A circuit board aperture online detection system and a detection method thereof

Abstract

This invention discloses an online inspection system and method for printed circuit board (PCB) apertures, specifically comprising: a platform for undisturbed, uniform-speed transmission, used to carry and drive the PCB through the inspection station at a constant linear velocity; a high-resolution incremental photoelectric encoder coaxially and directly connected to the drive shaft end of the platform, used to output differential pulse signals that strictly correspond to the absolute displacement of the PCB in real time; a non-transparent, dual-telecentric optical module, vertically mounted directly above the inspection station, with an optical distortion rate of less than 0.05%; the optical module is equipped with a large-area global exposure industrial camera, used to eliminate visible artifacts on the inner wall of the aperture at the edge of the field of view caused by the perspective angle; the printed circuit boards of this invention can flow through the inspection machine at extremely high speeds. The production capacity of one machine is equivalent to ten traditional stop-and-go automatic optical inspection machines, greatly reducing the depreciation costs of factory equipment.

Description

Technical Field

[0001] This invention relates to the field of surface mount technology in electronic manufacturing, specifically to an online detection system and method for circuit board apertures. Background Technology

[0002] With the rapid development of 5G communication, aerospace electronics, and advanced process chip packaging substrate technologies, printed circuit boards (PCBs) are evolving towards ultra-high multilayer construction, miniaturized hole diameters, and extremely fine wiring. In the drilling and metallization processes of PCB manufacturing, the precision of the hole diameter directly determines the conductivity of interlayer electrical interconnections and the yield rate of subsequent component insertion. Any hole diameter that is too small, too large, or has burrs inside will render the entire motherboard unusable.

[0003] Therefore, 100% inspection of the tens of thousands of micro-vias on printed circuit boards is a rigid requirement for electronics manufacturing plants. However, existing automated optical aperture inspection equipment reveals extremely serious underlying physical and optical defects when dealing with modern high-speed production lines: First, there's the inefficiency and vibration barrier caused by stop-and-go mechanical movement. Traditional automated optical inspection equipment must transport the printed circuit board to below the lens, stop the machine, take a picture, and then move to the next position to stop and take another picture. This frequent mechanical acceleration and deceleration not only results in extremely low throughput, making it unsuitable for high-speed production lines exceeding 1 meter per second, but more critically, each emergency stop triggers microscopic residual vibrations in the machine and lens. When measuring apertures at the 50-micron level, any mechanical micro-vibrations are amplified by the optical system, causing blurred image edges and completely losing the measurement reference.

[0004] Secondly, there's the perspective distortion problem with ordinary optical lenses. To improve inspection efficiency, equipment manufacturers typically use ordinary industrial lenses with a large field of view. However, according to the perspective principle of optical imaging, the viewing angle of an ordinary lens is conical with the optical center as its apex. This means that, apart from the central aperture directly facing the optical center, which appears perfectly circular, the micro-apertures at the edge of the field of view appear tilted to the camera. The camera sees not only the surface of the edge apertures but also even their inner walls. This perspective phenomenon causes the circular apertures at the edge of the field of view to appear as ellipses or trailing shapes in the image, making it impossible to extract the true center and diameter.

[0005] Finally, there's the optical conflict between motion blur and the strong reflectivity of the electroplated layer. Some manufacturers have attempted to use line scan cameras for continuous motion photography. However, under high-speed motion, continuous illumination from ordinary light sources causes the imaging sensor to record the movement trajectory of the object within the exposure time, resulting in fatal motion blur. Furthermore, printed circuit board apertures typically have immersion gold or electroplated copper layers, which produce intense specular reflections and highlight overflow under strong light, causing the aperture edges to be swallowed by halos. The aperture diameter extracted by the algorithm is often significantly smaller than the actual physical aperture diameter.

[0006] To address this, an online detection system and method for circuit board apertures are proposed. Summary of the Invention

[0007] The purpose of this invention is to provide an online detection system and method for circuit board apertures, aiming to solve one of the problems existing in the prior art.

[0008] Firstly, to solve the aforementioned technical problems, this application adopts a technical solution as follows: an online aperture detection system for circuit boards. This system abandons the inefficient stop-and-go mechanical stepping detection logic and achieves high-precision aperture measurement under high-speed continuous transmission based on a coupling mechanism of microsecond-level stroboscopic physical freezing and non-transparent orthographic projection. Specifically, it includes: The platform for undisturbed, uniform-speed transmission is used to carry and drive the circuit board through the inspection station at a constant linear speed; the drive shaft of the platform is coaxially and physically directly connected to a high-resolution incremental photoelectric encoder, which is used to output differential pulse signals that strictly correspond to the absolute displacement of the circuit board in real time. A non-transparent, telecentric optical module is vertically mounted directly above the inspection station, with an optical distortion rate of less than 0.05%. This optical module is equipped with a large-area global exposure industrial camera to eliminate visible artifacts on the inner wall caused by the perspective angle at the edge of the field of view. A transient nanosecond-level strobe illumination matrix is ​​arranged in a vertically opposed manner at the detection station, including a blue high-penetration backlight located below the circuit board and a red coaxial light source located above the circuit board and coaxial with the camera; the driving module of the light source is an ultra-high frequency constant current source based on silicon carbide switching devices, which has microsecond-level pulse lighting capability. The FPGA hard real-time synchronization engine is connected to the photoelectric encoder, industrial camera, and strobe illumination matrix via hardware pins. The synchronization engine is configured to divide and lock the encoder pulses. When the circuit board moves to the preset physical boundary of the optical field of view, without the intervention of any software system scheduling, the strobe illumination matrix is ​​directly triggered by the hardware gate circuit to emit an extremely short light pulse with a duration of 1 to 5 microseconds, and the global exposure of the camera is triggered synchronously at the microsecond level. The subpixel morphological edge measurement center acquires frozen images captured by the camera via gigabit Ethernet. Based on the physical diffraction edges generated by red-blue dual-wavelength interference in the image, the Zernike moment subpixel extraction algorithm is used to remove copper thickness reflection interference and calculate the true physical aperture of the circuit board micro-holes.

[0009] In one possible implementation, the support rail of the disturbance-free uniform speed transmission platform adopts an air-bearing guide rail; below the detection station, the support surface is replaced with a quartz glass support plate with extremely high light transmittance, and the blue high-transmittance backlight is arranged close to the bottom of the quartz glass support plate to ensure that the backlight can penetrate the through hole of the circuit board vertically and reach the camera above, forming a black and white silhouette with extremely high contrast.

[0010] In one possible implementation, the FPGA hard real-time synchronization engine is internally programmed with a motion blur physical cancellation algorithm. The pulse width that triggers the strobe illumination is strictly constrained by the line speed of the circuit board transmission and the system's calibrated pixel equivalent. Specifically, the product of the line speed of the circuit board transmission and the strobe pulse width of the light source must be less than or equal to half of the system's calibrated pixel equivalent. That is, in the extremely short time that the light pulse illuminates, the distance that the circuit board moves in physical space must be less than half the physical size of a pixel, thereby completely freezing motion blur at the optical sensor level.

[0011] In one possible implementation, the non-transparent dual telecentric optical module includes not only an object-side telecentric lens to eliminate the error of magnification varying with object distance, but also an image-side telecentric lens to ensure that light is perpendicularly incident on the microlens array of the camera sensor, thereby completely eliminating aperture measurement size drift caused by the slight vertical jitter of the circuit board during transmission.

[0012] In one possible implementation, the transient nanosecond-level stroboscopic illumination matrix uses a blue LED array with a wavelength of 450nm-470nm as the backlight source. The sharpest inner wall profile of the aperture is obtained by utilizing the physical characteristics of short wavelength and weak diffraction effect of blue light. The coaxial light source uses a red LED array with a wavelength of 620nm-630nm. The high reflectivity of red light to the gold or copper plating layer on the circuit board surface is used to extract the copper edge of the aperture. The two are combined to form a red-blue multispectral multiplexing.

[0013] Secondly, to solve the above-mentioned technical problems, another technical solution adopted in this application is: an online detection method for circuit board apertures using the system, comprising the following steps: Step S1, Rigid binding of physical displacement and optical exposure: The circuit board moves at high speed on the air-bearing platform, and the photoelectric encoder outputs a high-frequency pulse stream; the FPGA synchronization engine accurately counts the pulses, and whenever the cumulative number of pulses corresponds to the circuit board moving a physical distance of a complete camera field of view (minus the overlapping redundancy band), the underlying hardware immediately generates a trigger level. Step S2, Dual-spectrum microsecond-level flying freeze: The trigger level instantly reaches the stroboscopic constant current source and the camera, and the camera opens the global shutter; at the same time, the red coaxial light source and the blue back light source are excited with a transient current of thousands of amperes, emitting an extremely strong light burst that lasts for only 2 microseconds; the motion state of the circuit board is instantly optically frozen, and the camera target surface simultaneously captures the interference superposition image formed by the transmitted blue light and the reflected red light. Step S3, Dual-wavelength grayscale histogram stripping: After receiving the image, the sub-pixel morphological edge measurement center separates the R channel and the B channel; the high-contrast binarized image of the B channel is used to quickly locate the connected domain of the hole and define all regions of interest; then in the R channel, the grayscale gradient transition band caused by the uneven thickness of the electroplating layer on the hole wall is extracted for each region of interest. Step S4, Zernike Moment Subpixel Aperture Reconstruction: Within the locked grayscale transition zone, instead of relying on traditional integer pixel threshold segmentation, a two-dimensional Zernike spatial moment is used to calculate the continuous surface function of the image grayscale, and the real step physical edge inside the pixel is analyzed; the extracted subpixel edge point set is circle-fitted by the least squares method to output the absolute value of the physical aperture with micron-level precision. Step S5, Dynamic Adaptive Compensation for Thermal Expansion: The system's built-in infrared array reads the surface temperature of the circuit board currently passing through the detection station in real time. Based on the linear thermal expansion coefficient of different substrates, the physical aperture fitted in step S4 is subjected to temperature normalization inverse compensation, and the reference aperture at the standard temperature is output.

[0014] In one possible implementation, in step S2, the opening time window of the camera's global shutter is set to include and be slightly larger than the flicker time window of the light source; the camera's exposure sensitivity is entirely determined by the flicker pulse width of the light source, rather than by the camera's electronic shutter time, thereby completely eliminating the pollution of image contrast by stray light from the external environment.

[0015] In one possible implementation, during circle fitting in step S4, the system algorithm will remove outlier edge points that deviate from the center of the fitted circle by more than three standard deviations, thereby automatically filtering out false edge interference caused by residual dust or burrs in the hole and improving the repeatability accuracy of the hole diameter.

[0016] In one possible implementation, the system can simultaneously process more than 10,000 micro-holes in a single image captured by a flying camera on a circuit board. For each calculated hole diameter value, if it exceeds a preset upper or lower control limit, the system will map the physical coordinates of the hole back to the absolute coordinate system of the circuit board and directly drive the rejection / marking mechanism at the back end to physically mark the defective board through the lower-level computer.

[0017] Thirdly, to solve the above-mentioned technical problems, another technical solution adopted in this application is: a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the online detection method as described above.

[0018] The present invention has the following beneficial effects: 1. The printed circuit boards of this invention can flow through the inspection machine at extremely high speeds. The production capacity of one machine is equivalent to that of ten traditional stop-and-go automatic optical inspection machines, greatly reducing the factory's equipment depreciation costs.

[0019] 2. The dual telecentric system of this invention, combined with a sub-pixel algorithm, ensures that the measurement results are no longer based on spurious data stretched by calibration block software. Regardless of the aperture location or its minute size, the system outputs the absolute physical diameter based on optical orthographic projection, achieving a repeatability accuracy at the 1-micron level.

[0020] 3. The present invention, whether for highly reflective immersion gold boards, flexible printed circuit boards that are prone to moisture absorption and deformation, or substrates that expand slightly due to temperature, utilizes the combined effect of red and blue dual spectra and a thermodynamic compensation model to give the device an extremely strong material self-adaptive fault tolerance rate. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating an online detection method for circuit board apertures according to the present invention. Figure 2 This is a block diagram of an online circuit board aperture detection system according to the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Figure 1This is a flowchart illustrating an online detection method for circuit board aperture according to an embodiment of the present invention.

[0025] like Figure 2 The diagram shows a system module diagram for online detection of circuit board apertures according to the present invention. It should be noted that if substantially the same result is obtained, the method of this application is not based on... Figure 1 The sequence of processes shown is limited. Example 1

[0026] On a high-speed production line, a 600mm long packaging substrate covered with 200,000 micro-holes is speeding along an air-bearing guide at 800mm / s. If a traditional line scan camera were used at this time, because the line scan camera stitches the images line by line, any extremely small mechanical vibration during the transmission process would cause the stitched-up circular holes to become irregular shapes with jagged edges, completely losing the meaning of measurement.

[0027] This embodiment completely abandons the line scan scheme and adopts a large target area global shutter area array camera with dual telecentric lenses.

[0028] The physical function of a double telecentric lens: This is a large optical component with a complex array of multiple lenses that forcibly alters the direction of light refraction, allowing only the principal ray perpendicular to the lens surface to enter the camera. This means that the light spots projected onto the camera's target surface from the hundreds of thousands of micro-apertures on the substrate beneath the lens, whether at the very center or at the corner edges, are all perfectly vertical cylindrical projections. There is no observational bias due to the aperture walls, nor any optical distortion.

[0029] Extremely fast capture with hardware-level microsecond-level telephoto lenses: How can a high-speed moving substrate be photographed using a panel array camera? This relies on the precise time control of the FPGA's high-speed engine. An incremental photoelectric encoder, directly connected to the conveyor belt, outputs a high-frequency differential pulse stream as the belt rotates. These level signals are directly fed into the FPGA's hardware pins. The FPGA does not run any operating system; it only performs low-level gate circuit physical counting.

[0030] The system pre-calibrates the camera's field of view to a physical width of 100mm. For image stitching, it is set to take one picture every 90mm. This means that every time the FPGA completes a preset number of pulses, the voltage of the hardware circuit will instantly flip, generating an extremely fast trigger edge.

[0031] This trigger edge acts simultaneously on both the camera and the light source controller within a nanosecond timeframe. The camera's electronic shutter is fully open, awaiting light. At this moment, the ultra-high frequency constant current source, built on silicon carbide MOSFETs, inputs a transient pulse current of up to several hundred amperes into the red and blue dual-color LED array the instant the trigger edge is received. The LEDs burst forth with a dazzling burst of light, but the duration is strictly controlled within 2 microseconds.

[0032] Two microseconds later, the light array went out instantly, the camera shutter closed, and it began transmitting image data to the gigabit network cable.

[0033] The principle behind physical blur freezing: Within the 2 microseconds of light emission, the substrate moves a distance of 1.6 micrometers at a speed of 800 mm / s. The physical pixel equivalent of the camera and lens used in the system is calibrated to 3 micrometers / pixel. Since a physical displacement of 1.6 micrometers is only equivalent to half a pixel's worth of motion blur, it cannot cross the boundary of adjacent photosensitive pixels on the optical sensor. Therefore, although the substrate is moving at high speed, in the resulting megapixel image, all the micro-holes remain absolutely stationary with sharp edges, without the slightest motion blur. This is the core advantage of the flying camera technology. Example 2

[0034] Taking a clear photo is only the first step. The key at the algorithm level is how to remove interference from the image and extract the true physical aperture.

[0035] Physical separation of red and blue spectra: The blue light from the bottom of the substrate has extremely strong penetrating power and a small diffraction angle. It passes through the micro-holes directly to the camera, leaving high-contrast white bright spots in the B channel of the image against a pure black substrate. The red light from the top of the substrate is irradiated coaxially and perpendicularly, and is strongly reflected by the electroplated copper layer on the surface of the aperture. In the R channel of the image, the copper ring around the aperture appears bright, while the inside of the aperture appears black.

[0036] The advantage of this design is that when there are tiny foreign objects or burrs inside the hole, or when the electroplating layer on the hole wall is too thick, causing the actual effective hole diameter to become smaller, the blue light at the bottom will be physically blocked, and the light spot in channel B will accurately reflect the minimum light-passing bottleneck diameter inside the hole. The red light at the top is used to monitor whether there is copper layer peeling or misalignment on the hole surface.

[0037] Zernike subpixel-level limit calculation: After the system acquires the image, it does not use traditional Canny or Sobel operators to find edges because the edges found by these operators are composed of coarse square pixels with extremely poor accuracy. The sub-pixel morphological edge measurement center of this invention is introduced, which uses two-dimensional Zernike orthogonal moments in the grayscale transition zone where R / B channels overlap.

[0038] It treats the grayscale transition from black to white as a three-dimensional continuous grayscale topographic map. Through complex calculus matrix operations, the algorithm calculates the steepest tangent position of the grayscale surface at the boundary between two adjacent pixels. This position overcomes the physical limitations of the camera's photodiode grid, calculating the coordinates of the edge points to two decimal places.

[0039] Subsequently, the algorithm collects hundreds of sub-pixel edge points around the same hole and puts them into the least squares model to fit the equation of a perfect circle.

[0040] Robust noise rejection mechanism: During fitting, since the orifice walls may be contaminated with tiny dust particles, after each time the algorithm fits a circle's center and radius, it calculates the standard deviation of the distances from all edge points to the center. If the error of a point from the center exceeds three times the standard deviation, the algorithm classifies this point as dust or a burr, rejects it, and refits the circle equation. This iterative adaptive purification algorithm enables the system to output extremely robust measurement data even in dirty industrial environments. Example 3

[0041] When measurement accuracy reaches the 1-micron level, thermal expansion and contraction become a key factor affecting measurement accuracy. The substrate of printed circuit boards has a non-negligible linear coefficient of thermal expansion, typically ranging from 13 to 18 ppm per degree Celsius.

[0042] Testing the same board in the early morning of winter and the afternoon of summer, a substrate with a length of 500mm, will show significant physical expansion and contraction due to temperature differences. Hole diameters and spacing that were originally within acceptable limits may be mistakenly judged as out-of-tolerance defects without compensation due to changes in ambient temperature.

[0043] This invention constructs a thermophysical compensation defense line. At the front end of the inspection station, a non-contact medical-grade infrared temperature measurement array is deployed to continuously scan the real-time temperature distribution field on the surface of the printed circuit board entering the machine.

[0044] After the algorithm calculates the subpixel diameter and absolute physical coordinates of a certain hole, it does not output them immediately. The system inputs this dimensional data into the thermal expansion dynamic adaptive compensation model. The model extracts the linear thermal expansion coefficient parameter library of the substrate of the printed circuit board, combines it with the current transient board surface temperature transmitted back by the infrared probe, and uses reverse thermal shrinkage calculation to restore the measured size that has increased due to high temperature expansion to the absolute reference size at the international standard detection temperature.

[0045] This thermodynamic compensation design completely eliminates the seasonal accuracy drift problem caused by alternating ambient temperature to micron-level high-precision equipment, giving the system absolute calibration-free measurement stability 365 days a year, in all weather conditions and temperature ranges.

[0046] For other details regarding the implementation techniques of each module in the above-described system, please refer to the description of an online detection method for circuit board apertures in Embodiment 1 above, which will not be repeated here.

[0047] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system-type embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0048] like Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. It illustrates a structural schematic diagram suitable for implementing the electronic device in the embodiment of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0049] like Figure 3 As shown, the electronic device includes a processor, a memory, and a communication interface. The memory stores a computer program, and when the processor executes the computer program, it implements the online detection method for circuit board apertures according to the aforementioned embodiments of this disclosure. The electronic device can exchange data with other devices or systems through the communication interface.

[0050] The processor in the aforementioned electronic device serves as its core, responsible for executing the computer program stored in the memory to implement various functions of the online detection method for circuit board apertures. The processor can employ a high-performance multi-core CPU or a dedicated chip to meet the demands of complex calculations and real-time processing. The memory stores the operating system, application programs, data, and computer programs. In this embodiment, the memory stores the computer program implementing the online detection method for circuit board apertures. The memory can be RAM, ROM, Flash memory, or other types of non-volatile memory. The communication interface connects the electronic device to other devices or networks, enabling data transmission and exchange. In this embodiment, the communication interface supports various communication protocols and interface standards, such as Wi-Fi, Bluetooth, USB, and Ethernet, to meet communication needs in different scenarios.

[0051] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0052] According to embodiments of the present disclosure, a computer-readable storage medium stores a computer program, which, when executed by a processor, implements the functions of the online detection method for circuit board apertures described in the foregoing embodiments of the present disclosure.

[0053] The aforementioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0054] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here. The above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A circuit board aperture on-line inspection system, characterized by, Specifically, it includes: A platform is used to carry and drive the circuit board through the inspection station at a constant linear speed; the drive shaft of the platform is physically connected to a photoelectric encoder on the same axis. An optical module is vertically mounted directly above the inspection station, with an optical distortion rate of less than 0.05%; the optical module is equipped with an industrial camera for large-area global exposure. A strobe illumination matrix is ​​arranged in a vertically opposed manner at the detection station. The strobe illumination matrix includes a blue light penetrating backlight located below the circuit board and a red light coaxial light source located above the circuit board and coaxial with the camera. The synchronization engine is connected to the photoelectric encoder, industrial camera, and strobe illumination matrix via hardware pins; the synchronization engine is configured to perform frequency division and phase locking on the encoder pulses. The subpixel morphological edge measurement center acquires frozen images captured by the camera via gigabit Ethernet. Based on the physical diffraction edges generated by red-blue dual-wavelength interference in the image, the Zernike moment subpixel extraction algorithm is used to remove copper thickness reflection interference and calculate the true physical aperture of the circuit board micro-holes.

2. The line board hole diameter on-line inspection system according to claim 1, wherein, The support rail of the platform adopts an air-bearing guide rail; below the testing station, the support surface of the platform is a quartz glass support plate, and the blue light penetrating backlight is arranged close to the bottom of the quartz glass support plate.

3. The system according to claim 1, wherein The synchronization engine is internally programmed with a motion blur physical cancellation algorithm, and the pulse width that triggers the strobe illumination is physically constrained by the circuit board transmission line speed and the system's calibrated pixel equivalent.

4. The online circuit board aperture detection system according to claim 1, characterized in that, The optical module includes an object-side telecentric lens and a microlens array.

5. The online circuit board aperture detection system according to claim 1, characterized in that, In the strobe lighting matrix, the blue light penetrating backlight uses a blue LED array with a wavelength of 450nm-470nm; the red light coaxial light source uses a red LED array with a wavelength of 620nm-630nm.

6. A method for online detection of circuit board aperture using the system described in any one of claims 1 to 5, characterized in that, Includes the following steps: Step S1, Rigid binding of physical displacement and optical exposure: The circuit board moves on the air-bearing platform, and the photoelectric encoder outputs a high-frequency pulse stream; The FPGA synchronization engine counts the pulses. Whenever the accumulated pulse count corresponds to the physical distance that the circuit board has moved by a complete camera field of view, the underlying hardware immediately generates a trigger level. Step S2, Dual-spectrum microsecond-level aerial photography freeze: The trigger level reaches the stroboscopic constant current source and the camera, and the camera opens the global shutter; at the same time, the red coaxial light source and the blue back light source are excited, emitting a strong light burst lasting for 2 microseconds; the camera target surface simultaneously captures the interference superposition image formed by the transmitted blue light and the reflected red light; Step S3, Dual-wavelength grayscale histogram stripping: After receiving the image, the sub-pixel morphological edge measurement center separates the R channel and the B channel; the high-contrast binarized image of the B channel is used to coarsely locate the connected components of the hole and define all regions of interest; then in the R channel, the grayscale gradient transition band caused by the uneven thickness of the electroplating layer on the hole wall is extracted for each region of interest. Step S4, Zernike Moment Subpixel Aperture Reconstruction: Within the locked grayscale transition zone, the continuous surface function of the image grayscale is calculated using two-dimensional Zernike spatial moments to resolve the real step physical edges inside the pixels; the extracted subpixel edge point set is circle-fitted using the least squares method to output the absolute value of the physical aperture with micron-level precision. Step S5, Dynamic Adaptive Compensation for Thermal Expansion: The system's built-in infrared array reads the surface temperature of the circuit board currently passing through the detection station in real time. Based on the linear thermal expansion coefficient of different substrates, the physical aperture fitted in step S4 is subjected to temperature normalization inverse compensation, and the reference aperture at the standard temperature is output.

7. The online detection method for circuit board aperture according to claim 6, characterized in that, In step S2, the opening time window of the camera's global shutter is longer than the flicker time window of the light source.

8. The online detection method for circuit board aperture according to claim 6, characterized in that, In step S4, when performing circle fitting, outlier edge points whose center deviation from the fitted circle exceeds three times the standard deviation are included.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the detection method as described in any one of claims 6-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the detection method as described in any one of claims 6-8.

Images