Method and system for detecting uniformity of conductive substance in silk-screen printing

Abstract

The invention provides a silk-screen printing conductive substance uniformity detection method and system, and relates to the technical field of electronic material manufacturing and detection.The method comprises the steps that a conductive pattern image is acquired through an image acquisition device, surface topography feature data is obtained through phase profilometry processing, and the surface resistance value measured by a four-probe array is combined; establishing a resistance-density data pair, calculating a resistance change rate and a difference coefficient, generating a conductive uniformity quantitative index value, and identifying a conductive performance deviation region based on the conductive uniformity quantitative index value so as to optimize printing process parameters. The uniformity of the conductive substance can be accurately detected, and the accuracy of printing quality control is improved.

Description

Technical Field

[0001] The present invention relates to the manufacturing and detection technologies of electronic materials, and particularly to a method and system for detecting the uniformity of screen-printed conductive substances. Background Art

[0002] Currently, due to its advantages such as low cost, simple process, and wide application range, screen printing technology is widely used in the fields of electronic circuits, photovoltaic modules, and flexible electronics. However, the uniformity of conductive substance products directly affects their conductive performance and reliability. Especially in high-precision electronic devices, insufficient uniformity of conductive patterns can lead to abnormal local resistance, unstable device performance, and even failure. Traditional conductive performance detection methods mostly adopt single-point measurement methods, which are difficult to comprehensively reflect the three-dimensional distribution characteristics of conductive substances, and there are problems of one-sided detection results and large errors.

[0003] Existing detection means for screen-printed conductive substances usually focus on the evaluation of two-dimensional surface topography, while ignoring the accurate acquisition of three-dimensional parameters such as the thickness and volume density of conductive substances. At the same time, the method based only on surface resistance measurement is difficult to reveal the spatial resistance distribution characteristics of conductive patterns and their correlation with volume distribution density, making it have great limitations in identifying and quantifying regions with conductive uniformity deviation. In addition, the method of optimizing and adjusting printing process parameters based on detection results has not been systematized, lacking an effective feedback mechanism to improve conductive uniformity.

[0004] With the continuous growth of the demand for high-performance electronic devices, the market urgently needs a method for detecting conductive uniformity that can combine three-dimensional topography features and electrical performance measurement, which can accurately identify and quantitatively evaluate conductive performance deviations, and combine a process parameter optimization feedback system to improve the uniformity of screen-printed conductive substances and product quality. Therefore, there is an urgent need for a high-precision and systematic method and system for detecting the uniformity of screen-printed conductive substances. Summary of the Invention

[0005] Embodiments of the present invention provide a method and system for detecting the uniformity of screen-printed conductive substances, which can solve the problems in the prior art.

[0006] In the first aspect of the embodiments of the present invention, A method for detecting the uniformity of screen-printed conductive substances is provided, including: Collect the conductive pattern image of the screen-printed conductive material product as the input image through an image acquisition device, process the input image using phase profilometry to obtain the surface topography feature data of the conductive pattern, divide the surface area of the surface topography feature data into multiple detection grids, collect the three-dimensional position data of each detection grid, process the three-dimensional position data through an image segmentation algorithm to obtain the edge contour data of the conductive material, calculate the volume distribution density value of the conductive material in each detection grid based on the edge contour data, and construct a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value. The three-dimensional spatial distribution model includes the thickness value, area coverage coefficient, and spatial distribution parameters of the conductive material; Apply a predetermined voltage to multiple detection grids, measure the surface resistance measurement value using a four-probe array, generate the resistance distribution data of the detection grid according to the surface resistance measurement value, establish a spatial correspondence relationship between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional spatial distribution model, extract the resistance value and the corresponding volume distribution density value at each grid position to form a resistance-density data pair, calculate the ratio of the difference in resistance values between adjacent detection grids to their average resistance value to obtain the resistance change rate value, calculate the ratio of the difference in volume distribution density between adjacent detection grids to their average density value to obtain the difference coefficient value, generate a quantitative index value of conductivity uniformity according to the resistance change rate value and the difference coefficient value, identify the conductivity performance deviation area based on the quantitative index value of conductivity uniformity, and calculate the position coordinate value, deviation amplitude, and regional correlation coefficient of the conductivity performance deviation area; Input the position coordinate value, deviation amplitude, and regional correlation coefficient of the conductivity performance deviation area into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameters, generate the screen tension adjustment value and the squeegee pressure compensation value according to the printing pressure distribution parameters, input the screen tension adjustment value and the squeegee pressure compensation value into the printing equipment control system for parameter setting, perform the rework process of printing the conductive material after the parameter setting is completed, and repeat the detection and evaluation process for the conductive pattern after the rework process. When it is detected that the resistance change rate value between adjacent detection grids is less than the predetermined change rate threshold, confirm that the uniformity of the conductive material meets the process requirements.

[0007] In an alternative embodiment, Processing the input image using phase profilometry to obtain the surface topography feature data of the conductive pattern, dividing the surface area of the surface topography feature data into multiple detection grids, and collecting the three-dimensional position data of each detection grid, and processing the three-dimensional position data through an image segmentation algorithm to obtain the edge contour data of the conductive material includes: An image acquisition device is used to acquire the conductive pattern image of the screen-printed conductive material product as the input image. The conductive material product is irradiated by a collimated laser, and the incident parameters of the collimated laser and the optical parameters of the image acquisition device are adjusted to obtain the interference fringe images of the conductive pattern at four phase points. The four-step phase-shifting method in phase profilometry is used to process the interference fringe images at four phase points to obtain phase data. An adaptive phase compensation function is constructed based on the gray-scale difference between adjacent phase point images. The compensation coefficient of the phase compensation function is iteratively optimized to establish the mapping relationship between the compensation parameter and the environmental vibration frequency. The phase data is compensated according to the mapping relationship to obtain the compensated continuous phase distribution data. Based on the continuous phase distribution data, the average curvature and Gaussian curvature of each sampling point on the surface of the conductive pattern are calculated to obtain the surface topography feature data. A double-curvature threshold judgment criterion is constructed based on the average curvature and Gaussian curvature. A curvature gradient field is established based on the double-curvature threshold judgment criterion. The curvature mutation points are marked in the curvature gradient field. The surface area of the surface topography feature data is divided into multiple detection grids, and the grid size of the detection grid is determined according to the distribution position of the curvature mutation points, so that the grid size of the detection grid at the curvature mutation point is smaller than that of other regions. For each detection grid, an orthogonal structural element pair is constructed. The orthogonal structural element pair is respectively applied to the horizontal and vertical directions of the detection grid to obtain a bidirectional morphological gradient operator. The multi-scale decomposition layer number is determined according to the response value of the bidirectional morphological gradient operator. The morphological gradient operator is convolved with a Gaussian kernel function to construct a feature pyramid, and the three-dimensional position data of each detection grid is collected from the feature pyramid. Based on the three-dimensional position data and the surface topography feature data, an edge extraction processing matrix is established. The detection grid is divided into a target area and a background area. The gray-scale means of the target area and the background area are calculated. The initial segmentation threshold is determined according to the gray-scale means. Starting from the initial segmentation threshold, the dynamic segmentation thresholds of the target area and the background area are calculated in an iterative manner. The dynamic segmentation threshold is applied to the three-dimensional position data to obtain the initial edge contour points of the conductive material. The curvature of the initial edge contour points is calculated, and the edge contour points with abnormal curvature are removed. The remaining edge contour points are connected by cubic spline interpolation to obtain the edge contour data of the conductive material.

[0008] In an optional embodiment, Calculating the volume distribution density value of the conductive material in each detection grid according to the edge contour data, and constructing a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value includes: Extract the three-dimensional coordinate values of the edge contour points within each of the detection grids from the edge contour data. After inputting the three-dimensional coordinate values into a Gaussian filter to eliminate outliers, extract the difference between the maximum height value and the minimum height value within the detection grid as the local thickness value; Taking the center point of the detection grid as a reference, determine the intersection points of the edge contour and the boundary of the detection grid, calculate the actual covered area enclosed by the connection lines of the intersection points, and divide the product of the local thickness value and the actual covered area by the nominal volume of the grid to obtain the volume distribution density value, where the nominal volume of the grid is the product of the planar projected area of the detection grid and the standard height value; Calculate the average value and the standard deviation of the local thickness values of all the detection grids respectively to obtain the local thickness mean value and the local thickness standard deviation, and output the ratio of the local thickness standard deviation to the local thickness mean value as the thickness value of the conductive material; Calculate the sum of the actual covered areas of all the detection grids to obtain the total actual covered area, output the ratio of the total actual covered area to the theoretical area of the detection region as the area coverage coefficient, calculate the autocorrelation coefficient at different spatial displacement distances for the volume distribution density value as the spatial distribution parameter, and based on the volume distribution density value, construct a three-dimensional spatial distribution model including the thickness value, the area coverage coefficient, and the spatial distribution parameter.

[0009] In an alternative embodiment, Apply a predetermined voltage to multiple detection grids, measure the surface resistance measurement value using a four-probe array, generate the resistance distribution data of the detection grids according to the surface resistance measurement value, establish a spatial correspondence relationship between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional spatial distribution model, extract the resistance value and the corresponding volume distribution density value at each grid position, and form a resistance-density data pair including: Set positive and negative electrode pairs at the adjacent boundaries of multiple detection grids, apply an alternating predetermined voltage with an increasing amplitude and a fixed phase difference through the electrode pairs to form an initial electric field distribution of the detection grids, and determine the optimal predetermined voltage amplitude according to the current response between the electrode pairs. Move the four-probe array along a spiral scanning path within the detection grids with the predetermined voltage applied, and collect the probe contact force signal and the probe displacement signal in real time. Calculate the contact stability coefficient between the probe and the surface of the conductive material based on the fluctuation amplitude and the spectral characteristics of the signals; Dynamically adjust the measurement voltage amplitude and frequency of the outer probe pair of the four-probe array according to the contact stability coefficient, so that the measurement voltage and the predetermined voltage exhibit orthogonal characteristics in the frequency domain. Collect the voltage drop sequence and the phase information of the inner probe pair, separate the response characteristics of the predetermined voltage and the measurement voltage through orthogonal filtering, and construct a local equivalent circuit model of the conductive material considering the surface scattering effect in combination with the geometric parameters of the four-probe array, and extract the effective resistance component in the impedance characteristics as the surface resistance measurement value of the detection grid; Perform multi-scale decomposition on the measured value of surface resistance to extract the measurement noise characteristics, establish the mapping relationship between the noise characteristics and the contact stability coefficient, adaptively determine the number of repeated measurements, and use a weighted average algorithm considering signal stability to obtain the resistance distribution data of the detection grid; Associate the resistance distribution data with the two-dimensional measurement coordinates to construct a resistance distribution surface, extract the local curvature and gradient characteristics of the surface, increase the measurement points according to the adaptive density in the feature mutation region, and iteratively measure until the resistance change rate between adjacent measurement points meets the convergence condition to obtain an optimized high-precision resistance distribution surface; Based on the optimized resistance distribution surface, use a registration algorithm of feature point matching to establish the mapping relationship between the two-dimensional measurement coordinate system and the grid coordinate system of the three-dimensional space distribution model, calculate the equivalent resistance value of the grid position according to the mapping relationship, combine the volume distribution density value at the grid position, construct an initial resistance-density data pair considering spatial correlation, establish a measurement quality evaluation model based on the spatial distribution characteristics and weight distribution of the initial resistance-density data pair, calculate the data reliability index, and use a dynamic screening threshold adapted to the surface characteristics to finally output the resistance-density data pair that meets the quality requirements.

[0010] In an optional embodiment, Generate a quantitative index value of conductivity uniformity according to the resistance change rate value and the difference coefficient value, identify the conductivity performance deviation region based on the quantitative index value of conductivity uniformity, and calculate the position coordinate value, deviation amplitude, and regional correlation coefficient of the conductivity performance deviation region, including: Establish a grid space mapping matrix according to the three-dimensional coordinates and volume parameters of the detection grid, and calculate the grid shape eigenvalue; combine the grid shape eigenvalue with the resistance change rate value and the density difference coefficient value to construct a grid feature fusion model; perform normalization processing on the resistance change rate value and the density difference coefficient value based on the grid feature fusion model to obtain the grid feature normalized value; Establish a spatial distance matrix of adjacent detection grids, construct a distance weight calculation model based on the spatial distance matrix, input the grid feature normalized value into the distance weight calculation model to obtain the distance weighting factor between grids, and use a non-linear weighting method to fuse the grid feature normalized value and the distance weighting factor to generate a quantitative index value of conductivity uniformity; Construct a dynamic scanning window at the detection grid position where the quantitative index value of conductivity uniformity exceeds the preset index threshold, and calculate the index value gradient field within the dynamic scanning window; determine the search direction and step size based on the index value gradient field, and expand the dynamic scanning window until a gradient reverse point is obtained; determine the connected region between the gradient reverse points as the conductivity performance deviation region; Extract the features of the detection grid within the region with deviation in electrical conductivity performance, and establish a centroid calculation model with the quantitative index value of electrical conductivity uniformity as the weight coefficient; calculate the position coordinates of the region with deviation in electrical conductivity performance through an iterative optimization method; Calculate the difference between the maximum value of the index within the region with deviation in electrical conductivity performance and the average value of the indices of the adjacent grids outside the region boundary as the deviation amplitude; Analyze the topological connection relationship of the detection grids within the region with deviation in electrical conductivity performance, establish a correlation calculation model in combination with the spatial distribution characteristics of the quantitative index value of electrical conductivity uniformity, and quantitatively evaluate the region with deviation in electrical conductivity performance based on the correlation calculation model to output the regional correlation coefficient.

[0011] In an alternative embodiment, Extract the features of the detection grids within the region with deviation in electrical conductivity performance, and establish a centroid calculation model with the quantitative index value of electrical conductivity uniformity as the weight coefficient; calculating the position coordinates of the region with deviation in electrical conductivity performance through an iterative optimization method includes: Extract the resistance value, size, and position characteristic parameters of the detection grids within the region with deviation in electrical conductivity performance respectively, calculate the resistance gradient value based on the resistance value, and calculate the conductivity based on the resistance value and the size characteristic parameters; Normalize the resistance value, resistance gradient value, and conductivity of the detection grids to obtain standardized characteristic values, and calculate the quantitative index value of electrical conductivity uniformity based on the standardized characteristic values; Take the quantitative index value of electrical conductivity uniformity as the weight coefficient, establish a weighted centroid calculation model with the position characteristic parameters of the detection grids, and calculate the initial centroid coordinates of the region with deviation in electrical conductivity performance; Establish an optimization calculation region centered on the initial centroid coordinates, calculate the quantitative index value of electrical conductivity uniformity of the detection grids within the optimization calculation region based on the resistance value, resistance gradient value, and conductivity, substitute the calculated quantitative index value of electrical conductivity uniformity as the new weight coefficient into the weighted centroid calculation model, and calculate the new centroid coordinates; Calculate the difference between the new centroid coordinates and the initial centroid coordinates. When the difference is greater than the preset coordinate difference threshold, take the new centroid coordinates as the initial centroid coordinates and repeat the step of establishing the optimization calculation region until the difference is less than the preset coordinate difference threshold, and determine the finally calculated centroid coordinates as the position coordinates of the region with deviation in electrical conductivity performance.

[0012] In an alternative embodiment, Input the position coordinate values, deviation amplitudes, and regional correlation coefficients of the conductive property deviation regions into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameters. Generating the screen tension adjustment value and the squeegee pressure compensation value based on the printing pressure distribution parameters includes: Input the position coordinate values, deviation amplitudes, and regional correlation coefficients of the conductive property deviation regions into the printing defect analysis unit of the process parameter optimization system. Perform normalization processing based on the position coordinate values, and use a bivariate Gaussian distribution model to fit the position coordinates to obtain the spatial probability density function of the printing pressure distribution. The mean value of the spatial probability density function represents the pressure center offset of the printing pressure distribution parameters, and the covariance matrix represents the anisotropic characteristics of the printing pressure distribution parameters. Establish a weighted combination model of the pressure center offset and the anisotropic characteristics to obtain the pressure distribution characteristic quantity. Based on the deviation amplitude, establish a printing pressure compensation model, and use the piecewise linear interpolation method to construct a mapping function between the deviation amplitude and the pressure distribution adjustment amount. Calculate the initial pressure adjustment amount of the conductive property deviation region according to the mapping function, construct a topological relationship matrix based on the regional correlation coefficient, and perform spatial smoothing processing on the initial pressure adjustment amount based on the topological relationship matrix to obtain the pressure compensation characteristic quantity. Construct a weight coefficient matrix based on the pressure distribution characteristic quantity and a weight coefficient matrix based on the pressure compensation characteristic quantity. Multiply the weight coefficient matrix by the corresponding characteristic quantity to obtain the weighted characteristic quantity. Perform superposition operation on the weighted characteristic quantity to obtain the overall printing pressure distribution parameters. Calculate the covariance matrix of the overall printing pressure distribution parameters, perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalue sequence and the eigenvector sequence. Sort the eigenvector sequence according to the eigenvalue sequence size, select the eigenvector with the cumulative contribution rate reaching the preset contribution threshold as the change vector of the printing pressure distribution, construct a nonlinear mapping equation between the change vector and the screen tension, introduce the stress-strain curve parameters of the screen material into the nonlinear mapping equation, and establish a screen tension objective function. Use the gradient descent method to optimize and solve the tension objective function to obtain the optimal screen tension adjustment value. Perform projection transformation on the overall printing pressure distribution parameters in the squeegee length direction to obtain the pressure distribution curve. Perform piecewise polynomial fitting on the pressure distribution curve to establish a pressure compensation reference equation. Introduce a deformation compensation term related to the elastic modulus and thickness of the squeegee material into the pressure compensation reference equation to obtain the squeegee pressure compensation equation. Combine the squeegee pressure compensation equation with the motion trajectory equation including the squeegee moving speed and the squeegee inclination angle to calculate the squeegee pressure compensation value.

[0013] In the second aspect of the embodiments of the present invention, Provided is a system for detecting the uniformity of a screen-printed conductive material, comprising: A first unit, configured to collect an image of a conductive pattern of a screen-printed conductive material product through an image acquisition device as an input image, process the input image using phase profilometry to obtain surface topography feature data of the conductive pattern, divide the surface area of the surface topography feature data into a plurality of detection grids, collect three-dimensional position data of each detection grid, process the three-dimensional position data through an image segmentation algorithm to obtain edge contour data of the conductive material, calculate the volume distribution density value of the conductive material within each detection grid according to the edge contour data, and construct a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value. The three-dimensional spatial distribution model includes the thickness value, area coverage coefficient, and spatial distribution parameters of the conductive material; A second unit, configured to apply a predetermined voltage to a plurality of detection grids, measure a surface resistance measurement value using a four-probe array, generate resistance distribution data of the detection grids according to the surface resistance measurement value, establish a spatial correspondence relationship between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional spatial distribution model, extract the resistance value and the corresponding volume distribution density value at each grid position to form a resistance-density data pair, calculate the ratio of the difference in resistance values between adjacent detection grids to their average resistance value to obtain a resistance change rate value, calculate the ratio of the difference in volume distribution density between adjacent detection grids to their average density value to obtain a difference coefficient value, generate a quantitative index value of conductive uniformity according to the resistance change rate value and the difference coefficient value, identify a region with a conductive performance deviation based on the quantitative index value of conductive uniformity, and calculate the position coordinate value, deviation amplitude, and region correlation coefficient of the region with the conductive performance deviation; A third unit, configured to input the position coordinate value, deviation amplitude, and region correlation coefficient of the region with the conductive performance deviation into a printing defect analysis unit of a process parameter optimization system to calculate printing pressure distribution parameters, generate a screen tension adjustment value and a squeegee pressure compensation value according to the printing pressure distribution parameters, input the screen tension adjustment value and the squeegee pressure compensation value into a printing equipment control system for parameter setting, perform a rework process on the printed conductive material after the parameter setting is completed, and repeatedly execute the detection and evaluation process on the conductive pattern after the rework process. When it is detected that the resistance change rate value between adjacent detection grids is less than a predetermined change rate threshold, it is confirmed that the uniformity of the conductive material meets the process requirements.

[0014] In a third aspect of the embodiments of the present invention, Provided is an electronic device, comprising: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to call the instructions stored in the memory to execute the method described above.

[0015] In the fourth aspect of the embodiments of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the foregoing method is implemented.

[0016] In this embodiment, an image acquisition device is used to obtain an image of a conductive pattern, and phase profilometry is used to process and obtain surface topography feature data. By combining the three-dimensional position data analysis of the detection grid, the volume distribution density of the conductive material is analyzed, and a three-dimensional spatial distribution model is constructed, realizing an accurate quantitative characterization of the distribution state of the conductive material, providing a reliable data basis for subsequent uniformity evaluation. The surface resistance value of the detection grid is measured by a four-probe array. By establishing the correspondence between the resistance distribution data and the three-dimensional spatial distribution model, the resistance change rate and the density difference coefficient are calculated, and a quantitative index value of the conductive uniformity is generated, which can accurately identify the regions with conductive performance deviation, effectively evaluate the uniformity degree of the conductive material, and improve the accuracy and reliability of the detection results. Based on the characteristic parameters of the regions with conductive performance deviation, the printing pressure distribution parameters are calculated by a process parameter optimization system, and the screen tension adjustment value and the squeegee pressure compensation value are generated, realizing the automatic optimization and adjustment of the printing process parameters. Through rework processing and repeated detection, the uniformity of the conductive material is ensured to meet the process requirements, improving the production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic flow chart of the method for detecting the uniformity of a conductive material in screen printing according to the embodiments of the present invention; Figure 2 is a schematic structural diagram of the system for detecting the uniformity of a conductive material in screen printing according to the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] The technical solutions of the present invention will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0020] Figure 1 is a schematic flow chart of the method for detecting the uniformity of a conductive material in screen printing according to the embodiments of the present invention, as Figure 1As shown, the method includes: S101. Collect the conductive pattern image of the screen-printed conductive material product through an image acquisition device as the input image, process the input image using phase profilometry to obtain the surface topography feature data of the conductive pattern, divide the surface area of the surface topography feature data into multiple detection grids, collect the three-dimensional position data of each detection grid, process the three-dimensional position data through an image segmentation algorithm to obtain the edge contour data of the conductive material, calculate the volume distribution density value of the conductive material in each detection grid based on the edge contour data, and construct a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value. The three-dimensional spatial distribution model includes the thickness value, area coverage coefficient, and spatial distribution parameters of the conductive material; S102. Apply a predetermined voltage to multiple detection grids, measure the surface resistance measurement value using a four-probe array, generate the resistance distribution data of the detection grid according to the surface resistance measurement value, establish a spatial correspondence relationship between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional spatial distribution model, extract the resistance value and the corresponding volume distribution density value at each grid position to form a resistance-density data pair, calculate the ratio of the difference in resistance values between adjacent detection grids to their average resistance value to obtain the resistance change rate value, calculate the ratio of the difference in volume distribution density between adjacent detection grids to their average density value to obtain the difference coefficient value, generate a quantitative index value of conductivity uniformity according to the resistance change rate value and the difference coefficient value, identify the conductivity performance deviation area based on the quantitative index value of conductivity uniformity, and calculate the position coordinate value, deviation amplitude, and regional correlation coefficient of the conductivity performance deviation area; S103. Input the position coordinate value, deviation amplitude, and regional correlation coefficient of the conductivity performance deviation area into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameters, generate the screen tension adjustment value and the squeegee pressure compensation value according to the printing pressure distribution parameters, input the screen tension adjustment value and the squeegee pressure compensation value into the printing equipment control system for parameter setting, perform the rework process of printing the conductive material after the parameter setting is completed, repeat the detection and evaluation process for the conductive pattern after the rework process, and when it is detected that the resistance change rate value between adjacent detection grids is less than the predetermined change rate threshold, confirm that the uniformity of the conductive material meets the process requirements.

[0021] Among them, phase profilometry is a three-dimensional topography measurement method based on optical interference. By modulating the phase information of light waves, it can accurately obtain the height information of the object surface, thereby obtaining the surface topography feature data of the conductive pattern. The surface topography feature data refers to the three-dimensional structure information of the conductive pattern surface, including parameters such as height, curvature, and texture, which are used to describe the microscopic geometric features of the surface.

[0022] A four-probe array is a device used to measure sheet resistance, usually consisting of four equally spaced probes, capable of obtaining resistance values in a non-contact manner.

[0023] Rework processing refers to the process of reprinting or correcting a conductive pattern that does not meet the process requirements after printing. The detection and evaluation process is a process of repeatedly detecting and analyzing the conductive pattern after rework processing to confirm the effect of process parameter adjustment.

[0024] The change rate threshold is a standard value used to determine whether the resistance change rate meets the process requirements. When the resistance change rate of adjacent grids is lower than this value, it indicates that the uniformity meets the standard.

[0025] In an optional implementation, the surface topography feature data of the conductive pattern is obtained by processing the input image using phase profilometry. The surface area of the surface topography feature data is divided into multiple detection grids, the three-dimensional position data of each detection grid is collected, and the edge contour data of the conductive material is obtained by processing the three-dimensional position data through an image segmentation algorithm, including: An image acquisition device is used to acquire an image of the conductive pattern of the screen-printed conductive material product as the input image. The conductive material product is irradiated by a collimated laser, and the incident parameters of the collimated laser and the optical parameters of the image acquisition device are adjusted to obtain interference fringe images of the conductive pattern at four phase points; The four-step phase-shifting method in phase profilometry is used to process the interference fringe images at the four phase points to obtain phase data. An adaptive phase compensation function is constructed based on the gray difference between adjacent phase point images. The compensation coefficient of the phase compensation function is iteratively optimized to establish a mapping relationship between the compensation parameter and the environmental vibration frequency. The phase data is compensated according to the mapping relationship to obtain compensated continuous phase distribution data; Based on the continuous phase distribution data, the mean curvature and Gaussian curvature of each sampling point on the surface of the conductive pattern are calculated to obtain surface topography feature data. A double-curvature threshold judgment criterion is constructed based on the mean curvature and Gaussian curvature. A curvature gradient field is established based on the double-curvature threshold judgment criterion. Curvature mutation points are marked in the curvature gradient field. The surface area of the surface topography feature data is divided into multiple detection grids, and the grid size of the detection grid is determined according to the distribution position of the curvature mutation points, so that the grid size of the detection grid at the curvature mutation point is smaller than that of other regions; For each detection grid, an orthogonal structure element pair is constructed. The orthogonal structure element pair is respectively applied to the horizontal and vertical directions of the detection grid to obtain a bidirectional morphological gradient operator. The multi-scale decomposition layer number is determined according to the response value of the bidirectional morphological gradient operator. A feature pyramid is constructed by performing convolution operations on the morphological gradient operator using a Gaussian kernel function. The three-dimensional position data of each detection grid is collected from the feature pyramid; An edge extraction processing matrix is established based on the three-dimensional position data and surface topography feature data. The detection grid is divided into a target region and a background region. The gray mean values of the target region and the background region are calculated. An initial segmentation threshold is determined according to the gray mean values. Starting from the initial segmentation threshold, a dynamic segmentation threshold of the target region and the background region is calculated in an iterative manner. The dynamic segmentation threshold is applied to the three-dimensional position data to obtain the initial edge contour points of the conductive material. The curvature of the initial edge contour points is calculated, and the edge contour points with abnormal curvature are removed. The remaining edge contour points are connected by cubic spline interpolation to obtain the edge contour data of the conductive material.

[0026] Exemplarily, first, an image of the conductive pattern of the screen-printed conductive material product is collected as an input image. Then, the conductive material product is irradiated by a collimated laser to obtain the interference fringe images at four phase points of the conductive pattern.

[0027] Specifically, a high-resolution CCD camera is used as the image acquisition device, with a pixel resolution of 4096×3072 and a frame rate of 60fps. The collimated laser uses a helium-neon laser with a wavelength of 632.8nm and an output power of 5mW. The incident angle of the collimated laser is adjusted to 45 degrees, and the optical axis of the CCD camera is kept consistent with the normal direction of the sample surface. By adjusting the exposure time and gain parameters of the CCD camera, the gray value distribution of the interference fringe image is between 0-255 to obtain the best image contrast.

[0028] Next, the four-step phase-shifting method in phase profilometry is used to process the interference fringe images at four phase points to obtain phase data. The specific steps are as follows: First, the four interference fringe images are grayscale processed, and then the wrapped phase of each pixel point is calculated using the four-step phase-shifting algorithm. To improve the accuracy of phase calculation, a phase calculation method with an adaptive window size is adopted, and the size of the calculation window is dynamically adjusted according to the texture features of the local image. Generally, the window size is between 5×5 and 15×15 pixels.

[0029] To eliminate the influence of environmental vibration on phase measurement, an adaptive phase compensation function is constructed according to the gray difference between adjacent phase point images. The specific method is: calculate the gray difference map of adjacent two interference fringe images, and construct a compensation function using the statistical features of the gray difference map. The compensation function adopts a polynomial form, and the number of coefficients is determined according to the complexity of the vibration frequency, generally 3-5.

[0030] Iteratively optimize the compensation coefficients of the phase compensation function. Use the least squares method to optimize the compensation coefficients. Set the number of iterations to 100 times and the convergence threshold to 0.001. Establish the mapping relationship between the compensation parameters and the environmental vibration frequency. Through experimental tests, obtain the optimal compensation parameters at different vibration frequencies and establish a look-up table. According to the real-time monitored environmental vibration frequency, select the corresponding compensation parameters from the look-up table to compensate the phase data and obtain the compensated continuous phase distribution data.

[0031] Calculate the mean curvature and Gaussian curvature of each sampling point on the surface of the conductive pattern based on the continuous phase distribution data to obtain the surface topography feature data. Use the moving least squares method to fit the local surface, calculate the principal curvature of the surface, and then obtain the mean curvature and Gaussian curvature. Set the fitting window size to 11×11 pixels and the polynomial order to 2.

[0032] Construct a double-curvature threshold judgment criterion based on the mean curvature and Gaussian curvature. The selection of the threshold is based on the statistical characteristics of the curvature histogram. Generally, take the 95% quantile of the curvature distribution as the threshold. Establish a curvature gradient field based on the double-curvature threshold judgment criterion. Use the Sobel operator to calculate the curvature gradient, and the size of the gradient operator is 3×3. Mark the curvature mutation points in the curvature gradient field, and mark the points with curvature gradient values greater than the threshold as mutation points.

[0033] Divide the surface area of the surface topography feature data into multiple detection grids, and determine the grid size of the detection grids according to the distribution positions of the curvature mutation points. Adopt an adaptive grid division strategy. In the area where the curvature mutation points are dense, set the grid size to 5×5 pixels, and in other areas, set the grid size to 15×15 pixels.

[0034] Construct orthogonal structure element pairs for each detection grid. The shape of the structure element is linear and the length is 5 pixels. Apply the orthogonal structure element pairs to the horizontal and vertical directions of the detection grid respectively to obtain a bidirectional morphological gradient operator. Determine the multi-scale decomposition layer number according to the response value of the bidirectional morphological gradient operator, generally set to 3 - 5 layers.

[0035] Use the Gaussian kernel function to perform convolution operations on the morphological gradient operator to construct a feature pyramid. Set the standard deviation of the Gaussian kernel function to 1.6 and the kernel size to 5×5. Collect the three-dimensional position data of each detection grid from the feature pyramid, and use the interpolation method to improve the sampling accuracy. The interpolation method selected is bicubic interpolation.

[0036] An edge extraction processing matrix is established based on three-dimensional position data and surface topography feature data, and the size of the matrix is the same as the image size. The detection grid is divided into a target area and a background area, and the Otsu method is used to calculate the initial segmentation threshold. Starting from the initial segmentation threshold, an iterative method is used to calculate the dynamic segmentation thresholds of the target area and the background area. The number of iterations is set to 50, and the convergence threshold is set to 0.01.

[0037] The dynamic segmentation thresholds are applied to the three-dimensional position data to obtain the initial edge contour points of the conductive material. The curvature of the initial edge contour points is calculated, and the three-point method is used to calculate the curvature of discrete points. The edge contour points with abnormal curvature are removed, and the abnormal judgment criterion is that the curvature value exceeds 3 times the standard deviation of the average curvature. Cubic spline interpolation is used to connect the remaining edge contour points, and the number of interpolation points is set to 2 times the original number of points to obtain the edge contour data of the conductive material.

[0038] In this embodiment, through the phase profilometry and the adaptive phase compensation technology, the measurement accuracy and anti-vibration ability of the surface topography of the conductive pattern are effectively improved. The four-step phase-shifting method and the adaptive window phase calculation method are adopted to improve the calculation accuracy of the phase data. By constructing an adaptive phase compensation function and establishing the mapping relationship between the compensation parameters and the environmental vibration frequency, the influence of environmental vibration on the measurement results is effectively suppressed, so that high-quality surface topography data can be obtained even in a complex industrial environment. The adaptive grid division strategy based on the double curvature threshold is adopted to realize the efficient sampling of the surface of the conductive pattern. By analyzing the distribution characteristics of the mean curvature and the Gaussian curvature, the characteristic regions of the surface are accurately identified, and smaller grid sizes are used for fine sampling in these regions. This strategy not only ensures the measurement accuracy of the key regions but also improves the overall calculation efficiency. Combining the morphological gradient operator and the multi-scale decomposition technology, the accurate extraction of the edge of the conductive material is realized. By constructing a feature pyramid and using the dynamic segmentation thresholds, the noise and uneven illumination problems in the image are effectively processed. The curvature analysis and cubic spline interpolation technologies are used to further optimize the smoothness and continuity of the edge contour. This comprehensive processing method significantly improves the accuracy and robustness of the edge detection of the conductive pattern, providing a reliable data basis for subsequent quality control and defect detection.

[0039] In an alternative embodiment, calculating the volume distribution density value of the conductive material in each detection grid according to the edge contour data, and constructing a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value includes: Extracting the three-dimensional coordinate values of the edge contour points in each detection grid from the edge contour data, inputting the three-dimensional coordinate values into a Gaussian filter to eliminate abnormal points, and then extracting the difference between the maximum height value and the minimum height value in the detection grid as the local thickness value; Taking the center point of the detection grid as a reference, determine the intersection points of the edge contour and the boundary of the detection grid, calculate the actual covered area enclosed by the connecting line of the intersection points, divide the product of the local thickness value and the actual covered area by the nominal volume of the grid to obtain the volume distribution density value, where the nominal volume of the grid is the product of the planar projected area of the detection grid and the standard height value; Calculate the average value and standard deviation of the local thickness values of all detection grids respectively to obtain the local thickness mean value and the local thickness standard deviation, and output the ratio of the local thickness standard deviation to the local thickness mean value as the thickness value of the conductive material; Calculate the sum of the actual covered areas of all detection grids to obtain the total actual covered area, output the ratio of the total actual covered area to the theoretical area of the detection region as the area coverage coefficient, calculate the autocorrelation coefficient of different spatial displacement distances for the volume distribution density value as the spatial distribution parameter, and construct a three-dimensional spatial distribution model including the thickness value, the area coverage coefficient and the spatial distribution parameter based on the volume distribution density value.

[0040] Exemplarily, first, it is necessary to obtain the edge contour data within the detection grid. The detection grid adopts a regular rectangular grid division method, and the grid size can be set to 5 mm × 5 mm. For each detection grid, extract the three-dimensional coordinate values of all contour points within the grid from the edge contour data. To eliminate the abnormal points generated during the measurement process, use a Gaussian filter to smooth the coordinate values. The kernel size of the Gaussian filter is set to 3×3, and the standard deviation value is taken as 1.5. After the filtering process, extract the maximum height value and the minimum height value of the contour points within each grid, and calculate the difference between the two as the local thickness value of the conductive material within the grid.

[0041] When determining the actual covered area of the conductive material, taking the center point of the detection grid as the reference point, calculate the intersection coordinates of the edge contour and the four boundaries of the grid. When there are multiple intersection points, connect these intersection points in a clockwise direction in turn to form a closed polygon. Use the ray method to calculate the area of the polygon, which is the actual covered area of the conductive material within the grid. The nominal volume of the grid is equal to the product of the planar projected area of the grid and the standard height value. The standard height value can be set according to the actual application scenario, for example, taking 10 mm. Divide the product of the local thickness value and the actual covered area by the nominal volume of the grid to obtain the volume distribution density value of the conductive material within the grid.

[0042] Conduct a statistical analysis on the local thickness values of all detection grids, calculate the average value as the local thickness mean value, and at the same time calculate the standard deviation as the local thickness standard deviation. Taking an actual case as an example, a total of 400 grids are divided within the detection region. The calculated local thickness mean value is 8.5 mm, and the local thickness standard deviation is 1.2 mm. The ratio of the two is approximately 0.141, and this value reflects the uniformity of the thickness of the conductive material.

[0043] When calculating the area coverage feature, the actual coverage areas of all detection grids are accumulated to obtain the total sum. The theoretical area of the detection region is the sum of the areas of all grids. Taking the above case as an example, the total actual coverage area is 8500 square millimeters, and the theoretical area of the detection region is 10000 square millimeters. The calculated area coverage coefficient is 0.85, indicating the coverage degree of the conductive substance. Calculate the autocorrelation coefficient at different displacement distances for the volume distribution density value. The displacement distance can be selected as an integer multiple of the grid size, such as 5 millimeters, 10 millimeters, 15 millimeters, etc. The obtained autocorrelation coefficient reflects the spatial correlation of the distribution of the conductive substance.

[0044] Finally, based on the calculated volume distribution density value, combined with the thickness value, area coverage coefficient, and spatial distribution parameters, a three-dimensional spatial distribution model of the conductive substance is constructed. This model can be used to characterize the distribution characteristics of the conductive substance in space and provide a basis for subsequent analysis and optimization.

[0045] In this embodiment, by using Gaussian filtering to eliminate abnormal points and extract local thickness values, the accuracy and reliability of the thickness measurement of the conductive substance are improved, and the influence of measurement noise on the results is effectively reduced. The actual coverage area is calculated by using the grid center point reference and the connection line of the boundary intersection points, which accurately reflects the distribution range of the conductive substance and overcomes the problem of inaccurate area calculation in the traditional method. The three-dimensional spatial distribution model constructed based on the volume distribution density value comprehensively considers the thickness uniformity, area coverage degree, and spatial correlation, and can comprehensively characterize the distribution characteristics of the conductive substance, providing reliable technical support for related process optimization and quality control.

[0046] In an alternative implementation, a predetermined voltage is applied to multiple detection grids, and the surface resistance measurement value is obtained by using a four-probe array. Resistance distribution data of the detection grids are generated according to the surface resistance measurement value. A spatial correspondence relationship is established between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional spatial distribution model, and the resistance value and the corresponding volume distribution density value at each grid position are extracted to form a resistance-density data pair, including: Positive and negative electrode pairs are set at the adjacent boundaries of multiple detection grids. An alternating predetermined voltage with an increasing amplitude and a fixed phase difference is applied through the electrode pairs to form an initial electric field distribution of the detection grids, and the optimal predetermined voltage amplitude is determined according to the current response between the electrode pairs. The four-probe array is moved within the detection grids with the applied predetermined voltage along a spiral scanning path, and the probe contact force signal and the probe displacement signal are collected in real time. The contact stability coefficient between the probe and the surface of the conductive substance is calculated based on the fluctuation amplitude and spectral characteristics of the signals; Dynamically adjust the measurement voltage amplitude and frequency of the outer probe pair of the four-probe array according to the contact stability coefficient, so that the measurement voltage and the predetermined voltage exhibit orthogonal characteristics in the frequency domain. Collect the voltage drop sequence and phase information of the inner probe pair, separate the response characteristics of the predetermined voltage and the measurement voltage through orthogonal filtering, construct a local equivalent circuit model of the conductive material considering the surface scattering effect in combination with the geometric parameters of the four-probe array, and extract the effective resistance component in the impedance characteristics as the surface resistance measurement value of the detection grid; Perform multi-scale decomposition on the surface resistance measurement value to extract the measurement noise characteristics, establish the mapping relationship between the noise characteristics and the contact stability coefficient, adaptively determine the number of repeated measurements, and use the weighted average algorithm considering signal stability to obtain the resistance distribution data of the detection grid; Associate the resistance distribution data with the two-dimensional measurement coordinates to construct a resistance distribution surface, extract the local curvature and gradient characteristics of the surface, increase the measurement points according to the adaptive density in the feature mutation area, and iteratively measure until the resistance change rate between adjacent measurement points meets the convergence condition to obtain an optimized high-precision resistance distribution surface; Based on the optimized resistance distribution surface, use the registration algorithm of feature point matching to establish the mapping relationship between the two-dimensional measurement coordinate system and the grid coordinate system of the three-dimensional space distribution model. Calculate the equivalent resistance value of the grid position according to the mapping relationship, combine the volume distribution density value at the grid position, construct the initial resistance-density data pair considering spatial correlation, establish a measurement quality evaluation model based on the spatial distribution characteristics and weight distribution of the initial resistance-density data pair, calculate the data reliability index, and use a dynamic screening threshold adapted to the surface characteristics to finally output the resistance-density data pair that meets the quality requirements.

[0047] Exemplarily, first arrange electrode pairs on the detection grid and apply a predetermined voltage to the detection grid using an AC voltage source. The electrode pairs are set at the adjacent grid boundaries, and an initial electric field distribution is formed by gradually increasing the voltage amplitude (starting from 0.1V and increasing in steps of 0.1V) and maintaining a 90-degree phase difference. By monitoring the current response between the electrode pairs, when the signal-to-noise ratio of the current response reaches above 20dB, determine this voltage amplitude as the optimal predetermined voltage.

[0048] The four-probe array moves within the detection grid along an Archimedean spiral path. The starting point of the spiral path is set at the center of the grid, and the pitch is 1.5 times the probe spacing. During the movement, the probe contact force signal and displacement signal are collected in real time, and the sampling frequency is set to 1kHz. Perform time-frequency analysis on the collected signals, calculate the standard deviation and spectral energy distribution of the signals. When the standard deviation is less than 0.05 and the energy ratio of the main frequency component exceeds 85%, it indicates that the probe is in stable contact with the surface of the conductive material.

[0049] Dynamically adjust the measurement voltage of the outer probe pair according to the contact stability. When the contact stability is high, the measurement voltage frequency is selected as 1.5 times the predetermined voltage frequency, and the amplitude is taken as 0.8 times the predetermined voltage; when the contact stability decreases, the measurement voltage amplitude is correspondingly reduced and the frequency is increased to ensure that the measurement signal is orthogonal to the predetermined voltage in the frequency domain. A band-stop filter bank is used to separate the response characteristics of the two voltages. The center frequencies of the filters correspond to the frequencies of the predetermined voltage and the measurement voltage respectively, and the bandwidth is 10% of the center frequency.

[0050] Perform wavelet decomposition on the measurement data, select the db4 wavelet basis function, and perform 5-layer decomposition. Extract the measurement noise characteristics according to the variance distribution characteristics of each scale coefficient. When the noise variance exceeds 5% of the signal variance, increase the number of repeated measurements. Use the exponentially weighted average method based on signal stability to synthesize the resistance distribution data, and the weight coefficient is positively correlated with the signal stability.

[0051] When constructing the resistance distribution surface, use the radial basis function interpolation method, and select the multiquadric function as the kernel function. Calculate the Gaussian curvature and gradient values of each point on the surface. When the curvature change rate exceeds 30% or the gradient value exceeds the preset threshold, increase the measurement points in this area, and the measurement point spacing is reduced to half of the original. Repeat the measurement until the resistance change rate between adjacent points is less than 2%.

[0052] Finally, use the registration algorithm based on SIFT features to establish the mapping relationship between the two-dimensional measurement coordinates and the three-dimensional grid coordinates. Extract the surface feature points, match them with the surface feature points of the three-dimensional model, and use the RANSAC algorithm to remove the mismatched points. Calculate the equivalent resistance value of the grid position according to the mapping relationship, and construct the initial data pair in combination with the volume distribution density.

[0053] In this embodiment, by dynamically adjusting the measurement parameters and the multiple repeated measurement strategy, the accuracy and reliability of the resistance measurement are significantly improved. The AC voltage measurement method effectively suppresses external interference and improves the signal-to-noise ratio of the measurement signal. The measurement scheme based on the spiral scanning path and the adaptive encryption strategy realizes the efficient coverage of the detection area, and obtains higher spatial resolution in the area where the resistance value changes violently, ensuring the spatial continuity of the measurement results. The data processing method considering contact stability establishes a reliable resistance density correspondence relationship, providing high-quality basic data for subsequent analysis. The accurate docking of the two-dimensional and three-dimensional coordinate systems is achieved through feature point matching, ensuring the spatial consistency of the data.

[0054] In an alternative embodiment, a quantitative index value of conductivity uniformity is generated according to the resistance change rate value and the difference coefficient value, and a region with conductivity performance deviation is identified based on the quantitative index value of conductivity uniformity. Calculating the position coordinate value, deviation amplitude and regional correlation coefficient of the region with conductivity performance deviation includes: Based on the three-dimensional coordinates and volume parameters of the detection grid, establish a grid space mapping matrix and calculate the grid shape eigenvalue; combine the grid shape eigenvalue with the resistance change rate value and the density difference coefficient value to construct a grid feature fusion model; perform normalization processing on the resistance change rate value and the density difference coefficient value based on the grid feature fusion model to obtain the grid feature normalized value; Establish a spatial distance matrix for adjacent detection grids, construct a distance weight calculation model based on the spatial distance matrix, input the grid feature normalized value into the distance weight calculation model to obtain the distance weighting factor between grids, and use a non-linear weighting method to fuse the grid feature normalized value with the distance weighting factor to generate a quantitative index value of conductivity uniformity; Construct a dynamic scanning window at the position of the detection grid where the quantitative index value of conductivity uniformity exceeds the preset index threshold, and calculate the index value gradient field within the dynamic scanning window; determine the search direction and step size based on the index value gradient field, and expand the dynamic scanning window until a gradient reversal point is obtained; determine the connected region between the gradient reversal points as the region with conductivity performance deviation; Extract the features of the detection grids within the region with conductivity performance deviation, and establish a centroid calculation model with the quantitative index value of conductivity uniformity as the weight coefficient; calculate the position coordinates of the region with conductivity performance deviation through an iterative optimization method; Calculate the difference between the maximum value of the index within the region with conductivity performance deviation and the average value of the indices of the adjacent grids outside the region boundary as the deviation amplitude; Analyze the topological connection relationship of the detection grids within the region with conductivity performance deviation, establish a correlation calculation model in combination with the spatial distribution characteristics of the quantitative index value of conductivity uniformity, and perform quantitative evaluation on the region with conductivity performance deviation based on the correlation calculation model to output the regional correlation coefficient.

[0055] Exemplarily, first perform three-dimensional spatial modeling on the detection grid, collect volume parameters such as the length, width, and height of each grid, and establish the spatial mapping relationship of the grid. By analyzing the geometric shape of the grid, extract the shape features of the grid, including parameters such as the volume ratio, surface area ratio, and side length ratio of the grid. Combine these shape features with the resistance change rate and the density difference coefficient to construct a feature fusion model. Perform normalization processing on the combined features to make features with different dimensions comparable. For example, the volume ratio of a certain detection grid is 0.85, the surface area ratio is 0.92, the side length ratio is 0.88, the resistance change rate is 0.15, and the density difference coefficient is 0.12. After normalization processing, the grid feature normalized value is 0.78.

[0056] Next, calculate the spatial distance between adjacent detection grids and construct a distance weight model. Use a Gaussian kernel function to weight the distance, where the closer the grid, the greater the weight. Non-linearly combine the normalized grid features with the distance weights to generate a quantitative index of conductivity uniformity. For example, the distance weights of a certain detection grid from its adjacent grids are 0.9, 0.8, and 0.7 respectively. Combining with the normalized value of the grid feature 0.78, the conductivity uniformity index value of this grid is finally obtained as 0.82.

[0057] Construct a dynamic scanning window at the grid position where the conductivity uniformity index value exceeds a preset threshold (such as 0.75), and calculate the gradient change of the index values within the window. Determine the search path according to the gradient direction, and gradually expand the scanning window until the position where the gradient reverses is found. Determine the connected region between these gradient reversal points as the region with conductivity performance deviation. For example, the index values of 5 adjacent grids are detected to exceed 0.75 within a certain region, and the boundary range of this region is determined through dynamic scanning.

[0058] Extract the features of the determined deviation region. Use the conductivity uniformity index value as the weight coefficient and calculate the centroid position of the region in an iterative manner. For example, a certain deviation region contains 10 grid points. According to the index value weights of each point, the centroid coordinates of the region are calculated as (12.5, 8.3, 4.2).

[0059] Calculate the difference between the maximum index value within the deviation region and the average index value of the adjacent grids outside the region boundary to obtain the deviation amplitude. For example, the maximum index value within a certain deviation region is 0.95, and the average index value of the adjacent grids outside the boundary is 0.65, then the deviation amplitude is 0.3.

[0060] Finally, analyze the topological connection relationship of the grids within the deviation region. Combine with the spatial distribution characteristics of the conductivity uniformity index to establish a correlation degree calculation model. Quantitatively evaluate the deviation region by calculating the connectivity and index similarity between grids, and output the regional correlation coefficient. For example, the grid connectivity of a certain deviation region is 0.88, and the index similarity is 0.92. Finally, the regional correlation coefficient is obtained as 0.90.

[0061] In this embodiment, by establishing a grid feature fusion model and a distance weight calculation model, the accurate quantitative evaluation of the conductivity performance is realized, and the accuracy and reliability of the conductivity uniformity detection are improved. By using the dynamic scanning window and gradient search method, the boundary of the region with conductivity performance deviation can be adaptively determined, enhancing the flexibility and adaptability of the deviation region identification. Based on the centroid calculation and correlation degree analysis, the precise positioning and correlation evaluation of the deviation region are realized, providing reliable data support and decision-making basis for the optimization of the conductivity performance.

[0062] In an alternative embodiment, feature extraction is performed on the detection grids within the region of deviation in conductive performance, and a centroid calculation model is established using the quantitative index value of conductive uniformity as a weight coefficient; calculating the position coordinates of the region of deviation in conductive performance through an iterative optimization method includes: Resistance values, dimensions, and position characteristic parameters are respectively extracted from the detection grids within the region of deviation in conductive performance. Based on the resistance values, resistance gradient values are calculated, and based on the resistance values and dimension characteristic parameters, conductivity is calculated; The resistance values, resistance gradient values, and conductivity of the detection grids are normalized to obtain standardized eigenvalue, and based on the standardized eigenvalue, a quantitative index value of conductive uniformity is calculated; Using the quantitative index value of conductive uniformity as a weight coefficient, a weighted centroid calculation model is established with the position characteristic parameters of the detection grids, and the initial centroid coordinates of the region of deviation in conductive performance are calculated; An optimization calculation region is established with the initial centroid coordinates as the center. Based on the resistance values, resistance gradient values, and conductivity, the quantitative index value of conductive uniformity of the detection grids within the optimization calculation region is calculated, and the calculated quantitative index value of conductive uniformity is substituted as a new weight coefficient into the weighted centroid calculation model to calculate new centroid coordinates; Calculate the difference between the new centroid coordinates and the initial centroid coordinates. When the difference is greater than a preset coordinate difference threshold, use the new centroid coordinates as the initial centroid coordinates and repeat the step of establishing the optimization calculation region until the difference is less than the preset coordinate difference threshold. Determine the finally calculated centroid coordinates as the position coordinates of the region of deviation in conductive performance.

[0063] During the detection of the region of deviation in conductive performance, feature extraction is first performed on the detection grids. This includes extracting the resistance value, dimension, and position characteristic parameters of each detection grid. The resistance value is obtained through a high-precision resistance measuring instrument. The dimension characteristics include the length and width of the grid, and the position characteristic is the coordinate of the grid in the entire detection region.

[0064] Exemplarily, first, based on the obtained resistance values, resistance gradient values are calculated. The resistance gradient value reflects the rate of change of resistance in space and can be obtained by dividing the resistance difference between adjacent grids by the grid spacing. For example, if the resistance values of two adjacent grids are 10 ohms and 12 ohms respectively, and the grid spacing is 1 cm, then the resistance gradient value is 2 ohms / cm.

[0065] Next, conductivity is calculated using the resistance value and dimension characteristics. Conductivity is the reciprocal of resistivity and can be calculated from the resistance value, length, and cross-sectional area of the grid. Suppose the resistance of a certain grid is 5 ohms, the length is 2 cm, and the cross-sectional area is 0.1 square cm, then its conductivity is 4 siemens / m.

[0066] To make different features comparable, it is necessary to normalize the resistance value, resistance gradient value, and conductivity to obtain standardized feature values. Normalization can be performed using the min-max scaling method to map each feature value between 0 and 1. For example, if the resistance value ranges from 5 ohms to 15 ohms, the standardized value of 10 ohms is 0.5.

[0067] Based on the standardized feature values, calculate the quantitative index value of conductivity uniformity. This can be obtained by taking the weighted average of the standardized resistance value, resistance gradient value, and conductivity. The weights can be adjusted according to the specific application scenario. For example, weights of 0.4, 0.3, and 0.3 can be assigned respectively.

[0068] Use the calculated quantitative index value of conductivity uniformity as the weight coefficient to establish a weighted centroid calculation model with the position characteristic parameters of the detection grid. This model takes into account the position of each grid and the importance of its conductivity. For example, if the position coordinates of a certain grid are (3, 4) and its quantitative index value of conductivity uniformity is 0.8, then in the centroid calculation, the contribution of this grid is (2.4, 3.2).

[0069] Through the weighted centroid calculation model, the initial centroid coordinates of the conductivity deviation region can be obtained. This coordinate represents the central position of the region considering the conductivity.

[0070] With the initial centroid coordinates as the center, establish an optimized calculation region. This region can be a circular or square region centered on the initial centroid, and its size can be adjusted according to the actual situation. For example, it can be taken as 1 / 4 of the entire detection region.

[0071] Within the optimized calculation region, recalculate the quantitative index value of conductivity uniformity of the detection grid. The purpose of this step is to evaluate the conductivity within a more precise range. The calculation method is the same as before, but only the grids within the optimized calculation region are considered.

[0072] Substitute the newly calculated quantitative index value of conductivity uniformity as the new weight coefficient into the weighted centroid calculation model to obtain the new centroid coordinates. This new coordinate reflects the central position of the region considering the conductivity within a more precise range.

[0073] Calculate the difference between the new centroid coordinates and the initial centroid coordinates. This difference reflects the degree of optimization. Set a preset coordinate difference threshold, such as 0.1 cm. If the difference is greater than this threshold, it indicates that there is still room for further optimization.

[0074] When the difference is greater than the preset coordinate difference threshold, use the new centroid coordinates as the initial centroid coordinates and repeat the step of establishing the optimized calculation region. This process will continuously shrink the calculation region and improve the positioning accuracy.

[0075] Repeat the above steps until the difference between the new and old centroid coordinates is less than the preset coordinate difference threshold. At this time, the finally calculated centroid coordinates can be determined as the final position coordinates of the region with poor electrical conductivity performance.

[0076] In this embodiment, through multi-dimensional feature extraction and normalization processing of the detection grid, all aspects of the electrical conductivity performance are comprehensively considered, improving the accuracy and comprehensiveness of the evaluation. This method not only considers the resistance value, but also introduces the resistance gradient and conductivity, making the evaluation of the electrical conductivity performance more comprehensive and accurate. The iterative optimization method is used to locate the region with poor electrical conductivity performance, greatly improving the positioning accuracy. By continuously shrinking the calculation area and re-evaluating, it is possible to gradually approach the true deviation center, avoiding the errors that may be brought by single calculation, and making the final positioning result more reliable and accurate. The introduction of the quantitative index of electrical conductivity uniformity as a weight coefficient enables the full consideration of the influence of electrical conductivity performance during the positioning process. This weighted method ensures that the regions with poor electrical conductivity performance have a greater influence in the positioning calculation, thus more accurately reflecting the position of the true region with poor electrical conductivity performance. This not only improves the positioning accuracy, but also provides more valuable reference information for subsequent quality improvement and production optimization.

[0077] In an alternative embodiment, input the position coordinate value, deviation amplitude, and region correlation coefficient of the region with poor electrical conductivity performance into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameters. Generating the screen tension adjustment value and the squeegee pressure compensation value according to the printing pressure distribution parameters includes: Input the position coordinate value, deviation amplitude, and region correlation coefficient of the region with poor electrical conductivity performance into the printing defect analysis unit of the process parameter optimization system. Based on the position coordinate value, perform normalization processing, and use the bivariate Gaussian distribution model to fit the position coordinates to obtain the spatial probability density function of the printing pressure distribution; where the mean value of the spatial probability density function represents the pressure center offset of the printing pressure distribution parameters, and the covariance matrix represents the anisotropic characteristics of the printing pressure distribution parameters; establish a weighted combination model of the pressure center offset and the anisotropic characteristics to obtain the pressure distribution characteristic quantity; Based on the deviation amplitude, establish a printing pressure compensation model, and use the piecewise linear interpolation method to construct a mapping function between the deviation amplitude and the pressure distribution adjustment amount; calculate the initial pressure adjustment amount of the region with poor electrical conductivity performance according to the mapping function, construct a topological relationship matrix based on the region correlation coefficient, and perform spatial smoothing processing on the initial pressure adjustment amount based on the topological relationship matrix to obtain the pressure compensation characteristic quantity; Construct a weight coefficient matrix based on pressure distribution feature quantities and a weight coefficient matrix based on pressure compensation feature quantities; multiply the weight coefficient matrix by the corresponding feature quantities respectively to obtain weighted feature quantities; perform a superposition operation on the weighted feature quantities to obtain an overall printing pressure distribution parameter; Calculate the covariance matrix of the overall printing pressure distribution parameter, perform eigenvalue decomposition on the covariance matrix, obtain an eigenvalue sequence and an eigenvector sequence, sort the eigenvector sequence according to the size of the eigenvalue sequence, select the eigenvectors with the cumulative contribution rate reaching a preset contribution threshold as the change vectors of the printing pressure distribution, construct a non-linear mapping equation between the change vectors and the screen tension, introduce the stress-strain curve parameters of the screen material into the non-linear mapping equation, and establish a screen tension objective function; use the gradient descent method to optimize and solve the tension objective function to obtain the optimal screen tension adjustment value; Perform a projection transformation on the overall printing pressure distribution parameter in the doctor blade length direction to obtain a pressure distribution curve, perform piecewise polynomial fitting on the pressure distribution curve, establish a pressure compensation reference equation, introduce a deformation compensation term related to the elastic modulus and thickness of the doctor blade material into the pressure compensation reference equation to obtain a doctor blade pressure compensation equation, and combine the doctor blade pressure compensation equation with the motion trajectory equation including the doctor blade moving speed and doctor blade inclination angle to calculate the doctor blade pressure compensation value.

[0078] Exemplarily, first obtain the position coordinate values, deviation amplitudes, and regional correlation coefficients of the conductive performance deviation regions, and input these data into the printing defect analysis unit of the process parameter optimization system. Perform standardized preprocessing on the position coordinate values, and normalize the coordinate values to the range of zero to one. Then use a bivariate Gaussian distribution model to fit the standardized position coordinates, and obtain the spatial probability density function of the printing pressure distribution through the maximum likelihood estimation method. Specifically, when implementing, the number of sample points can be selected as 1000, the number of iteration times is set to 500, and the convergence threshold is set to 0.001.

[0079] After obtaining the spatial probability density function, extract its mean value as the offset of the pressure center, and this offset reflects the unevenness of the printing pressure distribution. At the same time, extract the covariance matrix to characterize the anisotropic characteristics of the pressure distribution. By setting the offset weight coefficient to 0.6 and the anisotropic characteristic weight coefficient to 0.4, establish a weighted combination model to obtain the pressure distribution feature quantity.

[0080] Next, a printing pressure compensation model is established based on the deviation amplitude. Using the piecewise linear interpolation method, the deviation amplitude is divided into multiple intervals, with the span of each interval set to 0.1, and a linear mapping relationship is established within each interval. When the deviation amplitude is between 0 and 0.1, the pressure adjustment amount is calculated according to a linear ratio; when the deviation amplitude is between 0.1 and 0.2, the adjustment amount increases by 50%; when the deviation amplitude is greater than 0.2, the adjustment amount increases by 100%.

[0081] Construct a topological relationship matrix according to the regional correlation coefficient. The correlation coefficient threshold is set to 0.8. When the correlation coefficient between two regions is greater than the threshold, the corresponding position in the matrix is marked as 1, otherwise it is marked as 0. Use this matrix to perform spatial smoothing processing on the initial pressure adjustment amount, and the smoothing window size is set to 3×3 to obtain the pressure compensation feature quantity.

[0082] When constructing the weight coefficient matrix, the weight of the pressure distribution feature quantity is set to 0.7, and the weight of the pressure compensation feature quantity is set to 0.3. Multiply the weight matrix by the corresponding feature quantity and superimpose them to obtain the overall printing pressure distribution parameter.

[0083] Calculate the covariance matrix of the overall printing pressure distribution parameter and perform eigenvalue decomposition. Set the preset contribution threshold to 85%, and select the eigenvectors whose cumulative contribution rate reaches this threshold as the change vectors of the printing pressure distribution. Combine the stress-strain curve parameters of the stencil material to establish a stencil tension objective function. Use the gradient descent method to optimize and solve, with the learning rate set to 0.01 and the maximum number of iterations set to 1000 times to obtain the optimal stencil tension adjustment value.

[0084] Finally, projectively transform the overall printing pressure distribution parameter in the length direction of the squeegee to obtain the pressure distribution curve. Use the cubic spline interpolation method to perform piecewise fitting on the curve and establish a pressure compensation reference equation. Introduce a deformation compensation term related to the elastic modulus and thickness of the squeegee material, and combine the motion trajectory equation of the squeegee moving speed and squeegee inclination angle to calculate the squeegee pressure compensation value.

[0085] In this embodiment, by establishing a bivariate Gaussian distribution model and a pressure compensation model, an accurate description and effective compensation of the printing pressure distribution are achieved, improving the uniformity and stability of the printing quality. By combining eigenvalue decomposition and a non-linear mapping equation, an optimization model of the stencil tension is established, achieving precise adjustment of the stencil tension and enhancing the controllability of the printing process. Through piecewise fitting of the pressure distribution curve and deformation compensation, a squeegee pressure compensation equation is established, achieving dynamic adjustment of the squeegee pressure and improving the adaptability and reliability of the printing process.

[0086] Figure 2 For the structural schematic diagram of a silk screen printing conductive material uniformity detection system according to an embodiment of the present invention, as Figure 2As shown, the system includes: A first unit configured to collect, via an image acquisition device, an image of a conductive pattern of a screen-printed conductive material product as an input image, process the input image using phase profilometry to obtain surface topography feature data of the conductive pattern, divide the surface area of the surface topography feature data into a plurality of detection grids, collect three-dimensional position data of each detection grid, process the three-dimensional position data through an image segmentation algorithm to obtain edge contour data of the conductive material, calculate a volume distribution density value of the conductive material within each detection grid based on the edge contour data, and construct a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value. The three-dimensional spatial distribution model includes a thickness value, an area coverage coefficient, and a spatial distribution parameter of the conductive material. A second unit configured to apply a predetermined voltage to a plurality of detection grids, measure a surface resistance measurement value using a four-probe array, generate resistance distribution data of the detection grids based on the surface resistance measurement value, establish a spatial correspondence relationship between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional spatial distribution model, extract the resistance value and the corresponding volume distribution density value at each grid position to form a resistance-density data pair, calculate a ratio of a difference in resistance values between adjacent detection grids to their average resistance value to obtain a resistance change rate value, calculate a ratio of a difference in volume distribution densities between adjacent detection grids to their average density value to obtain a difference coefficient value, generate a quantitative index value of conductivity uniformity based on the resistance change rate value and the difference coefficient value, identify a region with a deviation in conductivity performance based on the quantitative index value of conductivity uniformity, and calculate the position coordinate value, deviation amplitude, and regional correlation coefficient of the region with a deviation in conductivity performance. A third unit configured to input the position coordinate value, deviation amplitude, and regional correlation coefficient of the region with a deviation in conductivity performance into a printing defect analysis unit of a process parameter optimization system to calculate printing pressure distribution parameters, generate a screen tension adjustment value and a squeegee pressure compensation value based on the printing pressure distribution parameters, input the screen tension adjustment value and the squeegee pressure compensation value into a printing equipment control system for parameter setting, perform a rework process for printing the conductive material after the parameter setting is completed, repeat the detection and evaluation process for the conductive pattern after the rework process, and when it is detected that the resistance change rate value between adjacent detection grids is less than a predetermined change rate threshold, confirm that the uniformity of the conductive material meets the process requirements.

[0087] In a third aspect of the embodiments of the present invention, An electronic device is provided, including: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to call the instructions stored in the memory to execute the method described above.

[0088] In a fourth aspect of the embodiments of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the foregoing method is implemented.

[0089] The present invention may be a method, an apparatus, a system, and / or a computer program product. The computer program product may include a computer-readable storage medium, on which computer-readable program instructions for executing various aspects of the present invention are carried.

[0090] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting uniformity of screen-printed conductive material, characterized in that: include: The conductive pattern image of the screen-printed conductive material product is collected as an input image by an image acquisition device, and the input image is processed by phase profilometry to obtain surface morphology feature data of the conductive pattern, and the surface area of ​​the surface morphology feature data is divided into a plurality of detection grids, and the three-dimensional position data of each detection grid is collected, and the three-dimensional position data is processed by an image segmentation algorithm to obtain edge contour data of the conductive material, and the volume distribution density value of the conductive material in each detection grid is calculated according to the edge contour data, and a three-dimensional spatial distribution model of the conductive material is constructed based on the volume distribution density value, wherein the three-dimensional spatial distribution model includes a thickness value, an area coverage factor and a spatial distribution parameter of the conductive material; Apply a predetermined voltage to a plurality of detection grids, use a four-probe array to measure and obtain a surface resistance measurement value, generate resistance distribution data of the detection grid according to the surface resistance measurement value, establish a spatial correspondence between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional space distribution model, extract the resistance value and the corresponding volume distribution density value of each grid position to form a resistance-density data pair, calculate the ratio of the difference between the resistance values ​​of adjacent detection grids and their average resistance value to obtain a resistance change rate value, calculate the ratio of the difference between the volume distribution density of adjacent detection grids and their average density value to obtain a difference coefficient value, generate a conductive uniformity quantitative index value according to the resistance change rate value and the difference coefficient value, identify a conductive performance deviation area based on the conductive uniformity quantitative index value, and calculate the position coordinate value, deviation amplitude and regional correlation coefficient of the conductive performance deviation area; The position coordinate values, deviation amplitude and area correlation coefficient of the conductive performance deviation area are input into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameters, and the screen tension adjustment value and the scraper pressure compensation value are generated according to the printing pressure distribution parameters. The screen tension adjustment value and the scraper pressure compensation value are input into the printing equipment control system for parameter setting. After the parameter setting is completed, the printed conductive material is reworked, and the detection and evaluation process is repeated for the conductive pattern after the rework. When it is detected that the resistance change rate value of adjacent detection grids is less than the predetermined change rate threshold, it is confirmed that the uniformity of the conductive material meets the process requirements.

2. The method according to claim 1, characterized in that The input image is processed by phase profilometry to obtain surface morphology feature data of the conductive pattern, the surface area of ​​the surface morphology feature data is divided into a plurality of detection grids, the three-dimensional position data of each detection grid is collected, and the edge profile data of the conductive material is obtained by processing the three-dimensional position data by an image segmentation algorithm. The method includes: An image acquisition device is used to acquire a conductive pattern image of a screen-printed conductive material product as an input image, a collimated laser is used to irradiate the conductive material product, and the incident parameters of the collimated laser and the optical parameters of the image acquisition device are adjusted to acquire interference fringe images of the conductive pattern at four phase points; The interference fringe images of four phase points are processed by the four-step phase shift method in phase profilometry to obtain phase data, an adaptive phase compensation function is constructed according to the grayscale difference of the adjacent phase point images, the compensation coefficient of the phase compensation function is iteratively optimized, a mapping relationship between the compensation parameter and the ambient vibration frequency is established, the phase data is compensated according to the mapping relationship, and the compensated continuous phase distribution data is obtained; Calculating the average curvature and Gaussian curvature of each sampling point on the surface of the conductive pattern based on the continuous phase distribution data to obtain surface morphology feature data, constructing a hyperbolic threshold judgment criterion based on the average curvature and Gaussian curvature, establishing a curvature gradient field based on the hyperbolic threshold judgment criterion, marking a curvature mutation point in the curvature gradient field, dividing the surface area of ​​the surface morphology feature data into a plurality of detection grids, and determining the grid size of the detection grid according to the distribution position of the curvature mutation point, so that the detection grid size at the curvature mutation point is smaller than the detection grid size in other areas; An orthogonal structural element pair is constructed for each detection grid, and the orthogonal structural element pair is respectively applied to the horizontal direction and the vertical direction of the detection grid to obtain a bidirectional morphological gradient operator, and the number of multi-scale decomposition layers is determined according to the response value of the bidirectional morphological gradient operator, and a convolution operation is performed on the morphological gradient operator using a Gaussian kernel function to construct a feature pyramid, and three-dimensional position data of each detection grid is collected from the feature pyramid; An edge extraction processing matrix is ​​established based on the three-dimensional position data and the surface morphology feature data, the detection grid is divided into a target area and a background area, the grayscale means of the target area and the background area are calculated, an initial segmentation threshold is determined according to the grayscale mean, and the dynamic segmentation threshold of the target area and the background area is calculated in an iterative manner with the initial segmentation threshold as the starting point. The dynamic segmentation threshold is applied to the three-dimensional position data to obtain initial edge contour points of the conductive material; curvature calculation is performed on the initial edge contour points, edge contour points with abnormal curvature are removed, and the remaining edge contour points are connected using cubic spline interpolation to obtain edge contour data of the conductive material.

3. The method according to claim 1, characterized in that Calculating the volume distribution density value of the conductive material in each detection grid according to the edge profile data, and constructing a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value includes: Extracting the three-dimensional coordinate value of each edge contour point in the detection grid from the edge contour data, inputting the three-dimensional coordinate value into a Gaussian filter to eliminate abnormal points, and extracting the difference between the maximum height value and the minimum height value in the detection grid as the local thickness value; Taking the center point of the detection grid as a reference, determine the intersection of the edge contour and the boundary of the detection grid, calculate the actual coverage area surrounded by the intersection line, and divide the product of the local thickness value and the actual coverage area by the nominal volume of the grid to obtain the volume distribution density value, wherein the nominal volume of the grid is the product of the plane projection area of ​​the detection grid and the standard height value; Calculate the average value and standard deviation of the local thickness values ​​of all detection grids respectively to obtain the local thickness mean value and the local thickness standard deviation, and output the ratio of the local thickness standard deviation to the local thickness mean value as the thickness value of the conductive material; The sum of the actual coverage areas of all detection grids is calculated to obtain the total actual coverage area, and the ratio of the total actual coverage area to the theoretical area of ​​the detection area is output as the area coverage coefficient. The autocorrelation coefficient of different spatial displacement distances is calculated for the volume distribution density value as the spatial distribution parameter. Based on the volume distribution density value, a three-dimensional spatial distribution model including thickness value, area coverage coefficient and spatial distribution parameter is constructed.

4. The method according to claim 1, characterized in that: Applying a predetermined voltage to a plurality of detection grids, using a four-probe array to measure and obtain a surface resistance measurement value, generating resistance distribution data of the detection grid according to the surface resistance measurement value, establishing a spatial correspondence between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional space distribution model, extracting the resistance value and the corresponding volume distribution density value of each grid position, and forming a resistance-density data pair including: Positive and negative electrode pairs are arranged at adjacent boundaries of a plurality of detection grids, and a predetermined alternating voltage with increasing amplitude and fixed phase difference is applied through the electrode pairs to form an initial electric field distribution of the detection grid, and the optimal predetermined voltage amplitude is determined according to the current response between the electrode pairs, and the four-probe array is moved along a spiral scanning path within the detection grid to which the predetermined voltage is applied, and the probe contact force signal and the probe displacement signal are collected in real time, and the contact stability coefficient between the probe and the surface of the conductive material is calculated based on the fluctuation amplitude and spectrum characteristics of the signal; The measured voltage amplitude and frequency of the outer probe pair of the four-probe array are dynamically adjusted according to the contact stability coefficient, so that the measured voltage and the predetermined voltage present orthogonal characteristics in the frequency domain, the voltage drop sequence and phase information of the inner probe pair are collected, and the response characteristics of the predetermined voltage and the measured voltage are separated by orthogonal filtering. The local equivalent circuit model of the conductive material considering the surface scattering effect is constructed in combination with the geometric parameters of the four-probe array, and the effective resistance component in the impedance characteristic is extracted as the surface resistance measurement value of the detection grid; The surface resistance measurement value is decomposed at multiple scales to extract the measurement noise characteristics, and the mapping relationship between the noise characteristics and the contact stability coefficient is established. The number of repeated measurements is adaptively determined, and the weighted average algorithm considering the signal stability is used to obtain the resistance distribution data of the detection grid. The resistance distribution data is associated with the two-dimensional measurement coordinates to construct a resistance distribution surface, and the local curvature and gradient characteristics of the surface are extracted. In the feature mutation area, measurement points are added according to the adaptive density, and iterative measurements are performed until the resistance change rate of adjacent measurement points meets the convergence condition, thereby obtaining an optimized high-precision resistance distribution surface. Based on the optimized resistance distribution surface, a registration algorithm of feature point matching is adopted to establish the mapping relationship between the two-dimensional measurement coordinate system and the three-dimensional spatial distribution model grid coordinate system. The equivalent resistance value of the grid position is calculated according to the mapping relationship. Combined with the volume distribution density value at the grid position, the initial resistance-density data pair considering the spatial correlation is constructed. Based on the spatial distribution characteristics and weight distribution of the initial resistance-density data pair, a measurement quality assessment model is established, the data reliability index is calculated, and a dynamic screening threshold adapted to the surface characteristics is adopted to finally output the resistance-density data pair that meets the quality requirements.

5. The method according to claim 1, characterized in that Generating a conductive uniformity quantitative index value according to the resistance change rate value and the difference coefficient value, identifying a conductive performance deviation area based on the conductive uniformity quantitative index value, and calculating a position coordinate value, a deviation amplitude, and a regional correlation coefficient of the conductive performance deviation area includes: According to the three-dimensional coordinates and volume parameters of the detection grid, a grid space mapping matrix is ​​established to calculate the grid shape characteristic value; the grid shape characteristic value is combined with the resistance change rate value and the density difference coefficient value to construct a grid feature fusion model; based on the grid feature fusion model, the resistance change rate value and the density difference coefficient value are normalized to obtain a grid feature normalized value; Establishing a spatial distance matrix of adjacent detection grids, constructing a distance weight calculation model based on the spatial distance matrix, inputting the grid feature normalization value into the distance weight calculation model to obtain the distance weighting factor between grids, and using a nonlinear weighting method to fuse the grid feature normalization value with the distance weighting factor to generate a quantitative index value of conductive uniformity; A dynamic scanning window is constructed at the detection grid position where the quantitative index value of the conductive uniformity exceeds the preset index threshold, and the index value gradient field within the dynamic scanning window is calculated; a search direction and a step length are determined based on the index value gradient field, and the dynamic scanning window is expanded until a gradient reversal point is obtained; and a connected area between the gradient reversal points is determined as a conductive performance deviation area; Extracting features of the detection grid within the conductive performance deviation area, using the conductive uniformity quantitative index value as a weight coefficient to establish a centroid calculation model; calculating the position coordinates of the conductive performance deviation area by an iterative optimization method; Calculate the difference between the maximum value of the index in the conductive performance deviation area and the average value of the index of the adjacent grids outside the area boundary as the deviation amplitude; The topological connection relationship of the detection grid in the conductive performance deviation area is analyzed, and a correlation calculation model is established in combination with the spatial distribution characteristics of the conductive uniformity quantitative index value. The conductive performance deviation area is quantitatively evaluated based on the correlation calculation model, and the regional correlation coefficient is output.

6. The method according to claim 5, characterized in that Extracting features of the detection grids within the conductive performance deviation area, and using the conductive uniformity quantitative index value as a weight coefficient to establish a centroid calculation model; Calculating the position coordinates of the conductive performance deviation area by an iterative optimization method includes: Extracting resistance value, size and position characteristic parameters of the detection grids in the conductive performance deviation area respectively, calculating the resistance gradient value based on the resistance value, and calculating the conductivity based on the resistance value and the size characteristic parameters; Normalizing the resistance value, resistance gradient value and conductivity of the detection grid to obtain a standardized characteristic value, and calculating a conductive uniformity quantitative index value based on the standardized characteristic value; The quantitative index value of the conductive uniformity is used as a weight coefficient, and a weighted centroid calculation model is established with the position characteristic parameters of the detection grid to calculate the initial centroid coordinates of the conductive performance deviation area; An optimization calculation region is established with the initial centroid coordinate as the center, and a conductive uniformity quantitative index value of the detection grid in the optimization calculation region is calculated based on the resistance value, the resistance gradient value and the conductivity, and the calculated conductive uniformity quantitative index value is substituted into the weighted centroid calculation model as a new weight coefficient to calculate a new centroid coordinate; Calculate the difference between the new center of mass coordinates and the initial center of mass coordinates. When the difference is greater than a preset coordinate difference threshold, use the new center of mass coordinates as the initial center of mass coordinates to repeatedly execute the steps of establishing an optimized calculation area until the difference is less than the preset coordinate difference threshold, and determine the last calculated center of mass coordinates as the position coordinates of the conductive performance deviation area.

7. The method according to claim 1, characterized in that Inputting the position coordinate value, deviation amplitude and regional correlation coefficient of the conductive performance deviation area into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameter, and generating the screen tension adjustment value and the scraper pressure compensation value according to the printing pressure distribution parameter includes: Inputting the position coordinate values, deviation amplitude and regional correlation coefficient of the conductive performance deviation area into the printing defect analysis unit of the process parameter optimization system, performing standardization processing based on the position coordinate values, and fitting the position coordinates using a bivariate Gaussian distribution model to obtain a spatial probability density function of the printing pressure distribution; wherein the mean of the spatial probability density function represents the pressure center offset of the printing pressure distribution parameters, and the covariance matrix represents the anisotropic characteristics of the printing pressure distribution parameters; establishing a weighted combination model of the pressure center offset and the anisotropic characteristics to obtain a pressure distribution feature quantity; A printing pressure compensation model is established based on the deviation amplitude, and a mapping function between the deviation amplitude and the pressure distribution adjustment amount is constructed by a piecewise linear interpolation method; an initial pressure adjustment amount of a conductive performance deviation area is calculated according to the mapping function, a topological relationship matrix is ​​constructed based on the regional correlation coefficient, and the initial pressure adjustment amount is spatially smoothed based on the topological relationship matrix to obtain a pressure compensation feature amount; Constructing a weight coefficient matrix based on pressure distribution feature quantity and a weight coefficient matrix based on pressure compensation feature quantity; multiplying the weight coefficient matrix with the corresponding feature quantity to obtain weighted feature quantity; performing superposition operation on the weighted feature quantity to obtain the overall printing pressure distribution parameter; Calculate the covariance matrix of the overall printing pressure distribution parameters, perform eigenvalue decomposition on the covariance matrix, obtain an eigenvalue sequence and an eigenvector sequence, sort the eigenvector sequence according to the size of the eigenvalue sequence, select the eigenvector whose cumulative contribution rate reaches a preset contribution threshold as the change vector of the printing pressure distribution, construct a nonlinear mapping equation between the change vector and the screen tension, introduce the stress-strain curve parameters of the screen material into the nonlinear mapping equation, and establish a screen tension target function; use the gradient descent method to optimize and solve the tension target function to obtain the optimal screen tension adjustment value; The overall printing pressure distribution parameters are projected and transformed in the length direction of the scraper to obtain a pressure distribution curve, and the pressure distribution curve is fitted with a piecewise polynomial to establish a pressure compensation reference equation. A deformation compensation term related to the elastic modulus of the scraper material and the scraper thickness is introduced into the pressure compensation reference equation to obtain a scraper pressure compensation equation. The scraper pressure compensation equation is combined with a motion trajectory equation including the scraper movement speed and the scraper inclination angle to calculate the scraper pressure compensation value.

8. A screen-printed conductive material uniformity detection system, used to implement the method according to any one of claims 1 to 7, characterized in that: include: The first unit is used to collect a conductive pattern image of a screen-printed conductive material product as an input image through an image acquisition device, process the input image using phase profilometry to obtain surface morphology feature data of the conductive pattern, divide a surface area of ​​the surface morphology feature data into a plurality of detection grids, collect three-dimensional position data of each detection grid, process the three-dimensional position data through an image segmentation algorithm to obtain edge contour data of the conductive material, calculate a volume distribution density value of the conductive material in each detection grid according to the edge contour data, and construct a three-dimensional spatial distribution model of the conductive material based on the volume distribution density value, wherein the three-dimensional spatial distribution model includes a thickness value, an area coverage factor, and a spatial distribution parameter of the conductive material; The second unit is used to apply a predetermined voltage to a plurality of detection grids, obtain a surface resistance measurement value by using a four-probe array to measure, generate resistance distribution data of the detection grid according to the surface resistance measurement value, establish a spatial correspondence between the measurement coordinates of the resistance distribution data and the grid coordinates of the three-dimensional space distribution model, extract the resistance value and the corresponding volume distribution density value of each grid position to form a resistance-density data pair, calculate the ratio of the difference between the resistance values ​​of adjacent detection grids and their average resistance value to obtain a resistance change rate value, calculate the ratio of the difference between the volume distribution density of adjacent detection grids and their average density value to obtain a difference coefficient value, generate a conductive uniformity quantitative index value according to the resistance change rate value and the difference coefficient value, identify a conductive performance deviation area based on the conductive uniformity quantitative index value, and calculate the position coordinate value, deviation amplitude and regional correlation coefficient of the conductive performance deviation area; The third unit is used to input the position coordinate value, deviation amplitude and area correlation coefficient of the conductive performance deviation area into the printing defect analysis unit of the process parameter optimization system to calculate the printing pressure distribution parameters, generate the screen tension adjustment value and the scraper pressure compensation value according to the printing pressure distribution parameters, input the screen tension adjustment value and the scraper pressure compensation value into the printing equipment control system for parameter setting, perform rework of the printed conductive material after the parameter setting is completed, repeat the detection and evaluation process for the conductive pattern after the rework, and when it is detected that the resistance change rate value of adjacent detection grids is less than a predetermined change rate threshold, it is confirmed that the uniformity of the conductive material meets the process requirements.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images