter of a target LED at fixed time intervals, and a light intensity change curve and a color coordinate change curve are formed in a preset sampling period; Z2, the photochromic recognition module analyzes a plurality of continuous sampling points corresponding to the light intensity change curve and the color coordinate change curve in Z1, and judges whether the following convergence conditions are met or not respectively: Z21, light intensity convergence, wherein the light intensity change rate of a plurality of continuous sampling points is smaller than a set light intensity change rate threshold; z22, color coordinate convergence, wherein the color coordinate change amplitude of a plurality of continuous sampling points is smaller than a set change amplitude threshold value; when the two conditions Z21 and Z22 are satisfied simultaneously, it is determined that the LED has reached a stable light emission state.
4. The system of claim 1, wherein the system dynamically adjusts the delay check time when the maximum Tmax and the minimum Tmax in the dynamic range interval [ Tmin, tmax ] vary, comprising: k1, the system acquires a stable time set through a sample LED, extracts an original minimum stable time Tmin 0 and an original maximum stable time Tmax 0 , and constructs an initial delay time interval [ Tmin 0 ,Tmax 0 ]; K2, when the kth LED is detected, the system calculates the delay checking time DCT k by adopting the following formula; DCT k =Tmin+(Tmax-Tmin)×(k/N); n is the upper number of allowable adjustments; K3, as more LEDs complete stability detection, the system continuously updates Tmin and Tmax, respectively denoted as Tmin n 、Tmax n , and calculates the relative change percentage: Δmin=(Tmin n -Tmin 0 )/Tmin 0 ; Δmax=(Tmax n -Tmax 0 )/Tmax 0 ; K4, the system adjusts the current delay checking time DCT k based on the amplitude of Delamin and Delamax to obtain the corrected delay checking time DCT k ': DCT k ′=DCT k ×[1+α·Δmin+β·Δmax]; wherein alpha and beta are adjustable coefficients, and the range is 0.3-0.5, and the adjustable coefficients are used for controlling the correction proportion.
5. The system of claim 1, wherein the light color acquisition module further comprises a plurality of optical acquisition channels respectively arranged at a plurality of angular positions along the light emitting direction of the target LED; At each sampling moment, the system synchronously acquires spectrum data from different angles through the channels to form a multi-angle light color data set; The multi-angle light color data comprise a main wavelength value, CIE color coordinates (x, y), fitting color temperature values and light intensity values in a plurality of directions, wherein each direction corresponds to a sampling channel and a number; The system gives weight to the collected values in all directions, and carries out weighted fusion on the light color parameters in all directions; The fused light color parameters are used as representative input data of the current sampling time and are used by a subsequent time sequence modeling, steady state judgment and light color grade evaluation module.
6. The system of claim 1, wherein the method for classifying and determining the light color data by the processing module comprises: m1, after a target LED reaches a stable state, the system receives all effective sampling point data in the stable state from the light color identification module, extracts light color parameters of each frame, and calculates the average value of all effective frames as a light color feature vector of the target LED; m2, a preset multidimensional photochromic feature model consists of M training samples, and each sample vector is as follows: J=1, 2, m; one label for each sample vector Indicating the light color grade thereof; m3, calculating the light color grade by adopting minimum Euclidean distance matching or K neighbor classification; By passing through Calculating a color difference evaluation index 。
7. A method for detecting and analyzing light color of an LED, the method employing the analysis system according to any one of claims 1 to 6, comprising: Collecting spectrum data of a target LED in the lighting process, wherein the spectrum data comprise a dominant wavelength, CIE color coordinates (x, y), a fitting color temperature value and a light intensity value; Collecting a plurality of spectrum data points at fixed time intervals in a continuous preset time period to form a light color parameter sequence, and constructing a time sequence model to describe the change trend of the light color parameter along with time; Judging whether the LED reaches a stable luminous state or not according to a preset light color change amplitude and a trend threshold based on a time sequence model and a parameter sequence; Step four, after the stable state is identified, extracting all effective light color parameters in the time period, and calculating an average value as a stable light color feature vector; And fifthly, projecting the average light color vector into a multidimensional light color feature model, judging the light color grade through a minimum distance matching or K neighbor classification algorithm, and outputting a color difference evaluation index.

Description

Light color detection and analysis method and system for LED Technical Field The invention relates to the technical field of LED detection, in particular to a method and a system for detecting and analyzing the light color of an LED. Background The LED is used as a high-efficiency and long-service-life light-emitting device, and the consistency and stability of light color are key factors in quality control. Along with the wide application of LEDs in the fields of illumination, display, car lights and the like, different application scenes provide detection requirements with higher precision and more representative judgment on parameters such as dominant wavelength, CIE color coordinates, fitting color temperature values and the like. Therefore, in the links of LED production and delivery, the construction of standardized and quantifiable light color detection processes has become an important direction in the industry. Through retrieval, chinese patent (publication No. CN 112200200B) discloses a method for detecting the color of LED lamp light, which comprises the steps of fixing an image acquisition device, setting an LED lamp light acquisition position, placing LED lamps with various colors at the LED lamp light acquisition position, carrying out color digital image acquisition of the LED lamp light through the image acquisition device, sending the color digital image acquisition to a processor for processing, calculating to obtain the value ranges of color characteristic parameters of the LED lamp light with various colors, and forming a database of the value ranges of the color characteristic parameters. And placing the LED lamp to be detected at an LED lamp light acquisition position and lighting the LED lamp, carrying out color digital image acquisition on the LED lamp light to be detected by the image acquisition equipment, sending the LED lamp light to a processor for processing, calculating to obtain a color characteristic parameter value of the LED lamp light to be detected, and matching the color characteristic parameter value range database to obtain the color of the LED lamp light to be detected. In practical applications, since there is a thermal response process at the initial stage of the LED lighting, the light intensity and color coordinates of the LED tend to fluctuate in a short time. In addition, the LED emits light with a certain spatial light distribution angle difference, and color difference judgment is inaccurate when a single angle is sampled. However, the conventional static sampling method generally does not consider the time evolution trend of the light color parameters, and lacks a spatial fusion mechanism for obtaining information from multiple angles, and it may be difficult to cover an LED device with a rapid change of the light emission characteristics or an asymmetric structure, so the present invention provides a light color detection and analysis method and system for an LED light emitting diode. Disclosure of Invention The invention aims to provide a method and a system for detecting and analyzing light color of an LED, which are used for solving the problems in the background art. The invention can be realized by the following technical scheme that the system for detecting and analyzing the light color of the LED comprises a light color acquisition module, a time sequence modeling module, a light color identification module and a processing module; The light color acquisition module is used for acquiring spectrum data generated in the lighting process of the target LED, wherein the spectrum data comprises a main wavelength value, CIE color coordinates (x, y), a fitting color temperature value and a light intensity value; the time sequence modeling module is used for collecting a plurality of spectrum data points at fixed time intervals in a continuous preset time period after a target LED starts to emit light, calculating and obtaining a dominant wavelength, a color coordinate and a fitting color temperature value based on each spectrum data point, and forming a multi-time light color parameter sequence, namely a light color data set arranged according to time; The time sequence modeling module builds a time sequence model based on the light color parameter sequence, is used for describing the trend of the light color parameters such as dominant wavelength, color coordinates, color temperature and the like along with the time change, and outputs the time sequence model and a corresponding original parameter sequence; The light color identification module is used for receiving the time sequence model and the light color parameter sequence output by the time sequence modeling module; Identifying whether the LED enters a stable light-emitting state according to a preset judgment standard; The processing module is used for receiving the light color data of the target LED after the light color recognition module confirms the stable state, extracting the average value of t