PCB processing deviation correction method and system based on visual calibration

Abstract

The invention relates to the technical field of PCB production, in particular to a PCB processing deviation correction 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 points in real time; obtaining a multi-source deviation correction influence factor; constructing a deviation correction compensation function based on the physical deformation of the PCB; according to the method, the correction compensation function comprehensively considering various factors is introduced, so that the PCB position and size deviation caused by factors such as temperature change and physical deformation can be compensated; and the corresponding influence coefficient is calculated for each visual calibration point through the multi-source deviation correction influence factors, so that the more reliable visual calibration points have greater influence on the result, the PCB processing precision is improved, and the product quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of PCB production, and in particular to a PCB board processing correction method and system based on visual calibration. Background Art

[0002] Accurate position and size calibration is crucial in the manufacturing, testing, and assembly of PCBs. However, existing technologies have many shortcomings when dealing with complex and changing processing environments.

[0003] On the one hand, temperature changes cause PCB materials to expand and contract due to thermal fluctuations. Due to uneven local temperature distribution, this thermal expansion or contraction often exhibits non-uniformity, which in turn causes the positions of various points on the PCB to deviate from their theoretical positions. Currently, most existing methods are unable to accurately compensate for these dimensional and positional changes caused by temperature changes, which significantly affects processing accuracy. On the other hand, mechanical stress is also a key factor affecting PCB accuracy. During the production, transportation, and use of PCBs, they are inevitably subjected to various mechanical stresses such as bending, stretching, and extrusion. These stresses can cause physical deformations of the PCB, such as warping and twisting. When dealing with physical deformations caused by mechanical stress, existing technologies are unable to comprehensively and accurately calculate and compensate for deformation caused by stress. As a result, the actual size and position of the PCB are significantly different from the designed values, seriously affecting product quality and production efficiency. Summary of the Invention

[0004] The main purpose of the present invention is to provide a PCB board processing correction method and system based on visual calibration, aiming to solve the technical problems raised in the above background technology.

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

[0006] Select PCB visual calibration points based on feature significance;

[0007] Collect multi-source feature information of PCB visual calibration points in real time, and obtain visual confidence of the visual calibration points, temperature influence attenuation value of the visual calibration points, and vibration influence attenuation value of the visual calibration points based on the multi-source feature information of the PCB visual calibration points;

[0008] fusing 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;

[0009] Construct a correction compensation function based on the physical deformation of the PCB;

[0010] Constructing a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function;

[0011] Solve the correction compensation parameter optimization model through the optimization algorithm to obtain the optimal correction compensation parameters;

[0012] Performing time series smoothing processing on the correction compensation parameters;

[0013] Generate x-axis translation distance, y-axis translation distance and rotation angle compensation instructions based on the smoothed correction compensation parameters;

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

[0015] The present invention is further configured such that the step of obtaining 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 according to the multi-source feature information of the PCB visual calibration point comprises:

[0016] Obtaining coordinate information of the PCB visual calibration point and PCB image information, and obtaining visual confidence of the visual calibration point based on the coordinate information of the PCB visual calibration point and the PCB image information;

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

[0018] Obtaining the coordinates of the PCB fixed point and the overall vibration acceleration value, and obtaining the vibration impact attenuation value of the visual calibration point according to the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point and the overall vibration acceleration value;

[0019] 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 are fused to obtain a 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 according to the coordinate information of the PCB visual calibration point and the PCB image information comprises:

[0021] Obtaining a grayscale mean value of a visual calibration point according to the PCB image information;

[0022] Obtaining a standard deviation of a background area corresponding to a visual calibration point according to the PCB image information;

[0023] Calculating a signal-to-noise ratio of the visual calibration point according to the grayscale mean of the visual calibration point and the standard deviation of the background area corresponding to the visual calibration point;

[0024] Obtaining a Gaussian fitting standard deviation of a visual calibration point edge and a maximum Gaussian fitting standard deviation of an edge according to the PCB image information;

[0025] Obtaining a spatial relationship coefficient of the visual calibration point according to the coordinate information of the PCB visual calibration point;

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

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

[0028] Obtaining the angle between the visual calibration point and the adjacent visual calibration point according to the coordinate information of the PCB visual calibration point;

[0029] Obtaining the distance between the visual calibration point and the adjacent visual calibration point according to the coordinate information of the PCB visual calibration point;

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

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

[0032] Calculating the angular similarity between the angle of the visual calibration point and the adjacent visual calibration point and the theoretical angle;

[0033] Calculating the distance similarity between the distance between the visual calibration point and the adjacent visual calibration point and the theoretical distance;

[0034] The spatial relationship coefficient of the visual calibration points is calculated according to the angle similarity and the distance similarity.

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

[0036] Acquire the temperature deviation of the visual calibration point according to the PCB surface temperature distribution information;

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

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

[0039] Get the system's sensitivity to temperature;

[0040] The temperature influence attenuation value of the visual calibration point is obtained according to the temperature deviation of the visual calibration point, the thermal expansion coefficient of the PCB, the reference temperature of the visual calibration point and the sensitivity coefficient of the system to temperature.

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

[0042] Acquire the actual coordinates of the visual calibration points according to the coordinate information of the PCB visual calibration points;

[0043] Obtaining the coordinates of the PCB fixed points;

[0044] Obtaining a distance influence coefficient according to a difference between the coordinates of the PCB fixed point and the coordinates of the PCB fixed point;

[0045] Obtaining a root mean square value of the vibration acceleration of the visual calibration point according to the overall vibration acceleration value;

[0046] Get vibration sensitivity coefficient;

[0047] The vibration impact attenuation value of the visual calibration point is obtained according to 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 a correction compensation function based on the physical deformation of the PCB includes:

[0049] Obtaining the elastic modulus of the PCB substrate and the stress of the visual calibration point, and calculating 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 according to the elastic modulus of the PCB substrate and the stress of the visual calibration point;

[0050] Acquire the actual coordinates of the visual calibration points according to the coordinate information of the PCB visual calibration points;

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

[0052] Obtaining the PCB surface temperature distribution information;

[0053] A correction compensation function is constructed according to 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 PCB thermal expansion coefficient, the PCB substrate elastic modulus and the PCB surface temperature distribution information.

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

[0055] Get the theoretical coordinates of the visual calibration points;

[0056] Obtaining the correction compensation function;

[0057] Obtaining the multi-source correction influencing factor;

[0058] Get the last correction compensation parameters;

[0059] Obtain the impact value of the adjustment amplitude of the correction compensation parameter;

[0060] A correction compensation parameter optimization model is constructed according to the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influencing factor, the last correction compensation parameter and the correction compensation parameter adjustment amplitude influence value.

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

[0062] Get the vibration influence coefficient;

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

[0064] The deviation correction compensation parameter adjustment amplitude influence value is calculated according to the initial deviation correction compensation parameter adjustment amplitude influence value and the vibration influence coefficient.

[0065] The present invention also discloses a PCB processing correction system based on visual calibration, comprising:

[0066] Selection module, used to select PCB visual calibration points based on feature significance;

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

[0068] a fusion module, configured 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 a correction compensation function based on the physical deformation of the PCB;

[0070] A second calculation module is used to construct a correction compensation parameter optimization model based on the multi-source correction influencing factor and the correction compensation function;

[0071] The third calculation module is used to solve the correction compensation parameter optimization model through an optimization algorithm to obtain the optimal correction compensation parameters;

[0072] a fourth calculation module, configured to perform time series smoothing processing on the correction compensation parameters;

[0073] A first control module is configured to generate an x-axis translation distance, a y-axis translation distance, and a rotation angle compensation instruction based on the smoothed correction compensation parameter;

[0074] The second control module is used to transmit the x-axis translation distance, y-axis translation distance and rotation angle compensation instruction 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 instruction.

[0075] The beneficial effects of the present invention are as follows: the present 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, and calculates the corresponding influence coefficient for each visual calibration point through multi-source correction influencing factors, so that when calculating the correction compensation parameters, these more reliable visual calibration points have a greater impact on the results, thereby effectively improving the accuracy of the overall calculation, improving the precision of PCB processing, and effectively improving product quality. BRIEF DESCRIPTION OF THE DRAWINGS

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

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

[0078] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0079] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

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

[0081] S1, selects PCB visual calibration points based on feature significance. These include mark points, hole positions, edge contours, points with high contrast with surrounding areas, and points with unique shapes. These points are also determined based on specific processing techniques. For example, in the SMT process, to ensure that the chip can be accurately mounted in the specified position, the edge or center of the pad corresponding to the chip pin is selected as a visual calibration point to accurately adjust the position and angle of the chip during the SMT process.

[0082] S2, collecting multi-source feature information of the PCB visual calibration points in real time, and obtaining the visual confidence of the visual calibration points, the temperature influence attenuation value of the visual calibration points, and the vibration influence attenuation value of the visual calibration points according to the multi-source feature information of the PCB visual calibration points;

[0083] S3, fusing 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 calculation formula of which is:

[0084]

[0085] Where w i represents the multi-source correction influence factor of the i-th visual calibration point, C i represents the visual confidence of the i-th visual calibration point, ∑C i1 Represents the sum of the visual confidence of all visual calibration points, ω i Indicates the temperature attenuation value of the i-th visual calibration point, g i represents the vibration impact attenuation value of the i-th visual calibration point;

[0086] S4, constructs a correction compensation function based on the physical deformation of the PCB;

[0087] S5, constructing a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function;

[0088] S6, solving the correction compensation parameter optimization model through the optimization algorithm to obtain the optimal correction compensation parameter; in the iterative process, continuously updating the correction compensation parameter so that the value of the correction compensation parameter gradually decreases until the convergence condition is met (for example, the correction compensation parameter is optimal when the change in the correction compensation parameter is less than a preset threshold);

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

[0090] S8, generating an x-axis translation distance, a y-axis translation distance, and a rotation angle compensation instruction based on the smoothed correction compensation parameter;

[0091] S9, transmitting the x-axis translation distance, y-axis translation distance and rotation angle compensation instruction 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 instruction.

[0092] By introducing a correction compensation function that comprehensively considers multiple factors, the present invention can accurately compensate for PCB position and size deviations caused by factors such as temperature changes and physical deformation. In the correction compensation function, not only the scaling factor, thermal expansion coefficient and PCB translation factor (described below) are considered, but also physical deformation compensation terms are performed through the calculation of the stress-strain relationship (described below), so that the coordinates after the projection transformation are closer to the theoretical coordinates. At the same time, the present invention also uses multi-source correction influencing factors to calculate the corresponding influence coefficient for each visual calibration point according to the confidence of the visual calibration point, the temperature deviation of the area where it is located, and the overall vibration situation. Different visual calibration points are given different importance, so that when calculating the projection transformation and optimizing the correction compensation parameters, more reliable visual calibration points have a greater impact on the results, thereby effectively improving the accuracy of the overall calculation. This method can improve the environmental adaptability and robustness of the PCB processing process, while maintaining computational efficiency, significantly improving processing accuracy.

[0093] The step of obtaining 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 according to the multi-source feature information of the PCB visual calibration point comprises:

[0094] Obtaining coordinate information of the PCB visual calibration point and PCB image information, and obtaining visual confidence of the visual calibration point based on the coordinate information of the PCB visual calibration point and the PCB image information;

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

[0096] Obtaining the coordinates of the PCB fixed point and the overall vibration acceleration value, and obtaining the vibration impact attenuation value of the visual calibration point according to the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point and the overall vibration acceleration value;

[0097] 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 are fused to obtain a multi-source correction influence factor.

[0098] The purpose of integrating the visual confidence level of the visual calibration points, the temperature influence attenuation value of the visual calibration points, and the vibration influence attenuation value of the visual calibration points is to comprehensively consider the impact of multiple factors such as vision, temperature, and vibration on the PCB board processing accuracy. Ultimately, the multi-source correction influence factor is obtained. This source correction influence factor can be used to guide the correction operation during the PCB board processing process and determine the importance of the visual calibration points. By comprehensively considering these multi-source factors, more precise processing control can be achieved, thereby improving the processing quality and accuracy of the PCB boards.

[0099] The step of obtaining the visual confidence of the visual calibration point according to the coordinate information of the PCB visual calibration point and the PCB image information includes:

[0100] Obtaining a grayscale mean value of a visual calibration point according to the PCB image information;

[0101] Obtaining a standard deviation of a background area corresponding to a visual calibration point according to the PCB image information;

[0102] The signal-to-noise ratio of the visual calibration point is calculated according to the grayscale mean of the visual calibration point and the standard deviation of the background area corresponding to the visual calibration point. The calculation formula is:

[0103]

[0104] Where, SNR i The signal-to-noise ratio of the i-th visual calibration point, u i represents the grayscale mean of the i-th visual calibration point, q i Represents the standard deviation of the background area corresponding to the i-th visual calibration point;

[0105] Obtaining a Gaussian fitting standard deviation of a visual calibration point edge and a maximum Gaussian fitting standard deviation of an edge according to the PCB image information;

[0106] Obtaining a spatial relationship coefficient of the visual calibration point according to the coordinate information of the PCB visual calibration point;

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

[0108]

[0109] Where C i Represents the visual confidence of the i-th visual calibration point, SNR i represents the signal-to-noise ratio of the i-th visual calibration point, Indicates the standard deviation of the Gaussian fitting of the edge of the i-th visual calibration point, Indicates the maximum marginal Gaussian fitting standard deviation, z i Represents the spatial relationship coefficient of visual calibration points.

[0110] There may be interference factors such as lighting changes, stains, and wear in the PCB production environment, which will affect the recognition of visual calibration points. Various processing procedures such as drilling and patching require accurate position information. Therefore, in the PCB processing process, accurately determining the position of the PCB visual calibration points is crucial to ensuring production quality and efficiency. The present invention can improve the accuracy and reliability of the PCB visual inspection system through the visual confidence of the PCB visual calibration points. In the present invention, the visual confidence of the PCB visual calibration points is calculated by the signal-to-noise ratio of the visual calibration points, the spatial relationship coefficient of the visual calibration points, the Gaussian fitting standard deviation of the visual calibration points edge, and the maximum edge Gaussian fitting standard deviation. A higher SNR means better image quality and more reliable PCB visual calibration points. In a clear and noise-free image, the signal of the PCB visual calibration point can be clearly highlighted, which is conducive to accurate identification. Its position, edge sharpness part, the standard deviation is calculated by Gaussian fitting the gradient distribution of the edge. The smaller the value, the sharper the edge. After normalization, the sharpness is converted to a value between 0 and 1, and the maximum value corresponds to the sharpest edge, ensuring that the calibration point with clear edge obtains a higher weight in the confidence calculation. In order to convert the confidence of a single PCB vision calibration point into a weight relative to all PCB vision calibration points, normalized weight distribution is adopted to ensure that the sum of the weights of all PCB vision calibration points is 1. In this way, when the SNR of a PCB vision calibration point is high and the edge is sharp, its weight is significantly improved, and the weight of the PCB vision calibration point with low confidence due to occlusion, blur and other reasons is automatically reduced, effectively reducing its negative impact on the calculation results. By calculating the visual confidence of the PCB vision calibration point, the environmental adaptability and reliability of the PCB processing correction system are effectively improved.

[0111] The step of obtaining the spatial relationship coefficient of the visual calibration point according to the coordinate information of the PCB visual calibration point includes:

[0112] Obtaining the angle between the visual calibration point and the adjacent visual calibration point according to the coordinate information of the PCB visual calibration point;

[0113] Obtaining the distance between the visual calibration point and the adjacent visual calibration point according to the coordinate information of the PCB visual calibration point;

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

[0115] Obtain the theoretical distance between the visual calibration point and the adjacent visual calibration point; the theoretical angle and theoretical distance are obtained from the PCB design drawing;

[0116] Calculating the angular similarity between the angle of the visual calibration point and the adjacent visual calibration point and the theoretical angle;

[0117] Calculating the distance similarity between the distance between the visual calibration point and the adjacent visual calibration point and the theoretical distance;

[0118] The spatial relationship coefficient of the visual calibration points is calculated according to the angle similarity and the distance similarity.

[0119] The angle similarity and distance similarity are calculated using similarity formulas such as cosine similarity or Pearson correlation coefficient. The average value of the angle similarity and distance similarity is the spatial relationship coefficient of the visual calibration points.

[0120] Features on a PCB are often not isolated; there are certain spatial relationships between them. For example, features such as circuits and solder joints are densely packed and interrelated. Considering the relationship between adjacent visual calibration points helps determine the location of the point more accurately. PCBs may be affected by various factors, such as oil, dust, and uneven lighting, in different production environments and usage conditions. Considering the relationship between adjacent visual calibration points can enhance the adaptability of the detection system to these complex environments. By verifying the mutual presence of adjacent visual calibration points, the interference of environmental factors on the judgment of a single PCB visual calibration point can be reduced, thereby improving the stability of detection. The relationship between adjacent visual calibration points can be used to optimize the detection process. If the distribution of adjacent visual calibration points conforms to normal PCB design rules, the confidence of the PCB visual calibration point will be increased, otherwise it will be reduced. This allows the confidence calculation to be further optimized from the perspective of the spatial distribution of visual calibration points.

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

[0122] Acquire the temperature deviation of the visual calibration point according to the PCB surface temperature distribution information;

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

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

[0125] Get the system's sensitivity to temperature

[0126] The temperature influence attenuation value of the visual calibration point is obtained according to the temperature deviation of the visual calibration point, the thermal expansion coefficient of the PCB, the reference temperature of the visual calibration point, and the sensitivity coefficient of the system to temperature. The calculation formula is:

[0127]

[0128] Where, ω i represents the temperature attenuation value of the i-th visual calibration point, exp represents the natural exponential function, T i represents the temperature of the i-th visual calibration point, δ(T i ) represents the thermal expansion coefficient of the PCB at the temperature T of the i-th visual calibration point, δ1 represents the thermal expansion coefficient of the PCB at the reference temperature, ΔTi represents the temperature deviation of the i-th visual calibration point, T ref Indicates the reference temperature, λ indicates the system's sensitivity to temperature (λ is used to control the rate at which the attenuation value of the temperature effect of the visual calibration point decreases with temperature deviation. In some common PCB manufacturing or detection industrial scenarios, if the temperature change is relatively mild, the system's sensitivity to temperature is moderate, and λ is between (0.1-1). For example, in a conventional PCB patch production 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 take a smaller value, so that the temperature factor plays a moderate regulatory role in the calculation of the attenuation value of the temperature effect of the visual calibration point. For high-precision PCB manufacturing processes, such as advanced semiconductor packaging, which are very sensitive to temperature changes, more significant adjustments are required. To clearly reflect the impact of temperature factors on the weight coefficient, λ can take a value between (1-10). For example, in some packaging processes that require extremely high chip position accuracy, a slight change in temperature may cause the chip position to shift. Therefore, a larger λ value is required to more strictly adjust the weight according to the 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 tests or specific high-temperature industrial production environments, if the temperature change range is large and the change speed is fast, in order to enable the system to adjust the temperature influence attenuation value of the visual calibration point in a timely and accurate manner according to the temperature change, increasing λ can enhance the role of the temperature factor in the calculation of the temperature influence attenuation value of the visual calibration point).

[0129] Considering the impact of temperature on the importance of PCB visual calibration points, temperature changes will cause thermal expansion or contraction of PCB materials, which in turn causes the position of visual calibration points to shift. The temperature impact is quantified by the formula, where ΔT i =T(x i ,y i )-T ref is the local temperature of the visual calibration point obtained by the infrared thermal imager, T ref is the reference temperature, which is usually set to the standard ambient temperature for PCB processing, such as 25°C. λ controls the rate at which the weight decreases with temperature deviation. This formula imposes a nonlinear penalty on temperature deviation. When |ΔT i As | increases, the weight of the relevant visual calibration points decreases exponentially. Normalization processing is performed to avoid parameter incompatibility due to different reference temperatures. At the same time, taking into account the different sensitivities of different PCB materials to temperature, in high-precision PCB manufacturing processes, for materials that are extremely sensitive to temperature changes, slight changes in temperature may cause large deviations in processing accuracy. In this case, the PCB thermal expansion coefficient can more significantly adjust the impact of temperature on the weight, so that temperature factors can be more accurately considered in subsequent processing. In practical applications, when the temperature difference of the visual calibration point is small, its weight is close to the theoretical maximum value, dominating the calculation of related parameters to ensure high-precision positioning. When the temperature difference of the visual calibration point is large, the weight is significantly reduced, effectively avoiding unreliable data interfering with the calibration. The system can adapt to changes in temperature distribution during the processing process, such as local heating caused by heat generated by drilling. Under complex working conditions, it automatically focuses on reliable PCB visual calibration points, greatly improving the robustness of the system.

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

[0131] Acquire the actual coordinates of the visual calibration points according to the coordinate information of the PCB visual calibration points;

[0132] Obtaining the coordinates of the PCB fixed points;

[0133] The distance influence coefficient is obtained according to the difference between the coordinates of the PCB fixed point and the coordinates of the PCB fixed point, and the calculation formula is:

[0134]

[0135] Where, d i Indicates the distance influence coefficient of the i-th visual calibration point, P1 i Indicates the actual coordinates of the i-th visual calibration point, P3 i2represents the coordinates of the i2th PCB fixed point, n2 represents the number of PCB fixed point coordinates, and v represents the distance influence conversion coefficient. During the PCB processing process, a fixture is required to clamp and fix the PCB from a specific part, and the coordinates of the PCB fixed point are obtained through the clamped part of the PCB. During the PCB processing process, the drilling equipment or the patch equipment contacts the PCB. The closer to the coordinates of the PCB fixed point, the smaller the vibration amplitude is due to the clamping restriction, and the farther away from the coordinates of the PCB fixed point, the smaller the vibration amplitude is. Therefore, by calculating the distance between the coordinates of the visual calibration point and the coordinates of the PCB fixed point, the impact of vibration on the entire system can be more accurately evaluated. 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 of the visual calibration point is obtained according to the overall vibration acceleration value, and the calculation formula is:

[0137]

[0138] Where a i represents the root mean square value of the vibration acceleration of the i-th visual calibration point, a(t) represents the vibration acceleration of the visual calibration point at time t, and n1 represents the number of samples at time t;

[0139] Get vibration sensitivity coefficient;

[0140] The vibration impact attenuation value of the visual calibration point is obtained according to the distance influence coefficient, the overall vibration acceleration value and the vibration sensitivity coefficient, and the calculation formula is:

[0141]

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

[0143] During the PCB processing, vibration is an interference factor that cannot be ignored. The operation of the equipment, the vibration of the mechanical structure, etc. will cause the PCB to vibrate, and this vibration will affect the position accuracy of the visual calibration points on the PCB. For example, when performing high-precision patch operations, even tiny vibrations may cause deviations in the placement of components. In order to more reasonably consider the influence of vibration factors during the processing correction process, the present invention sets a vibration-related item to adjust the weight of each visual calibration point, so as to more accurately perform processing correction compensation. Therefore, the visual calibration point vibration influence attenuation value is introduced. The visual calibration point vibration influence attenuation value can dynamically adjust the weight of the visual calibration point according to the overall vibration acceleration of the PCB. When the root mean square value of the vibration acceleration of the visual calibration point is large, that is, the PCB vibrates violently, the value of the visual calibration point vibration influence attenuation value decreases, which means that the corresponding visual calibration point in the subsequent calculation The weight is reduced. In this way, in the processing correction calculation, the contribution of the visual calibration points affected by greater vibration to the final result will be reduced, thereby reducing the impact of vibration on processing accuracy. Conversely, when the vibration is small, the adjustment effect of the vibration attenuation value of the visual calibration point on the weight is weak, and the weight of the visual calibration point is relatively stable, ensuring that under normal circumstances, the processing correction calculation can be based on reliable visual information. Among them, γ is used to adjust the influence of the vibration factor on the weight coefficient. Its value can be adjusted according to the actual PCB processing scenario and the system's sensitivity to vibration. For example, in high-precision PCB processing technology that is 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 the visual calibration point when the vibration is severe. In ordinary PCB processing scenarios that are relatively less sensitive to vibration, the value of γ can be reduced to avoid excessive weight adjustment. i It can comprehensively reflect the severity of PCB vibration over a period of time, providing an accurate vibration intensity measurement indicator for the vibration attenuation item, so that the weight adjustment can be closely based on the actual vibration conditions. By setting the vibration impact attenuation value of the visual calibration point, the system can more effectively deal with vibration interference during the processing correction process, thereby improving the accuracy and stability of PCB processing.

[0144] The step of constructing a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function includes:

[0145] Obtaining the elastic modulus of the PCB substrate and the stress of the visual calibration point, and calculating 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 according to the elastic modulus of the PCB substrate and the stress of the visual calibration point;

[0146] Acquire the actual coordinates of the visual calibration points according to the coordinate information of the PCB visual calibration points;

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

[0148] Obtaining the PCB surface temperature distribution information;

[0149] A correction 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 PCB thermal expansion coefficient, the PCB substrate elastic modulus, and the PCB surface temperature distribution information. The projection coordinates of the visual calibration point are obtained through the correction compensation function. The formula is:

[0150]

[0151] Where P1 i represents the actual coordinates of the i-th visual calibration point, θ j represents the jth correction compensation parameter (initial correction compensation parameter), s x Indicates the scaling factor x of the PCB in the x-axis direction i The coordinate component of the actual coordinate of the i-th visual calibration point in the x-axis direction, Δx represents the x-axis translation distance of the visual calibration point, δ(T i ) represents the PCB thermal expansion coefficient of the i-th visual calibration point at temperature T, ΔT i represents the temperature deviation of the i-th visual calibration point, σ x(i) represents the stress of the i-th visual calibration point in the x-axis direction, E represents the elastic modulus of the PCB substrate, represents the strain of the i-th visual calibration point in the x-axis direction, s y Indicates the scaling factor y of the i-th visual calibration point in the y-axis direction i The coordinate component of the actual coordinate of the i-th visual calibration point in the y-axis direction, Δy represents the y-axis translation distance of the visual calibration point, σ y(i) represents the stress of the i-th visual calibration point in the y-axis direction, 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 of the visual calibration point in the x-axis direction and the stress of the visual calibration point in the y-axis direction can be detected and calculated by optical methods, thermal imaging methods or electronic methods. The following shows one of the acquisition methods:

[0153] The stress calculation formula of the visual calibration point in the x-axis direction is:

[0154] σ x(i) =E*o x *h;

[0155] Where σ x(i)It represents the stress of the visual calibration point in the x-axis direction, E represents the elastic modulus of the PCB substrate,

[0156] The stress calculation formula of the visual calibration point in the y-axis direction is:

[0157] σ y(i) =E*o y *h;

[0158] Where σ x(i) represents the stress of the i-th visual calibration point in the x-axis direction, σ y(i) represents the stress of the i-th visual calibration point in the y-axis direction, o x represents the real-time warping component of the i-th visual calibration point in the x-axis direction, o y represents the real-time warpage component of the i-th visual calibration point in the y-axis direction, E represents the elastic modulus of the PCB substrate, and h represents the PCB thickness;

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

[0160] The real-time warpage calculation formula is:

[0161]

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

[0163] The correction compensation function is used to accurately map the coordinates of the actual measurement points on the PCB to the theoretical coordinate system to achieve accurate calibration and correction of the PCB position and size. During the actual PCB processing process, various factors can cause deviations between the points on the PCB and the theoretical positions. Temperature changes can cause the PCB material to expand or contract thermally. This expansion or contraction may be uniform overall or non-uniform due to local temperature differences. Mechanical stress can cause physical deformation of the PCB, such as warping and stretching. At the same time, errors in the imaging system, positioning device, etc. may also occur, causing the coordinates of the measurement points to differ from the theoretical values. Therefore, the present invention provides a correction compensation function that comprehensively considers these factors, corrects the coordinates of the actual measurement points, and accurately reflects the actual position and size of the PCB, providing accurate data for subsequent processing correction. Based on the data, the correction compensation function can significantly improve the accuracy of PCB processing correction. In practical applications, it can effectively compensate for the PCB position and size deviations caused by factors such as temperature changes and physical deformation, so that the coordinates after projection transformation are closer to the theoretical coordinates, thereby reducing errors in the processing process and improving product quality and production efficiency. In the high-precision chip packaging process, it can ensure that the chip is accurately mounted at the specified position on the PCB and reduce the scrap rate. The correction compensation function includes a scaling coefficient to compensate for overall uniform thermal expansion or contraction and size scaling caused by factors such as the imaging system; the thermal expansion coefficient is combined with the local temperature deviation to compensate for the non-uniform thermal expansion caused by local temperature changes; Δx and Δy are used to adjust the offset of the visual calibration point on the x-axis and y-axis; and the 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 comprehensively handle various deviations that may occur in the PCB during actual processing. The correction compensation function can dynamically adjust parameters based on real-time monitoring data (such as temperature, vibration, physical deformation, etc.) to adapt to different processing environments and PCB conditions.

[0164] The step of constructing a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function includes:

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

[0166] Obtaining the correction compensation function;

[0167] Obtaining the multi-source correction influencing factor;

[0168] Get the last correction compensation parameters;

[0169] Obtain the impact value of the adjustment amplitude of the correction compensation parameter;

[0170] A correction compensation parameter optimization model is constructed based on the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influencing factor, the previous correction compensation parameter, and the correction compensation parameter adjustment amplitude influence value, and its formula is:

[0171]

[0172] Where θ j represents the jth correction compensation parameter, w i P1 represents the multi-source correction influence factor of the i-th visual calibration point, i represents the actual coordinates of the i-th visual calibration point, f proj (P1 i ,θ j ) represents the correction compensation function of the actual coordinates of the i-th visual calibration point with respect to the j-th correction compensation parameter, which is used to map the actual coordinates of the visual calibration point to the theoretical coordinate system, P2 i represents the theoretical coordinates of the i-th visual calibration point, θ j-1 represents the j-1th correction compensation parameter, and β represents the influence value of the correction compensation parameter adjustment amplitude;

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

[0174] In the complex scenario of PCB processing correction, it is necessary to comprehensively consider multiple factors to determine the optimal correction compensation parameters to achieve accurate calibration of the target position during PCB processing. There are deviations between the actual coordinates of the PCB visual calibration points obtained by actual measurement and the theoretical coordinates of the visual calibration points. These deviations are caused by the combined effects of multiple factors such as temperature changes, vibrations, and physical deformations. The present invention measures the size of these deviations through a correction compensation parameter optimization model, and finds the correction compensation parameters that minimize the deviations by optimizing this model. At the same time, in order 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. , it is possible to find a set of optimal correction compensation parameters so that the actual coordinates of the PCB visual calibration points after projection transformation are as close as possible to the theoretical coordinates, thereby improving the accuracy of PCB processing. For example, in the high-precision PCB patch process, the error of the patch position can be controlled within a very small range, greatly reducing the scrap rate. The correction compensation parameter optimization model takes into account a variety of influencing factors, including visual confidence, temperature deviation and vibration, and calculates and processes different PCB visual calibration points through multi-source correction influencing factors. This allows the system to maintain good performance in the face of complex and changeable processing environments and various interference factors, thereby improving the robustness of the system; β*||θ j -θ j-1 || 2 The setting of the term constrains the range of variation of the correction compensation parameters, avoids excessive fluctuations in the parameters, and enables 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. The correction compensation parameter optimization model is iteratively optimized using optimization algorithms (such as Levenberg-Marquardt, gradient descent method, and Newton method). 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 the optimal solution is found.

[0175] The step of obtaining the influence value of the adjustment amplitude of the correction compensation parameter includes:

[0176] Get the vibration influence coefficient;

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

[0178] The deviation correction compensation parameter adjustment amplitude influence value is calculated according to the initial deviation correction compensation parameter adjustment amplitude influence value and the vibration influence coefficient, and the calculation formula is:

[0179] β=β0*(1+γ*d i *a i );

[0180] Where, β represents the influence value of the adjustment range of the correction compensation parameter, β0 represents the influence value of the initial correction compensation parameter adjustment range, (1+γ*d i *a i ) represents the vibration influence coefficient; by introducing the vibration data collected by the acceleration sensor, the adaptability of the system in the vibration environment can be improved. The vibration influence coefficient can comprehensively reflect the energy of the vibration signal over a period of time and can effectively characterize the severity of the environmental vibration. β0 is a basic value set when the vibration influence is not considered, which is used in the balance correction compensation parameter optimization model. The term and β*||θ j -θ j-1 || 2 The relative importance of the term, when the vibration intensifies, that is (1+γ*d i *a i ) will increase, thus increasing β. After β increases, in the correction compensation parameter optimization model, β*||θ j -θ j-1 ‖ 2 Increasing the weight of the item will force the difference between the current correction compensation parameter and the previous correction compensation parameter to become smaller, making the parameter change smoother. The purpose of this is to suppress the interference of high-frequency noise caused by vibration on the parameters, avoid drastic fluctuations in the correction compensation parameters caused by 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] Get the gain coefficient;

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

[0184] Get the last correction compensation parameters;

[0185] The time series smoothing process is performed on the correction compensation parameter to be optimized according to the gain coefficient and the last correction compensation parameter, and the calculation formula is:

[0186] θ j+1 =K*θ j +(1-K)*θ j-1 ;

[0187] Where θ j+1 represents the j+1th correction compensation parameter after time series smoothing, θ j represents the jth correction compensation parameter, K represents the gain coefficient (for example, Kalman gain), θ j-1 Indicates the j-1th correction compensation parameter (the previous correction compensation parameter).

[0188] In PCB processing scenarios, the original correction compensation parameters (the jth correction compensation parameters) are often affected by various interference factors (such as vibration, temperature changes, and sensor noise), resulting in 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 enhancing the precision and reliability of operations such as PCB processing correction. K uses the Kalman gain, which is between 0 and 1. It is used 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 based on factors such as uncertainty during the PCB processing process. When the measurement noise is small, the Kalman gain will be closer to 1, indicating that the original correction compensation parameters are more reliable. When the measurement noise is large and the uncertainty is high, the Kalman gain will be closer to 0, indicating greater reliance on the previous correction compensation parameters. The present invention comprehensively considers the original correction compensation parameters and the previous correction compensation parameters, making it more reliable and stable than simply the original correction compensation parameters.

[0189] like Figure 2 As shown, the present invention also discloses a PCB processing correction system based on visual calibration, comprising:

[0190] Selection module 1, for selecting PCB visual calibration points based on feature significance;

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

[0192] A fusion module 3 is configured 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] A first calculation module 4 is used to construct a correction compensation function based on the physical deformation of the PCB;

[0194] A second calculation module 5 is used to construct a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function;

[0195] The third calculation module 6 is used to solve the correction compensation parameter optimization model through an optimization algorithm to obtain the optimal correction compensation parameters;

[0196] A fourth calculation module 7, configured to perform time series smoothing processing on the correction compensation parameters;

[0197] A first control module 8 is configured to generate an x-axis translation distance, a y-axis translation distance, and a rotation angle compensation instruction based on the smoothed correction compensation parameter;

[0198] The second control module 9 is used to transmit the x-axis translation distance, y-axis translation distance and rotation angle compensation instructions 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 instructions.

[0199] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0200] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A PCB board processing correction method based on visual calibration, characterized in that: include: Select PCB visual calibration points based on feature significance; Collect multi-source feature information of PCB visual calibration points in real time, and obtain visual confidence of the visual calibration points, temperature influence attenuation value of the visual calibration points, and vibration influence attenuation value of the visual calibration points based on the multi-source feature information of the PCB visual calibration points; fusing 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; Construct a correction compensation function based on the physical deformation of the PCB; Constructing a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function; Solve the correction compensation parameter optimization model through the optimization algorithm to obtain the optimal correction compensation parameters; Performing time series smoothing processing on the correction compensation parameters; Generate x-axis translation distance, y-axis translation distance and rotation angle compensation instructions based on the smoothed correction compensation parameters; The x-axis translation distance, y-axis translation distance and rotation angle compensation instructions are transmitted to a PCB board processing device, and the PCB board processing device adjusts a target processing position based on the x-axis translation distance, y-axis translation distance and rotation angle compensation instructions.

2. A PCB board processing correction method based on visual calibration according to claim 1, characterized in that: The step of obtaining 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 according to the multi-source feature information of the PCB visual calibration point comprises: Obtaining coordinate information of the PCB visual calibration point and PCB image information, and obtaining visual confidence of the visual calibration point based on the coordinate information of the PCB visual calibration point and the PCB image information; Acquire PCB surface temperature distribution information and PCB thermal expansion coefficient, and acquire a temperature influence attenuation value of a visual calibration point based on the PCB surface temperature distribution information and the PCB thermal expansion coefficient; Obtaining the coordinates of the PCB fixed point and the overall vibration acceleration value, and obtaining the vibration impact attenuation value of the visual calibration point according to the coordinate information of the PCB visual calibration point, the coordinates of the PCB fixed point and the overall vibration acceleration value; 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 are fused to obtain a multi-source correction influence factor.

3. The PCB board processing correction method based on visual calibration according to claim 2 is characterized in that: The step of obtaining the visual confidence of the visual calibration point according to the coordinate information of the PCB visual calibration point and the PCB image information includes: Obtaining a grayscale mean value of a visual calibration point according to the PCB image information; Obtaining a standard deviation of a background area corresponding to a visual calibration point according to the PCB image information; Calculating a signal-to-noise ratio of the visual calibration point according to the grayscale mean of the visual calibration point and the standard deviation of the background area corresponding to the visual calibration point; Obtaining a Gaussian fitting standard deviation of a visual calibration point edge and a maximum Gaussian fitting standard deviation of an edge according to the PCB image information; Obtaining a spatial relationship coefficient of the visual calibration point according to the coordinate information of the PCB visual calibration point; The visual confidence of the visual calibration point is obtained according to the signal-to-noise ratio of the visual calibration point, the spatial relationship coefficient of the visual calibration point, the Gaussian fitting standard deviation of the visual calibration point edge and the maximum edge Gaussian fitting standard deviation.

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

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

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

7. The PCB board processing and correction method based on visual calibration according to claim 2 is characterized in that: The step of constructing a correction compensation function based on the physical deformation of the PCB includes: Obtaining the elastic modulus of the PCB substrate and the stress of the visual calibration point, and calculating 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 according to the elastic modulus of the PCB substrate and the stress of the visual calibration point; Acquire the actual coordinates of the visual calibration points according to the coordinate information of the PCB visual calibration points; Obtaining the thermal expansion coefficient of the PCB; Obtaining the PCB surface temperature distribution information; A correction compensation function is constructed according to 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 PCB thermal expansion coefficient, the PCB substrate elastic modulus and the PCB surface temperature distribution information.

8. The PCB board processing correction method based on visual calibration according to claim 7 is characterized in that: The step of constructing a correction compensation parameter optimization model based on the multi-source correction influencing factors and the correction compensation function includes: Get the theoretical coordinates of the visual calibration points; Obtaining the correction compensation function; Obtaining the multi-source correction influencing factor; Get the last correction compensation parameters; Obtain the impact value of the adjustment amplitude of the correction compensation parameter; A correction compensation parameter optimization model is constructed according to the theoretical coordinates of the visual calibration point, the correction compensation function, the multi-source correction influencing factor, the last correction compensation parameter and the correction compensation parameter adjustment amplitude influence value.

9. The PCB board processing and correction method based on visual calibration according to claim 8, characterized in that: The step of obtaining the influence value of the adjustment amplitude of the correction compensation parameter includes: Get the vibration influence coefficient; Obtain the impact value of the initial correction compensation parameter adjustment amplitude; The deviation correction compensation parameter adjustment amplitude influence value is calculated according to the initial deviation correction compensation parameter adjustment amplitude influence value and the vibration influence coefficient.

10. A PCB processing correction system based on visual calibration, characterized in that: include: Selection module, used to select PCB visual calibration points based on feature significance; An acquisition module is used to acquire multi-source feature information of PCB visual calibration points in real time, and obtain visual confidence of the visual calibration points, temperature influence attenuation value of the visual calibration points, and vibration influence attenuation value of the visual calibration points based on the multi-source feature information of the PCB visual calibration points; a fusion module, configured 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 construct a correction compensation function based on the physical deformation of the PCB; A second calculation module is used to construct a correction compensation parameter optimization model based on the multi-source correction influencing factor and the correction compensation function; The third calculation module is used to solve the correction compensation parameter optimization model through an optimization algorithm to obtain the optimal correction compensation parameters; a fourth calculation module, configured to perform time series smoothing processing on the correction compensation parameters; A first control module is configured to generate an x-axis translation distance, a y-axis translation distance, and a rotation angle compensation instruction based on the smoothed correction compensation parameter; The second control module is used to transmit the x-axis translation distance, y-axis translation distance and rotation angle compensation instruction 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 instruction.

Images