PCB processing deviation correction method and system based on visual calibration

Abstract

The present application relates to the technical field of PCB production, and particularly relates to a PCB processing deviation rectification method based on visual calibration, which comprises the following steps: selecting a PCB visual calibration point based on feature saliency; collecting multi-source feature information of the PCB visual calibration point in real time; obtaining multi-source deviation rectification influence factors; constructing a deviation rectification compensation function based on PCB physical deformation; and constructing a deviation rectification compensation parameter optimization model based on the multi-source deviation rectification influence factors and the deviation rectification compensation function. The present application introduces a deviation rectification compensation function which comprehensively considers multiple factors, can compensate for the position and size deviations of the PCB caused by factors such as temperature changes and physical deformation, and calculates the corresponding influence coefficients for each visual calibration point through the multi-source deviation rectification influence factors, so that the more reliable visual calibration points have a greater impact on the results, thereby improving the precision of PCB processing and enhancing product quality.

Description

Technical Field

[0001] This invention relates to the field of PCB manufacturing technology, and in particular to a method and system for correcting deviations in PCB manufacturing based on visual calibration. Background Technology

[0002] Precise position and dimension calibration is crucial in the manufacturing, inspection, and assembly of PCB boards. However, existing technologies have many shortcomings when dealing with complex and ever-changing processing environments.

[0003] On the one hand, temperature changes cause PCB materials to expand and contract. Due to uneven local temperature distribution, this expansion or contraction is often non-uniform, resulting in deviations between the actual and theoretical positions of various points on the PCB. Currently, most existing methods are unable to accurately compensate for these dimensional and positional changes caused by temperature variations, significantly impacting processing accuracy. On the other hand, mechanical stress is also a key factor affecting PCB accuracy. During the production, transportation, and use of PCBs, various mechanical stresses, such as bending, stretching, and compression, are unavoidable. These stresses cause physical deformations in the PCB, such as warping and twisting. Existing technologies cannot comprehensively and accurately calculate and compensate for the deformation caused by mechanical stress, resulting in significant errors between the actual dimensions and positions of the PCB and the design values, seriously affecting product quality and production efficiency. Summary of the Invention

[0004] The main objective of this invention is to provide a PCB board processing correction method and system based on visual calibration, aiming to solve the technical problems mentioned in the background art.

[0005] This invention proposes a PCB board processing correction method based on visual calibration, comprising:

[0006] PCB visual calibration points are selected based on feature saliency.

[0007] Real-time acquisition of multi-source feature information of PCB visual calibration points; and acquisition of visual confidence, temperature influence attenuation value, and vibration influence attenuation value of visual calibration points based on the multi-source feature information of PCB visual calibration points.

[0008] The visual confidence level of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point are fused to obtain a multi-source correction influence factor.

[0009] Constructing a correction and compensation function based on PCB physical deformation;

[0010] An optimization model for correction compensation parameters is constructed based on the multi-source correction influencing factors and the correction compensation function.

[0011] The optimal correction compensation parameters are obtained by solving the optimization model of the correction compensation parameters through optimization algorithm.

[0012] The correction compensation parameters are subjected to time-series smoothing.

[0013] Based on the smoothed correction and compensation parameters, x-axis translation distance, y-axis translation distance, and rotation angle compensation commands are generated.

[0014] The x-axis translation distance, y-axis translation distance, and rotation angle compensation commands are transmitted to the PCB board processing equipment, which adjusts the target processing position based on the x-axis translation distance, y-axis translation distance, and rotation angle compensation commands.

[0015] The present invention is further configured such that the step of obtaining the visual confidence level, the temperature influence attenuation value, and the vibration influence attenuation value of the visual calibration point based on the multi-source feature information of the PCB visual calibration point includes:

[0016] Obtain the coordinate information of the PCB visual calibration points and the PCB image information, and obtain the visual confidence of the visual calibration points based on the coordinate information of the PCB visual calibration points and the PCB image information;

[0017] Obtain PCB surface temperature distribution information and PCB thermal expansion coefficient, and obtain the temperature influence attenuation value of visual calibration point based on the PCB surface temperature distribution information and PCB thermal expansion coefficient;

[0018] Obtain the coordinates of the PCB fixed point and the overall vibration acceleration value. Based on the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point and the overall vibration acceleration value, obtain the vibration influence attenuation value of the visual calibration point.

[0019] The visual confidence level of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point are fused to obtain the multi-source correction influence factor.

[0020] The present invention is further configured such that the step of obtaining the visual confidence of the visual calibration point based on the coordinate information of the PCB visual calibration point and the PCB image information includes:

[0021] The average grayscale value of the visual calibration points is obtained based on the PCB image information.

[0022] The standard deviation of the background area corresponding to the visual calibration point is obtained based on the PCB image information.

[0023] The signal-to-noise ratio of the visual calibration point is calculated based on the mean gray value of the visual calibration point and the standard deviation of the background region corresponding to the visual calibration point.

[0024] Based on the PCB image information, obtain the standard deviation of the Gaussian fitting at the visual calibration point edge and the standard deviation of the Gaussian fitting at the maximum edge.

[0025] The spatial relationship coefficients of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points.

[0026] The visual confidence level of the visual calibration point is obtained based on the signal-to-noise ratio of the visual calibration point, the spatial relationship coefficient of the visual calibration point, the standard deviation of the Gaussian fitting of the edge of the visual calibration point, and the standard deviation of the Gaussian fitting of the maximum edge.

[0027] The present invention is further configured such that the step of obtaining the spatial relationship coefficient of the visual calibration point based on the coordinate information of the PCB visual calibration point includes:

[0028] The angle between the visual calibration point and adjacent visual calibration points is obtained based on the coordinate information of the PCB visual calibration point;

[0029] The distance between the visual calibration point and the adjacent visual calibration point is obtained based on the coordinate information of the PCB visual calibration point;

[0030] Obtain the theoretical angle between the visual calibration point and adjacent visual calibration points;

[0031] Obtain the theoretical distance between the visual calibration point and its adjacent visual calibration points;

[0032] Calculate the angular similarity between the visual calibration point and the adjacent visual calibration points and the theoretical angle;

[0033] Calculate the distance similarity between the visual calibration point and its adjacent visual calibration points and the theoretical distance;

[0034] The spatial relationship coefficients of the visual calibration points are calculated based on the angular similarity and distance similarity.

[0035] The present invention is further configured such that the step of obtaining the attenuation value of the temperature influence at the visual calibration point based on the PCB surface temperature distribution information and the PCB thermal expansion coefficient includes:

[0036] The temperature deviation of the visual calibration point is obtained based on the PCB surface temperature distribution information.

[0037] Obtain the reference temperature of the visual calibration point;

[0038] Obtain the coefficient of thermal expansion of the PCB;

[0039] Obtain the system's temperature sensitivity coefficient;

[0040] The attenuation value of the temperature effect at the visual calibration point is obtained based on the temperature deviation at the visual calibration point, the thermal expansion coefficient of the PCB, the reference temperature at the visual calibration point, and the system's temperature sensitivity coefficient.

[0041] The present invention is further configured such that the step of obtaining the vibration influence attenuation value of the visual calibration point based on the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point, and the overall vibration acceleration value includes:

[0042] The actual coordinates of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points.

[0043] Obtain the coordinates of the PCB fixed point;

[0044] The distance influence coefficient is obtained based on the difference between the coordinates of the PCB fixed point and the coordinates of the PCB fixed point.

[0045] The root mean square value of the vibration acceleration at the visual calibration point is obtained based on the overall vibration acceleration value.

[0046] Obtain the vibration sensitivity coefficient;

[0047] The vibration influence attenuation value of the visual calibration point is obtained based on the distance influence coefficient, the overall vibration acceleration value, and the vibration sensitivity coefficient.

[0048] The present invention is further configured such that the step of constructing the correction compensation function based on PCB physical deformation includes:

[0049] Obtain the elastic modulus of the PCB substrate and the stress at the visual calibration point. Calculate the strain at the visual calibration point in the x-axis direction and the strain at the visual calibration point in the y-axis direction based on the elastic modulus of the PCB substrate and the stress at the visual calibration point.

[0050] The actual coordinates of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points.

[0051] Obtain the coefficient of thermal expansion of the PCB;

[0052] Obtain the temperature distribution information on the PCB surface;

[0053] A correction and compensation function is constructed based on the strain of the visual calibration point in the x-axis direction, the strain of the visual calibration point in the y-axis direction, the actual coordinates of the visual calibration point, the thermal expansion coefficient of the PCB, the elastic modulus of the PCB substrate, and the temperature distribution information of the PCB surface.

[0054] The present invention is further configured such that the step of constructing the correction compensation parameter optimization model based on the multi-source correction influence factor and the correction compensation function includes:

[0055] Obtain the theoretical coordinates of the visual calibration point;

[0056] Obtain the correction compensation function;

[0057] Obtain the multi-source correction influence factor;

[0058] Retrieve the previous correction compensation parameters;

[0059] Obtain the impact value of the adjustment range of the correction compensation parameters;

[0060] An optimization model for the correction compensation parameters is constructed based on the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influence factor, the previous correction compensation parameters, and the influence value of the adjustment range of the correction compensation parameters.

[0061] The present invention is further configured such that the step of obtaining the influence value of the adjustment range of the correction compensation parameter includes:

[0062] Obtain the vibration influence coefficient;

[0063] Obtain the impact value of the initial correction compensation parameter adjustment range;

[0064] The influence value of the adjustment range of the correction compensation parameters is calculated based on the influence value of the initial correction compensation parameter adjustment range and the vibration influence coefficient.

[0065] This invention also discloses a PCB manufacturing deviation correction system based on visual calibration, comprising:

[0066] The selection module is used to select PCB visual calibration points based on feature saliency.

[0067] The acquisition module is used to acquire multi-source feature information of PCB visual calibration points in real time, and obtain the visual confidence level, temperature influence attenuation value and vibration influence attenuation value of visual calibration points based on the multi-source feature information of PCB visual calibration points.

[0068] The fusion module is used to fuse the visual confidence of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point to obtain a multi-source correction influence factor.

[0069] The first calculation module is used to construct the correction compensation function based on the physical deformation of the PCB.

[0070] The second calculation module is used to construct an optimization model for correction compensation parameters based on the multi-source correction influence factor and the correction compensation function.

[0071] The third calculation module is used to solve the optimization model of the correction compensation parameters through optimization algorithms to obtain the optimal correction compensation parameters.

[0072] The fourth calculation module is used to perform time-series smoothing processing on the correction compensation parameters;

[0073] The first control module is used to generate x-axis translation distance, y-axis translation distance and rotation angle compensation commands based on the smoothed correction compensation parameters;

[0074] The second control module is used to transmit the x-axis translation distance, y-axis translation distance, and rotation angle compensation commands to the PCB board processing equipment, and the PCB board processing equipment adjusts the target processing position based on the x-axis translation distance, y-axis translation distance, and rotation angle compensation commands.

[0075] The beneficial effects of this invention are as follows: This invention introduces a correction compensation function that comprehensively considers multiple factors, which can accurately compensate for PCB position and size deviations caused by factors such as temperature changes and physical deformation. Furthermore, by calculating the corresponding influence coefficient for each visual calibration point through multi-source correction influence factors, these more reliable visual calibration points have a greater impact on the results when calculating correction compensation parameters, thereby effectively improving the overall calculation accuracy, increasing the precision of PCB processing, and effectively improving product quality. Attached Figure Description

[0076] Figure 1 This is a schematic diagram of a method flow according to an embodiment of this application.

[0077] Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application.

[0078] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0079] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0080] like Figure 1 As shown, this application provides a PCB board processing correction method based on visual calibration, including:

[0081] S1. Select PCB visual calibration points based on feature saliency. PCB visual calibration points with saliency include Mark points, hole positions, edge contours, points with high contrast to the surrounding area, and points with unique shapes. They also include points determined according to the specific manufacturing process. For example, in the surface mount process, it is necessary to ensure that the chip can be accurately mounted in the specified position. Therefore, the edge or center of the pad corresponding to the chip pin will be selected as the visual calibration point so as to accurately adjust the position and angle of the chip during the surface mount process.

[0082] S2, real-time acquisition of multi-source feature information of PCB visual calibration points, and acquisition of visual confidence, temperature influence attenuation value and vibration influence attenuation value of visual calibration points based on the multi-source feature information of PCB visual calibration points;

[0083] S3, the visual confidence level of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point are fused to obtain the multi-source correction influence factor, the calculation formula of which is:

[0084]

[0085] In the formula, w i C represents the multi-source correction influence factor for the i-th visual calibration point. i Let ∑C represent the visual confidence score of the i-th visual calibration point. i1 ω represents the sum of visual confidence scores for all visual calibration points. i g represents the attenuation value of temperature effect at the i-th visual calibration point. i This represents the attenuation value of the vibration effect at the i-th visual calibration point;

[0086] S4, constructing a correction and compensation function based on PCB physical deformation;

[0087] S5. Construct a correction compensation parameter optimization model based on the multi-source correction influence factor and the correction compensation function;

[0088] S6. Solve the optimization model of the correction compensation parameter through the optimization algorithm to obtain the optimal correction compensation parameter. During the iteration process, continuously update the correction compensation parameter so that the value of the correction compensation parameter gradually decreases until the convergence condition is met (such as the optimal correction compensation parameter is when the change of the correction compensation parameter is less than the preset threshold).

[0089] S7, perform time-series smoothing processing on the correction compensation parameters;

[0090] S8, Generate x-axis translation distance, y-axis translation distance and rotation angle compensation commands based on the smoothed correction compensation parameters;

[0091] S9, the x-axis translation distance, y-axis translation distance and rotation angle compensation command are transmitted to the PCB board processing equipment, and the PCB board processing equipment adjusts the target processing position based on the x-axis translation distance, y-axis translation distance and rotation angle compensation command.

[0092] This invention introduces a comprehensive correction and compensation function that considers multiple factors, enabling precise compensation for PCB position and size deviations caused by temperature changes, physical deformation, and other factors. The correction and compensation function considers not only scaling factors, thermal expansion coefficients, and PCB translation factors (described below), but also physical deformation compensation terms through stress-strain relationship calculations (described below), making the coordinates after projection transformation closer to the theoretical coordinates. Furthermore, this invention uses multi-source correction influence factors, calculating corresponding influence coefficients for each visual calibration point based on the confidence level of the visual calibration point, the temperature deviation of the area, and overall vibration conditions. Different visual calibration points are assigned different levels of importance, making more reliable visual calibration points have a greater impact on the results when calculating projection transformations and optimizing correction and compensation parameters, thereby effectively improving the overall calculation accuracy. This method enhances the environmental adaptability and robustness of PCB manufacturing, significantly improving processing accuracy while maintaining computational efficiency.

[0093] The steps of obtaining the visual confidence level, temperature-induced attenuation value, and vibration-induced attenuation value of the visual calibration point based on the multi-source feature information of the PCB visual calibration point include:

[0094] Obtain the coordinate information of the PCB visual calibration points and the PCB image information, and obtain the visual confidence of the visual calibration points based on the coordinate information of the PCB visual calibration points and the PCB image information;

[0095] Obtain PCB surface temperature distribution information and PCB thermal expansion coefficient, and obtain the temperature influence attenuation value of visual calibration point based on the PCB surface temperature distribution information and PCB thermal expansion coefficient;

[0096] Obtain the coordinates of the PCB fixed point and the overall vibration acceleration value. Based on the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point and the overall vibration acceleration value, obtain the vibration influence attenuation value of the visual calibration point.

[0097] The visual confidence level of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point are fused to obtain the multi-source correction influence factor.

[0098] The purpose of fusing the visual confidence level, temperature attenuation value, and vibration attenuation value of the visual calibration point is to comprehensively consider the influence of multiple factors such as vision, temperature, and vibration on the PCB board processing accuracy. The result is a multi-source correction influence factor, which can be used to guide the correction operation in the PCB board processing, determine the importance of the visual calibration point, and achieve more precise processing control by comprehensively considering these multi-source factors, thereby improving the processing quality and accuracy of the PCB board.

[0099] The step of obtaining the visual confidence level of the visual calibration point based on the coordinate information of the PCB visual calibration point and the PCB image information includes:

[0100] The average grayscale value of the visual calibration points is obtained based on the PCB image information.

[0101] The standard deviation of the background area corresponding to the visual calibration point is obtained based on the PCB image information.

[0102] The signal-to-noise ratio of the visual calibration point is calculated based on the mean gray value of the visual calibration point and the standard deviation of the background region corresponding to the visual calibration point. The calculation formula is as follows:

[0103]

[0104] In the formula, SNR i The signal-to-noise ratio of the i-th visual calibration point, u i Let q represent the mean gray value of the i-th visual calibration point. i This represents the standard deviation of the background region corresponding to the i-th visual calibration point;

[0105] Based on the PCB image information, obtain the standard deviation of the Gaussian fitting at the visual calibration point edge and the standard deviation of the Gaussian fitting at the maximum edge.

[0106] The spatial relationship coefficients of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points.

[0107] The visual confidence level of the visual calibration point is obtained based on the signal-to-noise ratio of the visual calibration point, the spatial relationship coefficient of the visual calibration point, the standard deviation of the Gaussian fitting of the edge of the visual calibration point, and the standard deviation of the Gaussian fitting of the maximum edge. The calculation formula is as follows:

[0108]

[0109] In the formula, C i SNR represents the visual confidence score of the i-th visual calibration point. i This represents the signal-to-noise ratio at the i-th visual calibration point. This represents the standard deviation of the Gaussian fit at the edge of the i-th visual calibration point. z represents the standard deviation of the maximum marginal Gaussian fit. i This represents the spatial relationship coefficient of the visual calibration points.

[0110] In the PCB manufacturing environment, factors such as changes in lighting, stains, and wear can interfere with the identification of visual calibration points. Since various processing steps, such as drilling and component placement, require precise location information, accurately determining the position of PCB visual calibration points is crucial for ensuring production quality and efficiency. This invention improves the accuracy and reliability of PCB visual inspection systems by utilizing the visual confidence score of PCB visual calibration points. In this invention, the visual confidence score of PCB visual calibration points is calculated using the signal-to-noise ratio (SNR) of the visual calibration points, the spatial relationship coefficient of the visual calibration points, the standard deviation of the Gaussian fit at the edges of the visual calibration points, and the standard deviation of the Gaussian fit at the maximum edges. A higher SNR indicates better image quality and more reliable PCB visual calibration points. In clear, noise-free images, the signals of PCB visual calibration points are clearly highlighted, facilitating accurate identification. Regarding the location and edge sharpness, the standard deviation is calculated by fitting the gradient distribution of the edge using Gaussian fitting. The smaller the value, the sharper the edge. After normalization, the sharpness is converted into a value between 0 and 1, with the maximum value corresponding to the sharpest edge. This ensures that calibration points with clear edges receive higher weights in the confidence calculation. To convert the confidence of a single PCB visual calibration point into a weight relative to all PCB visual calibration points, a normalized weight allocation is used to ensure that the sum of the weights of all PCB visual calibration points is 1. In this way, when a PCB visual calibration point has a high SNR and sharp edges, its weight is significantly increased. Conversely, PCB visual calibration points with low confidence due to occlusion, blurring, or other reasons have their weights automatically reduced, effectively minimizing their negative impact on the calculation results. By calculating the visual confidence of PCB visual calibration points, the environmental adaptability and reliability of the PCB processing correction system are effectively improved.

[0111] The step of obtaining the spatial relationship coefficients of the visual calibration points based on the coordinate information of the PCB visual calibration points includes:

[0112] The angle between the visual calibration point and adjacent visual calibration points is obtained based on the coordinate information of the PCB visual calibration point;

[0113] The distance between the visual calibration point and the adjacent visual calibration point is obtained based on the coordinate information of the PCB visual calibration point;

[0114] Obtain the theoretical angle between the visual calibration point and adjacent visual calibration points;

[0115] Obtain the theoretical distance between the visual calibration point and its adjacent visual calibration points; the theoretical angle and theoretical distance are obtained through PCB design drawings;

[0116] Calculate the angular similarity between the visual calibration point and the adjacent visual calibration points and the theoretical angle;

[0117] Calculate the distance similarity between the visual calibration point and its adjacent visual calibration points and the theoretical distance;

[0118] The spatial relationship coefficients of the visual calibration points are calculated based on the angular similarity and distance similarity.

[0119] Angular similarity and distance similarity are calculated using similarity formulas such as cosine similarity or Pearson correlation coefficient. The average of the angular similarity and distance similarity values ​​is the spatial relationship coefficient of the visual calibration point.

[0120] Features on a PCB are often not isolated; they have spatial relationships. For example, features such as circuits and solder joints are densely packed and interconnected. Considering the relationship between adjacent visual calibration points helps to more accurately determine the location of these points. PCBs may be affected by various factors under different production environments and usage conditions, such as oil stains, dust, and uneven lighting. Considering the relationship between adjacent visual calibration points can enhance the adaptability of the inspection system to these complex environments. By mutually verifying adjacent visual calibration points, the interference of environmental factors on the judgment of individual PCB visual calibration points can be reduced, improving the stability of the inspection. The relationship between adjacent visual calibration points can be used to optimize the inspection process. If the distribution of adjacent visual calibration points conforms to normal PCB design rules, the confidence of the PCB visual calibration point will increase; otherwise, it will decrease. Thus, the confidence calculation can be further optimized from the perspective of the spatial distribution of visual calibration points.

[0121] The step of obtaining the attenuation value of the visual calibration point temperature effect based on the PCB surface temperature distribution information and the PCB thermal expansion coefficient includes:

[0122] The temperature deviation of the visual calibration point is obtained based on the PCB surface temperature distribution information.

[0123] Obtain the reference temperature of the visual calibration point;

[0124] Obtain the coefficient of thermal expansion of the PCB;

[0125] Obtain the system's temperature sensitivity coefficient.

[0126] The attenuation value of the temperature influence at the visual calibration point is obtained based on the temperature deviation at the visual calibration point, the thermal expansion coefficient of the PCB, the reference temperature at the visual calibration point, and the system's temperature sensitivity coefficient. The calculation formula is as follows:

[0127]

[0128] In the formula, ω i Let T represent the attenuation value of temperature effect at the i-th visual calibration point, exp represent the natural exponential function, and T i Let δ(T) represent the temperature of the i-th visual calibration point. i ) represents the PCB thermal expansion coefficient at temperature T for the i-th visual calibration point, δ1 represents the PCB thermal expansion coefficient at the reference temperature, and ΔTi T represents the temperature deviation at the i-th visual calibration point. ref Here, λ represents the reference temperature, and λ represents the system's temperature sensitivity coefficient. (λ is used to control the rate at which the temperature effect attenuation value at the visual calibration point decreases with temperature deviation. In some common PCB manufacturing or inspection industrial scenarios, if temperature changes are relatively mild and the system's temperature sensitivity is moderate, λ can be between (0.1-1). For example, in a conventional PCB assembly workshop, the ambient temperature is relatively stable, and the impact of temperature changes on PCB processing accuracy is not particularly prominent. In this case, λ can be taken as a small value, allowing the temperature factor to play a moderate adjustment role in the calculation of the temperature effect attenuation value at the visual calibration point. For high-precision PCB manufacturing processes, such as advanced semiconductor packaging, which are highly sensitive to temperature changes, a more significant adjustment is needed.) To effectively reflect the influence of temperature on the weighting coefficient, λ can be taken between (1-10). For example, in some packaging processes with extremely high requirements for chip position accuracy, even a small change in temperature can cause chip position displacement. Therefore, a larger λ value is needed to more strictly adjust the weights according to temperature deviation. When the PCB is in a high-temperature environment with large temperature fluctuations, the value of λ also needs to be increased. For example, in some high-temperature testing or specific high-temperature industrial production environments, if the temperature change range is large and the change rate is fast, in order to enable the system to adjust the visual calibration point temperature influence attenuation value in a timely and accurate manner according to temperature changes, increasing λ can enhance the role of temperature factor in the calculation of the visual calibration point temperature influence attenuation value.

[0129] Considering the significant impact of temperature on PCB visual calibration points, temperature changes cause thermal expansion or contraction of the PCB material, leading to a shift in the position of the visual calibration points. A formula is used to quantify the effect of temperature, where ΔT... i =T(x i ,y i )-T ref It is the local temperature of the visual calibration point obtained through an infrared thermal imager, T ref The reference temperature is typically set to the standard ambient temperature for PCB manufacturing, such as 25°C. λ controls the rate at which the weight decreases with temperature deviation. This formula applies a non-linear penalty to the temperature deviation. i As the value increases, the weights of the relevant visual calibration points decrease exponentially, while simultaneously... Normalization is performed to avoid parameter incompatibility due to different reference temperatures. Considering the varying temperature sensitivities of different PCB materials, and given that even small temperature changes in high-precision PCB manufacturing can lead to significant deviations in processing accuracy for materials highly sensitive to temperature variations, the PCB thermal expansion coefficient significantly modulates the impact of temperature on weights. This allows for more accurate consideration of temperature factors in subsequent processing. In practical applications, when the temperature difference between visual calibration points is small, the weight approaches the theoretical maximum, dominating the calculation of relevant parameters and ensuring high-precision positioning. Conversely, when the temperature difference is large, the weight is significantly reduced, effectively preventing unreliable data from interfering with calibration. The system can adapt to changes in temperature distribution during processing, such as localized temperature rises caused by drilling heat. Under complex conditions, it automatically focuses on reliable PCB visual calibration points, greatly improving the system's robustness.

[0130] The step of obtaining the vibration attenuation value of the visual calibration point based on the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point, and the overall vibration acceleration value includes:

[0131] The actual coordinates of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points.

[0132] Obtain the coordinates of the PCB fixed point;

[0133] The distance influence coefficient is obtained based on the difference between the coordinates of the PCB fixed point and the coordinates of the PCB fixed point. The calculation formula is as follows:

[0134]

[0135] In the formula, d i P1 represents the distance influence coefficient of the i-th visual calibration point. i P3 represents the actual coordinates of the i-th visual calibration point. i2Let represent the coordinates of the 2nd PCB fixed point, n2 represent the number of PCB fixed point coordinates, and v represent the distance influence conversion coefficient. During PCB processing, fixtures are used to clamp and fix the PCB at specific locations. The coordinates of the PCB fixed points are obtained through the clamped parts of the PCB. During PCB processing, drilling or placement equipment comes into contact with the PCB. The closer the location is to the PCB fixed point coordinates, the smaller the vibration amplitude due to clamping restrictions. Conversely, the farther away the location is from the PCB fixed point coordinates, the smaller the vibration amplitude. Therefore, by calculating the distance between the coordinates of the visual calibration point and the coordinates of the PCB fixed points, the impact of vibration on the overall system can be assessed more accurately. The distance influence conversion coefficient is affected by the stiffness of the PCB material. When the stiffness of the PCB material is greater, the vibration amplitude may be smaller, and vice versa. Adjusting the size of the distance influence conversion coefficient according to the characteristics of the PCB material can further optimize the calculation of the vibration influence attenuation value of the visual calibration point.

[0136] The root mean square value of the vibration acceleration at the visual calibration point is obtained based on the overall vibration acceleration value, and the calculation formula is as follows:

[0137]

[0138] In the formula, a i Let represent the root mean square value of the vibration acceleration of the i-th visual calibration point, a(t) represent the vibration acceleration of the visual calibration point at time t, and n1 represent the number of samples at time t.

[0139] Obtain the vibration sensitivity coefficient;

[0140] The vibration influence attenuation value at the visual calibration point is obtained based on the distance influence coefficient, the overall vibration acceleration value, and the vibration sensitivity coefficient. The calculation formula is as follows:

[0141]

[0142] In the formula, g i a represents the attenuation value of vibration effect at the i-th visual calibration point. i d represents the root mean square value of the vibration acceleration at the i-th visual calibration point. i γ represents the distance influence coefficient of the i-th visual calibration point, and γ represents the vibration sensitivity coefficient.

[0143] Vibration is a significant disturbance factor in PCB manufacturing. Equipment operation and mechanical structure vibrations can all cause PCB vibration, which can affect the accuracy of visual calibration points on the PCB. For example, even minor vibrations can cause component placement deviations during high-precision surface mount technology (SMT) operations. To more effectively consider the impact of vibration during manufacturing correction, this invention introduces a vibration-related term to adjust the weights of each visual calibration point, thereby achieving more accurate manufacturing correction compensation. Therefore, a vibration impact attenuation value for visual calibration points is introduced. This attenuation value dynamically adjusts the weights of the visual calibration points based on the overall vibration acceleration of the PCB. When the root mean square value of the vibration acceleration of the visual calibration point is large, indicating severe PCB vibration, the vibration impact attenuation value decreases. This means that the corresponding visual calibration point will be more relevant in subsequent calculations. By reducing the weights, the contribution of visual calibration points significantly affected by vibration to the final result in the machining correction calculation decreases, thus reducing the impact of vibration on machining accuracy. Conversely, when vibration is small, the vibration attenuation value of the visual calibration point has a weaker effect on the weight adjustment, and the weight of the visual calibration point remains relatively stable. This ensures that machining correction calculations can be performed based on reliable visual information under normal circumstances. γ is used to adjust the degree of influence of vibration factors on the weight coefficients. Its value can be adjusted according to the actual PCB machining scenario and the system's sensitivity to vibration. For example, in high-precision PCB machining processes that are extremely sensitive to vibration, the value of γ can be appropriately increased to make the vibration attenuation term more sensitive to vibration changes and more significantly reduce the weight of visual calibration points when vibration is severe. In ordinary PCB machining scenarios that are relatively less sensitive to vibration, the value of γ can be decreased to avoid over-adjusting the weights. i It can comprehensively reflect the intensity of PCB vibration over a period of time, providing an accurate vibration intensity measurement index for vibration attenuation, so that the weight adjustment can be closely based on the actual vibration conditions. By setting the vibration influence attenuation value at the visual calibration point, the system can more effectively deal with vibration interference during the processing correction process, improving the accuracy and stability of PCB processing.

[0144] The steps for constructing the correction compensation parameter optimization model based on the multi-source correction influence factor and the correction compensation function include:

[0145] Obtain the elastic modulus of the PCB substrate and the stress at the visual calibration point. Calculate the strain at the visual calibration point in the x-axis direction and the strain at the visual calibration point in the y-axis direction based on the elastic modulus of the PCB substrate and the stress at the visual calibration point.

[0146] The actual coordinates of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points.

[0147] Obtain the coefficient of thermal expansion of the PCB;

[0148] Obtain the temperature distribution information on the PCB surface;

[0149] Based on the strain of the visual calibration point in the x-axis direction, the strain of the visual calibration point in the y-axis direction, the actual coordinates of the visual calibration point, the thermal expansion coefficient of the PCB, the elastic modulus of the PCB substrate, and the temperature distribution information of the PCB surface, a correction compensation function is constructed. The projected coordinates of the visual calibration point are obtained through the correction compensation function, and its formula is as follows:

[0150]

[0151] In the formula, P1 i θ represents the actual coordinates of the i-th visual calibration point. j Let s represent the correction compensation parameters for the j-th correction (initial correction compensation parameters). x The scaling factor x of the PCB in the x-axis direction. i The coordinate components of the actual coordinates of the i-th visual calibration point along the x-axis, where Δx represents the x-axis translation distance of the visual calibration point, and δ(T) i ) represents the coefficient of thermal expansion of the PCB at temperature T for the i-th visual calibration point, ΔT i σ represents the temperature deviation at the i-th visual calibration point. x(i) Let E represent the stress at the i-th visual calibration point along the x-axis, and let E represent the elastic modulus of the PCB substrate. s represents the strain of the i-th visual calibration point in the x-axis direction. y y represents the scaling factor of the i-th visual calibration point in the y-axis direction. i The coordinate components of the actual coordinates of the i-th visual calibration point along the y-axis, Δy represents the y-axis translation distance of the visual calibration point, and σ y(i) This represents the stress at the i-th visual calibration point along the y-axis. R(α) represents the strain of the i-th visual calibration point in the y-axis direction, and R(α) represents the rotation angle of the i-th visual calibration point.

[0152] The stress at the visual calibration point in the x-axis direction and the stress at the visual calibration point in the y-axis direction can be obtained by optical methods, thermal imaging methods, or electronic methods. The following shows one of the methods:

[0153] The formula for calculating the stress at the visual calibration point along the x-axis is:

[0154] σ x(i) =E*o x *h;

[0155] In the formula, σ x(i)The stress at the visual calibration point along the x-axis is represented by E, and E represents the elastic modulus of the PCB substrate.

[0156] The formula for calculating the stress at the visual calibration point in the y-axis direction is:

[0157] σ y(i) =E*o y *h;

[0158] In the formula, σ x(i) σ represents the stress at the i-th visual calibration point along the x-axis. y(i) This represents the stress at the i-th visual calibration point along the y-axis, o x o represents the real-time warp component of the i-th visual calibration point in the x-axis direction. y The real-time warpage component of the i-th visual calibration point in the y-axis direction is represented by E, which represents the elastic modulus of the PCB substrate, and h represents the PCB thickness.

[0159] The real-time warp component of the above-mentioned visual calibration point in the x-axis direction and the real-time warp component of the visual calibration point in the y-axis direction are obtained by curvature analysis of the surface method based on the real-time warp of the visual calibration point.

[0160] The formula for calculating the real-time warpage is:

[0161]

[0162] In the formula, o(x,y) represents the real-time warpage, and κ(x,y) represents the PCB surface height function.

[0163] The correction and compensation function is used to accurately map the coordinates of actual measured points on the PCB to the theoretical coordinate system, thereby achieving precise calibration and correction of the PCB's position and dimensions. In actual PCB manufacturing, various factors can cause deviations between the points on the PCB and their theoretical positions. Temperature changes cause thermal expansion or contraction of the PCB material, which may be uniform overall or non-uniform due to local temperature differences. Mechanical stress can cause physical deformation of the PCB, such as warping or stretching. Simultaneously, errors in imaging systems and positioning devices may also occur, leading to discrepancies between the measured point coordinates and the theoretical values. Therefore, this invention establishes a correction and compensation function that comprehensively considers these factors to correct the coordinates of the actual measured points, accurately reflecting the true position and dimensions of the PCB and providing accurate data for subsequent processing correction. Based on this, the correction and compensation function can significantly improve the accuracy of PCB manufacturing correction. In practical applications, it can effectively compensate for PCB position and size deviations caused by factors such as temperature changes and physical deformation, making the coordinates after projection transformation closer to the theoretical coordinates. This reduces errors in the manufacturing process, improves product quality and production efficiency, and ensures that chips are accurately mounted in the designated positions on the PCB in high-precision chip packaging processes, reducing scrap rates. The correction and compensation function includes a scaling factor to compensate for overall uniform thermal expansion or contraction and size scaling caused by factors such as the imaging system; a thermal expansion factor combined with local temperature deviations to compensate for non-uniform thermal expansion caused by local temperature changes; Δx and Δy to adjust the offset of the visual calibration point on the x-axis and y-axis; and a physical deformation compensation term. and By calculating local stress and elastic modulus, the physical deformation caused by mechanical stress is compensated. By comprehensively considering multiple factors, the correction compensation function can fully handle various deviations that may occur in the PCB during actual processing. The correction compensation function can dynamically adjust parameters based on real-time monitored data (such as temperature, vibration, physical deformation, etc.) to adapt to different processing environments and PCB conditions.

[0164] The steps for constructing the correction compensation parameter optimization model based on the multi-source correction influence factor and the correction compensation function include:

[0165] Obtain the theoretical coordinates of the visual calibration point;

[0166] Obtain the correction compensation function;

[0167] Obtain the multi-source correction influence factor;

[0168] Retrieve the previous correction compensation parameters;

[0169] Obtain the impact value of the adjustment range of the correction compensation parameters;

[0170] Based on the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influence factor, the previous correction compensation parameter, and the influence value of the adjustment range of the correction compensation parameter, an optimization model for the correction compensation parameter is constructed, and its formula is as follows:

[0171]

[0172] In the formula, θ j Let w represent the parameter for the j-th correction compensation. i P1 represents the multi-source correction influence factor for the i-th visual calibration point. i f represents the actual coordinates of the i-th visual calibration point. proj (P1 i ,θ j P2 represents the correction compensation function for the actual coordinates of the i-th visual calibration point with respect to the j-th correction compensation parameter, used to map the actual coordinates of the visual calibration point to the theoretical coordinate system. i Let θ represent the theoretical coordinates of the i-th visual calibration point. j-1 represents the correction compensation parameter for the (j-1)th time, and β represents the influence value of the adjustment range of the correction compensation parameter;

[0173] Calculate and optimize the correction compensation parameters (θ is solved iteratively using the Levenberg-Marquardt algorithm). j The correction compensation parameters include the x-axis translation distance, y-axis translation distance, and rotation angle of the visual calibration point, as well as the temperature difference scaling factor of the PCB in the x-axis direction and the temperature difference scaling factor of the PCB in the y-axis direction.

[0174] In the complex scenario of PCB manufacturing correction, multiple factors need to be considered to determine the optimal correction compensation parameters in order to achieve accurate calibration of the target position during PCB manufacturing. The actual coordinates of the visual calibration points obtained from actual measurements deviate from their theoretical coordinates. These deviations are caused by the combined effects of various factors such as temperature changes, vibration, and physical deformation. This invention uses a correction compensation parameter optimization model to measure the magnitude of these deviations and finds the correction compensation parameters that minimize these deviations by optimizing this model. Simultaneously, to avoid excessive fluctuations in the correction compensation parameters and ensure the stability and reliability of the system, the correction compensation parameter optimization model is continuously optimized. This method can find an optimal set of correction and compensation parameters to make the actual coordinates of the PCB visual calibration points after projection transformation as close as possible to the theoretical coordinates, thereby improving the accuracy of PCB processing. For example, in high-precision PCB placement processes, the error in the placement position can be controlled within a very small range, greatly reducing the scrap rate. The correction and compensation parameter optimization model considers various influencing factors, including visual confidence, temperature deviation, and vibration. It also calculates and processes different PCB visual calibration points through multi-source correction influencing factors, enabling the system to maintain good performance even when facing complex and changing processing environments and various interference factors, thus improving the robustness of the system. j -θ j-1 || 2 The setting of the parameters constrains the range of variation of the correction compensation parameters, avoiding excessive fluctuations in the parameters and enabling the system to maintain stable performance under different measurement data and environmental conditions. This helps to reduce the correction error caused by parameter mutations and improves the reliability and consistency of the system. Optimization algorithms (such as Levenberg-Marquardt, gradient descent, and Newton's method) are used to iteratively optimize the correction compensation parameter optimization model. These algorithms can continuously adjust the calibration parameter vector based on the gradient information of the correction compensation parameter optimization model, so that the value of the correction compensation parameter optimization model gradually decreases and finally finds the optimal solution.

[0175] The steps for obtaining the influence value of the adjustment range of the correction compensation parameter include:

[0176] Obtain the vibration influence coefficient;

[0177] Obtain the impact value of the initial correction compensation parameter adjustment range;

[0178] The influence value of the adjustment range of the correction compensation parameters is calculated based on the influence value of the initial correction compensation parameter adjustment range and the vibration influence coefficient. The calculation formula is as follows:

[0179] β=β0*(1+γ*d i *a i );

[0180] In the formula, β represents the influence value of the adjustment range of the correction compensation parameter, β0 represents the influence value of the initial adjustment range of the correction compensation parameter, (1+γ*d i *a i The vibration influence coefficient (β0) represents the vibration impact factor. By introducing vibration data collected by an accelerometer, the system's adaptability to vibration environments can be improved. The vibration influence coefficient comprehensively reflects the energy magnitude of the vibration signal over a period of time and effectively characterizes the severity of environmental vibration. β0 is a basic value set without considering the vibration influence and is used in the optimization model for balancing the correction and compensation parameters. Term and β*||θ j -θ j-1 || 2 The relative importance of the terms, when the vibration intensifies, i.e. (1+γ*d) i *a i The value of ) will increase, thus increasing β. After β increases, in the error correction compensation parameter optimization model, β*||θ j -θ j-1 || 2 Increasing the weight of an item forces the difference between the current correction compensation parameter and the previous correction compensation parameter to be smaller, making the parameter change smoother. The purpose of this is to suppress the interference of high-frequency noise generated by vibration on the parameters, avoid drastic fluctuations in the correction compensation parameters due to vibration, and ensure the stability and calibration accuracy of the system in a vibration environment.

[0181] The step of performing time-series smoothing processing on the correction compensation parameters includes:

[0182] Obtain the gain coefficient;

[0183] Obtain the correction compensation parameters to be optimized;

[0184] Get the parameters from the previous correction compensation;

[0185] Based on the gain coefficient and the previous correction compensation parameter, the correction compensation parameter to be optimized is subjected to time-series smoothing processing, and the calculation formula is as follows:

[0186] θ j+1 =K*θ j +(1-K)*θ j-1 ;

[0187] In the formula, θ j+1 θ represents the (j+1)th correction compensation parameter after time-series smoothing. j Let θ represent the correction compensation parameter for the j-th error, K represent the gain coefficient (e.g., using Kalman gain), and θ represent the error correction compensation parameter for the j-th error. j-1 This represents the correction compensation parameter for the (j-1)th correction (the previous correction compensation parameter).

[0188] In PCB manufacturing, the original correction compensation parameters (the j-th correction compensation parameter) are often affected by various interference factors (such as vibration, temperature changes, sensor noise, etc.), leading to inaccurate data. This filtering formula can effectively smooth the original correction compensation parameters, reduce the impact of noise and interference, and improve the accuracy and stability of the correction compensation parameters, thereby improving the precision and reliability of PCB manufacturing correction operations. K uses Kalman gain, which is between 0 and 1, to determine the weight of the original correction compensation parameters and the previous correction compensation parameters in the final output parameters. It is dynamically adjusted according to factors such as uncertainties in PCB manufacturing. When the measurement noise is low, the Kalman gain will be closer to 1, meaning that the original correction compensation parameters are trusted more. When the measurement noise is high and the uncertainty is high, the Kalman gain will be closer to 0, relying more on the previous correction compensation parameters. This invention comprehensively considers the original correction compensation parameters and the previous correction compensation parameters, making it more reliable and stable than the original correction compensation parameters alone.

[0189] like Figure 2 As shown, this invention also discloses a PCB manufacturing deviation correction system based on visual calibration, comprising:

[0190] Select module 1, which is used to select PCB visual calibration points based on feature saliency;

[0191] The acquisition module 2 is used to acquire multi-source feature information of PCB visual calibration points in real time, and obtain the visual confidence level, temperature influence attenuation value and vibration influence attenuation value of visual calibration points based on the multi-source feature information of PCB visual calibration points.

[0192] Fusion module 3 is used to fuse the visual confidence of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point to obtain a multi-source correction influence factor.

[0193] The first calculation module 4 is used to construct the correction compensation function based on the physical deformation of the PCB.

[0194] The second calculation module 5 is used to construct an optimization model for correction compensation parameters based on the multi-source correction influence factor and the correction compensation function.

[0195] The third calculation module 6 is used to solve the optimization model of the correction compensation parameters through optimization algorithms to obtain the optimal correction compensation parameters.

[0196] The fourth calculation module 7 is used to perform time-series smoothing processing on the correction compensation parameters;

[0197] The first control module 8 is used to generate x-axis translation distance, y-axis translation distance and rotation angle compensation commands based on the smoothed correction compensation parameters;

[0198] The second control module 9 is used to transmit the x-axis translation distance, y-axis translation distance and rotation angle compensation command to the PCB board processing equipment, and the PCB board processing equipment adjusts the target processing position based on the x-axis translation distance, y-axis translation distance and rotation angle compensation command.

[0199] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0200] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A PCB board processing correction method based on visual calibration, characterized in that, include: PCB visual calibration points are selected based on feature saliency. Real-time acquisition of multi-source feature information of PCB visual calibration points; and acquisition of visual confidence, temperature influence attenuation value, and vibration influence attenuation value of visual calibration points based on the multi-source feature information of PCB visual calibration points. The visual confidence level of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point are fused to obtain a multi-source correction influence factor. Obtain the elastic modulus of the PCB substrate and the stress at the visual calibration point. Calculate the strain at the visual calibration point in the x-axis direction and the strain at the visual calibration point in the y-axis direction based on the elastic modulus of the PCB substrate and the stress at the visual calibration point. The actual coordinates of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points. Obtain the coefficient of thermal expansion of the PCB; Obtain the temperature distribution information on the PCB surface; A correction and compensation function is constructed based on the strain of the visual calibration point in the x-axis direction, the strain of the visual calibration point in the y-axis direction, the actual coordinates of the visual calibration point, the thermal expansion coefficient of the PCB, the elastic modulus of the PCB substrate, and the temperature distribution information of the PCB surface. Obtain the theoretical coordinates of the visual calibration point; Obtain the correction compensation function; Obtain the multi-source correction influence factor; Retrieve the previous correction compensation parameters; Obtain the impact value of the adjustment range of the correction compensation parameters; An optimization model for the correction compensation parameters is constructed based on the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influence factor, the previous correction compensation parameters, and the influence value of the adjustment range of the correction compensation parameters. The optimal correction compensation parameters are obtained by solving the optimization model of the correction compensation parameters through optimization algorithm. The correction compensation parameters are subjected to time-series smoothing. Based on the smoothed correction and compensation parameters, x-axis translation distance, y-axis translation distance, and rotation angle compensation commands are generated. The x-axis translation distance, y-axis translation distance, and rotation angle compensation commands are transmitted to the PCB board processing equipment, which adjusts the target processing position based on the x-axis translation distance, y-axis translation distance, and rotation angle compensation commands.

2. The PCB board processing correction method based on visual calibration according to claim 1, characterized in that, The steps of obtaining the visual confidence level, temperature-induced attenuation value, and vibration-induced attenuation value of the visual calibration point based on the multi-source feature information of the PCB visual calibration point include: Obtain the coordinate information of the PCB visual calibration points and the PCB image information, and obtain the visual confidence of the visual calibration points based on the coordinate information of the PCB visual calibration points and the PCB image information; Obtain PCB surface temperature distribution information and PCB thermal expansion coefficient, and obtain the temperature influence attenuation value of visual calibration point based on the PCB surface temperature distribution information and PCB thermal expansion coefficient; Obtain the coordinates of the PCB fixed point and the overall vibration acceleration value. Based on the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point and the overall vibration acceleration value, obtain the vibration influence attenuation value of the visual calibration point. The visual confidence level of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point are fused to obtain the multi-source correction influence factor.

3. The PCB board processing correction method based on visual calibration according to claim 2, characterized in that, The step of obtaining the visual confidence level of the visual calibration point based on the coordinate information of the PCB visual calibration point and the PCB image information includes: The average grayscale value of the visual calibration points is obtained based on the PCB image information. The standard deviation of the background area corresponding to the visual calibration point is obtained based on the PCB image information. The signal-to-noise ratio of the visual calibration point is calculated based on the mean gray value of the visual calibration point and the standard deviation of the background region corresponding to the visual calibration point. Based on the PCB image information, obtain the standard deviation of the Gaussian fitting at the visual calibration point edge and the standard deviation of the Gaussian fitting at the maximum edge. The spatial relationship coefficients of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points. The visual confidence level of the visual calibration point is obtained based on the signal-to-noise ratio of the visual calibration point, the spatial relationship coefficient of the visual calibration point, the standard deviation of the Gaussian fitting of the edge of the visual calibration point, and the standard deviation of the Gaussian fitting of the maximum edge.

4. The PCB board processing correction method based on visual calibration according to claim 3, characterized in that, The step of obtaining the spatial relationship coefficients of the visual calibration points based on the coordinate information of the PCB visual calibration points includes: The angle between the visual calibration point and adjacent visual calibration points is obtained based on the coordinate information of the PCB visual calibration point; The distance between the visual calibration point and the adjacent visual calibration point is obtained based on the coordinate information of the PCB visual calibration point; Obtain the theoretical angle between the visual calibration point and its adjacent visual calibration points; Obtain the theoretical distance between the visual calibration point and its adjacent visual calibration points; Calculate the angular similarity between the visual calibration point and the adjacent visual calibration points and the theoretical angle; Calculate the distance similarity between the visual calibration point and its adjacent visual calibration points and the theoretical distance; The spatial relationship coefficients of the visual calibration points are calculated based on the angular similarity and distance similarity.

5. The PCB board processing correction method based on visual calibration according to claim 2, characterized in that, The step of obtaining the attenuation value of the visual calibration point temperature effect based on the PCB surface temperature distribution information and the PCB thermal expansion coefficient includes: The temperature deviation of the visual calibration point is obtained based on the PCB surface temperature distribution information. Obtain the reference temperature of the visual calibration point; Obtain the coefficient of thermal expansion of the PCB; Obtain the system's temperature sensitivity coefficient; The attenuation value of the temperature effect at the visual calibration point is obtained based on the temperature deviation at the visual calibration point, the thermal expansion coefficient of the PCB, the reference temperature at the visual calibration point, and the system's temperature sensitivity coefficient.

6. The PCB board processing correction method based on visual calibration according to claim 2, characterized in that, The step of obtaining the vibration attenuation value of the visual calibration point based on the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point, and the overall vibration acceleration value includes: The actual coordinates of the visual calibration points are obtained based on the coordinate information of the PCB visual calibration points. Obtain the coordinates of the PCB fixed point; The distance influence coefficient is obtained based on the difference between the coordinates of the PCB fixed point and the coordinates of the PCB fixed point. The root mean square value of the vibration acceleration at the visual calibration point is obtained based on the overall vibration acceleration value. Obtain the vibration sensitivity coefficient; The vibration influence attenuation value of the visual calibration point is obtained based on the distance influence coefficient, the overall vibration acceleration value, and the vibration sensitivity coefficient.

7. The PCB board processing correction method based on visual calibration according to claim 1, characterized in that, The steps for obtaining the influence value of the adjustment range of the correction compensation parameter include: Obtain the vibration influence coefficient; Obtain the impact value of the initial correction compensation parameter adjustment range; The influence value of the adjustment range of the correction compensation parameters is calculated based on the influence value of the initial correction compensation parameter adjustment range and the vibration influence coefficient.

8. A PCB manufacturing deviation correction system based on vision calibration, characterized in that, include: The selection module is used to select PCB visual calibration points based on feature saliency. The acquisition module is used to acquire multi-source feature information of PCB visual calibration points in real time, and obtain the visual confidence level, temperature influence attenuation value and vibration influence attenuation value of visual calibration points based on the multi-source feature information of PCB visual calibration points. The fusion module is used to fuse the visual confidence of the visual calibration point, the temperature influence attenuation value of the visual calibration point, and the vibration influence attenuation value of the visual calibration point to obtain a multi-source correction influence factor. The first calculation module is used to obtain the elastic modulus of the PCB substrate and the stress at the visual calibration point, calculate the strain of the visual calibration point in the x-axis direction and the strain of the visual calibration point in the y-axis direction based on the elastic modulus of the PCB substrate and the stress at the visual calibration point, and obtain the actual coordinates of the visual calibration point based on the coordinate information of the PCB visual calibration point. Obtain the thermal expansion coefficient of the PCB; obtain the surface temperature distribution information of the PCB; construct a correction compensation function based on the strain of the visual calibration point in the x-axis direction, the strain of the visual calibration point in the y-axis direction, the actual coordinates of the visual calibration point, the thermal expansion coefficient of the PCB, the elastic modulus of the PCB substrate, and the surface temperature distribution information of the PCB. The second calculation module is used to obtain the theoretical coordinates of the visual calibration point; obtain the correction compensation function; obtain the multi-source correction influence factor; obtain the previous correction compensation parameter; obtain the influence value of the adjustment range of the correction compensation parameter; and construct a correction compensation parameter optimization model based on the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influence factor, the previous correction compensation parameter, and the influence value of the adjustment range of the correction compensation parameter. The third calculation module is used to solve the optimization model of the correction compensation parameters through optimization algorithms to obtain the optimal correction compensation parameters. The fourth calculation module is used to perform time-series smoothing processing on the correction compensation parameters; The first control module is used to generate x-axis translation distance, y-axis translation distance and rotation angle compensation commands based on the smoothed correction compensation parameters; The second control module is used to transmit the x-axis translation distance, y-axis translation distance, and rotation angle compensation commands to the PCB board processing equipment, and the PCB board processing equipment adjusts the target processing position based on the x-axis translation distance, y-axis translation distance, and rotation angle compensation commands.

Images