LED detection method and system for braider

Abstract

The application relates to the technical field of bulk material sorting, in particular to an LED detection method and system of a taping machine, which comprises the following steps: when an LED patch element is detected, the spatial pose of an industrial camera is adjusted through an active optical compensation mechanism, and an initial LED lamp bead image is collected through the industrial camera; the initial LED lamp bead image is pretreated to obtain a target LED lamp bead image; the target LED lamp bead image is compared with a preset reference lamp bead image to generate an LED lamp bead detection result; if the LED lamp bead detection result is qualified, the LED patch element is moved to a carrier tape; when the number of times of generation of the LED lamp bead detection result exceeds a preset number threshold, the parameters of the active optical compensation mechanism are feedback-optimized through a PID algorithm based on an embedded standard reference lamp bead array to realize self-adaptive optimization of the detection system, and the compensation accuracy and system stability are improved.

Description

Technical Field

[0001] This invention relates to the field of bulk material sorting technology, and in particular to an LED detection method and system for a tape and reel machine. Background Technology

[0002] In the LED taping machine production process, high-speed and high-precision inspection technology is crucial for the quality control of LED chip components. Traditional inspection methods mainly rely on industrial cameras in a fixed position to capture images of LED beads, and then combine image processing algorithms for quality comparison and analysis. However, this static inspection method faces many challenges in practical applications, especially when LED components have slight offsets or angular deviations. Fixed imaging systems are difficult to adapt to dynamically changing inspection environments, and are prone to image quality degradation due to factors such as light fluctuations or mechanical vibrations, thus affecting the accuracy of the inspection results.

[0003] As equipment operating time increases, parameter drift issues in active optical compensation mechanisms gradually become apparent, while traditional methods lack an effective real-time feedback mechanism to dynamically adjust compensation parameters. This limitation prevents the system from automatically adapting to performance degradation after prolonged operation, leading to a gradual decrease in detection accuracy. Furthermore, during continuous batch testing, the accumulation of systematic errors further exacerbates the instability of test results, making it difficult for traditional static calibration methods to maintain the required level of detection accuracy, severely impacting production efficiency and product quality.

[0004] Existing LED detection methods for taping machines still rely on image comparison with fixed thresholds or periodic manual calibration, which cannot achieve adaptive optimization of the detection system. In the LED taping production process, there are problems such as insufficient optical compensation accuracy and parameter drift. Summary of the Invention

[0005] In view of this, this application provides an LED detection method and system for a tape and reel machine to solve the problems of insufficient optical compensation accuracy and parameter drift in the detection of existing LED tape and reel machines.

[0006] The LED detection method and system for a tape and reel machine provided in this application adopts the following technical solution: A method for detecting LEDs in a tape and reel machine, comprising: When an LED chip element is detected, the spatial pose of the industrial camera is adjusted by an active optical compensation mechanism, and the initial LED chip image is acquired by the industrial camera. The initial LED bead image is preprocessed to obtain the target LED bead image; Based on a preset reference LED image, the target LED image is compared to the reference image to generate LED detection results. If the LED chip test result is qualified, move the LED chip components to the carrier tape; When the number of times the LED bead detection results are generated exceeds the preset threshold, the parameters of the active optical compensation mechanism are optimized by feedback through a PID algorithm based on the embedded standard reference bead array.

[0007] By adopting the above technical solution, the industrial camera pose is dynamically adjusted by an active optical compensation mechanism to obtain a high-precision initial image. After preprocessing, the image is intelligently compared with a reference image to achieve quality inspection. Qualified products are automatically woven into the carrier tape. The standard reference LED array is combined with a PID algorithm to achieve periodic parameter feedback optimization, thereby effectively solving the problems of insufficient optical compensation and parameter drift in traditional detection methods, and significantly improving the detection accuracy and stability of LED tape production.

[0008] Preferably, when an LED chip element is detected, the spatial pose of the industrial camera is adjusted by an active optical compensation mechanism, and an initial image of the LED chip is acquired by the industrial camera. Prior to this, the method further includes: LED chip components are fed into the vibratory feeder of the tape and reel machine, and the vibratory feeder orients and arranges the LED chip components and transports them to the bottom of the industrial camera.

[0009] By adopting the above technical solution, and by adding a vibratory feeder orientation arrangement step before optical inspection, the LED patch element is accurately introduced and oriented to be directly below the industrial camera, ensuring that the subsequent active optical compensation mechanism can acquire the initial image in the optimal pose, thereby improving the positioning accuracy and imaging quality of the inspection system.

[0010] Preferably, the active optical compensation mechanism includes a focusable liquid lens assembly and a miniature servo motor, wherein adjusting the spatial pose of the industrial camera via the active optical compensation mechanism includes: When an LED chip element is detected, the adjustable liquid lens group performs continuous zooming to generate an LED position feedback signal. Based on the position feedback signal of the LED beads, the miniature servo motor adjusts the spatial pose of the industrial camera.

[0011] By adopting the above technical solution and using the collaborative working mechanism of adjustable focus liquid lens group and micro servo motor, when an LED patch element is detected, the liquid lens group generates a high-precision position feedback signal in real time through continuous zoom, and drives the servo motor to accurately adjust the spatial pose of the industrial camera, realizing sub-micron level dynamic optical compensation, which significantly improves the stability and accuracy of image acquisition.

[0012] Preferably, the initial LED bead image is preprocessed to obtain the target LED bead image, including: A Gaussian filter is used to denoise the initial LED bead image to obtain a denoised LED bead image; An edge enhancement operator is used to sharpen the edges of the denoised LED bead image to obtain an edge-enhanced LED bead image; A grayscale normalization algorithm is used to equalize the brightness of the edge-enhanced LED bead image to obtain the target LED bead image.

[0013] By adopting the above technical solution, the initial LED bead image is finely preprocessed through a three-level image processing flow of Gaussian filtering for noise reduction, edge enhancement and sharpening, and grayscale normalization and brightness equalization. This effectively eliminates environmental noise interference, enhances key feature edge information, and achieves brightness standardization processing, ultimately obtaining a target image with clear features and uniform contrast.

[0014] Preferably, based on a preset reference LED image, the target LED image is compared to the reference image to generate LED detection results, including: The slope parameter of the temperature-color temperature compensation curve is generated by analyzing the target LED bead image through a digital twin model. The color coordinates of the reference LED bead image are corrected by temperature compensation using a dynamic compensation algorithm to obtain the corrected reference color coordinates. The local feature windows of the target LED bead image and the compensated reference LED bead image are normalized using a structural similarity algorithm to obtain a similarity score. The color difference value is obtained by calculating the color difference between the target LED image and the compensated reference LED image using a color difference calculation algorithm. The LED bead detection results are generated based on the similarity score and color difference value.

[0015] By adopting the above technical solution, and using a technical path that combines digital twin models with dynamic compensation algorithms, the system achieves accurate color coordinate compensation and correction by analyzing the temperature-color temperature curve. Then, it employs a dual verification mechanism of structural similarity algorithm and color difference calculation. After normalization processing, it performs a comprehensive evaluation of local features and global color difference between the target image and the reference image, and finally generates the ED LED bead detection result. This effectively improves the adaptability of the detection system to environmental factors such as temperature changes and the accuracy of the detection results.

[0016] Preferably, if the LED chip test result is qualified, the LED chip component is moved to the carrier tape, including: If both the similarity score and the corresponding color difference value meet the corresponding preset detection threshold, the LED bead test result is qualified. Move the LED chip components that pass the LED bead test to the carrier tape.

[0017] By adopting the above technical solution and establishing a dual-judgment intelligent sorting mechanism, when the similarity score and color difference value simultaneously meet the preset detection threshold, the system automatically determines that the LED beads are qualified and triggers a precise mechanical transfer action to accurately place the qualified products into the carrier belt. This achieves closed-loop control of quality inspection and automated sorting, ensuring both the consistency of product quality and optimizing the continuity of the production process.

[0018] Preferably, it also includes: if the LED bead test result is unqualified, then the LED chip element is removed.

[0019] By adopting the above technical solution and setting up an automatic rejection mechanism, when the test results determine that the LED beads are unqualified, the system immediately triggers the rejection device to remove the defective components from the production line, preventing unqualified products from flowing into subsequent processes, thereby effectively ensuring the overall quality of the products and the purity of the production process.

[0020] Preferably, if the LED chip test result is unqualified, the LED chip components are removed, including: If either the similarity score or the corresponding color difference value fails to meet the preset detection threshold, the LED bead test result is unqualified. Remove LED chip components that fail the LED bead test.

[0021] By adopting the above technical solution, an anomaly detection mechanism is established. When either the similarity score or the color difference value does not meet the preset acceptable range, the LED bead is determined to be unqualified and is automatically removed. This ensures that only products that fully meet the standards can enter the subsequent processes, thereby effectively guaranteeing the consistency of product quality and the reliability of the production line.

[0022] Preferably, the parameters of the active optical compensation mechanism are optimized using a PID algorithm, and the process further includes: Establish an ambient light interference compensation model to collect ambient light intensity and color temperature data in real time; When the change in ambient light intensity exceeds a preset benchmark threshold, the ambient light fluctuation is predicted and modeled based on the Kalman filter algorithm, and the exposure parameters and gain value of the industrial camera are dynamically adjusted according to the prediction results.

[0023] By adopting the above technical solution and constructing an intelligent ambient light compensation system, after optimizing the PID parameters, the system monitors the changes in ambient light intensity and color temperature in real time. When the light intensity fluctuation exceeds the threshold, the camera exposure parameters are dynamically adjusted based on the Kalman filter prediction model, forming a closed-loop optimization system from mechanical compensation to optical parameter adjustment. This enables the detection system to actively adapt to changes in external lighting conditions, ensuring stable imaging quality under different environmental conditions.

[0024] An LED detection system for a tape and reel machine, comprising: The image acquisition module is used to adjust the spatial pose of the industrial camera through an active optical compensation mechanism when an LED chip element is detected, and to acquire an initial image of the LED chip through the industrial camera. The image processing module is used to preprocess the initial LED bead image to obtain the target LED bead image; The detection module is used to perform benchmark comparison on the target LED image based on a preset benchmark LED image, and generate LED detection results. The judgment module is used to move the LED chip component to the carrier tape if the LED chip test result is qualified. The feedback optimization module is used to optimize the parameters of the active optical compensation mechanism based on the embedded standard reference lamp array when the number of times the LED lamp bead detection results are generated exceeds the preset threshold.

[0025] By adopting the above technical solution, the image acquisition module achieves high-precision dynamic image acquisition, the image processing module completes feature enhancement preprocessing, the detection module performs benchmark comparison analysis, the judgment module controls the automatic tape production of qualified products, and the feedback optimization module realizes the periodic self-calibration of system parameters. The various functional modules work together to form a closed-loop control system, which solves the technical bottlenecks of traditional detection methods in terms of accuracy maintenance and stability, and significantly improves the detection efficiency and quality control level of LED tape production.

[0026] As can be seen from the above technical solutions, this application has the following advantages: When an LED chip element is detected, the spatial pose of the industrial camera is adjusted by an active optical compensation mechanism, and an initial LED chip image is acquired through the industrial camera. The initial LED chip image is preprocessed to obtain the target LED chip image. Based on a preset reference chip image, the target LED chip image is compared to generate an LED chip detection result. If the LED chip detection result is qualified, the LED chip element is moved to the carrier tape. When the number of times the LED chip detection result is generated exceeds a preset threshold, the parameters of the active optical compensation mechanism are optimized using a PID algorithm based on an embedded standard reference chip array. It is evident that by integrating three core technologies—dynamic optical compensation, intelligent image processing, and self-optimizing control—a closed-loop detection system covering the entire process from image acquisition and quality assessment to parameter optimization is constructed. This improves compensation accuracy and system stability, and solves the problems of insufficient optical compensation accuracy and parameter drift in existing LED taping machine detection. Attached Figure Description

[0027] Figure 1This is a flowchart of the steps of an LED detection method for a tape and reel machine provided in an embodiment of the present invention; Figure 2 This is a flowchart of another LED detection method for a tape and reel machine provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of an LED detection system for a tape and reel machine provided in an embodiment of the present invention. Detailed Implementation

[0028] Example 1: Please see Figure 1 , Figure 1 This is a flowchart illustrating the steps of an LED detection method for a tape and reel machine, as provided in an embodiment of the present invention.

[0029] This invention provides a method for detecting LEDs in a tape and reel machine, comprising: Step 101: When an LED chip element is detected, the spatial pose of the industrial camera is adjusted by an active optical compensation mechanism, and an initial LED chip image is acquired by the industrial camera.

[0030] LED taping machines are specifically designed to automatically and systematically load bulk LED surface mount components into carrier tape, which is then wound into a reel for automatic pickup and placement by SMT pick-and-place machines. The process includes feeding, inspection, carrier tape supply, and component loading; the specific steps are not detailed here.

[0031] It should be noted that LED surface mount components refer to LED beads that use surface mounted technology (SMT). They are usually miniaturized and standardized chip structures that can be rolled up by a tape and reel machine for automatic pick-up and mounting onto circuit boards by a pick-and-place machine.

[0032] Detecting an LED patch element means that after the LED patch element is oriented and transported to the detection position by the vibratory feeder, the PLC (Programmable Controllers) senses that the element has entered the detection field of view of the active optical compensation mechanism. In this invention, when an LED patch element is detected, optical adjustment and image acquisition actions are triggered to ensure the synchronization of the detection process and the component transport rhythm.

[0033] In this first embodiment, the active optical compensation mechanism includes a focusable liquid lens assembly and a miniature servo motor. The focusable liquid lens assembly serves as the front-end optical component of the industrial camera, adjusting the focal length (range 5-50mm) in real time via voltage regulation to ensure that the LED beads always present the optimal imaging plane on the camera sensor. When there is a ±0.1mm deviation in the position of the bulk LEDs conveyed by the vibratory feeder, the liquid lens can complete the focal length compensation within 10ms, avoiding the edge blurring problem caused by insufficient depth of field in traditional fixed-focal-length lenses.

[0034] A miniature servo motor is rigidly connected to an industrial camera, and the camera module's pitch / yaw angle is finely adjusted based on feedback from a position sensor. For example, when there is a cumulative error of 0.05mm in the carrier tape stepper mechanism, the servo motor drives the entire camera to shift, ensuring that the optical axis is always perpendicular to the LED center, thus eliminating the influence of perspective distortion on dimensional measurements.

[0035] In this first embodiment, at the beginning of each detection cycle, the PLC first identifies the position of the LED beads and drives the micro servo motor to automatically adjust the spatial pose of the industrial camera based on the feedback signal of the adjustable focus liquid lens group. This eliminates the measurement error caused by position deviation and ensures that the distance from the lens to the LED beads remains constant, so as to obtain an initial LED bead image with consistent size, clear and distortion-free image. This solves the problems of insufficient optical compensation accuracy and parameter drift.

[0036] Step 102: Preprocess the initial LED bead image to obtain the target LED bead image.

[0037] In this first embodiment, preprocessing refers to performing operations such as Gaussian filtering for noise reduction, edge enhancement, and grayscale normalization on the initial LED bead image, so that the output is a target LED bead image with no obvious noise, moderate contrast, and clearly distinguishable bead outlines.

[0038] Step 103: Based on the preset reference LED image, perform a benchmark comparison on the target LED image to generate LED detection results.

[0039] The preset reference LED image refers to the "standard template" used in LED taping machine testing to determine whether LEDs are qualified. Before the testing process starts, standardized image data is pre-stored after image acquisition and feature extraction of known qualified LEDs (reference LEDs). This ensures that the quality judgment of all LEDs to be tested has a unified and quantifiable standard, avoiding testing deviations caused by environmental changes, equipment drift, or subjective human judgment.

[0040] It is worth mentioning that if multiple types of LEDs are involved, multiple sets of reference images should be preset to avoid misjudgment across models. In this first embodiment, the reference LED beads are limited to key parameters such as size, color temperature, shape, and polarity marking, and the parameter settings depend on the specific situation.

[0041] It should be noted that the equipment needs to be debugged before acquiring the reference LED bead image: the reference LED bead is loaded into the vibratory feeder, and when the PLC detects the reference LED patch element, the position of the industrial camera is adjusted through the active optical compensation mechanism (to ensure that the focal length and angle are consistent with the actual production), and the initial image of the reference LED bead is acquired.

[0042] Furthermore, the initial reference image undergoes the same preprocessing procedure as the LED bead to be tested, generating a standardized reference LED bead image, which is then stored in the database. This ensures the uniformity and traceability of the testing standards, providing a stable quality benchmark for mass production.

[0043] In this first embodiment, based on a pre-set reference LED image, a detailed benchmark comparison analysis is performed on the acquired target LED image. The comparison generates LED detection and analysis results, thereby achieving an accurate assessment of the LED performance and status.

[0044] Step 104: If the LED bead test result is qualified, move the LED chip component to the carrier tape.

[0045] In this first embodiment, once the LED bead test result is deemed qualified, a qualified signal is sent to the PLC. The PLC then controls the precision positioning mechanism (which can be a suction nozzle or a push rod) to accurately grasp the LED chip element. Based on the synchronization signal of the carrier tape feeding mechanism, when the carrier tape moves to the preset loading position and pauses briefly, the LED chip element is placed smoothly and accurately in the groove. Throughout the process, the sensor provides real-time feedback on whether the element is in place, ensuring that the LED chip element is not offset or tilted in the carrier tape. This lays the foundation for subsequent cover tape encapsulation and reel winding. At the same time, this movement process is seamlessly connected with the testing process, avoiding a decrease in production efficiency due to waiting and ensuring the continuous and stable operation of the taping machine.

[0046] Step 105: When the number of times the LED bead detection results are generated exceeds the preset threshold, the parameters of the active optical compensation mechanism are optimized by feedback through the PID algorithm based on the embedded standard reference bead array.

[0047] In this first embodiment, when the cumulative number of LED bead detection results reaches a preset threshold, the system automatically triggers a calibration mechanism. At this time, the industrial camera acquires images of the embedded standard reference bead array and compares the acquired images with the corresponding preset benchmark images. By analyzing the deviations in key features such as geometric dimensions and color temperature / grayscale values, and combining this with a PID algorithm, the parameter correction amount of the active optical compensation mechanism is calculated. This allows for real-time feedback optimization of key parameters such as the focus adjustment voltage of the liquid lens and the displacement compensation amount of the micro servo motor in the active optical compensation mechanism. This dynamically offsets the impact of minor drifts caused by environmental changes or long-term operation of the mechanism on detection accuracy, ensuring that the active optical compensation mechanism always provides stable imaging conditions for the industrial camera. This guarantees the consistency and accuracy of subsequent LED bead detection in size measurement and color temperature judgment, maintaining high system reliability in complex environments. In this first embodiment, the standard reference bead array contains qualified bead samples with various typical color temperatures and sizes. Furthermore, the preset threshold can be set to 50 or 100 times, depending on the specific production situation.

[0048] It should be noted that the preset reference LED image is not fixed and is also linked with the standard reference LED array to achieve dynamic updates: after every 100 LEDs are detected, the system automatically calls the built-in standard reference LED array, collects its current image and compares it with the preset benchmark. If the image features of the reference LEDs are slightly deviated due to environmental changes (such as temperature and humidity fluctuations), the system dynamically corrects the feature parameters of the reference LED image through a PID algorithm to ensure that the benchmark always matches the actual production environment.

[0049] When an LED chip element is detected, this invention adjusts the spatial pose of an industrial camera through an active optical compensation mechanism and acquires an initial LED chip image through the industrial camera. The initial LED chip image is preprocessed to obtain a target LED chip image. Based on a preset reference chip image, the intermediate LED chip image is compared to generate an LED chip detection result. If the LED chip detection result is qualified, the LED chip element is moved to the carrier tape. When the number of times the LED chip detection result is generated exceeds a preset threshold, the parameters of the active optical compensation mechanism are optimized using a PID algorithm based on an embedded standard reference chip array. Therefore, by integrating three core technologies—dynamic optical compensation, intelligent image processing, and self-optimizing control—a closed-loop detection system covering the entire process from image acquisition and quality assessment to parameter optimization is constructed. This improves compensation accuracy and system stability, and solves the problems of insufficient optical compensation accuracy and parameter drift in existing LED taping machine detection.

[0050] Example 2: Please see Figure 2 , Figure 2 A flowchart illustrating the steps of another LED detection method for a tape and reel machine provided in an embodiment of the present invention.

[0051] Step 201: The LED chip components are introduced into the vibratory feeder of the tape and reel machine, and the vibratory feeder orients and arranges the LED chip components and transports them to the bottom of the industrial camera.

[0052] In this second embodiment, the worker pours the bulk LED chip components into the vibratory feeder of the tape and reel machine. After the vibratory feeder is powered on, it generates vibrations at a specific frequency. By utilizing the geometric features of the LED chip components themselves and the tilt angle and width design of the inner wall track of the vibratory feeder, the originally disordered stacked components are gradually adjusted in posture along the track under the vibration force, and finally arranged in an orderly manner in a uniform direction. They then slide out of the vibratory feeder along the track and are accurately transported to the detection position directly below the industrial camera. This provides a stable and uniformly oriented detection object for the subsequent adjustment of the active optical compensation mechanism and image acquisition. It should be noted that the geometric features of the LED chip components themselves include, but are not limited to, rectangular structures and protrusions or depressions of polarity markings.

[0053] Step 202: When the LED chip element is detected, the adjustable focus liquid lens group is used to continuously zoom and generate an LED position feedback signal.

[0054] Step 203: Based on the LED bead position feedback signal, the micro servo motor adjusts the spatial pose of the industrial camera and acquires the initial LED bead image through the industrial camera.

[0055] In this second embodiment, when the sensor detects that the LED chip element is being transported to the detection area below the industrial camera, it triggers the adjustable-focus liquid lens assembly to operate. The adjustable-focus liquid lens assembly does not require the movement of mechanical parts. By adjusting the applied voltage, it achieves continuous and stepless change of focal length within the range of 5-50mm. During this process, it captures the changes in the image clarity of the LED bead in real time. Based on the focal length parameter when the image is clearest, it deduces the actual distance between the lens and the bead and converts this distance information into an LED bead position feedback signal, which is transmitted to the PLC. The micro servo motor precisely adjusts the lens position based on the feedback signal to eliminate position deviation and ensure that the lens always maintains a constant distance from the bead. After the adjustment is completed, the initial LED bead image is acquired through the industrial camera. The entire process effectively eliminates imaging errors caused by small displacements during component transportation because the liquid lens has no mechanical wear, responds quickly, and the servo motor is precisely adjusted.

[0056] Step 204: Preprocess the initial LED bead image to obtain the target LED bead image.

[0057] Furthermore, step 204 may include the following sub-steps: S41. Use a Gaussian filter to denoise the initial LED bead image to obtain a denoised LED bead image.

[0058] In this second embodiment, after acquiring the initial LED bead image, a Gaussian filter is used to denoise the image. By calculating the weighted average of a pixel and its neighboring pixels, where the weight decreases in a Gaussian distribution as the distance increases, the key details of the LED bead are preserved to the greatest extent while smoothing the image noise. Finally, a denoised LED bead image with clear edges and no obvious noise is obtained, eliminating random noise caused by ambient light fluctuations, camera sensor noise, or vibration.

[0059] S42. The edge enhancement operator is used to sharpen the edges of the denoised LED bead image to obtain the edge-enhanced LED bead image.

[0060] In this second embodiment, edge enhancement operators (including but not limited to Sobel operator and Laplacian operator) are used to sharpen the image edges. By calculating the gradient change of pixel gray values, the gray value difference between adjacent pixels is enhanced, making the originally blurry edge lines clear and distinguishable, thus forming an edge-enhanced LED light bead image.

[0061] S43. Use a grayscale normalization algorithm to perform brightness equalization on the edge-enhanced LED bead image to obtain the target LED bead image.

[0062] In this second embodiment, the LED bead image after edge enhancement processing may have brightness differences due to uneven ambient light, such as local reflections or shadows. In this case, a grayscale normalization algorithm is used to perform brightness equalization processing on the image. By mapping the grayscale values ​​of the image to a preset standard range, the brightness deviation of different areas is eliminated, making the grayscale distribution of the bead surface, pins and background more uniform, ensuring that the brightness characteristics of each part of the image are comparable, and finally obtaining a target LED bead image with consistent grayscale and clear details.

[0063] Step 205: Based on the preset reference LED image, perform a benchmark comparison on the target LED image to generate LED detection results.

[0064] Furthermore, step 205 may include the following sub-steps: S51. Parameter analysis of the target LED bead image is performed using a digital twin model to generate the slope parameter of the temperature-color temperature compensation curve.

[0065] A digital twin model refers to a pre-built digital twin model of an LED chip that integrates a database of the optical characteristics of the chip under different temperature environments.

[0066] In this second embodiment, by analyzing the characteristics of the LED color temperature distribution and gray value changes in the target image, and combining the real-time collected ambient temperature data, the dynamic relationship between color temperature and temperature change under the current temperature condition is simulated and calculated, thereby generating the slope parameter of the temperature-color temperature compensation curve. This parameter can quantify the degree of influence of temperature fluctuation on the color temperature detection result, so as to ensure that the LED color temperature can still be accurately judged as to whether it meets the standard when the temperature changes.

[0067] S52. The color coordinates of the reference LED bead image are corrected by temperature compensation using a dynamic compensation algorithm to obtain the corrected reference color coordinates.

[0068] In this second embodiment, based on the slope parameter of the temperature-color temperature compensation curve, the original color coordinates of the preset reference LED image are corrected in real time through a dynamic compensation algorithm. Specifically, the dynamic compensation algorithm calculates the compensation amount of the color coordinates based on the difference between the actual temperature of the current detection environment and the reference temperature, combined with the slope parameter, and adjusts the X and Y components of the reference color coordinates accordingly, thereby obtaining the corrected reference color coordinates that match the current temperature conditions, avoiding misjudgment of color temperature caused by the mismatch between the reference color coordinates and the actual detection environment due to temperature fluctuations.

[0069] S53. The local feature windows of the target LED bead image and the compensated reference LED bead image are normalized using a structural similarity algorithm to obtain a similarity score.

[0070] In this second embodiment, based on the target LED image and the corrected reference LED image, the target LED image and the corrected reference LED image are first divided into multiple corresponding local feature windows. Then, the original matching degree values ​​of each window in terms of brightness, contrast and structural features are calculated by the Structural Similarity (SSIM) algorithm. These original values ​​are then normalized (that is, the data is mapped to a uniform scale of 0-1 to eliminate the dimensional deviation caused by feature differences between different windows). Finally, the normalized results of all windows are integrated to generate an overall similarity score. The similarity score intuitively quantifies the degree of matching between the target LED and the compensated reference LED in terms of local details and overall features.

[0071] It should be noted that the local feature window includes, but is not limited to, key areas such as the light-emitting area of ​​the LED, the edge of the pin, and the polarity mark.

[0072] S54. The color difference value is obtained by calculating the color difference between the target LED image and the compensated reference LED image using a color difference calculation algorithm.

[0073] In this second embodiment, based on the target LED image and the corrected reference LED image, the color coordinate data of the target LED image and the corrected reference LED image are quantified and calculated using a color difference calculation algorithm (e.g., CIEDE2000 color difference formula). Specifically, the RGB or CIELab color space parameters of the target LED and the reference LED in the same area are extracted and substituted into the algorithm formula to calculate the comprehensive difference value of the two in hue, brightness, and saturation, i.e., the color difference value ΔE. The color difference value can directly reflect the color temperature and color consistency between the target LED and the reference LED.

[0074] S55. Based on the similarity score and color difference value, generate the LED bead detection results.

[0075] Step 206: If the LED bead test result is qualified, move the LED chip component to the carrier tape.

[0076] Furthermore, step 206 may include the following sub-steps: S61. If both the similarity score and the corresponding color difference value meet the preset detection threshold, the LED bead detection result is qualified.

[0077] Preferably, the obtained similarity score is compared with a preset similarity threshold, and the obtained color difference value is compared with a set allowable color difference range. If the similarity score is not lower than the threshold and the color difference value is within the allowable range, the LED bead is deemed qualified. If any indicator fails to meet the requirements, it is deemed unqualified. Finally, an LED bead test result containing the qualified / unqualified result and specific out-of-tolerance items (such as insufficient similarity or excessive color difference value) is generated and transmitted to the PLC.

[0078] In this second embodiment, the similarity threshold is set to 0.95. If the similarity score is lower than 0.95, the LED bead is considered unqualified. Similarly, the color difference threshold is set to 2. If the color difference value is greater than 2, the LED bead is considered unqualified.

[0079] S62. Move the LED chip components that have passed the LED bead test to the carrier tape.

[0080] The implementation process of step S62 is similar to that of step 104, and will not be described again here.

[0081] Step 207: If the LED bead test result is unqualified, remove the LED chip components.

[0082] Furthermore, step 207 may include the following sub-steps: S71. If either the similarity score or the corresponding color difference value fails to meet the preset detection threshold, the LED bead test result is unqualified.

[0083] Step S71 has already been explained in detail in step S61, and will not be repeated here.

[0084] S72. Remove LED chip components whose LED bead test results are unqualified.

[0085] In this second embodiment, when the LED bead test result is determined to be unqualified, the system will send a trigger signal to the rejection mechanism. The rejection mechanism will drive the pneumatic push rod or vacuum nozzle to move the unqualified LED chip element to the rejection station, so that it is accurately removed from the conveying path and sent to the waste collection box. The whole process is synchronized with the conveying process of qualified elements and does not interfere with each other, ensuring that unqualified elements will not enter the subsequent carrier tape loading stage, thereby ensuring the overall quality of the tape and reel products.

[0086] Step 208: Establish an ambient light interference compensation model and collect ambient light intensity and color temperature data in real time.

[0087] In this second embodiment, multiple ambient light sensors are installed at the detection location. The ambient light sensors collect the light intensity and color temperature data of the current environment in real time, and input these dynamically changing data into a pre-built ambient light interference compensation model. The ambient light interference compensation model is trained based on a large amount of detection data under different ambient light conditions. It can quantitatively analyze the mapping relationship between ambient light intensity, color temperature and LED lamp bead image acquisition quality, and provide real-time parameter support for subsequent targeted optical compensation of target LED lamp bead images or reference lamp bead images, so as to reduce the interference of ambient light fluctuations on detection accuracy.

[0088] Step 209: When the change in ambient light intensity exceeds the preset benchmark threshold, the ambient light fluctuation is predicted and modeled based on the Kalman filter algorithm, and the exposure parameters and gain value of the industrial camera are dynamically adjusted according to the prediction results.

[0089] Preferably, when the ambient light sensor detects that the real-time change in ambient light intensity exceeds a preset benchmark threshold, the system immediately activates the Kalman filter algorithm. This algorithm, by fusing historical ambient light intensity data with the current measurement, constructs a predictive model for ambient light fluctuations. This model can accurately predict the changing trend of ambient light in the near future. Subsequently, based on the prediction results, the exposure time of the industrial camera and the sensor gain value are dynamically adjusted to ensure that even under conditions of drastic ambient light fluctuations, the industrial camera can still capture LED light bead images with stable brightness and clear details, avoiding overexposure or underexposure due to sudden changes in light, thereby maintaining consistent detection accuracy. In this second embodiment, the real-time change in ambient light intensity does not exceed 20%.

[0090] This invention integrates three core technologies—dynamic optical compensation, intelligent image processing, and self-optimizing control—to construct a closed-loop detection system covering the entire process from image acquisition and quality assessment to parameter optimization. This improves compensation accuracy and system stability, and solves the problems of insufficient optical compensation accuracy and parameter drift in existing LED taping machine detection.

[0091] Example 3: Please see Figure 3 , Figure 3 This is a structural block diagram of an LED detection system for a tape and reel machine, provided as an embodiment of the present invention.

[0092] The image acquisition module 301 is used to adjust the spatial pose of the industrial camera through an active optical compensation mechanism when an LED chip element is detected, and to acquire an initial image of the LED chip through the industrial camera.

[0093] The image processing module 302 is used to preprocess the initial LED bead image to obtain the target LED bead image.

[0094] The detection module 303 is used to perform a benchmark comparison on the target LED bead image based on a preset benchmark LED bead image, and generate LED bead detection results.

[0095] The judgment module 304 is used to move the LED chip component to the carrier tape if the LED chip test result is qualified.

[0096] The feedback optimization module 305 is used to optimize the parameters of the active optical compensation mechanism based on the embedded standard reference lamp array when the number of times the LED lamp bead detection result is generated exceeds the preset number threshold.

[0097] Since the above is a one-to-one LED detection system for a tape and reel machine, its implementation principle is similar to that of an LED detection system for a tape and reel machine, and will not be described in detail here.

[0098] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Identical components are represented by the same reference numerals. Therefore, all equivalent changes made to the structure, shape, and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for detecting LEDs in a tape and reel machine, characterized in that, include: When an LED chip element is detected, the spatial pose of the industrial camera is adjusted by an active optical compensation mechanism, and an initial image of the LED chip is acquired by the industrial camera. The initial LED bead image is preprocessed to obtain the target LED bead image; Based on a preset reference LED bead image, the target LED bead image is compared with the reference image to generate LED bead detection results; If the LED bead test result is qualified, the LED chip element is moved to the carrier tape; When the number of times the LED bead detection result is generated exceeds a preset threshold, based on the embedded standard reference bead array, if the image features of the standard reference bead shift, the parameters of the active optical compensation mechanism are optimized by feedback through a PID algorithm. The active optical compensation mechanism includes a focusable liquid lens assembly and a miniature servo motor. The adjustment of the spatial pose of the industrial camera via the active optical compensation mechanism includes: When an LED chip element is detected, the adjustable focus liquid lens group performs continuous zooming to generate an LED position feedback signal. Based on the LED bead position feedback signal, the micro servo motor adjusts the spatial pose of the industrial camera; The process of comparing the target LED image with a preset benchmark LED image to generate LED detection results includes: The target LED bead image is analyzed using a digital twin model to generate the slope parameter of the temperature-color temperature compensation curve. Based on the slope parameter, the color coordinates of the reference LED bead image are corrected by temperature compensation using a dynamic compensation algorithm to obtain the corrected reference color coordinates. Based on the corrected reference color coordinates and the corrected reference LED image, the local feature windows of the target LED image and the corrected reference LED image are calculated using a structural similarity algorithm, and the calculation results are normalized to obtain a similarity score. The color coordinate data of the target LED bead image and the corrected reference LED bead image are calculated using a color difference calculation algorithm to obtain the color difference value; Based on the similarity score and the color difference value, the LED bead detection result is generated.

2. The LED detection method for a tape and reel machine according to claim 1, characterized in that, When an LED chip element is detected, the spatial pose of the industrial camera is adjusted by an active optical compensation mechanism, and an initial image of the LED chip is acquired through the industrial camera. Prior to this, the process also includes: The LED chip elements are fed into the vibratory feeder of the tape and reel machine, and the vibratory feeder orients and arranges the LED chip elements and transports them to the area below the industrial camera.

3. The LED detection method for a tape and reel machine according to claim 1, characterized in that, The preprocessing of the initial LED bead image to obtain the target LED bead image includes: The initial LED image is denoised using a Gaussian filter to obtain a denoised LED image. An edge enhancement operator is used to sharpen the edges of the denoised LED bead image to obtain an edge-enhanced LED bead image. The edge-enhanced LED bead image is brightness-equalized using a grayscale normalization algorithm to obtain the target LED bead image.

4. The LED detection method for a tape and reel machine according to claim 1, characterized in that, If the LED bead test result is qualified, the LED chip element is moved to the carrier tape, including: If both the similarity score and the color difference value meet the corresponding preset detection threshold, then the LED bead detection result is qualified. The LED chip element whose test result is qualified is moved to the carrier tape.

5. The LED detection method for a tape and reel machine according to claim 1, characterized in that, Also includes: If the LED bead test result is unqualified, then the LED chip element shall be removed.

6. The LED detection method for a tape and reel machine according to claim 5, characterized in that, If the LED bead test result is unqualified, the LED chip element shall be removed, including: If either the similarity score or the color difference value fails to meet the corresponding preset detection threshold, the LED bead detection result is unqualified. Remove any LED chip components that fail the LED test.

7. The LED detection method for a tape and reel machine according to claim 1, characterized in that, The step of optimizing the parameters of the active optical compensation mechanism using a PID algorithm further includes: Establish an ambient light interference compensation model and collect ambient light intensity and color temperature data in real time; When the change in ambient light intensity exceeds a preset benchmark threshold, the ambient light fluctuation is predicted and modeled based on the Kalman filter algorithm, and the exposure parameters and gain value of the industrial camera are dynamically adjusted according to the prediction results.

8. An LED detection system for a tape and reel machine, characterized in that, include: The image acquisition module is used to adjust the spatial pose of the industrial camera through an active optical compensation mechanism when an LED chip element is detected, and to acquire an initial LED chip image through the industrial camera. The image processing module is used to preprocess the initial LED bead image to obtain the target LED bead image; The detection module is used to perform a benchmark comparison on the target LED bead image based on a preset benchmark LED bead image, and generate LED bead detection results; The judgment module is used to move the LED chip element to the carrier tape if the LED chip test result is qualified. The feedback optimization module is used to optimize the parameters of the active optical compensation mechanism by using a PID algorithm when the number of times the LED light bead detection result is generated exceeds a preset threshold, based on the embedded standard reference light bead array, if the image features of the standard reference light bead are offset. The active optical compensation mechanism includes a focusable liquid lens assembly and a miniature servo motor. The adjustment of the spatial pose of the industrial camera via the active optical compensation mechanism includes: When an LED chip element is detected, the adjustable focus liquid lens group performs continuous zooming to generate an LED position feedback signal. Based on the LED bead position feedback signal, the micro servo motor adjusts the spatial pose of the industrial camera; The process of comparing the target LED image with a preset benchmark LED image to generate LED detection results includes: The target LED bead image is analyzed using a digital twin model to generate the slope parameter of the temperature-color temperature compensation curve. Based on the slope parameter, the color coordinates of the reference LED bead image are corrected by temperature compensation using a dynamic compensation algorithm to obtain the corrected reference color coordinates. Based on the corrected reference color coordinates and the corrected reference LED image, the local feature windows of the target LED image and the corrected reference LED image are calculated using a structural similarity algorithm, and the calculation results are normalized to obtain a similarity score. The color coordinate data of the target LED bead image and the corrected reference LED bead image are calculated using a color difference calculation algorithm to obtain the color difference value; Based on the similarity score and the color difference value, the LED bead detection result is generated.

Images