Time-sharing stroboscopic AVI optical detection method

Abstract

The invention provides a time-sharing stroboscopic AVI optical detection method. The time-sharing stroboscopic AVI optical detection method comprises the following steps: step 1, time-sharing stroboscopic illumination control; step 2, image acquisition and separation; step 3, image fusion and enhancement; step 4, self-adapting to an application scene; and step 5, outputting a detection result. Starting from the dimension of light source control and image processing of PCB defect detection, a traditional detection mode of single light source imaging and fixed algorithm fusion is upgraded to a dual-light-source time-sharing stroboflash and multi-strategy self-adaptive AVI optical detection mode, a white light LED and an IR light source with the wavelength of 850 nm are adopted as detection light sources, and the detection accuracy is improved. The light source alternate lightening frequency and the image sensor line scanning frequency are both set to be 100 Hz, the frequency synchronization error is controlled within 0.05 Hz, and time-sharing capture of odd-numbered line white light images and even-numbered line IR images is achieved; the synthetic image is processed through a pixel-level splitting algorithm, the splitting and reconstruction precision reaches 99.95%, a pure white light image and a pure IR image without pixel crosstalk are output, and the detection efficiency is greatly improved.

Description

Technical Field

[0001] This invention belongs to the field of printed circuit board (PCB) manufacturing technology, and more specifically, relates to an AVI optical detection method for time-division stroboscopic flicker. Background Technology

[0002] As the core carrier of electronic devices, the surface and internal defects of PCB boards directly affect the operational stability and service life of electronic devices. As the electronics industry develops towards high precision and high density, the requirements for the accuracy and efficiency of PCB board defect detection are constantly increasing. Common defects include solder mask scratches, uneven ink, residual ink in holes, foreign objects on the hole walls, and buried via defects.

[0003] Currently, PCB defect detection mostly employs single-light source imaging technology. White light sources can clearly reveal surface texture and color defects, but their penetration is insufficient for deep defects such as those inside holes. Infrared light sources have strong penetration and can detect deep structural defects, but they struggle to identify surface details. Some detection schemes use both white and infrared light sources simultaneously, which easily leads to mixed illumination interference, resulting in a reduced image signal-to-noise ratio, blurred defect features, and an inability to accurately distinguish between surface and deep defects. For different types of PCBs, such as electroless gold / tin-plated boards and OSP / tin-plated boards, existing detection methods lack adaptive adjustment capabilities. During the detection process, metal reflection interference is significant, and the missed detection rate for defects inside holes is high, making it difficult to meet diverse detection needs. Furthermore, the image acquisition and processing flow of traditional detection methods is relatively independent, and the weight coefficients of the image fusion algorithm are fixed, making it impossible to dynamically optimize according to the defect type, further affecting detection accuracy and efficiency. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a time-division stroboscopic AVI optical inspection method. This method addresses the issues of PCB board defect inspection often employing single-light source imaging technology, the strong penetrating power of infrared light sources making it difficult to identify surface details, and the simultaneous illumination of white light and infrared light sources in some inspection schemes, which easily leads to mixed illumination interference. For different types of PCB boards such as electroless gold / tin plating boards and OSP / tin plating boards, existing inspection methods lack adaptive adjustment capabilities. The image acquisition and processing flow of traditional inspection methods is relatively independent, and the weight coefficients of image fusion algorithms are fixed, making it impossible to dynamically optimize according to defect types, further affecting inspection accuracy and efficiency.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for detecting time-division flicker defects in PCB boards using dual light sources, comprising the following steps: Step 1: Time-division strobe lighting control; Step 2: Image acquisition and separation; Step 3: Image fusion and enhancement; Step 4: Adapt to application scenarios; Step 5: Output the test results.

[0006] Furthermore, in step one, the specific operation of the time-division flicker lighting control is as follows: S1: White LED and IR light source are selected as the detection light source, wherein the wavelength of the IR light source is 850nm or 940nm; S2: Synchronize the light source trigger signal with the image sensor line scan signal to ensure that the light source alternation frequency is consistent with the image sensor line scan frequency, with a frequency synchronization error ≤0.1Hz; S3: Controls odd-numbered rows of images to be captured by white light illumination, and even-numbered rows of images to be captured by IR light illumination, ensuring that each row of images corresponds to a single illumination condition and avoiding interference from mixed illumination.

[0007] Furthermore, in step two, the specific operations for image acquisition and separation are as follows: S1: Control the image sensor to acquire images in a line scanning manner and generate a composite image composed of odd-numbered white light images and even-numbered IR images; S2: The synthesized image is transmitted to the image processing unit, and the synthesized image is processed by the pixel-level splitting algorithm to reconstruct the pure white light image and the pure IR image respectively; S3: Ensure that the image processing unit achieves a splitting and reconstruction accuracy of ≥99.9% for the synthesized image, and that there is no pixel crosstalk between the reconstructed pure white light image and the pure IR image.

[0008] Furthermore, in step three, the specific operations of image fusion and enhancement are as follows: S1: Perform surface texture and color feature analysis on pure white light images to identify surface defects such as scratches on the solder mask layer and uneven ink distribution on PCB boards. S2: Utilize the strong penetrating power of IR images to extract deep structural features such as residual ink inside the hole, foreign matter on the hole wall, and defects in buried holes; S3: Use weighted fusion or feature-level fusion algorithms to fuse pure white light images and pure IR images to generate a final detection image that retains both surface details and deep information. S4: The weighting coefficients of the weighted fusion algorithm can be adaptively adjusted according to the defect type of the PCB board. When detecting surface defects, the weight of the white light image accounts for 60%-80%, and when detecting deep defects, the weight of the IR image accounts for 60%-80%. S5: The specific process of the feature-level fusion algorithm is as follows: extract the surface texture features of the pure white light image and the deep structure features of the pure IR image respectively, establish a feature dataset, perform feature matching and fusion, and generate the final detection image.

[0009] Furthermore, in step four, the specific operations for the adaptive application scenario are as follows: S1: For electroless gold / plated tin boards, pure white light images are used to inspect the surface coverage integrity, combined with IR images to suppress metal reflection interference, ensuring a metal reflection suppression rate of ≥90%; S2: For OSP / soldering boards, IR imaging is used to enhance the detection of defects inside the holes, combined with pure white light imaging to assist in positioning, ensuring that the detection rate of defects inside the holes is ≥99%.

[0010] Furthermore, in step five, the specific operation for outputting the detection result is as follows: S1: Compare the final detected image with the preset defect threshold, and mark the location and type of the defect using an image recognition algorithm; S2: Generate a visual inspection report based on the marking results. The report includes a defect distribution map, defect type statistics, and inspection accuracy data.

[0011] Furthermore, steps one through five also include an intelligent detection and collaborative control system, used to control the orderly execution of steps one through five in the time-division stroboscopic AVI optical detection method. The intelligent detection and collaborative control system includes a light source synchronization control module, an image separation and reconstruction module, a scene adaptive matching module, and a detection result output module. The light source synchronization control module is used to acquire the line scanning frequency data of the image sensor in real time, precisely adjust the alternating lighting frequency of the white LED and the IR light source, ensuring a frequency synchronization error ≤0.1Hz, and achieving precise linkage between light source triggering and line scanning. The image separation and reconstruction module is used to receive the synthesized image transmitted from the image sensor and process the synthesized image using a pixel-level splitting algorithm. The system ensures a splitting and reconstruction accuracy of ≥99.9%, outputting pure white light and pure IR images without pixel crosstalk. The scene adaptive matching module is used to identify the type of PCB board to be inspected. For electroplated / tin-plated boards, it automatically matches a "white light priority + IR reflection suppression" detection strategy, and for OSP / electroplated boards, it automatically matches an "IR hole defect enhancement + white light assisted positioning" detection strategy, ensuring a metal reflection suppression rate of ≥90% and a hole defect detection rate of ≥99%. The detection result output module is used to receive the fused final detection image, compare and analyze it with a preset defect threshold, mark the defect location and type, and generate a visualized detection report containing a defect distribution map, defect type statistics, and detection accuracy data.

[0012] The present invention has at least the following beneficial effects: 1. This invention employs time-division strobe illumination using white light and IR light sources, and synchronizes it with the image sensor's line scanning control. This fundamentally avoids interference from mixed illumination, ensuring that each line of the image corresponds to a single illumination condition. This effectively improves the image signal-to-noise ratio and the clarity of defect features, enabling precise differentiation between surface and deep defects.

[0013] 2. This invention employs an image fusion strategy that combines weighted fusion and feature-level fusion. This strategy can adaptively adjust the weight coefficients according to the defect type, fully integrating the surface detail information of the white light image with the deep structural information of the IR image. The resulting final detection image takes into account multi-dimensional defect features, significantly improving detection accuracy.

[0014] 3. This invention designs an adaptive inspection scheme for different types of PCB boards, effectively suppressing the interference of metal reflection from electroplated / tin-plated boards, reducing the missed detection rate of in-hole defects in OSP / electroplated boards, meeting diverse inspection needs, and having a wider range of applications.

[0015] 4. This invention integrates time-division strobe illumination, image separation and reconstruction, adaptive fusion, and scene adaptation into one system to build a complete PCB board defect detection system. The detection process is highly efficient and collaborative, which can greatly improve detection efficiency and provide reliable quality assurance for high-precision PCB board production. Attached Figure Description

[0016] Figure 1 This is a flowchart of the PCB board dual-light source time-division stroboscopic defect detection process of the present invention. Detailed Implementation

[0017] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention. Example:

[0018] As attached Figure 1 As shown: This invention provides an AVI optical detection method based on time-division stroboscopic motion, comprising the following steps: Step 1: Time-division strobe lighting control; Step 2: Image acquisition and separation; Step 3: Image fusion and enhancement; Step 4: Adapt to application scenarios; Step 5: Output the test results.

[0019] In step one, the specific operation of the time-division strobe lighting control is as follows: S1: White LED and IR light source are selected as the detection light source, wherein the wavelength of the IR light source is 850nm or 940nm; S2: Synchronize the light source trigger signal with the image sensor line scan signal to ensure that the light source alternation frequency is consistent with the image sensor line scan frequency, with a frequency synchronization error ≤0.1Hz; S3: Controls odd-numbered rows of images to be captured by white light illumination, and even-numbered rows of images to be captured by IR light illumination, ensuring that each row of images corresponds to a single illumination condition and avoiding interference from mixed illumination.

[0020] In step two, the specific operations for image acquisition and separation are as follows: S1: Control the image sensor to acquire images in a line scanning manner and generate a composite image composed of odd-numbered white light images and even-numbered IR images; S2: The synthesized image is transmitted to the image processing unit, and the synthesized image is processed by the pixel-level splitting algorithm to reconstruct the pure white light image and the pure IR image respectively; S3: Ensure that the image processing unit achieves a splitting and reconstruction accuracy of ≥99.9% for the synthesized image, and that there is no pixel crosstalk between the reconstructed pure white light image and the pure IR image.

[0021] In step three, the specific operations of image fusion and enhancement are as follows: S1: Perform surface texture and color feature analysis on pure white light images to identify surface defects such as scratches on the solder mask layer and uneven ink distribution on PCB boards. S2: Utilize the strong penetrating power of IR images to extract deep structural features such as residual ink inside the hole, foreign matter on the hole wall, and defects in buried holes; S3: Use weighted fusion or feature-level fusion algorithms to fuse pure white light images and pure IR images to generate a final detection image that retains both surface details and deep information. S4: The weighting coefficients of the weighted fusion algorithm can be adaptively adjusted according to the defect type of the PCB board. When detecting surface defects, the weight of the white light image accounts for 60%-80%, and when detecting deep defects, the weight of the IR image accounts for 60%-80%. S5: The specific process of the feature-level fusion algorithm is as follows: extract the surface texture features of the pure white light image and the deep structure features of the pure IR image respectively, establish a feature dataset, perform feature matching and fusion, and generate the final detection image.

[0022] In step four, the specific operations for the adaptive application scenario are as follows: S1: For electroless gold / plated tin boards, pure white light images are used to inspect the surface coverage integrity, combined with IR images to suppress metal reflection interference, ensuring a metal reflection suppression rate of ≥90%; S2: For OSP / soldering boards, IR imaging is used to enhance the detection of defects inside the holes, combined with pure white light imaging to assist in positioning, ensuring that the detection rate of defects inside the holes is ≥99%.

[0023] In step five, the specific operation for outputting the detection results is as follows: S1: Compare the final detected image with the preset defect threshold, and mark the location and type of the defect using an image recognition algorithm; S2: Generate a visual inspection report based on the marking results. The report includes a defect distribution map, defect type statistics, and inspection accuracy data.

[0024] Steps one through five also include an intelligent detection and collaborative control system, used to control the orderly execution of steps one through five in the time-division stroboscopic AVI optical detection method. The intelligent detection and collaborative control system includes a light source synchronization control module, an image separation and reconstruction module, a scene adaptive matching module, and a detection result output module. The light source synchronization control module is used to acquire the line scanning frequency data of the image sensor in real time, precisely adjust the alternating lighting frequency of the white LED and the IR light source, ensuring a frequency synchronization error ≤0.1Hz, and achieving precise linkage between light source triggering and line scanning. The image separation and reconstruction module is used to receive the synthesized image transmitted from the image sensor and process the synthesized image using a pixel-level splitting algorithm. The system ensures a splitting and reconstruction accuracy of ≥99.9%, outputting pure white light and pure IR images without pixel crosstalk. The scene adaptive matching module is used to identify the type of PCB board to be inspected. For electroplated / tin-plated boards, it automatically matches a "white light priority + IR reflection suppression" detection strategy, and for OSP / electroplated boards, it automatically matches an "IR hole defect enhancement + white light assisted positioning" detection strategy, ensuring a metal reflection suppression rate of ≥90% and a hole defect detection rate of ≥99%. The detection result output module is used to receive the fused final detection image, compare and analyze it with a preset defect threshold, mark the defect location and type, and generate a visualized detection report containing a defect distribution map, defect type statistics, and detection accuracy data.

[0025] The specific usage and function of this embodiment are as follows: This invention addresses the issue of PCB board defect detection from the perspectives of light source control and image processing. It upgrades the traditional single-light source imaging and fixed algorithm fusion detection method to a dual-light source time-division stroboscopic and multi-strategy adaptive AVI optical detection method. It employs a white LED and an 850nm wavelength IR light source as the detection light source, setting both the alternating illumination frequency of the light source and the image sensor's line scanning frequency to 100Hz, with the frequency synchronization error controlled within 0.05Hz. This achieves time-division capture of odd-numbered rows of white light images and even-numbered rows of IR images. Pixel-level splitting calculations are then used to further enhance the detection process. The method processes the synthesized image, achieving a 99.95% splitting and reconstruction accuracy. It outputs pure white light and pure IR images free of pixel crosstalk. For surface defect detection, the white light image weight is set to 70%, and for deep defect detection, the IR image weight is set to 70%, achieving image fusion. For electroless gold plates, an automatic "white light priority + IR reflection suppression" strategy is applied, increasing the metal reflection suppression rate to 92%. For OSP plates, an automatic "IR hole defect enhancement + white light assisted positioning" strategy is applied, increasing the hole defect detection rate to 99.2%. Compared to traditional single-light source detection methods, this method improves the overall defect detection rate by 30%, and compared to mixed-light detection methods, the image signal-to-noise ratio is improved by 25%. It simplifies the original 12-step detection process of "multiple imaging - separate processing - manual comparison," significantly improving detection efficiency.

[0026] Any aspects of this invention not described in detail are well-known to those skilled in the art.

[0027] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. An AVI optical detection method based on time-division stroboscopic motion, characterized in that, The following steps are used: Step 1: Time-division strobe lighting control; Step 2: Image acquisition and separation; Step 3: Image fusion and enhancement; Step 4: Adapt to application scenarios; Step 5: Output the test results.

2. The AVI optical detection method for time-division stroboscopic distortion as described in claim 1, characterized in that, In step one, the specific operation of the time-division flicker lighting control is as follows: S1: White LED and IR light source are selected as the detection light source, wherein the wavelength of the IR light source is 850nm or 940nm; S2: Synchronize the light source trigger signal with the image sensor line scan signal to ensure that the light source alternation frequency is consistent with the image sensor line scan frequency, with a frequency synchronization error ≤0.1Hz; S3: Controls odd-numbered rows of images to be captured by white light illumination, and even-numbered rows of images to be captured by IR light illumination, ensuring that each row of images corresponds to a single illumination condition and avoiding interference from mixed illumination.

3. The AVI optical detection method for time-division stroboscopic distortion as described in claim 1, characterized in that, In step two, the specific operations for image acquisition and separation are as follows: S1: Control the image sensor to acquire images in a line scanning manner and generate a composite image composed of odd-numbered white light images and even-numbered IR images; S2: The synthesized image is transmitted to the image processing unit, and the synthesized image is processed by the pixel-level splitting algorithm to reconstruct the pure white light image and the pure IR image respectively; S3: Ensure that the image processing unit achieves a splitting and reconstruction accuracy of ≥99.9% for the synthesized image, and that there is no pixel crosstalk between the reconstructed pure white light image and the pure IR image.

4. The AVI optical detection method for time-division stroboscopic distortion as described in claim 1, characterized in that, In step four, the specific operations of the adaptive application scenario are as follows: S1: For electroless gold / plated tin boards, pure white light images are used to inspect the surface coverage integrity, combined with IR images to suppress metal reflection interference, ensuring a metal reflection suppression rate of ≥90%; S2: For OSP / soldering boards, IR imaging is used to enhance the detection of defects inside the holes, combined with pure white light imaging to assist in positioning, ensuring that the detection rate of defects inside the holes is ≥99%.

5. The AVI optical detection method for time-division stroboscopic distortion as described in claim 1, characterized in that, In step five, the specific operation for outputting the detection result is as follows: S1: Compare the final detected image with the preset defect threshold, and mark the location and type of the defect using an image recognition algorithm; S2: Generate a visual inspection report based on the marking results. The report includes a defect distribution map, defect type statistics, and inspection accuracy data.

6. The AVI optical detection method for time-division stroboscopic distortion as described in claim 1, characterized in that, Steps one through five also include an intelligent detection and collaborative control system, which is used to control the orderly execution of steps one through five in the time-division stroboscopic AVI optical detection method. The intelligent detection and collaborative control system includes a light source synchronization control module, an image separation and reconstruction module, a scene adaptive matching module, and a detection result output module. The light source synchronization control module is used to collect the line scanning frequency data of the image sensor in real time, and precisely adjust the alternating lighting frequency of the white LED and the IR light source to ensure that the frequency synchronization error is ≤0.1Hz, thereby achieving precise linkage between light source triggering and line scanning. The image separation and reconstruction module is used to receive the synthesized image transmitted by the image sensor, process the synthesized image through a pixel-level splitting algorithm, ensure the splitting and reconstruction accuracy is ≥99.9%, and output a pure white light image and a pure IR image without pixel crosstalk. The scene adaptive matching module is used to identify the type of PCB board to be inspected. For electroplated / tin-plated boards, it automatically matches the "white light priority + IR reflection suppression" detection strategy, and for OSP / electroplated boards, it automatically matches the "IR hole defect enhancement + white light assisted positioning" detection strategy, ensuring that the metal reflection suppression rate is ≥90% and the hole defect detection rate is ≥99%. The detection result output module is used to receive the fused final detection image, compare and analyze it with a preset defect threshold, mark the defect location and type, and generate a visualized detection report containing defect distribution map, defect type statistics and detection accuracy data.

Images