Method, device and system for detecting omnibearing defects of light-emitting element based on multi-view imaging

Abstract

The invention provides a light-emitting element omnibearing defect detection method, device and system based on multi-view imaging, and the method comprises the steps: arranging a to-be-detected display module on a defect detection tool, and obtaining the detection data of a light-emitting element in the display module and a plurality of spectral images of different views; performing preprocessing and feature extraction on each spectral image to obtain feature data; and judging according to the feature data and the detection data to obtain a detection result. According to the light-emitting element omni-directional defect detection method based on multi-view-angle imaging, multiple spectral images of different view angles of the light-emitting element are adopted, judgment is conducted in combination with detection data and feature data, one-time omni-directional defect detection is achieved, multiple key quality indexes such as appearance, photoelectric performance and uniformity are covered, the detection dimension is enriched, and the detection efficiency is improved. The detection efficiency is improved, the omission ratio is reduced, the spectral image is collected in all directions without dead angles, and the detection blind area of a single visual angle is eliminated.

Description

Technical Field

[0001] This application belongs to the field of lighting defect detection technology, and in particular relates to a method, device and system for omnidirectional defect detection of light-emitting elements based on multi-view imaging. Background Technology

[0002] Existing methods for defect detection in LED display modules include CCD image acquisition, manual visual inspection, single-view inspection, and point scanning. Traditional manual visual inspection identifies defects in the LED display module from the image, but this method suffers from low efficiency, high subjectivity, fatigue, inconsistent standards, and a high rate of missed defects, failing to meet the efficiency requirements of modern production lines. Single-view inspection involves photographing only the emitting surface of the LEDs from directly above the display module, failing to detect issues such as scratches or missing adhesive on the sides, electrode oxidation, and abnormal side light emission or uneven color distribution caused by internal gold wire breaks or die bonding defects. Single-view inspection also has blind spots and a high rate of missed defects. Point scanning inspection is inefficient and cannot perform rapid full inspection. Summary of the Invention

[0003] One embodiment of this application provides a method, apparatus, and system for omnidirectional defect detection of light-emitting elements based on multi-view imaging, in order to solve the problems of incomplete orientation, low detection efficiency, and easy omissions in existing defect detection methods for LED lamp bead display modules.

[0004] In a first aspect, one embodiment of this application provides a method for omnidirectional defect detection of light-emitting elements based on multi-view imaging, comprising the following steps:

[0005] The display module to be tested is placed on the defect detection fixture, and the detection data of the light-emitting element in the display module and multiple spectral images from different perspectives are obtained.

[0006] Each of the spectral images is preprocessed and its features are extracted to obtain feature data;

[0007] The detection result is obtained by judging based on the feature data and the detection data.

[0008] Optionally, the feature data includes appearance data, average image brightness value, image grayscale standard deviation, and color coordinates; the detection data includes forward current, reverse current, luminous flux, and luminous intensity. Based on the feature data and the detection data, a judgment is made to obtain the detection result, including:

[0009] If the appearance data contains any one or more of the following: scratches, black spots, impurities, and bubbles, the test result is that the appearance of the light-emitting element is unqualified.

[0010] If the luminous intensity and / or the luminous flux are not within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is unqualified;

[0011] If the reverse current is greater than the forward current, the detection result indicates that the light-emitting element is leaking current.

[0012] If the color coordinates are not within the target set area, the detection result is that the color of the light-emitting element is unqualified;

[0013] If the average brightness value of the image is not greater than the set lower brightness threshold or the standard deviation of the image grayscale is greater than the set grayscale threshold, then the detection result is that the grayscale distribution of the light-emitting surface of the light-emitting element is uneven.

[0014] Optionally, the multi-view imaging-based omnidirectional defect detection method for light-emitting elements further includes: acquiring a transmission image of the light-emitting element in the display module; if the transmission image shows gold wire breakage and / or chip die bonding voids, the detection result indicates that there is an internal defect in the light-emitting element.

[0015] Optionally, each of the spectral images is preprocessed and its features are extracted to obtain feature data including:

[0016] Each of the spectral images is subjected to grayscale conversion, noise reduction, and image enhancement processing to obtain a processed image corresponding to each of the spectral images;

[0017] All the processed images are localized and segmented into regions of interest to obtain image data containing the light-emitting area, electrode area, and package edge;

[0018] Feature data is obtained by extracting grayscale features, RGB color components, geometric features, and texture features from all the image data.

[0019] Optionally, acquiring multiple spectral images of the light-emitting element in the display module from different viewing angles includes: using a camera device to acquire multiple spectral images of the light-emitting element from above, side, and below in visible light mode or a specific wavelength mode; or, acquiring detection data of the light-emitting element in the display module includes: applying a rated operating current to the light-emitting element from the forward and reverse directions respectively to obtain forward current and reverse current; applying a standard light source to the light-emitting element to obtain luminous flux and luminous intensity.

[0020] Secondly, one embodiment of this application provides a multi-view imaging-based omnidirectional defect detection device for light-emitting elements, comprising: a data acquisition unit, a judgment and detection unit, and a data processing and feature extraction unit;

[0021] The data acquisition unit is used to set the display module to be inspected on the defect detection fixture, and acquire the detection data of the light-emitting element in the display module and multiple spectral images from different perspectives;

[0022] The data processing and feature extraction unit is used to preprocess and extract features from each of the spectral images to obtain feature data;

[0023] The judgment and detection unit is used to make a judgment based on the feature data and the detection data to obtain a detection result.

[0024] Optionally, the feature data includes appearance data, average image brightness value, image grayscale standard deviation, and color coordinates; the detection data includes forward current, reverse current, luminous flux, and luminous intensity; and the judgment detection unit includes:

[0025] If the appearance data contains any one or more of the following: scratches, black spots, impurities, and bubbles, the test result is that the appearance of the light-emitting element is unqualified.

[0026] If the luminous intensity and / or the luminous flux are not within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is unqualified;

[0027] If the reverse current is greater than the forward current, the detection result indicates that the light-emitting element is leaking current.

[0028] If the color coordinates are not within the target set area, the detection result is that the color of the light-emitting element is unqualified;

[0029] If the average brightness value of the image is not greater than the set lower brightness threshold or the standard deviation of the image grayscale is greater than the set grayscale threshold, then the detection result is that the grayscale distribution of the light-emitting surface of the light-emitting element is uneven.

[0030] Optionally, the data acquisition unit is further configured to acquire a transmission image of the light-emitting element in the display module, and the judgment and detection unit is further configured to determine that there is an internal defect in the light-emitting element if the transmission image shows gold wire breakage and / or chip die bonding voids.

[0031] Optionally, the data processing and feature extraction unit includes a preprocessing subunit, a segmentation subunit, and a feature extraction subunit;

[0032] The preprocessing subunit is used to perform grayscale conversion, noise reduction, and image enhancement processing on each of the spectral images to obtain a processed image corresponding to each of the spectral images;

[0033] The segmentation subunit is used to locate and segment the region of interest in all the processed images to obtain image data containing the light-emitting area, the electrode area, and the edge of the package.

[0034] The feature extraction subunit is used to extract grayscale features, RGB color components, geometric features, and texture features from all the image data to obtain feature data.

[0035] Secondly, one embodiment of this application provides a multi-view imaging-based omnidirectional defect detection system for light-emitting elements, including a defect detection fixture and an industrial control computer. The defect detection fixture is provided with a detection station for placing the light-emitting elements in the display module. The industrial control computer performs automated defect detection on the light-emitting elements in the detection station according to the multi-view imaging-based omnidirectional defect detection method described above.

[0036] One embodiment of this application provides a method, apparatus, and system for omnidirectional defect detection of light-emitting elements based on multi-view imaging. The method includes placing the display module to be inspected on a defect detection fixture, acquiring detection data of the light-emitting elements in the display module and multiple spectral images from different viewpoints; preprocessing and feature extraction of each spectral image to obtain feature data; and making a judgment based on the feature data and detection data to obtain the detection result. This omnidirectional defect detection method for light-emitting elements uses multiple spectral images of the light-emitting elements from different viewpoints, combined with detection data and feature data, to achieve one-time omnidirectional defect detection, covering multiple key quality indicators such as appearance, photoelectric performance, and uniformity. It enriches the detection dimensions, not only improving detection efficiency and reducing the false negative rate, but also achieving omnidirectional, blind-spot-free acquisition of spectral images, eliminating the detection blind spots of a single viewpoint, and solving the problems of incomplete orientation, low detection efficiency, and easy false negatives in existing defect detection methods for LED light bead display modules.

[0037] This multi-view imaging-based all-around defect detection device for light-emitting elements achieves automated detection of display modules from all angles (including top and side surfaces) and from multiple dimensions (including appearance, brightness, color, and uniformity) through a data acquisition unit, a judgment and detection unit, and a data processing and feature extraction unit. It can effectively detect potential internal defects, thereby improving detection accuracy and reducing the false negative rate.

[0038] This multi-view imaging-based all-around defect detection system for light-emitting elements can simultaneously detect appearance defects (such as scratches, contamination, and missing adhesive), photoelectric parameters (such as brightness, color temperature, and wavelength), and internal structural defects (such as broken gold wires, chip cracks, and poor die bonding) of light-emitting elements. It enables automated detection of display modules from all angles (including top and side surfaces) and from multiple dimensions (including appearance, brightness, color, and uniformity), and can effectively detect potential internal defects, thereby improving detection accuracy and reducing the false negative rate. Attached Figure Description

[0039] To more clearly illustrate the technical solution in one embodiment of this application, the accompanying drawings used in the description of the embodiment will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0040] To gain a more complete understanding of this application and its beneficial effects, the following description will be provided in conjunction with the accompanying drawings. In the following description, the same reference numerals denote the same parts.

[0041] Figure 1 A flowchart illustrating the steps of a method for omnidirectional defect detection of light-emitting elements based on multi-view imaging, provided as an embodiment of this application.

[0042] Figure 2 This is a schematic diagram of a multi-view imaging-based omnidirectional defect detection system for light-emitting elements, provided as an embodiment of this application.

[0043] Figure 3 A qualified front-emitting image of a multi-view imaging-based omnidirectional defect detection method for light-emitting elements provided in one embodiment of this application.

[0044] Figure 4 A frontal image of a defect with dark spots provided by a multi-view imaging-based omnidirectional defect detection method for light-emitting elements, as an embodiment of this application.

[0045] Figure 5 A schematic diagram of the framework of an all-around defect detection device for light-emitting elements based on multi-view imaging, provided as an embodiment of this application. Detailed Implementation

[0046] The technical solution of one embodiment of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0047] One embodiment of this application provides a method, apparatus, and system for omnidirectional defect detection of light-emitting elements based on multi-view imaging, in order to solve the problems of incomplete orientation, low detection efficiency, and easy omissions in existing defect detection methods for LED lamp bead display modules.

[0048] Patent terminology:

[0049] Region of Interest (ROI) is an image analysis method used in machine vision and image processing to delineate regions to be processed. It uses bounding boxes, circles, ellipses, or irregular polygons to outline specific areas, reducing computational load and focusing on key information processing. This region can be obtained through pixel matrix cropping or operator calculation, and is applied in fields such as face recognition, image segmentation, and object detection, supporting dynamic tracking and multi-region processing. ROI implementation methods include fixed-region encoding and dynamic tracking encoding. Combined with visual attention mechanisms, salient regions in stereoscopic videos can be extracted, and targets can be located through differences in features such as brightness, texture, and depth. Platforms such as Halcon, OpenCV, and Matlab provide slicing operations, mask definitions, and neural network fusion techniques to optimize boundary accuracy, and combine with the H.265 encoding standard to improve the efficiency of high-definition video transmission.

[0050] A standard light source is an artificial light source whose radiation is similar to that of a CIE standard illuminator, as specified by the CIE.

[0051] Example 1:

[0052] One embodiment of this application provides a method for omnidirectional defect detection of light-emitting elements based on multi-view imaging. For an example, please refer to [link to example]. Figure 1 , Figure 1 A flowchart illustrating the steps of a method for omnidirectional defect detection of light-emitting elements based on multi-view imaging, as provided in one embodiment of this application. Figure 2 This is a schematic diagram of a multi-view imaging-based omnidirectional defect detection system for light-emitting elements, provided as an embodiment of this application.

[0053] like Figure 1 As shown, this application invention provides a method for omnidirectional defect detection of light-emitting elements based on multi-view imaging, including the following steps:

[0054] S100. Set the display module to be inspected on the defect inspection fixture, and acquire the inspection data of the light-emitting elements in the display module and multiple spectral images from different perspectives;

[0055] S200. Preprocess and extract features from each spectral image to obtain feature data;

[0056] S300. Make a judgment based on the feature data and detection data to obtain the detection result.

[0057] To further explain, such as Figure 2 As shown, in step S100, the display module to be inspected is first installed on a defect detection fixture. Then, based on the defect detection fixture, detection data of the light-emitting elements in the display module and multiple spectral images from different viewing angles are acquired to provide data for subsequent steps. In step S200, each spectral image acquired in step S100 is preprocessed and feature extracted to obtain feature data, which provides data for subsequent step S300. In step S300, the detection data obtained in step S100 and the feature data obtained in step S200 are compared with preset standard thresholds or qualified sample data to obtain the detection result. The detection result includes the appearance judgment, photoelectric parameter judgment, colorimetric parameter judgment, and uniformity judgment of the light-emitting elements. The light-emitting elements can be selected as LED beads.

[0058] Furthermore, this multi-view imaging-based omnidirectional defect detection method for light-emitting elements acquires multiple spectral images of the light-emitting element from different perspectives through step S100, achieving omnidirectional, blind-spot-free spectral image acquisition and eliminating the detection blind spots of a single perspective. It can effectively detect side and internal defects. This method uses detection data and feature data for judgment, achieving one-time omnidirectional defect detection, covering multiple key quality indicators such as appearance, photoelectric performance, and uniformity, thus enriching the detection dimensions. The detection results obtained by this method also improve the accuracy of omnidirectional defect detection of light-emitting elements, with a low false negative rate. Automated omnidirectional defect detection of light-emitting elements using this multi-view imaging-based method significantly improves detection efficiency.

[0059] One embodiment of this application provides a method for omnidirectional defect detection of light-emitting elements based on multi-view imaging. The method includes placing the display module to be inspected on a defect detection fixture, acquiring detection data of the light-emitting elements in the display module and multiple spectral images from different viewpoints; preprocessing and feature extraction of each spectral image to obtain feature data; and making a judgment based on the feature data and detection data to obtain the detection result. This method for omnidirectional defect detection of light-emitting elements based on multi-view imaging uses multiple spectral images of the light-emitting elements from different viewpoints, combined with detection data and feature data, to achieve one-time omnidirectional defect detection. It covers multiple key quality indicators such as appearance, photoelectric performance, and uniformity, enriching the detection dimensions. This not only improves detection efficiency and reduces the false negative rate but also achieves omnidirectional, blind-spot-free acquisition of spectral images, eliminating blind spots in single-view detection. It solves the problems of incomplete orientation, low detection efficiency, and easy false negatives in existing defect detection methods for LED light bead display modules.

[0060] In one embodiment of the invention, the feature data includes appearance data, average image brightness value, image grayscale standard deviation, and color coordinates; the detection data includes forward current, reverse current, luminous flux, and luminous intensity. Based on the feature data and the detection data, a judgment is made to obtain the detection result, which includes:

[0061] If the appearance data shows any one or more of the following: scratches, black spots, impurities, or bubbles, the test result is that the appearance of the light-emitting element is unqualified.

[0062] If the luminous intensity and / or luminous flux are not within the set allowable range, the test result is that the photoelectric performance of the light-emitting element is unqualified;

[0063] If the reverse current is greater than the forward current, the detection result indicates that the light-emitting element is leaking current.

[0064] If the color coordinates are not within the target set area, the test result is that the color of the light-emitting element is unqualified;

[0065] If the average brightness value of the image is not greater than the set lower limit threshold or the standard deviation of the image grayscale is greater than the set grayscale threshold, the detection result is that the grayscale distribution of the light-emitting surface of the light-emitting element is uneven.

[0066] To further clarify, while the testing standards are physically consistent across different light-emitting elements, the specific parameter thresholds differ. During the appearance assessment of light-emitting elements, after preprocessing the spectral image, image algorithms (such as the grayscale standard deviation method for uniformity analysis and Blob analysis (connected component analysis) for defect localization, among other mature technologies, can be used to identify abnormal areas and obtain appearance data. Then, the appearance data is used to determine whether the light-emitting element's appearance is acceptable based on factors such as the integrity of the outline, the presence of scratches, black spots, impurities, and bubbles. For example, if the appearance data package contains incomplete outlines, scratches, black spots, impurities, and / or bubbles, the light-emitting element's appearance is deemed unacceptable; conversely, if the appearance data package contains complete outlines, no scratches, no black spots, no impurities, and no bubbles, the light-emitting element's appearance is deemed acceptable.

[0067] In the embodiments of the invention, during the determination of the photoelectric parameters of the light-emitting element, if the luminous intensity and / or luminous flux are not within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is unqualified. If both the luminous intensity and luminous flux are within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is qualified. The set allowable range can be set according to actual needs and is not specifically limited here.

[0068] In the embodiments of the invention, during the determination of the photoelectric parameters of the light-emitting element, if the reverse current is greater than the forward current, the detection result is that the light-emitting element is leaking current; if the reverse current is not greater than the forward current, the detection result is that the light-emitting element is not leaking current.

[0069] In the embodiments of the invention, during the determination of the colorimetric parameters of the light-emitting element, an elliptical tolerance range centered on the target chromaticity coordinates is set on the CIE1931 chromaticity diagram. If the color coordinates are not within the target range, the detection result is that the color of the light-emitting element is unqualified; if the color coordinates are within the target range, the detection result is that the color of the light-emitting element is qualified. The target range is automatically adjusted based on the light-emitting element to be tested.

[0070] In the embodiments of the invention, the purpose of determining the uniformity of the light-emitting element is to judge whether the gray-scale distribution of the light-emitting surface of the light-emitting element is uniform. In the process of determining the uniformity of the light-emitting element, if a monochrome camera and filter are used to acquire a spectral image, then if the obtained image gray-scale standard deviation is greater than a set gray-scale threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is non-uniform; if the obtained image gray-scale standard deviation is not greater than the set gray-scale threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is uniform. If a color camera is used to acquire a spectral image, then if the obtained image average brightness value is not greater than a set brightness lower limit threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is non-uniform; if the obtained image average brightness value is greater than the set brightness lower limit threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is uniform. The set brightness lower limit threshold and the set gray-scale threshold can be set according to actual needs, and no specific numerical limit is specified here.

[0071] In one embodiment of the invention, the multi-view imaging-based omnidirectional defect detection method for light-emitting elements further includes: acquiring a transmission image of the light-emitting element in the display module; if the transmission image shows gold wire breakage and / or chip die bonding voids, the detection result indicates that there is an internal defect in the light-emitting element.

[0072] To further explain, a light source is placed below the light-emitting element (LED), and an imaging device (such as a camera) is placed above it. Transmission images are acquired when the LED is not powered on, and these are used to detect shadow defects in internal structures such as chips and gold wires. The projected image from the front is combined with the transmission image from the bottom and then compared with a standard projected image of the LED to comprehensively infer the presence of internal defects such as broken gold wires or chip bonding voids. If the transmission image shows broken gold wires and / or chip bonding voids, the detection result indicates that there are internal defects in the LED. If the transmission image does not show broken gold wires or chip bonding voids, the detection result indicates that there are no internal defects in the LED.

[0073] In the embodiments of the invention, this multi-view imaging-based omnidirectional defect detection method for light-emitting elements does not judge each spectral image in isolation, but rather performs correlation analysis between frontal image features and side image features, comprehensively judges multiple indicators, and classifies defects, thereby achieving intelligent inference of internal and external defects. This multi-view imaging-based omnidirectional defect detection method for light-emitting elements combines multiple imaging modes such as active light emission (powered on) and passive transmission (external light source) to achieve the detection of different types of defects in light-emitting elements.

[0074] In the embodiments of the invention, each spectral image is preprocessed and its features are extracted to obtain feature data including:

[0075] Each spectral image is processed by grayscale conversion, noise reduction, and image enhancement to obtain a processed image corresponding to each spectral image;

[0076] All processed images are localized and segmented into regions of interest to obtain image data containing the light-emitting area, electrode area, and package edge;

[0077] Feature data is obtained by extracting grayscale features, RGB color components, geometric features, and texture features from all image data.

[0078] To further clarify, texture features refer to the use of Local Binary Pattern (LBP), which generates binary codes by comparing the grayscale differences between the center pixel and its neighboring pixels, suitable for texture classification. RGB color components directly extract the R / G / B values ​​of the original image as features, commonly used for color analysis or image enhancement. Geometric features include quantifying the degree to which a shape approximates a circle and obtaining the number of pixels and perimeter of the target region through contour detection, used for shape analysis. Grayscale features refer to calculating the grayscale mean from the R / G / B values ​​and representing the brightness value using the grayscale standard deviation. This multi-view imaging-based omnidirectional defect detection method for light-emitting elements employs digital image processing techniques such as grayscale conversion, noise reduction, image enhancement, and image segmentation for quantitative analysis, avoiding subjective human error, being sensitive to minute defects, improving detection accuracy, and reducing the false negative rate.

[0079] like Figure 2 As shown, in one embodiment of the invention, acquiring multiple spectral images of the light-emitting element in the display module from different viewing angles includes: using a camera device to acquire multiple spectral images of the light-emitting element from above, side, and below in visible light mode or a specific wavelength mode; or, acquiring detection data of the light-emitting element in the display module includes: applying a rated operating current to the light-emitting element from the forward and reverse directions respectively to obtain a forward current and a reverse current; applying a standard light source to the light-emitting element to obtain luminous flux and luminous intensity.

[0080] To further explain, the display module to be inspected is precisely placed on the inspection station of the defect inspection fixture, and its rated operating current is applied through probes or clamps, or different driving currents and duty cycles are set according to inspection requirements, to enable it to emit light normally, and the forward and reverse currents are detected. Using multiple high-resolution imaging devices (such as CCD cameras) arranged around the defect inspection fixture, an architecture employing at least one front-facing camera and one side-facing camera working in tandem, omnidirectional inspection is achieved, acquiring images of the light-emitting element synchronously or sequentially from different angles. The architecture includes at least: a front-facing camera: vertically aligned with the light-emitting surface of the light-emitting element, acquiring orthogonal spectral images for analyzing brightness distribution uniformity, color spots, dark spots, and light-emitting surface contamination. At least one side-facing camera: aligned with the side of the light-emitting element at a certain tilt angle (e.g., 30-60 degrees), acquiring oblique spectral images for detecting side packaging defects, electrode conditions, and observing the fit between the light-emitting chip and phosphor from the side. Transmitted light imaging equipment: A light source is placed below the light-emitting element, and an imaging device is placed above the light-emitting element. Transmitted images are acquired when the light-emitting element is not powered on, and are used to detect shadow defects in internal structures such as chips and gold wires. Visible light mode refers to using a color CCD camera to acquire RGB images for colorimetric analysis. Specific wavelength mode refers to adding filters (such as blue and red filters) in front of the camera lens, or using a monochrome CCD camera with a specific wavelength light source to separate the blue light excitation light from the phosphor excitation light of the light-emitting element, thereby more accurately locating defects in the chip or phosphor.

[0081] Figure 3 A qualified front-facing light-emitting image of a multi-view imaging-based omnidirectional defect detection method for light-emitting elements, provided as an embodiment of this application. Figure 4 A frontal image of a defect with dark spots provided by a multi-view imaging-based omnidirectional defect detection method for light-emitting elements, as an embodiment of this application.

[0082] In embodiments of the present invention, such as Figures 2 to 4As shown, a 2-megapixel color CCD camera is selected as the front-facing camera, and a 1-megapixel monochrome CCD camera is used as the side-facing camera (tilted at a 45-degree angle). The display module is delivered to the detection position by a conveyor belt driven by a stepper motor, and is fixed and powered by a pneumatic clamp. The detection process is as follows: After the light-emitting element (such as an LED bead) is in place, a 20mA forward current is applied. This triggers the front and side cameras to simultaneously expose and acquire images. The images are transmitted to an industrial control computer, where software (developed based on libraries such as Halcon / OpenCV) performs image processing. The software first binarizes the spectral image of the front, locates the area of ​​the light-emitting element, and calculates the average brightness value of the image (if it is lower than the set lower brightness threshold X, it is judged as unqualified in brightness); it also calculates the grayscale standard deviation within the area (if it is higher than the set upper grayscale threshold Y, it is judged as unqualified in uniformity). Subsequently, edge detection is performed on the spectral image of the side to check whether the package outline is complete and smooth. Finally, based on all the criteria, a "qualified" or "unqualified" signal is output.

[0083] Furthermore, this multi-view imaging-based omnidirectional defect detection method for light-emitting elements can simultaneously detect surface defects (such as scratches, contamination, and missing adhesive), photoelectric parameters (such as brightness, color temperature, and wavelength), and internal structural defects (such as broken gold wires, chip cracks, and poor die bonding) of light-emitting elements. This enables automated inspection of display modules from all angles (including top and side surfaces) and multiple dimensions (including appearance, brightness, color, and uniformity). It can also effectively identify potential internal defects that are difficult to detect using traditional methods, thereby improving detection accuracy and reducing the false negative rate. This multi-view imaging-based omnidirectional defect detection method for light-emitting elements can achieve automated, comprehensive, and high-precision inspection in the production process, thereby improving yield and reducing costs.

[0084] Example 2:

[0085] Figure 5 A schematic diagram of the framework of an all-around defect detection device for light-emitting elements based on multi-view imaging, provided as an embodiment of this application.

[0086] like Figure 5 As shown, the present invention also provides a multi-view imaging-based omnidirectional defect detection device for light-emitting elements, comprising: a data acquisition unit 100, a judgment and detection unit 200, and a data processing and feature extraction unit 300.

[0087] The data acquisition unit 100 is used to set the display module to be inspected on the defect detection fixture and acquire the detection data of the light-emitting element in the display module and multiple spectral images from different perspectives.

[0088] The data processing and feature extraction unit 200 is used to preprocess and extract features from each spectral image to obtain feature data;

[0089] The judgment and detection unit 300 is used to make judgments based on feature data and detection data to obtain detection results.

[0090] In the embodiments of the invention, the feature data includes appearance data, average image brightness value, image grayscale standard deviation, and color coordinates; the detection data includes forward current, reverse current, luminous flux, and luminous intensity; and the determination and detection unit includes:

[0091] If the appearance data shows any one or more of the following: scratches, black spots, impurities, or bubbles, the test result is that the appearance of the light-emitting element is unqualified.

[0092] If the luminous intensity and / or luminous flux are not within the set allowable range, the test result is that the photoelectric performance of the light-emitting element is unqualified;

[0093] If the reverse current is greater than the forward current, the detection result indicates that the light-emitting element is leaking current.

[0094] If the color coordinates are not within the target set area, the test result is that the color of the light-emitting element is unqualified;

[0095] If the average brightness value of the image is not greater than the set lower limit threshold or the standard deviation of the image grayscale is greater than the set grayscale threshold, the detection result is that the grayscale distribution of the light-emitting surface of the light-emitting element is uneven.

[0096] To further clarify, while the testing standards are physically consistent across different light-emitting elements, the specific parameter thresholds differ. During the appearance assessment of light-emitting elements, after preprocessing the spectral image, image algorithms (such as the grayscale standard deviation method for uniformity analysis and Blob analysis (connected component analysis) for defect localization, among other mature technologies, can be used to identify abnormal areas and obtain appearance data. Then, the appearance data is used to determine whether the light-emitting element's appearance is acceptable based on factors such as the integrity of the outline, the presence of scratches, black spots, impurities, and bubbles. For example, if the appearance data package contains incomplete outlines, scratches, black spots, impurities, and / or bubbles, the light-emitting element's appearance is deemed unacceptable; conversely, if the appearance data package contains complete outlines, no scratches, no black spots, no impurities, and no bubbles, the light-emitting element's appearance is deemed acceptable.

[0097] In the embodiments of the invention, during the determination of the photoelectric parameters of the light-emitting element, if the luminous intensity and / or luminous flux are not within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is unqualified. If both the luminous intensity and luminous flux are within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is qualified. The set allowable range can be set according to actual needs and is not specifically limited here.

[0098] In the embodiments of the invention, during the determination of the photoelectric parameters of the light-emitting element, if the reverse current is greater than the forward current, the detection result is that the light-emitting element is leaking current; if the reverse current is not greater than the forward current, the detection result is that the light-emitting element is not leaking current.

[0099] In the embodiments of the invention, during the determination of the colorimetric parameters of the light-emitting element, an elliptical tolerance range centered on the target chromaticity coordinates is set on the CIE1931 chromaticity diagram. If the color coordinates are not within the target range, the detection result is that the color of the light-emitting element is unqualified; if the color coordinates are within the target range, the detection result is that the color of the light-emitting element is qualified. The target range is automatically adjusted based on the light-emitting element to be tested.

[0100] In the embodiments of the invention, the purpose of determining the uniformity of the light-emitting element is to judge whether the gray-scale distribution of the light-emitting surface of the light-emitting element is uniform. In the process of determining the uniformity of the light-emitting element, if a monochrome camera and filter are used to acquire a spectral image, then if the obtained image gray-scale standard deviation is greater than a set gray-scale threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is non-uniform; if the obtained image gray-scale standard deviation is not greater than the set gray-scale threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is uniform. If a color camera is used to acquire a spectral image, then if the obtained image average brightness value is not greater than a set brightness lower limit threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is non-uniform; if the obtained image average brightness value is greater than the set brightness lower limit threshold, the detection result is that the gray-scale distribution of the light-emitting surface of the light-emitting element is uniform. The set brightness lower limit threshold and the set gray-scale threshold can be set according to actual needs, and no specific numerical limit is specified here.

[0101] In the embodiments of the invention, the data acquisition unit 100 is further configured to acquire a transmission image of the light-emitting element in the display module, and the judgment detection unit is further configured to determine that there is an internal defect in the light-emitting element if the transmission image shows a broken gold wire and / or a chip die-bonding void.

[0102] To further explain, a light source is placed below the light-emitting element (LED), and an imaging device (such as a camera) is placed above it. Transmission images are acquired when the LED is not powered on, and these are used to detect shadow defects in internal structures such as chips and gold wires. The projected image from the front is combined with the transmission image from the bottom and then compared with a standard projected image of the LED to comprehensively infer the presence of internal defects such as broken gold wires or chip bonding voids. If the transmission image shows broken gold wires and / or chip bonding voids, the detection result indicates that there are internal defects in the LED. If the transmission image does not show broken gold wires or chip bonding voids, the detection result indicates that there are no internal defects in the LED.

[0103] In the embodiments of the invention, the data processing and feature extraction unit 300 includes a preprocessing subunit, a segmentation subunit, and a feature extraction subunit.

[0104] The preprocessing subunit is used to perform grayscale conversion, noise reduction, and image enhancement on each spectral image to obtain a processed image corresponding to each spectral image;

[0105] The segmentation subunit is used to locate and segment the region of interest in all processed images to obtain image data containing the light-emitting area, electrode area, and package edge;

[0106] The feature extraction subunit is used to extract grayscale features, RGB color components, geometric features, and texture features from all image data to obtain feature data.

[0107] Example 3:

[0108] like Figure 2 As shown, the present invention also provides a multi-view imaging-based omnidirectional defect detection system for light-emitting elements, including a defect detection fixture and an industrial control computer. The defect detection fixture is provided with a detection station for placing light-emitting elements in the display module. The industrial control computer performs automated defect detection on the light-emitting elements in the detection station according to the above-mentioned multi-view imaging-based omnidirectional defect detection method for light-emitting elements.

[0109] Furthermore, the content of the multi-view imaging-based all-round defect detection method for light-emitting elements is described in Embodiment 1. This multi-view imaging-based all-round defect detection system for light-emitting elements can simultaneously detect appearance defects (such as scratches, contamination, and missing adhesive), photoelectric parameters (such as brightness, color temperature, and wavelength), and internal structural defects (such as broken gold wires, chip cracks, and poor die bonding) of light-emitting elements. It can realize the automated detection of display modules in all directions (including top surface and side surface) and in multiple dimensions (including appearance, brightness, color, and uniformity), and can effectively discover potential internal defects, thereby improving detection accuracy and reducing the false negative rate.

[0110] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more features.

[0111] The above provides a detailed description of an embodiment of the method for omnidirectional defect detection of light-emitting elements based on multi-view imaging. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for omnidirectional defect detection of light-emitting elements based on multi-view imaging, characterized in that, Includes the following steps: The display module to be tested is placed on the defect detection fixture, and the detection data of the light-emitting element in the display module and multiple spectral images from different perspectives are obtained. Each of the spectral images is preprocessed and its features are extracted to obtain feature data; The detection result is obtained by judging based on the feature data and the detection data.

2. The method for omnidirectional defect detection of light-emitting elements based on multi-view imaging according to claim 1, characterized in that, The feature data includes appearance data, average image brightness value, image grayscale standard deviation, and color coordinates. The detection data includes forward current, reverse current, luminous flux, and luminous intensity. Based on the feature data and the detection data, a judgment is made to obtain the following detection results: If the appearance data contains any one or more of the following: scratches, black spots, impurities, and bubbles, the test result is that the appearance of the light-emitting element is unqualified. If the luminous intensity and / or the luminous flux are not within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is unqualified; If the reverse current is greater than the forward current, the detection result indicates that the light-emitting element is leaking current. If the color coordinates are not within the target set area, the detection result is that the color of the light-emitting element is unqualified; If the average brightness value of the image is not greater than the set lower brightness threshold or the standard deviation of the image grayscale is greater than the set grayscale threshold, then the detection result is that the grayscale distribution of the light-emitting surface of the light-emitting element is uneven.

3. The method for omnidirectional defect detection of light-emitting elements based on multi-view imaging according to claim 1, characterized in that, Also includes: Obtain a transmission image of the light-emitting element in the display module. If the transmission image shows broken gold wires and / or chip die bonding voids, the detection result indicates that there is an internal defect in the light-emitting element.

4. The method for omnidirectional defect detection of light-emitting elements based on multi-view imaging according to any one of claims 1-3, characterized in that, Each of the aforementioned spectral images undergoes preprocessing and feature extraction to obtain feature data including: Each of the spectral images is subjected to grayscale conversion, noise reduction, and image enhancement processing to obtain a processed image corresponding to each of the spectral images; All the processed images are localized and segmented into regions of interest to obtain image data containing the light-emitting area, electrode area, and package edge; Feature data is obtained by extracting grayscale features, RGB color components, geometric features, and texture features from all the image data.

5. The method for omnidirectional defect detection of light-emitting elements based on multi-view imaging according to any one of claims 1-3, characterized in that, Acquiring multiple spectral images of the light-emitting element in the display module from different viewing angles includes: using a camera device to acquire multiple spectral images of the light-emitting element from above, side, and below in visible light mode or a specific wavelength mode; or, acquiring detection data of the light-emitting element in the display module includes: applying a rated operating current to the light-emitting element from the forward and reverse directions respectively to obtain forward current and reverse current; applying a standard light source to the light-emitting element to obtain luminous flux and luminous intensity.

6. A device for omnidirectional defect detection of light-emitting elements based on multi-view imaging, characterized in that, include: Data acquisition unit, data processing and feature extraction unit, and judgment and detection unit; The data acquisition unit is used to set the display module to be inspected on the defect detection fixture, and acquire the detection data of the light-emitting element in the display module and multiple spectral images from different perspectives; The data processing and feature extraction unit is used to preprocess and extract features from each of the spectral images to obtain feature data; The judgment and detection unit is used to make a judgment based on the feature data and the detection data to obtain a detection result.

7. The omnidirectional defect detection device for light-emitting elements based on multi-view imaging according to claim 6, characterized in that, The feature data includes appearance data, average image brightness value, image grayscale standard deviation, and color coordinates; the detection data includes forward current, reverse current, luminous flux, and luminous intensity; and the judgment detection unit includes: If the appearance data contains any one or more of the following: scratches, black spots, impurities, and bubbles, the test result is that the appearance of the light-emitting element is unqualified. If the luminous intensity and / or the luminous flux are not within the set allowable range, the detection result is that the photoelectric performance of the light-emitting element is unqualified; If the reverse current is greater than the forward current, the detection result indicates that the light-emitting element is leaking current. If the color coordinates are not within the target set area, the detection result is that the color of the light-emitting element is unqualified; If the average brightness value of the image is not greater than the set lower brightness threshold or the standard deviation of the image grayscale is greater than the set grayscale threshold, then the detection result is that the grayscale distribution of the light-emitting surface of the light-emitting element is uneven.

8. The omnidirectional defect detection device for light-emitting elements based on multi-view imaging according to claim 6, characterized in that, The data acquisition unit is further configured to acquire a transmission image of the light-emitting element in the display module, and the judgment and detection unit is further configured to determine that there is an internal defect in the light-emitting element if the transmission image shows a broken gold wire and / or a chip die-bonding void.

9. The omnidirectional defect detection device for light-emitting elements based on multi-view imaging according to claim 6, characterized in that, The data processing and feature extraction unit includes a preprocessing subunit, a segmentation subunit, and a feature extraction subunit; The preprocessing subunit is used to perform grayscale conversion, noise reduction, and image enhancement processing on each of the spectral images to obtain a processed image corresponding to each of the spectral images; The segmentation subunit is used to locate and segment the region of interest in all the processed images to obtain image data containing the light-emitting area, the electrode area, and the edge of the package. The feature extraction subunit is used to extract grayscale features, RGB color components, geometric features, and texture features from all the image data to obtain feature data.

10. A multi-view imaging-based omnidirectional defect detection system for light-emitting elements, characterized in that, The device includes a defect detection fixture and an industrial control computer. The defect detection fixture is provided with a detection station for placing light-emitting elements in a display module. The industrial control computer performs automated defect detection on the light-emitting elements in the detection station according to the multi-view imaging-based omnidirectional defect detection method for light-emitting elements as described in any one of claims 1-5.

Images