LED circuit board testing device and testing method based on spectrum matching and correction

Abstract

An LED circuit board testing device and testing method based on spectrum matching and correction, including S1: scanning a code to obtain the number on the LED module circuit board; S2: placing the LED module circuit board and fixing it; S3: selecting and executing an automated testing process; S4: the control system in the control terminal issues a test instruction and executes the automated testing process; S5: image acquisition and data processing, S6: quality judgment is made based on the result of the optical signal image data processing in S5, and the product is classified; S7: the operator removes the LED module circuit board and puts it into the original antistatic packaging. The present invention effectively eliminates non-measurement stray light interference, improves measurement contrast, and enhances the accuracy of chromaticity testing by setting a filter that matches the spectrum of the LED lamp to be tested; radiation calibration can ensure the accuracy and consistency of the measurement data and improve the reliability of the test results; the fully automatic and customizable test process simplifies the operation process and maintenance process.

Description

Technical Field

[0001] The invention belongs to the field of testing technology, and in particular relates to an LED circuit board testing device and a testing method based on spectrum matching and correction. Background Art

[0002] With the rapid development of LED technology, LED lighting products have been widely used in various fields. In order to ensure the quality and performance of LED products, their optical performance needs to be accurately tested. The traditional LED optical performance test method has the following problems: the chromaticity test accuracy is not high enough and is easily affected by ambient light and equipment errors; the test process is cumbersome and requires manual input of LED position parameters, which is inefficient;

[0003] The technical solution described in the authorized Chinese patent CN117274405B is a method for detecting the working color of LED lights based on machine vision. Figure 7 As shown in FIG. 1 , the method includes collecting an LED light working image in RGB color space, converting it into an image in HSI color space, and roughly segmenting the LED light color using the DBSCAN clustering algorithm. The clustering result is then corrected according to edge features and halo features, and finally a selected cluster is obtained, thereby improving the accuracy and precision of LED light working color detection.

[0004] The technical solution described in the authorized Chinese patent CN117058150B is a method and device for detecting defects in LED lamp beads. Figure 8 , Fig. 9 The solution shown in the figure collects pictures and current data after the lamp panel is powered on, inputs them into the defect detection network for processing, detects pseudo defects based on the aspect ratio and divides the picture into defective and non-defective areas, then collects a second picture and detects the second pseudo defect. This solution solves the problem of inaccurate lamp bead defect detection in traditional technology and improves detection accuracy and product quality.

[0005] In addition to the above-mentioned common LED optical performance test devices, the current market also includes visual, spectrometer and photometer test methods. The test principles and methods of these devices are as follows:

[0006] Visual: Visual testing is a method of testing the optical performance of LEDs, which involves manually visually inspecting the performance of LED light source products after they are powered on. Testers will observe the brightness, color, uniformity and other characteristics of LED light source products, and record relevant data and judgment results. In visual testing, testers will use their naked eyes to observe and evaluate the appearance and optical performance of LED light source products. They may check whether the brightness of the LED light source is uniform, whether the color is consistent, and whether there are any obvious defects or abnormalities. Through visual testing, testers can preliminarily evaluate the quality and performance of LED light source products, but this method may be affected by subjective factors, and the test results may not be accurate and consistent enough. Therefore, in actual production, other more sophisticated testing devices and methods are usually combined to conduct a comprehensive evaluation and verification of LED optical performance.

[0007] Spectrometer: Spectrometer is a high-precision testing equipment that can provide accurate and reliable LED optical performance data. It is widely used in LED lighting, display screens and other LED application fields. By splitting the light of the LED light source through a spectrometer, the spectral distribution curve of the LED can be obtained. In this method, the light emitted by the LED light source passes through the spectrometer's spectrometer and is decomposed into spectral components of different wavelengths. The spectrometer then records and measures the light intensity at each wavelength, and finally generates a normalized spectral distribution curve of the LED. Through spectrometer testing, the spectral characteristics of the LED light source in different wavelength ranges can be obtained, including parameters such as spectral peak, color temperature, and color uniformity, so as to fully understand the optical performance of the LED light source. It can help evaluate the spectral quality and color performance of the LED light source, and detect whether the light source meets the design requirements, which is of great significance to the research and development, production and quality control of LED products.

[0008] Photometer: The principle of photometer is based on photoelectric effect and optical measurement technology. Photometer usually includes a photosensitive detector, filter and signal processing circuit. When the light emitted by LED light source shines on the photosensitive detector of photometer, the photon will excite the electrons in the photosensitive detector and generate charge. The photosensitive detector converts the light signal into an electrical signal. Photometer is usually equipped with a filter to select light of a specific wavelength range for measurement. This helps to accurately measure parameters such as light intensity, illumination and color temperature of LED light source. The signal processing circuit of the photometer will amplify, filter and convert the electrical signal received from the photosensitive detector to obtain accurate photometer measurement results. Finally, the light intensity, luminous flux, illumination and other information of LED light source can be obtained, and the brightness of LED light source can be detected to see whether it meets the design requirements, and whether the optical performance of LED products meets the standard requirements can be verified.

[0009] However, in actual use, it can be found that the existing detection devices are complicated to operate, require a professional environment, and are costly: in the field of high-precision measurement, such as using integrating spheres, spectrometers and other equipment for testing, it is necessary to test in a specific laboratory environment, including constant temperature and humidity conditions, and professional knowledge and skills are required for operation and maintenance. Professional training is required for correct operation, which increases the cost of use and the technical threshold. This may limit the quality control and testing capabilities of LED products with high precision requirements.

[0010] Insufficient test accuracy: The existing LED test devices on the market have not been calibrated for brightness and color, nor have they eliminated stray light factors that affect measurement accuracy. The test accuracy is limited and cannot meet the testing needs of high-precision LED products.

[0011] Low test efficiency: Conventional visual and non-automatic positioning and customizable process testing methods require a long test time to obtain accurate test results. The test efficiency is low and cannot meet the testing needs of large-scale LED products.

[0012] Therefore, the existing LED circuit board testing technology needs to be improved. Summary of the invention

[0013] Purpose of the invention: In order to overcome the above shortcomings, the purpose of the present invention is to provide an LED circuit board testing device based on spectrum matching and correction, which has a simple structure, reasonable design, easy production, and effectively improves the detection efficiency.

[0014] Technical solution: In order to achieve the above-mentioned purpose, the present invention provides an LED circuit board testing device based on spectrum matching and correction, comprising: a testing machine, on which an LED module circuit board is placed, and on which at least one LED lamp is provided;

[0015] An integrating sphere, which is installed on one side of the working table of the testing machine;

[0016] A beam splitter, used to reflect the light emitted by the integrating sphere into the measuring optical path; the beam splitter is connected to the back plate of the test machine through a mounting frame, and the beam splitter cooperates with the integrating sphere;

[0017] An imaging lens, used for focusing light; the imaging lens is connected to the back plate of the test machine through a mounting frame and is arranged above the beam splitter;

[0018] A filter wheel mechanism is used to eliminate the influence of ambient stray light and improve the measurement contrast. The filter wheel mechanism is installed on the back plate of the test machine through a mounting bracket. The filter wheel mechanism includes a filter wheel and a rotation drive mechanism for driving the filter wheel to rotate. The rotation drive mechanism is connected to the back plate of the test machine through a mounting bracket, and its output end cooperates with the filter wheel. The filter wheel is provided with at least one filter and at least one hollow hole. The filter matches the spectrum of the LED light to be tested.

[0019] A camera, used to receive and measure optical signals, the camera is connected to the back plate of the test machine through a camera mounting frame, and the filter wheel is arranged between the imaging lens and the camera;

[0020] The control terminal is used to control the lighting of the LED lamp on the LED module circuit board, the rotation of the filter wheel and the camera to obtain an image; the LED module circuit board, the integrating sphere, the rotation drive mechanism and the camera are all connected to the control terminal.

[0021] The present invention provides a testing method for an LED circuit board testing device based on spectrum matching and correction, and the specific testing method comprises the following steps:

[0022] S1): Scan the barcode to obtain the number on the LED module circuit board, that is, use a scanner to scan the barcode or QR code on the LED module circuit board to obtain its unique number;

[0023] S2): placing and fixing the LED module circuit board, that is, placing the LED module circuit board on the workbench of the test machine, ensuring that the pins on the LED module circuit board are accurately aligned with the connection points of the test equipment, and ensuring that the LED module circuit board remains stable during the test;

[0024] S3): Select and execute the automated test process, that is, the user selects a preset automated test process or customizes a new automated test process through the control terminal according to the test requirements. The control terminal can start the spectral radiation calibration in the control system, obtain traceable high-precision brightness and chromaticity test values ​​through measurement equation calculation, and is equipped with an integrating sphere as a standard radiation source calibrator, so that absolute spectral radiation calibration can be performed at any time;

[0025] S4): The control system in the control terminal issues a test instruction to execute the automated test process, that is, the filter wheel control module in the control system commands the rotary drive mechanism to drive the filter wheel to rotate to the filter position corresponding to the current working color of the LED lamp on the LED module circuit board, ensuring that the center of the filter is aligned with the center of the optical path composed of the imaging lens and the camera, so as to perform the corresponding optical property test. At the same time, the control terminal issues an instruction to light up the LED lamp on the LED module circuit board and adjust it to the color component and working voltage to be tested;

[0026] S5): Image acquisition and data processing, i.e., the light beam emitted by the LED lamp is collected by an optical system consisting of a beam splitter, an imaging lens, a filter and a camera, and the light signal image is acquired according to the response of the light intensity on the camera image plane, and the acquired light signal image data is processed, and the optical performance index of the LED lamp is calculated;

[0027] S6): according to the result of the optical signal image data processing in S5), quality judgment is performed and products are classified, that is, whether the LED lamp meets the requirements of high-precision optical applications, good products and bad products are classified, bad products are eliminated, and the next test task of the customized process is performed on good products until all test processes are completed and all optical performance parameters required for all LED lamps are obtained;

[0028] S7): The operator removes the LED module circuit board and puts it into the original anti-static packaging. The test of this product is completed and the next product is tested.

[0029] The present invention discloses a testing method for an LED circuit board testing device based on spectrum matching and correction. In the S3), an integrating sphere is provided as a standard radiation source calibrator, and absolute spectrum radiation calibration can be performed at any time. The specific process is as follows: S31): initial installation and calibration, that is, installing the integrating sphere on one side of the test light path, and installing the beam splitter on one side of the integrating sphere, to ensure that the outgoing light of the integrating sphere can be reflected into the light path of the imaging lens, the filter wheel and the camera; fine-adjusting the center positions of the integrating sphere and the beam splitter in a plane perpendicular to the light path, respectively, to ensure that the light path is aligned and the light signal of the light path can be normally received by the camera. After the adjustment, the positions of the fixing integrating sphere and the beam splitter will not change;

[0030] S32): Initial calibration, switch to calibration mode through the control system in the control terminal: turn on the integrating sphere, record the output signal S of the camera std (λ), according to the known spectral radiation L of the integrating sphere std (λ), calculate the camera response function R(λ):

[0031]

[0032] The calculated response function R(λ) is saved, and the absolute spectral radiometric calibration coefficient in the measurement equation is updated for subsequent calculations of the camera response value to photometric and chromaticity parameters with actual physical meanings. In the intervals between normal measurement processes, the absolute spectral radiometric calibration can be performed at any time to ensure the accuracy and consistency of the measurement results.

[0033] S33): After the absolute radiation calibration is completed, the control system in the control terminal will switch back to the measurement mode.

[0034] The present invention discloses a testing method for an LED circuit board testing device based on spectrum matching and correction, wherein S5): the camera collects image data and processes the collected image data, and the specific processing process is as follows:

[0035] S51): the camera control module in the control terminal commands the camera to start collecting luminous image data of the LED lamp on the LED module circuit board, and saves the collected luminous image data to a specified path;

[0036] S52): automatically reading the luminous image data from the specified path in S51) through the data processing system in the control terminal, pre-processing the luminous image area, that is, converting the RGB color domain of the luminous image into the HSV color domain, and performing denoising, binarization, and expansion processing on the luminous image, and then using the circle recognition algorithm based on Hough transform to automatically locate the luminous area position of the LED lamp;

[0037] S53): Calculate the mean values ​​of the hue component and the intensity component within the luminous range of the LED lamp, and bring the calculated data into the measurement equation, and finally obtain the accurate brightness value and chromaticity value of the measured LED lamp according to the measurement equation result.

[0038] The present invention discloses a testing method for an LED circuit board testing device based on spectral matching and correction, wherein the circle recognition algorithm based on Hough transform in S52) uses the parametric equation of the circle to map the circle in the image space to the parameter space, and uses an accumulator to count possible circle centers and radii, and finds the area in the image that is most likely to be a circle by voting in the parameter space, so that the position and size of the circle can be effectively detected and located from an image, and good robustness can be shown even in the case of more noise or partial obstruction of the circle.

[0039] The standard equation of a circle in Cartesian coordinates is: (xa) 2 +(yb) 2 =r 2

[0040] Among them, (a, b) is the coordinate of the center of the circle, r is the radius of the circle, and (x, y) is the edge point;

[0041] In the Hough transform, the problem in the image space needs to be converted to the parameter space. For circle detection, each circle can be represented by three parameters: the coordinates of the center of the circle (a, b) and the radius r of the circle. Therefore, the parameter space will be a three-dimensional space.

[0042] The specific process of the circle recognition algorithm based on Hough transform is as follows:

[0043] S521): First, edge detection algorithm is used to extract edge points (x, y) in the luminous image. For each edge point (x, y) in the luminous image, all possible radii r are considered;

[0044] For each circle with radius r, the center coordinates (a, b) must satisfy the following formula:

[0045] a=x-rcos(θ), b=y-rsin(θ), θ∈[0, 2π];

[0046] S522): create a three-dimensional array (a, b, r) as an accumulator, whose dimensions correspond to the possible values ​​of a, b and r;

[0047] For each edge point (x, y) and each polar coordinate angle value θ, calculate the corresponding a and b, and add 1 to the corresponding (a, b, r) position in the accumulator;

[0048] After completing the voting of all edge points in the luminous image, the maximum value in the accumulator indicates the location where a circle may exist in the luminous image. By finding the local maximum value in the accumulator, the most likely center position and radius of the circle can be determined, thereby determining the location of the LED light and the luminous area.

[0049] In the testing method of the LED circuit board testing device based on spectrum matching and correction described in the present invention, the following strategies are also adopted in the circle recognition algorithm based on Hough transform to improve the calculation efficiency and accuracy, and the specific strategies are as follows:

[0050] S5201): using the gradient direction as an estimate of the polar coordinate angle value θ to reduce unnecessary calculations of the polar coordinate angle value θ;

[0051] S5202): Threshold processing is performed on the accumulator, and only candidate circles whose votes exceed the software setting and are verified by previous experiments to be suitable for the threshold of the current LED light being tested are considered;

[0052] S5203): Non-maximum suppression is used on the accumulator to ensure that the positions of the detected candidate circles are not too close, so that the circle recognition algorithm based on Hough transform can effectively detect circular objects from noisy image data.

[0053] In the testing method of the LED circuit board testing device based on spectrum matching and correction described in the present invention, the specific process of converting the RGB color gamut into the HSV color gamut in S52) is as follows:

[0054] The specific process of converting the RGB color gamut into the HSV color gamut in S52) is as follows:

[0055] Assume (DN R , D.N.G , D.N. B ) are the red component, green component and blue component in the RGB color value, and (H, S, V) are the hue, saturation and brightness in the converted HSV color value;

[0056] First, normalize the RGB values: R , D.N. G , D.N. B ) value from 0-255 range to 0-1 range (DN R1 , D.N. G1 , D.N. B1 ), and calculate DN R1 , D.N. G1 , D.N. B1 The maximum value max and the minimum value min in the component;

[0057] DN R1 =DN R / 255.0

[0058] DN G1 =DN G / 255.0

[0059] DN B1 =DN B / 255.0

[0060] max=max(DN R1 , D.N. G1 , D.N. B1 )

[0061] min=min(DN R1 , D.N. G1 , D.N. B1 )

[0062] Then, calculate the brightness value V: V = max

[0063] Next, calculate the saturation value S:

[0064] When V≠0, S=(V-min) / V

[0065] When V=0, then S=0. Again, calculate the hue value H:

[0066] If max = min, then H = 0

[0067] If max = DN R1 And DN G1 ≥DN B1 , then H=60*(DN G1 -DN B1 ) / (max-min)

[0068] If max = DN R1 And DN G1 <DN B1 , then H=60*(DN G1 -DN B1 ) / (max-min)+360

[0069] If max = DN G1 , then H=60*(DN B1 -DN R1 ) / (max-min)+120

[0070] If max = DN B1 , then H=60*(DN R1 -DN G1 ) / (max-min)+240

[0071] Finally, after the above calculations, the values ​​of hue H, saturation S, and brightness V in the HSV color domain are obtained.

[0072] In the testing method of the LED circuit board testing device based on spectrum matching and correction described in the present invention, the specific process of denoising, binarization and expansion of the luminous image in S52) is as follows:

[0073] S52.1): Read image data of the luminescent image and convert the image data into a grayscale image;

[0074] S52.2): Select an appropriate denoising method to smooth the acquired grayscale image to remove high-frequency noise;

[0075] S52.3): Binarize the smoothed grayscale image using an adaptive threshold binarization method to convert the grayscale image into a black-and-white binary image, i.e., segmentation is achieved by adaptive threshold segmentation, where the adaptive threshold is obtained by taking a compromise position based on the response value of the LED lamp under test and the response value of the environmental background;

[0076] S52.4): Perform a morphological dilation operation on the binary black-and-white binary image. The dilation can fill small holes and discontinuous lines in the target area, making the target area more complete and continuous.

[0077] The expansion operation is performed using a structural element determined by an experimental method that is two orders of magnitude smaller than the size of the LED lamp being tested. The kernel size is selected experimentally based on the actual LED lamp being tested. The specific process is as follows:

[0078] S52.41): define dilation kernel;

[0079] S52.42): Apply a dilation operation.

[0080] The present invention discloses a testing method for an LED circuit board testing device based on spectral matching and correction, wherein the calibration parameters in the measurement equation (S53) are obtained by precise spectral radiation calibration of the LED circuit board testing device based on spectral matching and correction, and the specific process of the spectral radiation calibration is as follows:

[0081] The basic principle of spectral radiometric calibration is to use a standard radiation source with known spectral radiation characteristics, and to establish a mapping relationship between the camera output signal and the actual spectral radiation brightness by measuring the camera's response to these known spectral radiations.

[0082] Assuming that the known spectral radiance of the standard light source is L0(λ), the output signal of the camera is S0(λ), and the spectral radiance response function of the camera is R(λ), then:

[0083] L0(λ)=R(λ)S0(λ)

[0084] Therefore, by measuring the known spectral radiance L0(λ) of the standard light source and the output signal S0(λ) of the camera, the spectral radiance response function R(λ) of the camera can be obtained:

[0085]

[0086] Since the camera's response to different known spectral radiation measurements has different signal-to-noise ratios, multiple standard radiation sources are used to perform multi-point spectral radiation calibration to improve the accuracy and reliability of the spectral radiation response function calibration curve;

[0087] Select multiple standard spectral radiation sources, and for each standard spectral radiation source, record the camera output signal S T1 (λ), S T2 (λ) and S T3 (λ);

[0088] According to the known spectral radiation L of each standard spectral radiation source T1 (λ), L T2 (λ) and L T3 (λ), calculate the camera's response function to each standard spectral radiation source:

[0089]

[0090] Fit the response function data of multiple standard spectral radiation sources to generate a comprehensive calibration curve R T (λ), and the fitting coefficients a0, a1, …, a are obtained by the least squares fitting method. n ;

[0091] After the spectral radiation calibration is completed, for the measurement of the unknown spectral radiation brightness to be measured, assuming that the output signal of the unknown spectral radiation source is S(λ), the unknown spectral radiation brightness L(λ) to be measured can be calculated by the following formula:

[0092] L(λ)=R T (λ)S(λ)

[0093] Among them, S(λ) is the output signal of the camera for the unknown spectrum, R T (λ) is the response function obtained through calibration, and L(λ) is the unknown spectral radiation brightness to be measured, so that the output signal of the camera can be accurately converted into spectral radiation brightness information with clear physical meaning, thereby realizing spectral radiation calibration correction.

[0094] In the testing method of the LED circuit board testing device based on spectrum matching and correction described in the present invention, the specific process of judging whether the LED lamp meets the requirements of high-precision optical applications and classifying good products and bad products in S6) is as follows:

[0095] If the LED light meets the requirements of high-precision optical applications, the good indicator light will be turned on; if the LED light does not meet the requirements of high-precision optical applications, the defect indicator light will be turned on;

[0096] For products with defective indicators, the control system will mark the defect based on the specific defect type of brightness or chromaticity and the number of the LED module circuit board, and then take the defective products to the temporary storage area for defective products;

[0097] For good products, the data processing system returns the test results to the control system to continue to execute the next test task of the customized process until all test processes are completed and all the optical performance parameters required for all LED lights are obtained;

[0098] The operator removes the LED module circuit board and puts it into the original anti-static packaging. The test of this product is completed and the next product is tested.

[0099] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0100] 1. The LED circuit board test device based on spectrum matching and correction described in the present invention simplifies its operation process by optimizing the structure of the test device. By setting a filter that matches the spectrum of the LED lamp to be tested on the filter wheel, it can effectively eliminate the interference of non-measured stray light, improve the measurement contrast, enhance the accuracy of chromaticity testing, and effectively isolate the light of non-target light sources to ensure that only the light from the LED itself is measured, thereby reducing the chromaticity measurement error caused by ambient light pollution; at the same time, radiation calibration is also set to ensure the accuracy and consistency of the measurement data. Spectral radiation calibration can not only trace the brightness measurement results to the standards of the Institute of Metrology, but also help correct the nonlinear and unstable response deviations in the instrument, thereby improving the reliability of the test results.

[0101] 2. The testing method of the LED circuit board testing device based on spectral matching and correction described in the present invention simplifies the operation process and maintenance process and reduces the need for professional knowledge through a fully automatic and customizable testing process, so that users can operate the equipment without complex training, further reducing operating costs.

[0102] 3. The present invention uses a shape recognition algorithm based on Hough transform to automatically locate the position and range of the LED by circle recognition. It does not require input of position parameters each time. Unlike most fixed position value methods on the market, it eliminates the value error caused by LED position offset caused by circuit board processing errors. At the same time, it saves the workload caused by changing models and re-measuring the value position and area.

[0103] 4. The LED circuit board testing device based on spectral matching and correction described in the present invention is traceable and has high precision: the testing equipment is spectral radiation calibrated, and traceable high-precision brightness and chromaticity test values ​​can be obtained through measurement equation calculation; and it is equipped with a standard spectral radiation source as a calibrator, so that absolute spectral radiation calibration can be performed at any time.

[0104] 5. A fully automatic test process that can be customized during the test of an LED circuit board test device based on spectral matching and correction in the present invention: the test equipment utilizes the control of the power supply and filter wheel of the LED module circuit board to be tested, and can carry out automatic testing according to the user-defined process, and can be customized to measure all LED lights on the circuit board at the same time or measure individual LED lights one by one. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] Figure 1 This is a flow chart of LED optical performance testing in a quality inspection environment using an LED circuit board testing device based on spectrum matching and correction in the present invention;

[0106] Figure 2 A schematic diagram of the optical path of the LED circuit board testing device based on spectrum matching and correction in the present invention;

[0107] Figure 3 This is a schematic structural diagram of the LED circuit board testing device based on spectrum matching and correction according to the present invention;

[0108] Figure 4 It is a partial schematic diagram of the installation of the filter wheel mechanism in the present invention;

[0109] Figure 5 It is a partial schematic diagram of the installation of the rotary drive mechanism and the filter wheel in the present invention;

[0110] Figure 6 It is a partial schematic diagram of the filter wheel in the present invention;

[0111] Figure 7 It is a detection flow chart of the authorized Chinese patent CN117274405B in the background technology of the present invention;

[0112] Figure 8 It is a flow chart of the method for detecting lamp bead defects in the authorized Chinese patent CN117058150B in the background technology of the present invention;

[0113] Fig. 9 It is the overall detection flow chart of the authorized Chinese patent CN117058150B in the background technology of the present invention;

[0114] In the figure: 1 test machine, 2 LED module circuit board, 3 LED lamp, 4 integrating sphere, 5 beam splitter, 6 imaging lens, 7 filter wheel mechanism, 71 filter wheel, 72 rotation drive mechanism, 73 filter, 74 hollow hole, 75 outer shell, 721 rotation drive motor, 722 worm, 751 light hole, 752 cover plate, 8 camera, 9 control terminal. DETAILED DESCRIPTION

[0115] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments.

[0116] Example

[0117] like Figures 1 to 6 The LED circuit board testing device based on spectrum matching and correction shown in the figure comprises: a testing machine 1, on which an LED module circuit board 2 is placed, and on which at least one LED lamp 3 is arranged;

[0118] An integrating sphere 4, wherein the integrating sphere 4 is installed on one side of the workbench of the testing machine 1;

[0119] The beam splitter 5 is used to reflect the light emitted by the integrating sphere 4 into the measuring optical path; the beam splitter 5 is connected to the back plate of the test machine 1 through a mounting frame, and the beam splitter 5 cooperates with the integrating sphere 4;

[0120] An imaging lens 6, used for focusing light; the imaging lens 6 is connected to the back plate of the test machine 1 through a mounting frame and is arranged above the beam splitter 5;

[0121] The filter wheel mechanism 7 is used to eliminate the influence of ambient stray light and improve the measurement contrast. The filter wheel mechanism 7 is installed on the back plate of the test machine 1 through a mounting bracket. The filter wheel mechanism 7 includes a filter wheel 71 and a rotation drive mechanism 72 for driving the filter wheel 71 to rotate. The rotation drive mechanism 72 is connected to the back plate of the test machine 1 through a mounting bracket, and its output end cooperates with the filter wheel 71. The filter wheel 71 is provided with at least one filter 73 and at least one hollow hole 74. The filter 73 matches the spectrum of the LED lamp 3.

[0122] The camera 8 is used to receive and measure the optical signal. The camera 8 is connected to the back plate of the test machine 1 through a camera mounting frame. The filter wheel 71 is arranged between the imaging lens 6 and the camera 8;

[0123] The control terminal 9 is used to control the lighting of the LED lamp 3 on the LED module circuit board 2, the rotation of the filter wheel 71 and the image acquisition operation of the camera 8; the LED module circuit board 2, the integrating sphere 4, the rotation drive mechanism 72 and the camera 8 are all connected to the control terminal 9.

[0124] It should be noted that the center of the light outlet of the integrating sphere 4 is highly consistent with the center of the beam splitter 5. This ensures that the optical path is aligned during operation, further improving the accuracy of the detection.

[0125] like Figure 5 The rotary drive mechanism 72 shown in the figure comprises a rotary drive motor 721, and a worm 722 is provided at the output end of the rotary drive motor 721. The helical teeth on the worm 722 mesh with the teeth on the outer side of the filter wheel 71, as shown in FIG. Figure 4 The filter wheel mechanism 7 shown also includes an outer shell 75, the filter wheel 71 is arranged in the outer shell 75, and the outer shell 75 is provided with a group of light holes 751, and the light holes 751 are provided with a cover plate 752. When working, one of the light holes 751 is selected for use according to needs, and the unnecessary light holes 751 can be blocked by the cover plate 752.

[0126] It should be noted that the filter that matches the spectrum of the LED light to be tested can effectively eliminate the interference of stray light in measurement, improve contrast, and enhance the accuracy of colorimetry testing.

[0127] In actual work, when the blue component of the LED lamp 3 is tested, the filter wheel 71 turns to the blue filter that matches the spectrum of the LED to be tested, or when the red component of the LED is tested, the filter wheel 71 turns to the red filter that matches the spectrum of the LED lamp to be tested, or when the green component of the LED lamp is tested, the filter wheel 71 turns to the green filter that matches the spectrum of the LED lamp to be tested, which can prevent interference from stray light, improve contrast, and thus improve measurement accuracy.

[0128] When testing the color combination after the three-color fusion of the LED lamp, the filter wheel 71 rotates to the hollow hole 74 to measure the chromaticity parameters.

[0129] A testing method for an LED circuit board testing device based on spectrum matching and correction, characterized in that the specific testing method comprises the following steps:

[0130] S1): Scan the barcode to obtain the number on the LED module circuit board 2, that is, use a scanner to scan the barcode or QR code on the LED module circuit board 2 to obtain its unique number;

[0131] S2): placing and fixing the LED module circuit board 2, that is, placing the LED module circuit board 2 on the workbench of the test machine 1, ensuring that the pins on the LED module circuit board 2 are accurately aligned with the connection points of the test equipment, and ensuring that the LED module circuit board 2 remains stable during the test;

[0132] S3): Select and execute the automated test process, that is, the user selects a preset automated test process or customizes a new automated test process through the control terminal 9 according to the test requirements. The control terminal 9 can start the spectral radiation calibration in the control system, obtain traceable high-precision brightness and chromaticity test values ​​through measurement equation calculation, and is equipped with an integrating sphere 4 as a standard radiation source calibrator, so that absolute spectral radiation calibration can be performed at any time;

[0133] S4): The control system in the control terminal 9 issues a test instruction to execute the automated test process, that is, the filter wheel control module in the control system commands the rotary drive mechanism 72 to drive the filter wheel 71 to rotate to the position of the filter 73 corresponding to the current working color of the LED lamp 3 on the LED module circuit board 2, ensuring that the center of the filter 73 is aligned with the center of the optical path composed of the imaging lens 6 and the camera 8, so as to perform the corresponding optical characteristic test. At the same time, the control terminal 9 issues an instruction to light up the LED lamp 3 on the LED module circuit board 2 and adjust it to the color component and working voltage to be tested;

[0134] S5): Image acquisition and data processing, i.e., the light beam emitted by the LED lamp 3 is collected by an optical system consisting of a beam splitter 5, an imaging lens 6, a filter 73 and a camera 8, and the light signal image is acquired according to the response of the light intensity on the image plane of the camera 8, the acquired light signal image data is processed, and the optical performance index of the LED lamp 3 is calculated;

[0135] S6): Perform quality judgment based on the result of the optical signal image data processing in S5, and classify the products, that is, judge whether the LED lamp 3 meets the requirements of high-precision optical applications, classify good products and bad products, remove bad products, and perform the next test task of the customized process on good products until all test processes are completed to obtain all the optical performance parameters required by all LED lamps 3;

[0136] S7): The operator removes the LED module circuit board 2 and puts it into the original anti-static packaging. The test of this product is completed and the test of the next product is carried out.

[0137] In the embodiment described in S3), the integrating sphere 4 is equipped as a standard radiation source calibrator, and absolute spectral radiation calibration can be performed at any time. The specific process is as follows: S31): initial installation and calibration, that is, the integrating sphere 4 is installed on one side of the test light path, and the beam splitter 5 is installed on one side of the integrating sphere 4 to ensure that the outgoing light of the integrating sphere 4 can be reflected into the light path of the imaging lens 6, the filter wheel 7 and the camera 8; the center positions of the integrating sphere 4 and the beam splitter 5 are respectively slightly adjusted in the plane perpendicular to the light path to ensure that the light path is aligned and the light signal of the light path can be normally received by the camera 8. After the adjustment, the positions of the fixing integrating sphere 4 and the beam splitter 5 will not change;

[0138] S32): Initial calibration, switch to calibration mode by the control system in the control terminal 9: turn on the integrating sphere 4, record the output signal S of the camera 8 std (λ), according to the known spectral radiation L of the integrating sphere std (λ), calculate the camera response function R(λ):

[0139]

[0140] The calculated response function R(λ) is saved, and the absolute spectral radiometric calibration coefficient in the measurement equation is updated for subsequent calculation of the camera 8 response value to photometric and chromaticity parameters with actual physical meaning; absolute spectral radiometric calibration can be performed at any time during the normal measurement process to ensure the accuracy and consistency of the measurement results;

[0141] S33): After the absolute radiation calibration is completed, the control system in the control terminal 9 will switch back to the measurement mode.

[0142] It should be noted that the integrating sphere 4, as a standard radiation source calibrator, has passed the reference transfer and has known radiation characteristics.

[0143] In the test method of the LED circuit board test device based on spectrum matching and correction described in this embodiment, the specific process of the camera S5) collecting image data and processing the collected image data is as follows:

[0144] S51): The camera control module in the control terminal 9 commands the camera 8 to start collecting the luminous image data of the LED lamp 3 on the LED module circuit board 2, and saves the collected luminous image data to a specified path;

[0145] S52): automatically reading the luminous image data from the specified path in S51 through the data processing system in the control terminal 9, preprocessing the luminous image area, converting the RGB color domain of the luminous image into the HSV color domain, and performing denoising, binarization, and expansion processing on the luminous image, and then using the circle recognition algorithm based on Hough transform to automatically locate the luminous area position of the LED lamp 3;

[0146] S53): Calculate the mean values ​​of the hue component and the intensity component within the luminous range of the LED lamp 3, and bring the calculated data into the measurement equation, and finally obtain the accurate brightness value and chromaticity value of the measured LED lamp 3 according to the measurement equation results.

[0147] The circle recognition algorithm based on Hough transform in S52) in this embodiment uses the parametric equation of the circle to map the circle in the image space to the parameter space, and uses an accumulator to count the possible center and radius of the circle, and finds the area in the image that is most likely to be a circle by voting in the parameter space. It can effectively detect and locate the position and size of the circle from an image, and can show good robustness even in the case of more noise or partial occlusion of the circle.

[0148] The standard equation of a circle in Cartesian coordinates is: (xa) 2 +(yb) 2 =r 2

[0149] Among them, (a, b) is the coordinate of the center of the circle, r is the radius of the circle, and (x, y) is the edge point;

[0150] In the Hough transform, the problem in the image space needs to be converted to the parameter space. For circle detection, each circle can be represented by three parameters: the coordinates of the center of the circle (a, b) and the radius r of the circle. Therefore, the parameter space will be a three-dimensional space.

[0151] The specific process of the circle recognition algorithm based on Hough transform is as follows:

[0152] S521): First, edge detection algorithm is used to extract edge points (x, y) in the luminous image. For each edge point (x, y) in the luminous image, all possible radii r are considered;

[0153] For each circle with radius r, the center coordinates (a, b) must satisfy the following formula:

[0154] a=x-rcos(θ), b=y-rsin(θ), θ∈[0, 2π];

[0155] S522): create a three-dimensional array (a, b, r) as an accumulator, whose dimensions correspond to the possible values ​​of a, b and r;

[0156] For each edge point (x, y) and each polar coordinate angle value θ, calculate the corresponding a and b, and add 1 to the corresponding (a, b, r) position in the accumulator;

[0157] It should be noted that here the equation of the circle is changed from the Cartesian coordinate system (center coordinates, radius) to the polar coordinate system (center coordinates, radius, trigonometric functions of angles). The reason is that the parameter space of the circle is continuous in polar coordinates and can be easily converted into a lookup table, which significantly reduces the amount of calculation of the quadratic equation. In addition, the characteristics of the circle are easier to capture in polar coordinates, which reduces the impact of noise on parameter identification.

[0158] After completing the voting for all edge points in the luminous image, the maximum value in the accumulator indicates the position where a circle may exist in the luminous image. By finding the local maximum value in the accumulator, the most likely center position and radius of the circle can be determined, thereby determining the position and luminous area of ​​the LED lamp 3.

[0159] In the test method of the LED circuit board test device based on spectrum matching and correction described in this embodiment, the following strategies are also adopted in the circle recognition algorithm based on Hough transform in S52) to improve the calculation efficiency and accuracy. The specific strategies are as follows:

[0160] S5201): using the gradient direction as an estimate of the polar coordinate angle value θ to reduce unnecessary calculations of the polar coordinate angle value θ;

[0161] S5202): Threshold processing is performed on the accumulator, and only candidate circles whose votes exceed the software setting and are verified by previous experiments to be suitable for the current LED light 3 threshold under test are considered;

[0162] S5203): Non-maximum suppression is used on the accumulator to ensure that the positions of the detected candidate circles are not too close, so that the circle recognition algorithm based on Hough transform can effectively detect circular objects from noisy image data. It should be noted that the edge detection link in the algorithm will calculate the gradient vector, that is, the gradient direction of the image pixel value, and the gradient direction points to the direction where the center of the circle may be.

[0163] It should be noted that Non-Maximum Suppression, NMS: In a local area, the pixel with the maximum response value is retained, and other smaller values ​​around it within a certain range will be suppressed, that is, set to 0 or ignored. In circle detection, the accumulator is used to record the position of the possible center of the circle, and "non-maximum suppression" can be used to screen out the most representative center of the circle to avoid multiple detections of centers that are too close. The specific method is to find the local maximum point in the accumulator, determine the response value around the point, and if the response value of other points is larger than that of this point, the response value of this point is set to 0, thereby eliminating redundant circle detection results. This ensures that the final detected circle position is accurate and non-repetitive.

[0164] In the test method of the LED circuit board test device based on spectrum matching and correction described in this embodiment, the specific process of converting the RGB color gamut into the HSV color gamut in S52) is as follows:

[0165] Assume (DN R , D.N. G , D.N. B ) are the red component, green component and blue component in the RGB color value, and (H, S, V) are the hue, saturation and brightness in the converted HSV color value;

[0166] First, normalize the RGB values: R , D.N. G , D.N. B ) value from 0-255 range to 0-1 range (DN R1 , D.N. G1 , D.N. B1 ), and calculate DN R1 , D.N. G1 , D.N. B1 The maximum value max and the minimum value min in the component;

[0167] DN R1 =DN R / 255.0

[0168] DN G1 =DN G / 255.0

[0169] DN B1 =DN B / 255.0

[0170] max=max(DN R1 , D.N. G1 , D.N. B1 )

[0171] min=min(DN R1 , D.N. G1 , D.N. B1 )

[0172] Then, calculate the brightness value V: V = max

[0173] Next, calculate the saturation value S:

[0174] When V≠0, S=(V-min) / V

[0175] When V = 0, then S = 0. Again, calculate the hue value H:

[0176] If max = min, then H = 0

[0177] If max = DN R1 And DN G1 ≥DN B1 , then H=60*(DN G1 -DN B1 ) / (max-min)

[0178] If max = DN R1 And DN G1 <DN B1 , then H=60*(DN G1 -DN B1 ) / (max-min)+360

[0179] If max = DN G1 , then H=60*(DN B1 -DN R1 ) / (max-min)+120

[0180] If max = DN B1 , then H=60*(DN R1 -DN G1 ) / (max-min)+240

[0181] Finally, after the above calculations, the values ​​H, S, V of hue H, saturation S, and brightness V in the HSV color gamut are obtained.

[0182] In the test method of the LED circuit board test device based on spectrum matching and correction described in this embodiment, the specific process of denoising, binarization and expansion of the luminous image in S52) is as follows:

[0183] S52.1): Read image data of the luminescent image and convert the image data into a grayscale image;

[0184] S52.2): Select an appropriate denoising method to smooth the acquired grayscale image to remove high-frequency noise;

[0185] S52.3): Binarize the smoothed grayscale image using an adaptive threshold binarization method to convert the grayscale image into a black-and-white binary image, i.e., segmentation is achieved by adaptive threshold segmentation, where the adaptive threshold is obtained by taking a compromise position based on the response value of the LED lamp 3 under test and the response value of the environmental background;

[0186] S52.4): Perform a morphological dilation operation on the binary black-and-white binary image. The dilation can fill small holes and discontinuous lines in the target area, making the target area more complete and continuous.

[0187] The expansion operation is performed using a structural element determined by an experimental method that is two orders of magnitude smaller than the size of the LED lamp 3 under test. The core size is selected experimentally based on the actual situation of the LED lamp 3 under test. The specific process is as follows:

[0188] S52.41): define an expansion kernel, that is, select the size and shape of the expansion kernel through experiments according to the actual situation of the LED lamp 3 under test;

[0189] S52.42): Apply the dilation operation, and use the defined dilation kernel to apply the dilation operation to the image to be processed. It should be noted that in the field of image processing, dilation is a morphological operation used to expand the boundaries of objects in the image. The key to dilation is to use a small matrix called a dilation kernel to operate the image. This kernel determines the specific method and effect of the dilation operation: choosing a kernel with a shape similar to the object to be processed can obtain better processing effects (such as using a cross or rectangular structure element to process lines); a small kernel will retain more image details, while a larger kernel may cause loss of details; the larger the kernel, the greater the amount of calculation, so the size and shape of the kernel in actual processing needs to strike a balance between effect and efficiency. Therefore, I will not go into details here.

[0190] In the present embodiment, a test method for an LED circuit board test device based on spectrum matching and correction is described. The calibration parameters in the measurement equation (S53) are obtained by precise spectrum radiation calibration of the LED circuit board test device based on spectrum matching and correction. The specific process of the spectrum radiation calibration is as follows:

[0191] The basic principle of spectral radiometric calibration is to use a standard radiation source with known spectral radiation characteristics, and to establish a mapping relationship between the output signal of the camera 8 and the actual spectral radiation brightness by measuring the response of the camera 8 to these known spectral radiations;

[0192] Assuming that the known spectral radiance of the standard light source is L0(λ), the output signal of the camera is S0(λ), and the spectral radiance response function of the camera is R(λ), then:

[0193] L0(λ)=R(λ)S0(λ)

[0194] Therefore, by measuring the known spectral radiance L0(λ) of the standard light source and the output signal S0(λ) of the camera, the spectral radiance response function R(λ) of the camera can be obtained:

[0195]

[0196] Since the camera's response to different known spectral radiation measurements has different signal-to-noise ratios, multiple standard radiation sources are used to perform multi-point spectral radiation calibration to improve the accuracy and reliability of the spectral radiation response function calibration curve;

[0197] Select multiple standard spectral radiation sources, and for each standard spectral radiation source, record the camera output signal S T1 (λ), S T2 (λ) and S T3 (λ);

[0198] According to the known spectral radiation L of each standard spectral radiation source T1 (λ), L T2 (λ) and L T3 (λ), calculate the camera's response function to each standard spectral radiation source:

[0199]

[0200] Fit the response function data of multiple standard spectral radiation sources to generate a comprehensive calibration curve R T (λ), and the fitting coefficients a0, a1, …, a are obtained by the least squares fitting method. n ;

[0201] After the spectral radiation calibration is completed, for the measurement of the unknown spectral radiation brightness to be measured, assuming that the output signal of the unknown spectral radiation source is S(λ), the unknown spectral radiation brightness L(λ) to be measured can be calculated by the following formula:

[0202] L(λ)=R T (λ)S(λ)

[0203] Among them, S(λ) is the output signal of the camera for the unknown spectrum, R T (λ) is the response function obtained through calibration, and L(λ) is the unknown spectral radiation brightness to be measured, so that the output signal of the camera 8 can be accurately converted into spectral radiation brightness information with clear physical meaning, thereby realizing spectral radiation calibration correction.

[0204] It should be noted that L(λ)=R T (λ)S(λ) is the measurement equation, brightness = calibration coefficient * V component in color pixel value;

[0205] Hue = calibration coefficient * H component in color pixel value. The actual pixel value during operation is the integral value within the bandpass spectrum range of the standard filter; if it is a through hole, it is the integral value within the luminous spectrum range of the entire LED being tested. During the operation, the basis for judging good or bad products is also the integral value of brightness and hue in a specific range of the spectrum.

[0206] In the test method of the LED circuit board test device based on spectrum matching and correction described in this embodiment, the specific process of judging whether the LED lamp 3 meets the requirements of high-precision optical applications and classifying good products and bad products in S6) is as follows:

[0207] If the LED light meets the requirements of high-precision optical applications, the good indicator light is turned on; if the LED light 3 does not meet the requirements of high-precision optical applications, the defect indicator light is turned on;

[0208] For products with defective indicators lit, the control system will mark the defect according to the specific defect type of brightness or chromaticity and the number of the LED module circuit board 2, and then take the defective products to the temporary storage of defective products;

[0209] For good products, the data processing system returns the test results to the control system to continue to execute the next test task of the customized process until all test processes are completed and all the optical performance parameters required by all LED lamps 3 are obtained;

[0210] The operator removes the LED module circuit board 2 and puts it into the original anti-static packaging. The test of this product is completed and the next product is tested.

[0211] The above description is only a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements without departing from the principle of the present invention, and these improvements should also be regarded as within the protection scope of the present invention.

Claims

1. A testing method for an LED circuit board testing device based on spectrum matching and correction, characterized in that: The specific test method includes the following steps: S1): Scanning the barcode to obtain the number on the LED module circuit board (2), that is, using a scanner to scan the barcode or QR code on the LED module circuit board (2) to obtain its unique number; S2): placing the LED module circuit board (2) and fixing it, that is, placing the LED module circuit board (2) on the workbench of the test machine (1), ensuring that the pins on the LED module circuit board (2) are accurately aligned with the connection points of the test equipment, and ensuring that the LED module circuit board (2) remains stable during the test; S3): Select and execute an automated test process, that is, the user selects a preset automated test process or customizes a new automated test process through the control terminal (9) according to the test requirements, and the control terminal (9) can start spectral radiation calibration in the control system, obtain traceable high-precision brightness and chromaticity test values ​​through measurement equation calculation, and is equipped with an integrating sphere (4) as a standard radiation source calibrator, which is installed on one side of the test light path, and the beam splitter (5) is installed on one side of the integrating sphere (4), ensuring that the output light of the integrating sphere (4) can be reflected into the light path of the imaging lens (6), the filter wheel (71) and the camera (8), so that absolute spectral radiation calibration can be performed at any time; S4): the control system in the control terminal (9) issues a test instruction to execute an automated test process, that is, the filter wheel control module in the control system commands the rotary drive mechanism (72) to drive the filter wheel (71) to rotate to the position of the filter (73) corresponding to the current working color of the LED light (3) on the LED module circuit board (2), ensuring that the center of the filter (73) is aligned with the center of the optical path formed by the imaging lens (6) and the camera (8) so as to perform the corresponding optical characteristic test. At the same time, the control terminal (9) issues an instruction to light up the LED light (3) on the LED module circuit board (2) and adjust it to the color component and working voltage to be tested; S5): image acquisition and data processing, i.e., the light beam emitted by the LED lamp (3) is collected by an optical system consisting of a beam splitter (5), an imaging lens (6), a filter (73) and a camera (8), and light signal image acquisition is completed according to the response of the light intensity on the image plane of the camera (8), the collected light signal image data is processed, and the optical performance index of the LED lamp (3) is calculated; S6): according to the result of the optical signal image data processing in S5), quality judgment is made and products are classified, that is, whether the LED lamp (3) meets the requirements of high-precision optical applications is judged, good products and bad products are classified, bad products are eliminated, and the next test task of the customized process is performed on the good products, until all the test processes are completed, and all the optical performance parameters required for all LED lamps (3) are obtained; S7): The operator removes the LED module circuit board (2) and puts it into the original anti-static packaging. The test of this product is completed and the next product is tested.

2. A testing method for an LED circuit board testing device based on spectrum matching and correction according to claim 1, characterized in that: The integrating sphere (4) is provided as a standard radiation source calibrator in the S3), and absolute spectral radiation calibration can be performed at any time. The specific process is as follows: S31): initial installation and calibration, that is, installing the integrating sphere (4) on one side of the test light path; fine-adjusting the center positions of the integrating sphere (4) and the beam splitter (5) in a plane perpendicular to the light path to ensure that the light path is aligned and the optical signal of the light path can be normally received by the camera (8); after the adjustment, the positions of the integrating sphere (4) and the beam splitter (5) are fixed and no longer changed; S32): Initial calibration, switch to calibration mode through the control system in the control terminal (9): turn on the integrating sphere (4), and record the output signal S of the camera (8) std (λ), according to the known spectral radiation L of the integrating sphere (4) std (λ), calculate the response function R(λ) of the camera (8): The calculated response function R(λ) is saved, and the absolute spectral radiation calibration coefficient in the measurement equation is updated for subsequent calculation of the camera (8) response value to photometric and chromaticity parameters with actual physical meaning; In the intervals of normal measurement process, absolute spectral radiation calibration can be performed at any time to ensure the accuracy and consistency of the measurement results; S33): After the absolute radiation calibration is completed, the control system in the control terminal (9) will switch back to the measurement mode.

3. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 2, characterized in that: S5): The camera (8) collects image data and processes the collected image data. The specific processing process is as follows: S51): the camera control module in the control terminal (9) commands the camera (8) to start collecting luminous image data of the LED lamp (3) on the LED module circuit board (2), and saves the collected luminous image data to a specified path; S52): automatically reading the luminous image data from the specified path in S51) through the data processing system in the control terminal (9), pre-processing the luminous image area, that is, converting the RGB color domain of the luminous image into the HSV color domain, and performing denoising, binarization, and expansion processing on the luminous image, and then using a circle recognition algorithm based on Hough transform to automatically locate the luminous area position of the LED lamp (3); S53): Calculate the mean values ​​of the hue component and the intensity component within the luminous range of the LED lamp (3), and bring the calculated data into the measurement equation, and finally obtain the accurate brightness value and chromaticity value of the measured LED lamp (3) according to the measurement equation result.

4. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 3, characterized in that: The circle recognition algorithm based on Hough transform in 52) uses the parametric equation of the circle to map the circle in the image space to the parameter space, and uses an accumulator to count the possible center and radius of the circle. By voting in the parameter space, the area in the image that is most likely to be a circle is found. The position and size of the circle can be effectively detected and located from an image, and good robustness can be shown even in the case of more noise or partial occlusion of the circle. The standard equation of a circle in Cartesian coordinates is: (xa) 2 +(yb) 2 =r 2 Among them, (a, b) is the coordinate of the center of the circle, r is the radius of the circle, and (x, y) is the edge point; In the Hough transform, the problem in the image space needs to be converted to the parameter space. For circle detection, each circle can be represented by three parameters: the coordinates of the center of the circle (a, b) and the radius r of the circle. Therefore, the parameter space will be a three-dimensional space. The specific process of the circle recognition algorithm based on Hough transform is as follows: S521): First, edge detection algorithm is used to extract edge points (x, y) in the luminous image. For each edge point (x, y) in the luminous image, all possible circle radii r are considered; For each circle with radius r, the center coordinates (a, b) must satisfy the following formula: a=x-rcos(θ), b=y-rsin(θ), θ∈[0, 2π]; S522): create a three-dimensional array (a, b, r) as an accumulator, whose dimensions correspond to the possible values ​​of a, b and r; For each edge point (x, y) and each polar coordinate angle value θ, calculate the corresponding a and b, and add 1 to the corresponding (a, b, r) position in the accumulator; After completing the voting of all edge points in the luminous image, the maximum value in the accumulator indicates the position where a circle may exist in the luminous image. By finding the local maximum value in the accumulator, the most likely center position and radius of the circle can be determined, thereby determining the position and luminous area of ​​the LED lamp (3).

5. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 4, characterized in that: In the above S52), the following strategies are also adopted in the circle recognition algorithm process based on Hough transform to improve the calculation efficiency and accuracy. The specific strategies are as follows: S5201): using the gradient direction as an estimate of the polar coordinate angle value θ to reduce unnecessary calculations of the polar coordinate angle value θ; S5202): Threshold processing is performed on the accumulator, and only candidate circles whose votes exceed the software setting and are verified by previous experiments to be suitable for the threshold of the current LED light under test (3) are considered; S5203): Non-maximum suppression is used on the accumulator to ensure that the positions of the detected candidate circles are not too close, so that the circle recognition algorithm based on Hough transform can effectively detect circular objects from noisy image data.

6. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 3, characterized in that: The specific process of converting the RGB color gamut into the HSV color gamut in S52) is as follows: Set DN R , D.N. G , D.N. B They are the red component, green component and blue component in the RGB color value, respectively, and H, S, V are the hue, saturation and brightness in the converted HSV color value; First, normalize the RGB values: R , D.N. G , D.N. B Convert the value from 0-255 range to 0-1 range DN R1 , D.N. G1 , D.N. B1 , and calculate DN R1 , D.N. G1 , D.N. B1 The maximum value max and the minimum value min in the component; DN R1 =DN R / 255.0 DN G1 =DN G / 255.0 DN B1 =DN B / 255.0 max=max(DN R1 ,DN G1 ,DN B1 ) min=min(DN R1 ,DN G1 ,DN B1 ) Then, calculate the brightness value V: V = max Next, calculate the saturation value S: When V≠0, S=(V-min) / V When V=0, then S=0 Again, calculate the hue value H: If max = min, then H = 0 If max = DN R1 And DN G1 ≥DN B1 , then H=60*(DN G1 -DN B1 ) / (max-min) If max = DN R1 And DN G1 <DN B1 , then H=60*(DN G1 -DN B1 ) / (max-min)+360 If max = DN G1 , then H=60*(DN B1 -DN R1 ) / (max-min)+120 If max = DN G1 , then H=60*(DN R1 -DN G1 ) / (max-min)+240 Finally, after the above calculations, the values ​​(H, S, V) of hue H, saturation S, and brightness V in the HSV color domain are obtained.

7. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 3, characterized in that: The specific process of denoising, binarizing and dilating the luminous image in S52) is as follows: S52.1): Read image data of the luminescent image and convert the image data into a grayscale image; S52.2): Select an appropriate denoising method to smooth the acquired grayscale image to remove high-frequency noise; S52.3): using an adaptive threshold binarization method to binarize the smoothed grayscale image, and converting the grayscale image into a black and white binary image, that is, by adaptive threshold segmentation, the adaptive threshold is obtained by taking a compromise position based on the response value of the LED lamp (3) under test and the response value of the environmental background; S52.4): performing a morphological dilation operation on the binary black-and-white binary image, so that small holes and discontinuous lines in the target area can be filled by dilation, making the target area more complete and continuous; The expansion operation is performed using a structural element determined by an experimental method that is two orders of magnitude smaller than the size of the LED lamp (3) being tested, and a core size selected by experiment based on the actual situation of the LED lamp (3) being tested. The specific process is as follows: S52.41): define dilation kernel; S52.42): Apply a dilation operation.

8. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 3, characterized in that: The calibration parameters in the measurement equation of S53) are obtained by precise spectral radiation calibration of the LED circuit board test device based on spectral matching and correction. The specific process of the spectral radiation calibration is as follows: The basic principle of spectral radiometric calibration is to use a standard radiation source with known spectral radiation characteristics, and to establish a mapping relationship between the output signal of the camera (8) and the actual spectral radiation brightness by measuring the response of the camera (8) to these known spectral radiations; Assuming that the known spectral radiance of the standard light source is L0(λ), the output signal of camera (8) is S0(λ), and the response function of camera (8) is R(λ), then: L0(λ)=R(λ)S0(λ) Therefore, by measuring the known spectral radiance L0(λ) of the standard light source and the output signal S0(λ) of the camera (8), the spectral radiance response function R(λ) of the camera (8) can be obtained: Since the signal-to-noise ratio of the camera (8) in response to different known spectral radiation measurements is different, multiple standard radiation sources are used to perform multi-point spectral radiation calibration to improve the accuracy and reliability of the spectral radiation response function calibration curve; Select a plurality of standard spectral radiation sources, and for each standard spectral radiation source, record the output signal S of the camera (8) T1 (λ), S T2 (λ) and S T3 (λ); According to the known spectral radiation L of each standard spectral radiation source T1 (λ), L T2 (λ) and L T3 (λ), calculate the response function of the camera (8) to each standard spectral radiation source: Fit the response function data of multiple standard spectral radiation sources to generate a comprehensive calibration curve R T (λ), and the fitting coefficients a0, a1, …, a are obtained by the least squares fitting method. n ; After the spectral radiation calibration is completed, for the measurement of the unknown spectral radiation brightness to be measured, assuming that the output signal of the unknown spectral radiation source is S(λ), the unknown spectral radiation brightness L(λ) to be measured can be calculated by the following formula: L(λ)=R T (λ)S(λ) Where S(λ) is the output signal of the camera (8) for the unknown spectrum, R T (λ) is the response function obtained through calibration, and L(λ) is the unknown spectral radiation brightness to be measured, so that the output signal of the camera (8) can be accurately converted into spectral radiation brightness information with clear physical meaning, thereby realizing spectral radiation calibration correction.

9. The testing method of the LED circuit board testing device based on spectrum matching and correction according to claim 1, characterized in that: The specific process of judging whether the LED lamp (3) meets the requirements of high-precision optical applications and classifying good products and bad products in S6) is as follows: If the LED light (3) meets the requirements of high-precision optical applications, a good product indicator light is turned on; if the LED light (3) does not meet the requirements of high-precision optical applications, a defect indicator light is turned on; For products with defective indicator lights, the control system will mark the defect according to the specific defect type of brightness or chromaticity and the number of the LED module circuit board (2), and then take the defective products to a temporary storage area for defective products; For good products, the data processing system returns the test results to the control system to continue to execute the next test task of the customized process until all test processes are completed and all the optical performance parameters required for all LED lamps (3) are obtained; The operator removes the LED module circuit board (2) and puts it into the original anti-static packaging. The test of this product is completed and the test of the next product is carried out.

10. An LED circuit board testing device based on spectrum matching and correction, characterized in that: include: A testing machine (1), on which an LED module circuit board (2) is placed, and on which at least one LED lamp (3) is arranged; An integrating sphere (4), wherein the integrating sphere (4) is installed on one side of a workbench of the testing machine (1); A beam splitter (5) is used to reflect light emitted by the integrating sphere (4) into a measuring light path formed by the beam splitter (5), an imaging lens (6), a filter (73) and a camera (8); the beam splitter (5) is connected to a back plate of a test machine (1) via a mounting frame, and the beam splitter (5) cooperates with the integrating sphere (4); An imaging lens (6) for focusing light; the imaging lens (6) is connected to the back plate of the test machine (1) via a mounting frame and is arranged above the beam splitter (5); A filter wheel mechanism (7) is used to eliminate the influence of ambient stray light and improve the measurement contrast. The filter wheel mechanism (7) is mounted on the back plate of the test machine (1) via a mounting bracket. The filter wheel mechanism (7) comprises a filter wheel (71) and a rotation drive mechanism (72) for driving the filter wheel (71) to rotate. The rotation drive mechanism (72) is connected to the back plate of the test machine (1) via a mounting bracket, and its output end cooperates with the filter wheel (71). The filter wheel (71) is provided with at least one filter (73) and at least one hollow hole (74). The filter (73) matches the spectrum of the LED lamp (3) to be tested. A camera (8) for receiving and measuring optical signals, wherein the camera (8) is connected to a back plate of the test machine (1) via a camera mounting frame, and the filter wheel (71) is arranged between the imaging lens (6) and the camera (8); A control terminal (9) is used to control the lighting of an LED lamp (3) on an LED module circuit board (2), the rotation of a filter wheel (71) and the image acquisition operation of a camera (8); the LED module circuit board (2), the integrating sphere (4), the rotation drive mechanism (72) and the camera (8) are all connected to the control terminal (9); The light beam emitted by the LED lamp (3) is projected onto the camera (8) through a beam splitter (5), an imaging lens (6), and a filter (73), and an image is formed on the camera (8). The light signal image is collected based on the response of the light intensity on the image plane of the camera (8), the collected light signal image data is processed, and the optical performance index of the LED lamp (3) is calculated.

Images