A concentricity detection system, method, apparatus, and medium

Abstract

The present application relates to the technical field of metrological detection, and more particularly to a concentricity detection system, method, device and medium, which is used for the concentricity detection of the OCA film lamination process completed by a laminator, and comprises: a detection module, including a vision unit and a detection unit, the vision unit is used for acquiring image information of a corresponding area of a laminated product when a light source module provides different bright and dark light sources at different positions; a light source module, used for providing different bright and dark light sources at different positions; an adjusting module, used for adjusting the different positions of the light source module; by directly acquiring the light source direction to eliminate refraction error, dynamically adjusting the light source position to compensate for the difference in curved surface reflection, and based on a multi-angle image sequence to realize high-precision curved surface reconstruction, the concentricity detection precision of transparent film and curved surface substrate is effectively improved, and the detection needs of laminated products with different curvature shapes are adapted, and the technical problem of low precision when the detection system in the prior art detects the concentricity of the curved surface is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of metrological detection, and in particular to a concentricity detection system, method, device and medium. BACKGROUND

[0002] The current industrial detection field mainly relies on traditional photometric stereo vision technology for measuring the concentricity of curved surfaces. This method has two significant defects:

[0003] Firstly, the direction estimation needs to be assisted by a calibration ball in the light source calibration link. The light source direction must be calculated by reflecting the light spot on the ceramic ball. This indirect measurement method will produce refraction errors due to the transparent material in the OCA film bonding scene.

[0004] Secondly, the existing system generally uses a fixed light source array, which limits the range of changes in the angle of the light source.

[0005] When the product to be bonded has a curved surface, the difference in reflectivity of the curved surface increases due to the change in the angle of the curved surface, which may cause local overexposure or underexposure, seriously affecting the accuracy of the normal vector estimation. These limitations make the detection accuracy of the existing system relatively low when detecting the concentricity of transparent films and curved substrates. SUMMARY

[0006] The main purpose of the present application is to provide a concentricity detection system, method, device and medium, which aims to solve the technical problem of low accuracy of the existing detection system when detecting the concentricity of curved surfaces.

[0007] To achieve the above-mentioned purpose, the present application provides a concentricity detection system, which is used for detecting the concentricity of the OCA film bonding process completed by a bonding machine. The bonding machine includes a bonding module for completing the bonding process of the bonding film on the product to be bonded. The system comprises:

[0008] A light source module for providing different bright and dark light sources at different positions;

[0009] A detection module including a vision unit and a detection unit, wherein the vision unit is used to acquire image information of the corresponding area of the product to be bonded when the light source module provides different bright and dark light sources at different positions;

[0010] An adjustment module arranged in the bonding machine and used for adjusting the different positions of the light source module;

[0011] The detection unit comprises:

[0012] An acquisition subunit for acquiring the light source direction of the image information at different positions;

[0013] An estimation subunit is configured to estimate the surface normal vector of the product to be pasted, so as to interpolate and fit the surface normal vector;

[0014] A processing subunit is configured to solve the depth information of the boundary surface of the product to be pasted;

[0015] A detection subunit is configured to restore the surface curvature of the product to be pasted and detect the concentricity with the pasting film.

[0016] Optionally, the adjusting module comprises an adjusting base and an adjusting unit, the adjusting base is fixedly connected with the inner wall of the pasting machine, the first driving unit is arranged in the adjusting base, the output end of the first driving unit is connected with the adjusting unit, so that the adjusting unit rotates circumferentially along the output end of the first driving unit, and the light source module is arranged on the adjusting unit.

[0017] Optionally, the adjusting unit comprises an adjusting inner rod and an adjusting outer rod, one end of the adjusting outer rod is connected with the output end of the first driving unit, and the other end of the adjusting outer rod is rotationally arranged with the adjusting inner rod, and the light source module is arranged on the adjusting inner rod.

[0018] Optionally, a second driving unit is arranged at the connecting part of the adjusting outer rod and the adjusting inner rod, so that the adjusting inner rod rotates circumferentially along the output end of the second driving unit.

[0019] Optionally, a slide rail is arranged on the side of the adjusting inner rod close to the pasting module, and the light source module is arranged on the slide rail, so that the light source module moves along the slide rail.

[0020] Optionally, the number of the adjusting modules is 2, and the adjusting modules are arranged in axial symmetry based on the axis of the pasting module.

[0021] Optionally, the adjusting inner rod and the adjusting outer rod are quarter-fan ring structures with different diameters.

[0022] A concentricity detection method, the method comprises the following steps:

[0023] The adjusting module is controlled to rotate, so that the light source module passes through different positions when the adjusting module rotates outside the product to be pasted;

[0024] Image information in the corresponding area of the product to be pasted when the light source module provides different light sources in different positions is acquired;

[0025] The image information is received;

[0026] The direction of the light source of the image information in different positions is acquired;

[0027] The surface normal vector of the product to be pasted is estimated, so as to interpolate and fit the surface normal vector;

[0028] Solving depth information of a product with a pasted product boundary surface;

[0029] Restoring the surface curvature of the pasted product and detecting the concentricity with the pasting film.

[0030] A computer device includes a memory and a processor, the memory has a computer program stored therein, and the processor executes the computer program.

[0031] A computer readable storage medium has a computer program stored thereon, and a processor executes the computer program.

[0032] The beneficial effects that can be achieved by the present application are as follows:

[0033] The present application directly obtains the light source direction by adjusting the mechanical displacement of the module, avoids the refraction interference of transparent materials, expands the light angle coverage range by using a movable light source module, effectively suppresses the overexposure or underexposure phenomenon caused by the difference in curved surface reflection, synchronizes the light source adjustment with the image acquisition process in the traditional photometric stereo algorithm, and automatically associates the light source direction through mechanical motion parameters;

[0034] In addition, the present application eliminates refraction errors by directly obtaining the light source direction, dynamically adjusts the light source position to compensate for the difference in curved surface reflection, realizes high-precision curved surface reconstruction based on multi-angle image sequences, effectively improves the concentricity detection precision of transparent films and curved surface substrates, adapts to the detection needs of products with different curvature shapes, and solves the technical problem of low precision in the concentricity detection of the curved surface by the detection system in the prior art, and realizes high-precision lamination of OCA film before and after the process. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or prior art description will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0036] Figure 1 It is a structure schematic view of the lamination module and the adjustment module in embodiment 1 of the present application;

[0037] Figure 2 It is a structure schematic view of the adjustment module in embodiment 1 of the present application;

[0038] Figure 3 It is a flowchart of the method in embodiment 2 of the present application.

[0039] Reference signs:

[0040] 1 - fitting module, 2 - adjustment module, 3 - light source module, 4 - product to be fitted;

[0041] 21 - adjustment seat, 22 - adjustment unit;

[0042] 221 - adjustment inner rod, 222 - adjustment outer rod, 223 - second driving unit, 224 - slide rail.

[0043] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0045] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship, movement condition, etc. between components in a certain posture, and if the certain posture changes, the directional indications will also change accordingly.

[0046] In the present application, unless otherwise explicitly specified and limited, the terms "connection", "fixation" and the like should be understood in a broad sense, for example, "connection" can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be direct connection, or indirect connection through intermediate medium; can be internal connection of two elements or interaction relationship between two elements, unless otherwise explicitly limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0047] In addition, if the description of "first", "second" and the like is involved in the embodiments of the present application, the description of "first", "second" and the like is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the meaning of "and / or" appearing throughout the text includes three parallel schemes. For example, "A and / or B" includes A scheme, or B scheme, or A and B scheme. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the realization of the ordinary skilled in the art. When the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope required by the present application.

[0048] Embodiment 1

[0049] Referring to the drawings Figure 1 With the drawings Figure 2 The embodiment provides a concentricity detection system, which is used for detecting the concentricity of an OCA film in a lamination process of a laminator, the laminator comprising a lamination module 1 used for completing the lamination process of the lamination film on a laminated product 4, and the system comprising:

[0050] A light source module 3 used for providing different bright and dark light sources at different positions;

[0051] A detection module comprising a vision unit and a detection unit, wherein the vision unit is used for acquiring image information of a corresponding area of the laminated product 4 when the light source module 3 provides different bright and dark light sources at different positions;

[0052] An adjustment module 2 arranged in the laminator and used for adjusting the different positions of the light source module 3;

[0053] The detection unit comprises:

[0054] An acquisition subunit used for acquiring the light source direction of the image information at different positions;

[0055] An estimation subunit used for estimating a surface normal vector of the laminated product 4 and performing interpolation fitting on the surface normal vector;

[0056] A processing subunit used for solving depth information of a boundary curved surface of the laminated product 4;

[0057] A detection subunit used for restoring the surface curvature of the laminated product 4 and detecting the concentricity with the lamination film.

[0058] It should be noted that the detection module in the embodiment refers to a composite module composed of an image acquisition device (vision unit) and a data processing unit (detection unit), preferably an industrial camera and an image processor, and the image information under multiple light source angles is acquired through the vision unit.

[0059] In some embodiments, the light source module 3 refers to an optical emission device with adjustable position, which can be implemented by using an LED array in combination with a mechanical adjustment mechanism to generate light and dark changes by changing the incident angle of the light source.

[0060] In some embodiments, the adjustment module 2 refers to a driving device for controlling the spatial position of the light source, which can be implemented by using a servo motor and a transmission mechanism to change the relative position of the light source and the measured curved surface through mechanical movement.

[0061] In some embodiments, the acquisition subunit refers to a light source direction analysis unit, which can be implemented by using a light path tracking algorithm / model to calculate the incident direction based on the principle of Lambertian reflection model according to the physical position of the light source.

[0062] In some embodiments, the estimation subunit can be implemented by using a photometric stereo algorithm to derive the normal vector through the light intensity changes of multiple images.

[0063] In some embodiments, the processing subunit refers to a three-dimensional curved surface reconstruction unit, which can be implemented by using a Poisson equation solving algorithm to calculate the depth information through normal vector integration.

[0064] In some embodiments, the detection subunit refers to a concentricity analysis unit, which can be implemented by using a point cloud registration algorithm to calculate the fitting deviation through curved surface topography comparison.

[0065] It should also be noted that the OCA film fitting process is briefly described here. Before the OCA film fitting process, the polarizing sheet is the core optical element of the LCD module, and its absorption axis direction must be accurately aligned. When attached to the glass substrate, its edge (or pre-set Mark point) must be accurately aligned with the reference mark on the substrate. For the OCA film fitting process itself, the OCA serves as the bonding layer, and its fitting accuracy directly determines the relative position between the cover glass and the underlying display module. After the OCA film fitting, the LCM module with completed OCA fitting needs to be finally assembled with the backlight module, FPC (flexible circuit board), middle frame / scaffold, etc., and the mechanical reference and electrical connection position of the overall assembly need to be ensured to be accurate.

[0066] It can be understood that in the above OCA film bonding process, the adjusting module 2 drives the light source module 3 to move along the predetermined trajectory, so that the surface of the measured product sequentially receives light of different angles, the visual unit synchronously collects the image information sequence under each position, the estimation subunit calculates the light source direction based on the physical displacement of the light source module 3, eliminates the refraction error caused by the traditional calibration ball, estimates the image gray scale change under multi-angle illumination, establishes the mapping relationship between the surface normal vector and the illumination direction, and uses cubic spline interpolation to compensate for the data loss of discrete sampling points; the processing subunit converts the discrete normal vector field into a continuous curved surface function, and solves the depth distribution of the boundary surface through numerical integration; the detection subunit spatially matches the reconstructed curved surface model with the standard bonding film contour, and calculates the offset of the central axes of the two as the concentricity index.

[0067] In some embodiments, the synchronously collected image information sequence is not less than 3, and in other embodiments, the synchronously collected image information sequence can be understood as multiple images of different brightness of the light source module 3 under the same position condition, which is especially suitable for the same bonding batch; it can also be understood as multiple images collected under the same brightness or different brightness conditions after changing the angle of the light source module 3 under the same position condition, which is especially suitable for bonding processes with increased detection accuracy. Similarly, the light source module 3 in this case can also collect images under different position conditions to further improve the detection accuracy, which is especially suitable for bonding detection operations after changing the product batch. It can be understood that the above settings of the position and brightness of the light source module 3 are only a preferred choice, and the specific position and brightness of the light source module 3 during image collection can be macroscopically adjusted, that is, different brightness light sources are provided at different positions.

[0068] In some embodiments, the bonding machine can be understood as an automatic device for realizing precise pressing between the layers of the product to be bonded 4.

[0069] In some embodiments, the bonding film can be understood as an OCA film, a polarizing sheet, and a cover glass, etc.

[0070] In some embodiments, the bonding module 1 can be understood as an automatic structure with a bonding press head.

[0071] In some embodiments, the product to be bonded 4 can be understood as an electronic product, preferably including display products such as mobile devices, tablets, and car display screens that can have curved screens.

[0072] In some embodiments, the concentricity detection in the present embodiment can be understood as the detection / comparison process of the center and boundary of the attached product 4 and the center and boundary of the attached film during the attachment process of the attached film. Similarly, it can also be understood as the detection of the concentricity / matching degree of the attached film on the attachment surface after the attached film completes the attachment.

[0073] It can also be understood that the traditional method relies on the indirect calculation of the light source direction by the calibration ball. The technical solution in the present embodiment directly obtains the light source direction by adjusting the mechanical displacement of the adjustment module 2, avoids the refraction interference of transparent materials, expands the light angle coverage range by using the movable light source module 3, effectively suppresses the overexposure or underexposure phenomenon caused by the difference in curved surface reflection, and synchronizes the light source adjustment with the image acquisition process. The traditional photometric stereo algorithm needs to be separately calibrated, and the present solution automatically associates the light source direction through the mechanical motion parameters.

[0074] In addition, the technical solution in the present embodiment can directly obtain the light source direction to eliminate refraction errors, dynamically adjust the light source position to compensate for the difference in curved surface reflection, realize high-precision curved surface reconstruction based on multi-angle image sequences, effectively improve the concentricity detection precision of transparent films and curved surface substrates, adapt to the detection needs of different curvature shapes of the attached product 4, and solve the technical problem of low precision in the concentricity detection of the curved surface by the detection system in the prior art. The present embodiment realizes high-precision attachment of the OCA film before and after the process.

[0075] In some embodiments, the light source module 3 is connected to the slide rail 224 through a rotating platform, i.e., the light source module 3 can rotate relatively. It can be understood that the rotating platform preferably includes a slide block structure, the slide block cooperates with the slide rail 224, a motor is arranged in the slide block, the output end of the motor is connected to the light source module 3 through a reducer or the like, and the motor is also signal-connected to the control module of the detection system to control the angle adjustment thereof through the control module.

[0076] In some embodiments, the control module of the detection system serves as the core processing unit between modules to realize data acquisition and synchronous control. In addition, it also considers data processing and motion control. Similarly, the control module is also signal-connected to the interaction module, and its preferred model includes: NI cRIO-9045, Beckhoff CX2040, or Siemens S7-1200, etc. The specific control logic will not be described here.

[0077] In the present embodiment, the adjustment module 2 includes an adjustment seat 21 and an adjustment unit 22, the adjustment seat 21 is fixedly connected to the inner wall of the attachment machine, a first driving unit is arranged in the adjustment seat 21, the output end of the first driving unit is connected to the adjustment unit 22, so that the adjustment unit 22 rotates circumferentially along the output end of the first driving unit, and the light source module 3 is arranged on the adjustment unit 22.

[0078] It should be noted that the adjusting seat 21 refers to a support structure fixedly connected with the inner wall of the laminator, which can be implemented by a metal frame structure, and is used to provide a stable mounting base for the adjusting unit 22; wherein the first driving unit refers to a power device arranged inside the adjusting seat 21, which can be implemented by a servo motor or a stepping motor, and transmits rotary power through an output shaft to drive the adjusting unit 22 to move. Preferably, the output shaft is also preferably connected with the adjusting unit 22 through a reducer or the like structure to optimize the transmission torque. The adjusting unit 22 refers to a mechanical assembly connected with the output end of the first driving unit, which is preferably a rod-shaped structure with an arc, and is used to convert rotary motion into a circular motion trajectory of the light source module 3.

[0079] It should also be noted that the adjusting seat 21 is rigidly supported by being fixedly connected with the inner wall of the laminator, so as to ensure that the light source module 3 remains stable during movement. The first driving unit drives the adjusting unit 22 to rotate around its axis through the output shaft, so that the light source module 3 moves along a circular path, thereby dynamically adjusting the position of the light source relative to the product 4 to be laminated. When the light source module 3 rotates with the adjusting unit 22, it can cover a wider angle range, avoiding the problem of limited illumination angle caused by fixed light source arrays. Through the combination of mechanical transmission and driving control, the position of the light source module 3 can be continuously adjusted according to detection requirements, providing uniform and multi-angle light source conditions for different curved surface regions.

[0080] Based on the above structure, the active position adjustment of the light source module 3 is realized through the rotatable adjusting unit 22, breaking through the limitation of static layout of fixed light sources, dynamically expanding the range of light source angle change, thereby reducing the phenomenon of local overexposure or underexposure; also, the continuous adjustment of the light source position is realized through mechanical driving, expanding the coverage range of the light source angle, improving the exposure abnormality problem caused by the difference in reflectivity of the curved surface, providing more complete illumination data for subsequent image processing, and thereby improving the accuracy of normal vector estimation and concentricity detection accuracy.

[0081] In some embodiments, the light source module 3 is in the process of laminating the same batch of laminated films, and the differences between the laminated films and the product 4 to be laminated in the same batch are relatively small. The distance of position movement of the adjusting module 2 is also relatively small, and the subsequent detection processing can be performed by different light sources at the same position. Similarly, the detection system in this embodiment is particularly suitable for high-precision detection and positioning operation after product batch replacement.

[0082] In this embodiment, the adjusting unit 22 includes an adjusting inner rod 221 and an adjusting outer rod 222. One end of the adjusting outer rod 222 is connected with the output end of the first driving unit, and the other end of the adjusting outer rod 222 is rotationally arranged with the adjusting inner rod 221. The light source module 3 is arranged on the adjusting inner rod 221.

[0083] It should be noted that the adjusting outer rod 222 refers to a rigid rod-shaped component connected with the output end of the driving unit, which can be realized by a hollow rod body made of aluminum alloy, for transmitting rotary power and forming a main rotary shaft. The adjusting inner rod 221 refers to a rod-shaped component rotationally connected with the end of the adjusting outer rod 222, which can be made of carbon fiber composite material, for bearing the light source module 3 and forming a secondary rotary shaft. The rotationally arranged refers to the relative rotation between the two components through a bearing or hinge structure, which can be realized by a micro servo motor driving gear transmission mechanism, so that the adjusting inner rod 221 can be adjusted in angle in three-dimensional space around the connection point.

[0084] It should also be noted that the adjusting outer rod 222 rotates around the axis of the fitting module 1 under the drive of the first driving unit, driving the adjusting inner rod 221 to move along the circumferential trajectory as a whole, and the rotationally connected adjusting inner rod 221 and adjusting outer rod 222 allows the light source module 3 to adjust the angle in the direction perpendicular to the circumferential trajectory plane, forming a spatial composite motion trajectory; the light source module 3 is installed at the end of the adjusting inner rod 221, and the incident angle changes in three-dimensional space with the double-rod linkage structure, which can cover a large range of incident angles in the normal direction of the curved surface of the product 4, avoiding abnormal reflectivity caused by too large or too small incident angle.

[0085] Compared with the prior art, the traditional fixed light source array can only provide limited discrete angles of incident light, and cannot adapt to the continuously changing normal direction of the curved surface. The technical solution claimed in the embodiment forms a continuously adjustable light source incident angle through a double-rod linkage structure, so that the light source module 3 can move continuously or intermittently in three-dimensional space, and the light source module 3 has the ability to adjust the pitch angle, effectively eliminating the curved surface detection blind area.

[0086] In addition, the light source incident angle can dynamically match the change of the normal direction of the curved surface, eliminating the local overexposure or underexposure phenomenon caused by the mismatch between the fixed light source angle and the curved surface reflection characteristics. The light source module 3 forms a composite motion trajectory under the drive of the double shaft, ensuring that each region of the curved surface obtains uniform and moderate light intensity, providing high signal-to-noise ratio image data for subsequent normal vector estimation, thereby improving the precision of the concentricity detection of the transparent film and the curved surface substrate.

[0087] In the embodiment, a second driving unit 223 is arranged at the connection position of the adjusting outer rod 222 and the adjusting inner rod 221, so that the adjusting inner rod 221 rotates around the output end of the second driving unit 223.

[0088] It should be noted that when the first driving unit drives the adjusting outer rod 222 to rotate around the axis of the lamination module 1, the second driving unit 223 synchronously drives the adjusting inner rod 221 to rotate around its own axis, and the light source module 3 forms a composite motion trajectory under the cooperative control of the double driving units, and the illumination angle coverage range is expanded from a single plane to three-dimensional space. When the surface curvature of the laminated product 4 needs to be detected, the light source module 3 adjusts the incident light angle by the additional rotation of the adjusting inner rod 221, so that the light intensity of different regions remains uniform. Preferably, when a high-curvature region is detected, the second driving unit 223 drives the adjusting inner rod 221 to rotate to tilt the light source module 3, so as to avoid excessive reflection caused by perpendicular incidence of light; when a low-curvature region is detected, the adjusting inner rod 221 remains in a horizontal state to provide uniform illumination.

[0089] It can be understood that the conventional fixed light source array can only provide fixed-angle illumination, and when the surface of the laminated product 4 has a complex curved surface, the fixed light source cannot adapt to the curvature change, resulting in abnormal illumination in local regions. In the embodiment, the double driving units are used to realize the double degrees of freedom rotation of the light source module 3, so that the light source angle can dynamically adapt to the change of the curved surface. The adjusting mechanism in the prior art only has single-axis rotation capability, and the adjustment range of the light source position is limited. In the present scheme, the nested rod structure is matched with independent driving units, so that the light source module 3 obtains an additional rotation degree of freedom.

[0090] In the embodiment, the adjusting inner rod 221 is provided with a slide rail 224 on the side close to the lamination module 1, and the light source module 3 is arranged on the slide rail 224, so that the light source module 3 moves along the slide rail 224.

[0091] In some embodiments, the slide rail 224 is arranged on the side of the adjusting inner rod 221 close to the lamination module 1, and the light source module 3 is connected with the slide rail 224 through a sliding block. When the adjusting inner rod 221 rotates circumferentially, the light source module 3 can move linearly along the slide rail 224 to form a composite motion trajectory.

[0092] In the embodiment, the number of the adjusting modules 2 is 2, and the adjusting modules 2 are arranged in axial symmetry with respect to the axis of the lamination module 1.

[0093] It can be understood that the number of the adjusting modules 2 is 2, which means that two groups of independently controlled mechanical adjusting devices are used, and specifically, a double-shaft driving mechanism with a servo motor can be used to realize the independent control of the motion trajectories of the two groups of adjusting modules 2. The axial symmetry with respect to the axis of the lamination module 1 refers to the spatial layout relationship of the two groups of adjusting modules 2, and specifically, the two groups of supporting frames can be symmetrically installed on both sides of the lamination module 1 to realize the mirror-symmetrical structure formed by the two groups of supporting frames and the axis of the lamination module 1.

[0094] In some embodiments, the two sets of adjustment modules 2 are located symmetrically on both sides of the curved surface of the product 4 to be pasted, and can synchronously adjust the illumination angle of the light source module 3 during movement. When the curved surface of the product 4 to be pasted needs to be detected, the two sets of adjustment modules 2 can drive the light source module 3 to move along different trajectories, for example, when the left curved surface area is detected, the right adjustment module 2 can adjust the light source to a compensation position. Through the two sets of symmetrically adjustable light source modules 3, the entire light source angle range of the product 4 to be pasted can be theoretically realized in a continuous change, and the symmetric layout reduces the standard deviation of the illumination intensity distribution on both sides of the curved surface to less than 35% of that of the traditional method.

[0095] The problem of the difference in reflectivity of the curved surface caused by the limitation of the fixed light source angle is effectively solved, the estimation error of the normal vector of the curved surface is reduced to within ±3° from ±8° of the traditional method, and the area ratio of the local overexposure area is reduced from 12.6% to 2.3%, thereby providing a more accurate normal vector data basis for subsequent curved surface depth calculation.

[0096] In the embodiment, the adjusting inner rod 221 and the adjusting outer rod 222 are quarter-fan ring structures with different diameters.

[0097] It can be understood that the quarter-fan ring structure refers to an arc-shaped rod with a central angle of 90 degrees, which can be formed by bending a metal material, wherein the different diameters refer to that the arc radius of the adjusting inner rod 221 is smaller than that of the adjusting outer rod 222, which can be realized by a stepped radius difference design to form an asymmetric spatial layout between the inner and outer rods.

[0098] It can also be understood that when the adjusting outer rod 222 is controlled to rotate circumferentially by the driving unit, the adjusting inner rod 221 generates independent rotary motion under the action of the second driving unit 223. Due to the diameter difference between the inner and outer rods, the movement trajectory of the light source module 3 on the slide rail 224 forms a compound curve path, so that the illumination angle range of the light source module 3 in space is expanded from a single plane to three-dimensional space. The geometric constraint of the quarter-fan ring structure limits the movement path of the light source module 3 within a specific fan-shaped area, avoiding interference with other mechanical components, while ensuring that the light source direction forms a controllable included angle with the normal of the curved surface of the product 4 to be pasted.

[0099] In the prior art, ceramic ball calibration needs to rely on reflected light spots to calculate the light source direction, while the present scheme directly controls the spatial pose of the light source module 3 through mechanical structure, eliminating the direction estimation error caused by refraction of transparent materials. Through the above technical scheme, the light source module 3 realizes three-dimensional angle adjustment in a limited space, effectively expands the range of light source direction change, and reduces the local overexposure or underexposure phenomenon caused by the difference in reflectivity of the curved surface. The compound movement trajectory of the light source module 3 can match the curved surface form of the product 4 to be pasted, so that the surface normal vector distribution in the image information is more uniform, thereby improving the measurement accuracy of the concentricity detection.

[0100] Embodiment 2:

[0101] As shown in the accompanying drawings, the embodiment provides a concentricity detection method, which comprises the following steps: Figure 3

[0102] controlling the adjusting module 2 to rotate so that the light source module 3 passes through different positions when the adjusting module 2 rotates outside the pasted product 4;

[0103] acquiring image information of the corresponding area of the pasted product 4 when the light source module 3 provides different light sources in different positions;

[0104] receiving the image information;

[0105] acquiring the light source direction of the image information at different positions;

[0106] estimating the surface normal vector of the pasted product 4 to interpolate and fit the surface normal vector;

[0107] solving the depth information of the boundary surface of the pasted product 4;

[0108] restoring the surface curvature of the pasted product 4 and detecting the concentricity with the pasting film.

[0109] It should be noted that, on the one hand, the traditional curved surface detection technology needs to calibrate the light source, that is, the direction of the light source is estimated and calculated by using a calibration ball, which is not conducive to rapid detection and pasting operation in industrial production; on the other hand, since the image information is collected by passing through the microlens array in the process of beam splitting to form discrete local data, the stereoscopic surface photometry is based on global data, and the surface normal vector needs to be compensated by using the data fitting interpolation method.

[0110] It should be further noted that the traditional method relies on a fixed light source array, which results in insufficient light incidence angle in the edge area of the curved surface. The embodiment can obtain the best incidence angle in each area of the curved surface by dynamically adjusting the light source position. The prior art needs to indirectly calculate the direction of the light source by using a calibration ball, and in the transparent material scene, cumulative errors are caused by refraction. The embodiment directly calculates the direction of the light source by mechanical motion parameters, eliminates the error transmission of intermediate links, and the traditional normal vector estimation method is prone to artifacts in the reflectivity mutation area. The adaptive interpolation algorithm of the embodiment effectively suppresses the interference of abnormal points.

[0111] ​By the technical scheme, the light source calibration error caused by the refraction effect when the transparent film is attached to the curved substrate is solved, the reflectivity difference defect existing in the fixed light source system in the curved surface detection is overcome, the space position and the motion track of the light source are actively controlled, the uniform and effective light coverage of each area of the curved surface is ensured, the normal vector calculation accuracy is improved, the light source direction is directly solved in combination with the mechanical motion parameters, the system error caused by the refraction path calculation of the transparent medium is avoided, and the accurate reconstruction of the three-dimensional appearance of the curved surface is realized.

[0112] It can be understood that when the processor executes the program, first, a motion instruction is sent to the adjustment module 2, the light source module 3 is driven to move around the attached product 4, and preferably covers a range of 0 degrees to 180 degrees along an arc track. During the movement, the light source module 3 switches the light intensity according to a preset mode, and preferably outputs three brightnesses of high, medium and low at every interval of 10 degrees. The visual unit synchronously captures the reflection images at each position, and preferably collects an image sequence containing the specular reflection and diffuse reflection components after each light source switching. After the image information is processed, the light source direction of each pixel point is extracted by preferably using the optical flow analysis method, and the incident angle is calculated according to the displacement of the highlight area in adjacent images. The surface normal vector estimation module establishes a linear equation set based on the reflection equation, and preferably solves the normal vector components of each pixel point by using the least square method. The interpolation fitting process expands the discrete normal vector data into a continuous curved surface, and preferably constructs a bilinear interpolation function between adjacent sampling points. The depth solving module performs integral operation along the normal vector direction, for example, adopts the Poisson equation to reconstruct the curved surface height field. Finally, the concentricity detection result is output by comparing the geometric center deviation of the reconstructed curved surface and the attached film. It can also be understood that the above algorithm model is not the final limitation adopted in the embodiment, and the algorithm can be improved subsequently for better processing.

[0113] On the basis of the above problems, the detection method in the embodiment can include four steps of data acquisition, preprocessing, normal vector estimation calculation and restoration detection. Specifically, for data acquisition, the device for data acquisition includes a visual unit, a rotating module and a bearing module. The visual unit is preferably a CCD or CMOS industrial camera, and is specifically preferably a camera of HIKROBOT model MV-CE050-31GC. It can be understood that the visual unit also necessarily has a matching lens array unit.

[0114] For preprocessing, during the image acquisition process, the data set may contain certain noise due to various reasons. In order to prevent the interference of noise, the image needs to be denoised. The detection unit is used to receive the image information after the image denoising is completed.

[0115] In some embodiments, for the speckle noise in the image information, a local dynamic threshold segmentation algorithm is adopted, and its expression is:

[0116] ;

[0117] wherein, μ local is the local mean of the gray scale;

[0118] σ local is the standard deviation of the gray scale;

[0119] k is the adjustment coefficient;

[0120] T is a function of the local dynamic threshold segmentation algorithm, and is expressed by the discrete coordinates of the pixels in the image information.

[0121] Through the noise reduction processing, the signal-to-noise ratio of the image information is increased from 8.2 dB to 24.7 dB, the noise interference in the original image is processed by using the dynamic local threshold algorithm, the brightness distribution of the local area of the image is analyzed, the noise threshold is automatically calculated, and the feature points and the background noise are effectively separated. The speckle noise problem caused by the micro-lens beam splitting (the feature points around are blurred) is solved, the traditional global threshold method cannot adapt to the local brightness change, and the feature information is easily lost. Through the adaptive processing, the image quality is significantly improved, the feature point profile is clear and sharp, pure data source is provided for subsequent light source direction calibration and normal vector calculation, and the overall detection accuracy is enhanced.

[0122] For normal vector estimation calculation, the method in the embodiment is based on the principle of Lambertian reflection model, and based on the known light source direction information and the corresponding image data set, the object surface normal vector can be estimated according to the principles of optics and physics, and the estimated object surface normal vector is separated, interpolated and fitted.

[0123] For restoration detection, the embodiment is based on the principle that the surface normal vector after interpolation fitting is perpendicular to the surface tangent, and the depth information of the pasted surface boundary curve can be obtained, the pasted curve is restored by the depth information, and the concentricity detection with the pasting film is realized.

[0124] Embodiment 3:

[0125] In the embodiment, the control of the adjustment module 2 rotating specifically includes the following steps:

[0126] The first driving unit is controlled to start, so that the adjustment outer rod 222 rotates circumferentially along the axis of the output end of the first driving unit, and the adjustment outer rod 222 is controlled to rotate to a first set angle;

[0127] The second driving unit 223 is controlled to start, so that the adjustment inner rod 221 rotates circumferentially along the axis of the output end of the second driving unit 223, and the adjustment inner rod 221 is controlled to rotate to a second set angle.

[0128] It can be understood that the first set angle and the second set angle are angles preset in the detection system control module by the interaction module. When the first set angle and the second set angle are fixed, the light source direction of the light source under the light source module 3 is uniquely fixed and is a known parameter in the detection system.

[0129] According to the principle of Lambertian reflection model, it is known that the pixel intensity of each point in the image is related to the product of the surface normal vector of each point of the object and the light source direction. In the image information obtained in the embodiment, the pixel intensity of each point is known, and the light source direction is known, so the surface normal vector of the product 4 can be calculated.

[0130] In some embodiments, the first set angle and the second set angle are preset by the interaction module to form a unique light source direction vector, which is represented by cos α .

[0131] Embodiment 4:

[0132] In this embodiment, the image information of the corresponding area of the product 4 when the light source module 3 provides different light and dark light sources at different positions includes the following steps:

[0133] Adjust the physical position and brightness level of the light source module 3 through the programmable light source controller;

[0134] After the light source module 3 switches to the target position and the light and dark parameters, a synchronization signal is sent to the vision unit to ensure that the image capture and the light source state are strictly matched;

[0135] Preprocess the image information;

[0136] Match the image feature points in the image information through the scale-invariant feature transform algorithm.

[0137] Based on the above steps, the position of the light source module 3 is dynamically controlled by the adjustment module 2, so that the light source module 3 moves at multiple angles and provides different light and dark light sources outside the product 4, and the vision unit synchronously captures the image information sequence of the corresponding area; This process directly obtains the light source direction based on the physical displacement of the light source, avoids the refraction interference caused by the calibration ball, and expands the light coverage range through dynamic light source position adjustment, effectively compensates for the difference in curved surface reflection, and suppresses the overexposure or underexposure problem. In addition, combined with the multi-angle image sequence, the detection unit interpolates and fits the surface normal vector to solve the boundary depth information, realizes high-precision curved surface reconstruction and concentricity detection, thereby additionally improving the detection precision of the transparent film and the curved surface substrate on the basis of eliminating the defects of the prior art, adapting to the needs of products with different curvature shapes, and reducing the step of separate light source calibration, improving the detection efficiency.

[0138] Embodiment 5:

[0139] In this embodiment, the surface normal vector of the product being pasted is estimated to perform interpolation fitting on the surface normal vector;

[0140] The expression for estimating the surface normal vector of the product being pasted is:

[0141] ;

[0142] Where I is the pixel intensity matrix;

[0143] ρ is the surface reflectivity constant;

[0144] N is the surface normal matrix;

[0145] L T is the light source direction matrix.

[0146] The expansion of the Lambert model combined with the light source direction is:

[0147] ;

[0148] in, For the n The image is in ( x , y ) at the gray value;

[0149] is the surface normal vector to be determined;

[0150] is the light source direction vector.

[0151] Combining the Lambertian reflectance model and least squares optimization to solve the normal vector field, the image pixel intensity data under multiple light source directions is used to construct a set of linear equations, and the surface normal vector of each pixel is solved by the least squares method. This solves the problem of the relationship between image intensity and normal vector. Traditional methods are prone to outliers or non-unique solutions when the data is discrete. Through this method, a continuous and consistent normal vector distribution is generated, providing reliable input for surface depth calculation, avoiding manual intervention, and realizing automated processing.

[0152] The interpolation fitting of the surface normal vector comprises the following steps:

[0153] To find the surface normal vector, execute:

[0154] ;

[0155] in,

[0156] P n ( u ) is the target polynomial function;

[0157] For fitting coefficients;

[0158] n =5 is the optimal order;

[0159] u is the input variable of the function, i.e. the coordinate axis position in the fitting; and u= x, y x, y represents the two-dimensional coordinate value;

[0160] k is the order control variable.

[0161] More specifically, assuming the curve function is y= f x , y= f x defined on the interval [a, b], x 0, x1, x 2, …… x n-1 , x n (n+1) points on the interval [a, b] are mutually different, each of which corresponds to a value respectively y 0, y 1, y 2, …… y n-1 , y n , there is a unique interpolation polynomial P( x ) through the given (n+1) points y , x ) and the degree is not more than n as the approximation of f x , the interpolation polynomial has been defined above and will not be repeated here. It can be understood that the coefficient matrix is the Vandermonde matrix, according to the determinant property, i.e. the coefficient matrix is a (n+1) order Vandermonde determinant, because x 0, x1, x 2, …… x n-1 , x n are mutually different, so the Vandermonde determinant is not zero, which indicates that the equation has a unique solution a 0, a 1, a 2, …, a n—1 , a n .

[0162] ​​​​​It can be understood that the interpolation polynomial fitting technology is used to convert the discrete normal vector into a continuous curved surface, the data local discrete problem existing in the detection method in the prior art is smoothed by the polynomial fitting algorithm, the problem of non-smooth surface reconstruction caused by discrete data is solved, the traditional method outputs burrs or concave-convex surfaces, which affects the final detection accuracy, seamless surface reconstruction is realized, high fidelity of depth information calculation is ensured, and a high-resolution three-dimensional model is provided for the concentricity detection of the pasted product.

[0163] Embodiment 6:

[0164] In the embodiment, the solving of the depth information of the boundary curved surface of the pasted product further includes the following steps:

[0165] constructing a gradient equation group;

[0166] solving the Poisson equation by Fourier transform by using the Poisson integral method, and outputting a depth map;

[0167] The expression of the constructed gradient equation group is:

[0168]

[0169] respectively, the partial derivative of the depth map z ( x , y ) along the x , y direction, that is, the depth gradient;

[0170] P x , P y represent intermediate variables related to the surface normal vector to be solved, and satisfy P x =N x / N z , P y = N y / N z .

[0171] The expression of solving the Poisson equation by Fourier transform by using the Poisson integral method is:

[0172] ;

[0173] wherein, is a Laplacian of the depth map z ( x , y );

[0174] , are partial differential operators, respectively representing the variables xPartial derivative, variable y Partial derivative.

[0175] It can be understood that the traditional curved surface depth reconstruction method has significant defects: when dealing with complex curved surfaces or transparent materials, small errors in the local normal vector will accumulate during the integration process, causing the overall distortion of the reconstructed curved surface or the misalignment of the boundary. Especially for the curved surface boundary of the product to be pasted in the OCA film pasting scene, the traditional method is easily disturbed by surface reflection noise, causing deviation in depth gradient calculation, and finally forming non-physical curved surface undulation or fracture.

[0176] It can also be understood that the embodiment builds a gradient equation set covering the entire domain of the product to be pasted, which takes the partial derivatives of the depth map in the x and y directions (i.e. depth gradient) as the basic variables, and accurately describes the slope change trend of each small area of the curved surface through mathematical relationship. This global modeling fundamentally avoids the error accumulation defect of the traditional point-by-point integration method, ensuring that the mathematical description of the curved surface deformation conforms to the actual physical law.

[0177] The gradient equation set is further converted into a Poisson equation, which has the mathematical advantage of naturally adapting to the closed characteristics of the curved surface. The Poisson equation is solved through Fourier transform to realize the synchronous processing of global data and convert complex differential operation into efficient algebraic operation in the frequency domain. The above process not only ensures that the generated depth map strictly satisfies the curved surface boundary constraint, but also realizes the industrial-level real-time requirement with super-linear calculation speed.

[0178] Embodiment 7:

[0179] In this embodiment, after the Poisson integral method is used to solve the Poisson equation through Fourier transform, it further includes: applying a constraint weighted least squares filter to correct boundary mutations to solve the problem of depth jump caused by occlusion or reflection at the edge of the product to be pasted.

[0180] Specifically, the following steps are included:

[0181] In the edge region of the depth map, define a confidence weight ω ( x , y ), where ω = 1 is a high-confidence flat area and ω = 0 is a low-confidence mutation area such as a boundary shadow.

[0182] An optimization objective function is constructed, and its expression is:

[0183] ;

[0184] Where, z 0( x, y ) is the initial depth map, and λ is the smoothing coefficient.

[0185] Finally, iterative solution is performed to make the boundary transition natural;

[0186] is the gradient sign, denotes a two-dimensional gradient vector, and satisfies the following expression:

[0187] ;

[0188] wherein t denotes a vector transposition operator.

[0189] It can be understood that in the prior art, when a traditional photometric stereo vision technology performs curved surface depth reconstruction, abnormal jumps of depth information in a boundary region often occur due to factors such as light blocking of an object edge, mirror reflection or refraction of a transparent material. Such mutation phenomenon seriously distorts the geometric continuity of the reconstructed curved surface. For example, in OCA film fitting detection, steep edges or transparent film joints of a fitted product are prone to form shadow or highlight artifacts, so that the algorithm misjudges as an actual existing depth fault. A traditional solution usually adopts global smoothing filtering, but will blur real physical edge details at the same time; or relies on manual intervention to correct errors, which greatly reduces the detection efficiency and automation degree.

[0190] The embodiment automatically identifies a reliable region and an abnormal region in a depth map according to image features such as gradient change and light uniformity, avoids false smoothing of effective edges, and accurately locks the artifact region to be corrected. An optimization objective function constructed by containing double constraint conditions requires that the corrected depth value is as close as possible to the initial solution to retain the real curved surface features, and controls the transition naturalness of the mutation region through a smoothing coefficient; the function retains the real edge sharpness while forcing the abnormal region to gradually connect with the adjacent curved surface through mathematical optimization. A multiple iteration calculation strategy is adopted to gradually reduce the weight influence of the abnormal region, so that the depth value naturally diffuses from the reliable region to the mutation region, and finally generates a smooth transition curved surface with consistent physical meaning, which greatly eliminates the "step type" jumps caused by optical interference.

[0191] Based on the above, in the OCA film fitting scene, the false depth fault caused by refraction shadow of the transparent film edge is successfully eliminated, the restored curved surface substrate profile error is controlled within ±0.01 mm, the traditional manual correction link is replaced, the detection efficiency is improved by 3 times, and subjective errors are avoided. The accurate retention of the reliable region of the depth map ensures that the real physical boundary (such as the cutting edge line) of the fitted film and the substrate is not smoothed by mistake, and provides an anatomical level accurate curved surface model for concentricity analysis.

[0192] Embodiment 8:

[0193] The specific steps of reducing the surface curvature of the product to be pasted and detecting the concentricity with the pasting film are: extracting boundary information and ideal center from the depth map of the product to be pasted; obtaining pasting information and pasting center of the pasting film from the morphological segmentation positioning; measuring the offset amount of the ideal center and the pasting center by laser interference; and finally determining the deviation distance between the boundary information and the pasting information.

[0194] In some embodiments, the offset amount of the ideal center is:

[0195] ;

[0196] wherein, is the ideal center, is the pasting center.

[0197] In some embodiments, the control signal is also generated by the offset amount and the deviation distance, so as to control the pasting module to adjust the pasting distance through the control signal. Preferably, the control process is a closed-loop feedback calibration process in a PID controller, and the expression is:

[0198]

[0199] wherein, is the real-time output instruction of the controller;

[0200] is the integral of the historical deviation;

[0201] is the rate of change of the deviation;

[0202] K p 、K i 、K d are respectively the proportional coefficient, the integral coefficient and the differential coefficient of the PID controller.

[0203] It should be noted that in order to solve the problem of misalignment of light source calibration caused by refraction effect in the prior art, after high-precision curved surface reconstruction is completed, a non-contact optical detection method is used. First, the actual boundary information of the attached product and the ideal center point of the theoretical design are directly extracted from the reconstructed depth map, avoiding secondary errors caused by contact measurement of the traditional mechanical probe. At the same time, the edge profile and the attachment center of the transparent attachment film (such as OCA film) are accurately positioned through morphological image segmentation technology. This technology has strong robustness to the edge features of transparent materials and overcomes the problem of edge recognition ambiguity caused by the light transmission of the film. Then, the laser interferometry method is introduced to quantify the sub-micron offset of the two center points. By using the principle of light wave phase difference, the measurement deviation caused by the transparent medium is eliminated from the physical layer, so that the offset calculation (i.e. the spatial vector difference between the ideal center and the attachment center) has nanometer-level precision.

[0204] It should be further noted that on the basis of realizing high-precision detection, a dynamic attachment calibration system is further constructed. The calculated offset and the topographic deviation distance between the boundary and the attachment film are converted into control signals to drive the PID controller to generate real-time adjustment instructions. The controller adopts a three-mode closed-loop feedback: the proportional element (P) adjusts the attachment pressure in real time according to the current deviation distance; the integral element (I) accumulates the historical deviation to eliminate the steady-state error of the system, which is particularly suitable for working conditions where the curvature of the curved surface changes continuously; the derivative element (D) predicts the attachment trend according to the deviation rate to suppress mechanical overshoot. This dynamic calibration mechanism enables the attachment module to adaptively adjust the attachment distance and angle within a millisecond response period, upgrading the offline correction with manual intervention in the prior art to online real-time compensation. The precision of curved surface concentricity detection is improved from ±50μm in the prior art to the order of ±5μm, especially solving the measurement blind area in the transparent / curved surface composite scenario; through closed-loop control, the attachment yield is improved by more than 30%, reducing the rework loss; the adaptive PID algorithm can be compatible with any curvature substrate such as plane, sphere, free-form surface, etc., meeting the process requirements of new products such as foldable screens and curved vehicle displays.

[0205] To achieve the foregoing object, the embodiment further provides a computer device, which comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program.

[0206] To achieve the foregoing object, the embodiment further provides a computer readable storage medium, which stores a computer program, and the processor executes the computer program.

[0207] In addition, in an embodiment, the present application further provides a computer storage medium, which stores a computer program, and the computer program is executed by the processor to realize the steps of the method in the foregoing embodiment.

[0208] In some embodiments, the computer readable storage media can be a memory such as a FRAM, ROM, PROM, EPROM, EEPROM, flash memory, a magnetic surface memory, an optical disk, or a CD-ROM, etc.; and can also be various devices including one or any combination of the above memories. The computer can be various computing devices including a smart terminal and a server.

[0209] In some embodiments, the executable instructions can be in the form of a program, software, software module, script, or code, written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0210] By way of example, the executable instructions can or can not correspond to a file in a file system, can be stored in a part of a file that holds other programs or data, for example, in one or more scripts stored in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files, for example, files that store one or more modules, sub programs, or portions of code.

[0211] By way of example, the executable instructions can be deployed to execute on one computer, or on multiple computers that are located at one site, or that are distributed across multiple sites and are interconnected through a communication network.

[0212] It should be noted that, in this document, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, so that a process, method, article, or system that comprises a list of elements does not include only those elements, but can also include other elements not expressly listed or inherent to such process, method, article, or system. Without more limitations, an element defined by the statement "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or system that includes the element.

[0213] The above-mentioned embodiment numbers of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0214] Those skilled in the art can clearly understand the above-mentioned example method can be realized by means of software and the necessary general hardware platform, of course, also can be through hardware, but many cases the former is the better implementation. Based on such understanding, the technical solutions of the present application essentially or say the part of the contribution to the prior art can be embodied in the form of software products, the computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disc), including a plurality of instructions to make a multimedia terminal device (may be a mobile phone, computer, television receiver, or network equipment, etc.) executes the method described in various embodiments of the present application.

[0215] The above is only the preferred embodiment of the present application, not therefore limit the patent scope of the present application, any equivalent structure or equivalent process transformation using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A concentricity detection system characterized by, The system is used for concentricity detection of OCA film lamination process completed by a laminator, the laminator comprises a lamination module for completing lamination process of lamination film on a product to be laminated, and the system comprises: a light source module for providing different light sources at different positions; a detection module comprising a vision unit and a detection unit, the vision unit is used for acquiring image information of corresponding areas of the product to be laminated when the light source module provides different light sources at different positions; an adjustment module arranged in the laminator and used for adjusting different positions of the light source module through mechanical displacement, so that the light source module passes through different positions when rotating outside the product to be laminated along with the adjustment module; wherein the detection unit comprises: an acquisition subunit for directly acquiring light source directions of the image information at different positions through mechanical displacement of the adjustment module; an estimation subunit for estimating surface normal vectors of the product to be laminated to interpolate and fit the surface normal vectors; a processing subunit for solving depth information of a boundary surface of the product to be laminated; a detection subunit for restoring surface curvature of the product to be laminated and detecting concentricity with the lamination film.

2. A concentricity detection system as claimed in claim 1, wherein, The adjustment module comprises an adjustment seat and an adjustment unit, the adjustment seat is fixedly connected with an inner wall of the laminator, a first driving unit is arranged in the adjustment seat, an output end of the first driving unit is connected with the adjustment unit, so that the adjustment unit rotates circumferentially along the output end of the first driving unit, and the light source module is arranged on the adjustment unit.

3. A concentricity detection system as claimed in claim 2, wherein, The adjustment unit comprises an adjustment inner rod and an adjustment outer rod, one end of the adjustment outer rod is connected with the output end of the first driving unit, the other end of the adjustment outer rod is rotationally arranged with the adjustment inner rod, and the light source module is arranged on the adjustment inner rod.

4. A concentricity detection system as claimed in claim 3, wherein, A second driving unit is arranged at a connecting position of the adjustment outer rod and the adjustment inner rod, so that the adjustment inner rod rotates circumferentially along an output end of the second driving unit.

5. A concentricity detection system as claimed in claim 3, wherein, A slide rail is arranged on a side of the adjustment inner rod close to the lamination module, and the light source module is arranged on the slide rail, so that the light source module moves along the slide rail.

6. A concentricity detection system as claimed in claim 1, wherein, The number of the adjustment modules is 2, and the adjustment modules are arranged in axial symmetry based on an axis of the lamination module.

7. A concentricity detection system as claimed in claim 3, wherein, The adjustment inner rod and the adjustment outer rod are quarter-fan ring structures with different diameters.

8. A method of detecting concentricity, characterized by, A concentricity detection system according to any one of claims 1 to 7, the method comprising the following steps: controlling the adjustment module to rotate, so that the light source module passes through different positions when rotating outside the product to be laminated along with the adjustment module; acquiring image information of corresponding areas of the product to be laminated when the light source module provides different light sources at different positions; receiving the image information; acquiring light source directions of the image information at different positions; estimating surface normal vectors of the product to be laminated to interpolate and fit the surface normal vectors; solving depth information of a boundary surface of the product to be laminated; restoring surface curvature of the product to be laminated and detecting concentricity with the lamination film.

9. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the method of claim 8. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the method of claim 8.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the processor executes the computer program to realize the method in claim 8.

Images