A visual guiding calibration and alignment method suitable for display panel bonding process

Abstract

A visual guiding calibration and alignment method suitable for display panel bonding process of the application comprises the following steps: arranging left upper camera, left lower camera, right upper camera and right lower camera respectively at positions corresponding to four corners of the display panel above the UVW alignment platform, using a checkerboard combined calibration board integrated with ArUco markers at four corners, keeping the reference position of the rotation axis of the UVW platform unchanged, driving the calibration board to rotate around the rotation axis by a preset angle, collecting images of the four corner regions of the screen to be bonded and the back shell respectively, and roughly positioning two initial edge lines corresponding to the corner points through template matching; converting the pixel coordinates of each corner point of the screen and the back shell to the global calibration board coordinate system through the corresponding partition affine transformation matrix; adjusting the pose of the platform to compensate for the alignment deviation, and performing the down bonding operation after the pose compensation is completed. The application quickly and conveniently realizes the alignment offset calculation and high-precision alignment bonding of the screen and the back shell.

Description

Technical Field

[0001] This invention relates to the field of visual guidance technology for display panels, and specifically to a visual guidance calibration and alignment method suitable for display panel bonding processes. Background Technology

[0002] With the rapid iteration of display technology towards larger sizes and higher resolutions, large-size display panels have been widely used in smart TVs, commercial displays, automotive displays, and other fields. As a core process in the production of large-size display panels, the lamination process directly determines the display effect, product yield, and lifespan of the panel, making it one of the key bottlenecks restricting the mass production of large-size display panels.

[0003] Currently, the visual guidance calibration and alignment technologies used in the bonding process of large-size display panel screens and back covers are mainly divided into two categories: one is a calibration method based on a single calibration board (such as a pure checkerboard), combined with conventional corner point positioning algorithms to achieve alignment; the other relies on multi-camera local calibration, achieving multi-point local alignment through complex coordinate stitching. However, the above-mentioned existing technologies have the following significant drawbacks in practical applications, making it difficult to meet the high-precision and high-efficiency bonding requirements of large-size display panels. The specific drawbacks are as follows: First, the single checkerboard calibration board is easily affected by lighting and surface reflection, resulting in a large error in corner point extraction and a large mapping deviation between the camera pixel coordinate system and the calibration board coordinate system. Secondly, the rotation center calibration accuracy is low and the stability is poor: During the bonding process of large-size display panels, the screen and the back shell need to be adjusted around a fixed rotation axis. The calibration accuracy of the rotation center directly affects the accuracy of the alignment compensation. Existing technologies mostly use local sampling to fit the rotation center without considering the global positional relationship, which is easily affected by local error interference, resulting in a large positioning deviation of the rotation center. At the same time, the reference position of the rotation axis is easily affected by equipment vibration and temperature changes and may shift. Existing methods cannot achieve dynamic correction of the rotation center, resulting in insufficient stability and thus causing rotational offset errors during the bonding process. Third, the bonding and alignment accuracy requirements for large-size display panels typically reach the sub-pixel level (≤0.1 pixels). However, existing corner positioning methods are mostly based on pixel-level extraction without sub-pixel-level optimization. They are also easily affected by factors such as panel edge noise, stains, and reflections, resulting in large corner positioning errors. They cannot accurately obtain the corner position relationship between the screen and the back cover, leading to deviations in the calculation of translation compensation and rotation compensation, which ultimately affects bonding accuracy and causes a decrease in product yield. Summary of the Invention

[0004] The present invention proposes a visual guidance calibration and alignment method suitable for display panel bonding process, which can at least solve one of the technical problems in the background art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A visually guided calibration and alignment method suitable for display panel bonding processes includes the following steps: S1. Place the upper left camera, lower left camera, upper right camera and lower right camera at the four corners of the display panel above the UVW alignment platform to establish a global calibration board coordinate system and an independent pixel coordinate system for each camera. S2. A checkerboard-style calibration board with four corner integrated ArUco markers is placed at the reference zero position of the UVW platform. Four cameras respectively acquire images of the corresponding small calibration units at the corners, extract the pixel coordinate point set of all corner points in each small calibration unit, and combine the physical size parameters of the calibration board to obtain the global calibration board coordinate point set corresponding to each corner point. Based on the pixel coordinate point set of each partition and the calibration board coordinate point set, the affine transformation matrices of the four partitions (upper left, lower left, upper right, and lower right) are solved respectively to establish the mapping relationship from the local pixel coordinate system to the global calibration board coordinate system. S3. Keeping the rotation axis reference position of the UVW platform unchanged, drive the calibration board to rotate around the rotation axis by a preset angle. The four cameras will again acquire images of the small calibration units at each corner to obtain the set of pixel coordinates of the rotated corner points. Through the affine transformation matrix of the corresponding partition, the set of pixel coordinates of the rotated points is converted into the set of coordinates of the rotated points in the global calibration board coordinate system. Traverse all matching coordinate point pairs before and after rotation and calculate the equation of the perpendicular bisector of each pair of coordinate points. Construct a weighted least squares optimization objective function and solve for the optimal intersection point of all perpendicular bisectors, which is used as the rotation center coordinate of the UVW platform. S4. Collect images of the four corner areas of the screen and back cover to be bonded, and roughly locate the two initial edge lines corresponding to the corner points by template matching; generate N equally spaced rectangular measuring calipers along each initial edge line, with the long side of the calipers parallel to the edge line and the short side along the edge normal direction; perform sub-pixel grayscale sampling along the normal direction in each caliper, and use a Gaussian first derivative edge model to perform nonlinear fitting on the grayscale profile to obtain the sub-pixel edge points at each caliper position; obtain the precise edge line based on the fitting of all sub-pixel edge points, and obtain the sub-pixel level pixel coordinates of each corner point of the screen and back cover from the intersection of the two edge lines at the same corner point; S5. Using the affine transformation matrix of the corresponding partition, the pixel coordinates of each corner point of the screen and the back shell are uniformly transformed to the global calibration board coordinate system; calculate the difference between the X and Y coordinates of the four corner points and take the average value to obtain the overall translation offset; fit the two diagonal lines connecting the screen and the back shell respectively, calculate the angle between the two sets of diagonal lines and take the average value to obtain the angle offset. S6. Feedback the translational offset and angular offset to the motion control system of the UVW platform to drive the platform to adjust its posture to compensate for the alignment deviation. After the posture compensation is completed, perform the pressing and bonding operation.

[0006] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein the combined calibration board mentioned in step S2 specifically includes: The main body is a large-size checkerboard calibration board, whose physical size matches the size of the display panel to be bonded. A small calibration unit is set at each of the four corners of the large calibration board: the upper left corner, the lower left corner, the upper right corner, and the lower right corner. Each small calibration unit integrates a unique ID ArUco mark and a local checkerboard pattern. The specified corner point of the small calibration unit at the upper left corner of the large calibration board is used as the origin of the global calibration board coordinate system. Based on the physical size of the large and small calibration boards, the actual coordinate values ​​of all corner points in the global calibration board coordinate system are pre-generated.

[0007] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein: the affine transformation matrix in step S2 is represented in homogeneous coordinate form, and the mapping relationship satisfies:

[0008] in , To determine the coordinates of the corner point calibration plate, , For the corner pixel coordinates, the affine transformation matrix M is used. , , , and , express, , , , For the parameters of the linear transformation, , These are the translation parameters.

[0009] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein: in step S3, for any pair of matching coordinate points before and after rotation... and The general form of the equation of the perpendicular line The parameter calculation method is as follows:

[0010]

[0011]

[0012] Iterate through all matching coordinate pairs in all four partitions to generate the corresponding set of perpendicular bisector equations.

[0013] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein: the weighted least squares optimization objective function in step S3 is:

[0014] in, Let be the coordinates of the rotation center to be solved. , , Let be the equation parameters of the i-th perpendicular bisector, and be the weights. By solving for the minimum value of the objective function, the globally optimal coordinates of the rotation center are obtained.

[0015] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein: in step S4, the method for generating the measuring calipers specifically comprises: Let the two endpoints of the initial edge line obtained by rough positioning be... and The coordinates of the center point of the i-th caliper are:

[0016] The rotation angle of each caliper is consistent with the initial edge line angle, satisfying the requirement that... All calipers are evenly spaced along the edge line, covering the effective edge area near the corners.

[0017] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein: in step S4, the expression of the Gaussian first derivative edge model is:

[0018] in, The coordinates are along the direction of the caliper normal. For the sub-pixel precise position to be solved, A The amplitude of the edge, The Gaussian scaling parameter characterizes the degree of edge blurring; here, it represents the degree of edge blurring and is related to the lens's depth of field and focal length. B Offset the background grayscale value; The model is derived using the nonlinear least squares method. By fitting the actual grayscale sampling data, the solution is obtained. Then, transform it to the image coordinate system to obtain the sub-pixel edge points corresponding to the caliper positions.

[0019] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, wherein: in step S5, the calculation rules for the translational offset and angular offset are as follows: Calculate the coordinate differences of the corresponding corner points of the screen and the back cover in the X and Y directions respectively. Take the average of the X-direction differences of the four corner points to obtain the X-direction translation offset, and take the average of the Y-direction differences of the four corner points to obtain the Y-direction translation offset. Connect the top-left and bottom-right corners of the screen and the bottom-left and top-right corners to obtain two diagonal lines. Similarly, obtain two diagonal lines for the back cover. Calculate the angle between the two sets of corresponding diagonal lines and take the average of the two angles as the final angle offset.

[0020] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, the calibration stage further includes a dynamic correction step for the rotation center. During the batch bonding production process, the rotation center calibration process in step S3 is repeated every set number of production batches to update the rotation center coordinates; or before each bonding, a single calibration feature point is used to check the rotation center deviation. When the deviation exceeds the preset threshold, a full recalibration is triggered to offset the rotation axis offset caused by equipment vibration and temperature drift.

[0021] As a preferred embodiment of the visual guidance calibration and alignment method applicable to display panel bonding process described in this invention, step S4 further includes an image reflection suppression preprocessing step before performing sub-pixel corner point positioning: The acquired panel images are processed using a homomorphic filtering algorithm to separate the illumination and reflection components of the image, suppress specular reflection and uneven illumination on the panel surface, and then perform subsequent edge coarse positioning and sub-pixel fitting operations.

[0022] The beneficial effects of this invention are: This invention can achieve high-precision coordinate transformation between pixel coordinate system and calibration board coordinate system, and accurately calculate the translation compensation and rotation compensation between the relative positions of the screen and the back cover, thereby quickly and conveniently realizing the alignment offset calculation and high-precision alignment and bonding of the screen and the back cover. Attached Figure Description

[0023] Figure 1 This is a visual alignment model diagram of a display panel, which is applicable to the visual guidance calibration and alignment method of the present invention for display panel bonding process.

[0024] Figure 2 This is an ArUco calibration board diagram for large-size alignment of the visual guidance calibration and alignment method applicable to display panel bonding processes according to the present invention.

[0025] Figure 3 This is a model diagram of the calibration plate coordinate system for the visual guidance calibration and alignment method applicable to the display panel bonding process of the present invention.

[0026] Figure 4 This is a schematic diagram illustrating the platform rotation center calibration principle of the visual guidance calibration and alignment method applicable to the display panel bonding process of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0028] like Figures 1-4 As shown in the figure, a visual guidance calibration and alignment method applicable to display panel bonding process in this embodiment specifically includes the following steps: S1. When performing visually guided alignment and bonding, the alignment platform must first be visually calibrated. Common models of visual platforms... Figure 1 As shown, it mainly includes the top left camera, bottom left camera, top right camera, and bottom right camera that capture images of the four corners of the back cover and screen, as well as the corresponding image coordinate system of the top left camera. , and the lower left camera image coordinate system , and the upper right camera image coordinate system , and the lower right camera image coordinate system , The screen is attached to the camera position using a suction cup, while the back cover is located on the UVW platform. The screen and back cover are moved into the camera's field of view in sequence to take pictures. S2. Mapping calibration between pixel coordinates and calibration board coordinates: This invention designs an ArUco calibration board based on a combination of ArUco markers and a checkerboard calibration board, as follows: Figure 2 As shown.

[0029] The design concept is that the small calibration plates at the four corners of the calibration plate are designed with a combination of ArUco markings and a checkerboard pattern, while the larger calibration plates in other areas are designed with a checkerboard pattern. The coordinate model of the calibration plate is defined as follows: Figure 3 As shown: In the diagram, LU, LD, RU, and RD represent the four smaller calibration plates located at the top left, bottom left, top right, and bottom right corners of the large calibration plate. These correspond to the four cameras in the alignment vision model. The calibration plate coordinate system is represented as follows: , The Id=0 identifier marks the location of each small calibration plate where id=0. Based on the specific dimensions of the calibration plate, the corner point of each small calibration plate with id=0 can be calculated. , The coordinates of the calibration plate in the coordinate system, such as Figure 3 Winning the bid; During actual calibration, the UVW platform was kept in the reference position (the initial image position was not rotated). Four cameras respectively captured four small calibration patterns: LU, LD, RU, and RD. Based on the initial corner coordinates of id=0, an algorithm was developed to calculate the calibration board coordinate point set corresponding to the other corners. , , , The algorithm obtains the pixel coordinates of each small calibration board corner point. , , , Based on the pixel coordinates and calibration board coordinate point set data, and using the following model formula, calculate the affine transformation matrices at the four positions LU, LD, RU, and RD. , , , .

[0030]

[0031] in , To determine the coordinates of the corner point calibration plate, , For the corner pixel coordinates, the affine transformation matrix M is used. , , , and , express, , , , The parameters are those for the linear transformation. , These are the translation parameters.

[0032] S3. Platform rotation center calibration based on global positional relationships: S31. Coordinate Data Acquisition: Based on step S2, the ArUco calibration board is rotated a certain angle on the UVW platform, and images of the four positions LU, LD, RU, and RD are captured again using four cameras to obtain the pixel coordinate set corresponding to the corner points. , , , , S32. Based on the affine transformation matrix calculated in step S2 , , , Calculate the coordinate point set of the rotated corner calibration plate. , , ,

[0033] S33. Construct the perpendicular bisector: Iterate through all the matching coordinate pairs in all the collection groups. For each pair of matching coordinates ( , ),in The coordinates of the calibration plate before rotation, To determine the coordinates of the rotated calibration plate, calculate the perpendicular line according to the following model. .

[0034]

[0035]

[0036]

[0037] Based on the above calculation method, calculate the perpendicular bisector of all pairs of coordinate points before and after rotation.

[0038] S34. Solving for the optimal intersection point: The general equation of the i-th perpendicular bisector is defined as follows:

[0039] Based on the model for calculating the perpendicular bisector in S33, point To the The normalized distance of the perpendicular bisectors is expressed as:

[0040] The weighted least squares objective function is constructed as follows:

[0041] Among them, weight

[0042] The objective function is to find the point that minimizes the sum of the squared weighted distances of all perpendicular bisectors, which is the optimal center of rotation estimate.

[0043] The diagram shows... Figure 4 The figure shows the coordinates of an arbitrary calibration plate for each region. and the coordinates after rotation and calibration plate coordinates and the coordinates after rotation Taking this as an example, each region involved in the calculation contains multiple pairs of coordinate points before and after rotation, and the optimal rotation center coordinates are finally solved based on the least squares method.

[0044] S4. Based on the above calibration parameters, during visually guided alignment and bonding, it is necessary to use four cameras to acquire images of the four corner areas of the screen and the back cover, and obtain the coordinate system of the screen and back cover corner calibration plate. , Based on the coordinates below, this invention proposes a sub-pixel precision corner positioning method, as detailed below: S41, Obtaining Corner Edge Lines To address interference issues at screen edges in specific scenarios, this invention does not perform a full-image edge search. Instead, based on the screen's approximate location and orientation, it dynamically generates a set of structured "measuring calipers." Initial edge lines corresponding to corner points are roughly located using template matching. N equally spaced measuring calipers are then generated along the normal direction. Each caliper represents a rectangular region centered on line L, with its long side parallel to L and its short side parallel to the normal. The coordinates of the center point of the i-th caliper are... Represented as

[0045] The rotation angle of each caliper is the angle of the initial edge line:

[0046] S42. Sub-pixel edge localization algorithm based on Gaussian derivative model: Within each measuring caliper, the image grayscale values ​​are sampled to form a grayscale profile, which is then fitted using a continuous edge model.

[0047] For each caliper, grayscale values ​​are sampled along its shorter side (i.e., the normal direction) to obtain a one-dimensional grayscale signal g(x), where It is a continuous coordinate along the short side of the caliper (sub-pixel level sampling is achieved through bicubic interpolation).

[0048] This invention uses the first derivative of a Gaussian function as a model to fit this smooth edge. The model function... Represented as:

[0049] in, For the sub-pixel precise position to be solved, A The amplitude (contrast) of the edge. For Gauss The value here represents the degree of edge blur, which is related to the lens's depth of field and focal length.B This is the background grayscale value offset.

[0050] The model is derived using the nonlinear least squares method. The model is fitted to the actual acquired grayscale data g(x). The optimal parameters are solved by minimizing the sum of squared residuals S. .

[0051] The solution obtained This refers to the sub-pixel coordinates of the edge at the caliper's position. The coordinates of this normal direction... Transforming back to the image coordinate system yields a sub-pixel edge point. .

[0052] Subpixel edge points found by all N calipers { , , ..., The data are then combined to perform a final overall straight-line fitting.

[0053] Finally, based on the above method, the edge lines of the four corners of the screen and back cover are obtained. The intersection of the two edge lines at each corner is then calculated to obtain the precise pixel coordinates of the screen's corner point. Following the algorithm mentioned above, the pixel coordinates of the top-left, bottom-left, top-right, and bottom-right edge corners of the screen are obtained sequentially. , , , And the pixel coordinates of the top left and bottom right corners of the back cover. , , , .

[0054] S43, Coordinate Transformation Based on the pixel coordinate system to calibration board coordinate system mapping calibration matrix calculated in step S2 , , , screen pixel coordinates , , , and back shell pixel coordinate conversion , , , Convert to calibration plate coordinates , , , and , , , .

[0055] S5, Offset Calculation Calculate the relative coordinate differences based on the coordinates of the calibration board corresponding to the top left, bottom left, top right, and bottom right corners of the screen and back cover, and obtain the offset in the X direction of the four corners. , , , and the offset of the four corners in the Y direction , , , Then, calculate their average to obtain the final precise offset in the X direction. , Then, generate straight lines at the top left and bottom right corners of the screen respectively. A straight line is generated from the bottom left and top right diagonal points. Similarly, obtain the diagonal line of the posterior shell. , ,calculate , The angle between intersecting straight lines and , The angle between intersecting straight lines Finally, calculate and The average value is the angular offset. .

[0056] S6, Alignment and Fitting Based on offset , , The feedback is sent to the mechanical axis for posture adjustment to compensate for the relative posture of the screen or back cover. After compensation, the screen and back cover are pressed down to complete the alignment and bonding.

[0057] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0058] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A visual guidance calibration and alignment method suitable for display panel bonding processes, characterized in that, Includes the following steps: S1. Place the upper left camera, lower left camera, upper right camera and lower right camera at the four corners of the display panel above the UVW alignment platform to establish a global calibration board coordinate system and an independent pixel coordinate system for each camera. S2. A checkerboard-style calibration board with four corner integrated ArUco markers is placed at the reference zero position of the UVW platform. Four cameras respectively acquire images of the corresponding small calibration units at the corners, extract the pixel coordinate point set of all corner points in each small calibration unit, and combine the physical size parameters of the calibration board to obtain the global calibration board coordinate point set corresponding to each corner point. Based on the pixel coordinate point set of each partition and the calibration board coordinate point set, the affine transformation matrices of the four partitions (upper left, lower left, upper right, and lower right) are solved respectively to establish the mapping relationship from the local pixel coordinate system to the global calibration board coordinate system. S3. Keeping the rotation axis reference position of the UVW platform unchanged, drive the calibration board to rotate around the rotation axis by a preset angle. The four cameras will again acquire images of the small calibration units at each corner to obtain the set of pixel coordinates of the rotated corner points. Through the affine transformation matrix of the corresponding partition, the set of pixel coordinates of the rotated points is converted into the set of coordinates of the rotated points in the global calibration board coordinate system. Traverse all matching coordinate point pairs before and after rotation and calculate the equation of the perpendicular bisector of each pair of coordinate points. Construct a weighted least squares optimization objective function and solve for the optimal intersection point of all perpendicular bisectors, which is used as the rotation center coordinate of the UVW platform. S4. Collect images of the four corner areas of the screen and back cover to be bonded, and roughly locate the two initial edge lines corresponding to the corner points by template matching. N equally spaced rectangular measuring calipers are generated along each initial edge line. The long side of the caliper is parallel to the edge line, and the short side is along the edge normal direction. Sub-pixel grayscale sampling is performed in each caliper along the normal direction. The grayscale profile is nonlinearly fitted using a Gaussian first derivative edge model to solve for the sub-pixel edge points at each caliper position. Precise edge lines are obtained by fitting all sub-pixel edge points. Sub-pixel level pixel coordinates of each corner point of the screen and the back shell are obtained from the intersection of two edge lines at the same corner point. S5. Using the affine transformation matrix of the corresponding partition, the pixel coordinates of each corner point of the screen and the back shell are uniformly transformed to the global calibration board coordinate system; calculate the difference between the X and Y coordinates of the four corner points and take the average value to obtain the overall translation offset; fit the two diagonal lines connecting the screen and the back shell respectively, calculate the angle between the two sets of diagonal lines and take the average value to obtain the angle offset. S6. Feedback the translational offset and angular offset to the motion control system of the UVW platform to drive the platform to adjust its posture to compensate for the alignment deviation. After the posture compensation is completed, perform the pressing and bonding operation.

2. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: The combined calibration board mentioned in step S2 specifically includes: The main body is a large-size checkerboard calibration board, whose physical size matches the size of the display panel to be bonded. A small calibration unit is set at each of the four corners of the large calibration board: the upper left corner, the lower left corner, the upper right corner, and the lower right corner. Each small calibration unit integrates a unique ID ArUco mark and a local checkerboard pattern. The specified corner point of the small calibration unit at the upper left corner of the large calibration board is used as the origin of the global calibration board coordinate system. Based on the physical size of the large and small calibration boards, the actual coordinate values ​​of all corner points in the global calibration board coordinate system are pre-generated.

3. The visual guidance calibration and alignment method for display panel bonding process according to claim 2, characterized in that: The affine transformation matrix described in step S2 is represented in homogeneous coordinates, and the mapping relationship satisfies: in , To determine the coordinates of the corner point calibration plate, , For the corner pixel coordinates, the affine transformation matrix M is used. , , , and , express, , , , For the parameters of the linear transformation, , These are the translation parameters.

4. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: In step S3, for any pair of matching coordinate points before and after rotation... and The general form of the equation of the perpendicular line The parameter calculation method is as follows: Iterate through all matching coordinate pairs in all four partitions to generate the corresponding set of perpendicular bisector equations.

5. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: The weighted least squares optimization objective function described in step S3 is: in, Let be the coordinates of the rotation center to be solved. , , Let be the equation parameters of the i-th perpendicular bisector, and be the weights. By solving for the minimum value of the objective function, the globally optimal coordinates of the rotation center are obtained.

6. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: In step S4, the method for generating the measuring caliper is specifically as follows: Let the two endpoints of the initial edge line obtained by rough positioning be... and The coordinates of the center point of the i-th caliper are: The rotation angle of each caliper is consistent with the initial edge line angle, satisfying the requirement that... All calipers are evenly spaced along the edge line, covering the effective edge area near the corners.

7. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: In step S4, the expression for the Gaussian first derivative edge model is: in, The coordinates are along the direction of the caliper normal. For the sub-pixel precise position to be solved, A The amplitude of the edge, The Gaussian scaling parameter characterizes the degree of edge blurring; here, it represents the degree of edge blurring and is related to the lens's depth of field and focal length. B Offset the background grayscale value; The model is derived using the nonlinear least squares method. By fitting the actual grayscale sampling data, the solution is obtained. Then, transform it to the image coordinate system to obtain the sub-pixel edge points corresponding to the caliper positions.

8. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: In step S5, the calculation rules for the translational offset and angular offset are as follows: Calculate the coordinate differences of the corresponding corner points of the screen and the back cover in the X and Y directions respectively. Take the average of the X-direction differences of the four corner points to obtain the X-direction translation offset, and take the average of the Y-direction differences of the four corner points to obtain the Y-direction translation offset. Connect the top-left and bottom-right corners of the screen and the bottom-left and top-right corners to obtain two diagonal lines. Similarly, obtain two diagonal lines for the back cover. Calculate the angle between the two sets of corresponding diagonal lines and take the average of the two angles as the final angle offset.

9. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: The calibration phase also includes a dynamic correction step for the rotation center: During the batch bonding production process, the rotation center calibration process in step S3 is repeated every set number of production batches to update the rotation center coordinates; or before each bonding, a single calibration feature point is used to check the rotation center deviation. When the deviation exceeds the preset threshold, a full recalibration is triggered to offset the rotation axis offset caused by equipment vibration and temperature drift.

10. The visual guidance calibration and alignment method for display panel bonding process according to claim 1, characterized in that: Step S4, before sub-pixel corner location, also includes an image reflection suppression preprocessing step: The acquired panel images are processed using a homomorphic filtering algorithm to separate the illumination and reflection components of the image, suppress specular reflection and uneven illumination on the panel surface, and then perform subsequent edge coarse positioning and sub-pixel fitting operations.

Images