Automobile glass film removing method and matrix film removing system

Abstract

The invention relates to the technical field of automobile glass processing, in particular to an automobile glass film removing method and a matrix film removing system. The method comprises the steps that three-dimensional contour scanning is conducted on an area, to be subjected to film removal, of the automobile glass, a space coordinate grid of the surface of the glass is constructed based on scanning data, the area, to be subjected to film removal, is divided into a plurality of sub-area units with unique coordinate identifiers, and the precision of the coordinate grid is dynamically adjusted according to the curvature of the curved surface of the glass. The grid density of the area with the large curvature is higher than that of the plane area. According to the method, a dynamic coordinate grid division mechanism based on three-dimensional contour scanning is constructed, the grid density is intelligently adjusted according to the curvature characteristics of different areas of the curved glass, and it is ensured that the machining precision of the area with the large curvature is matched with the data resolution; the problem that in a traditional method, the fixed grid precision cannot adapt to the complex form of the curved surface is fundamentally solved.

Description

Technical Field

[0001] This invention relates to the field of automotive glass processing technology, specifically to a method for removing film from automotive glass and a matrix film removal system. Background Technology

[0002] To meet the installation requirements of in-vehicle electronic devices such as ETC devices and rain sensors, the metal coating on the surface of automotive glass needs to be precisely removed. Since modern automotive glass generally employs a double curvature or complex curved surface design to optimize aerodynamic performance and visibility, traditional laser coating removal technology faces the following problems when processing such curved surfaces: differences in the normal vectors of different areas of the curved surface make it difficult to precisely control the laser incident angle, easily leading to uneven energy distribution or localized thermal stress concentration; the film thickness may exhibit gradient differences in areas of curvature variation, and traditional fixed-parameter energy output modes cannot simultaneously ensure the integrity of the coating removal and the safety of the glass substrate; furthermore, the complexity of the curved surface contour makes trajectory planning unable to effectively adapt to the processing accuracy requirements of different curvature areas, easily resulting in edge residue or over-processing.

[0003] In existing technologies, the coating removal schemes for flat or low-curvature glass have problems such as insufficient coating removal accuracy and high glass damage rate when dealing with complex curved surfaces due to optical path calibration errors, energy adjustment lag, and coarse trajectory planning. Therefore, there is an urgent need for a coating removal method that can dynamically adapt to the curved surface features and accurately control the laser incident path and energy output to meet the fine processing requirements of high-end automotive glass. Summary of the Invention

[0004] This disclosure presents a method for removing film from automotive glass and a matrix film removal system, with the aim of overcoming at least one of the defects existing in the prior art.

[0005] To achieve the above objectives, the technical solution disclosed in this invention is as follows: According to one aspect of this disclosure, a method for removing film from automotive glass is provided, the method comprising the steps of: A three-dimensional contour scan is performed on the area of ​​the automotive glass to be coated. Based on the scan data, a spatial coordinate grid is constructed on the glass surface. The area to be coated is divided into multiple sub-region units with unique coordinate identifiers. The accuracy of the coordinate grid is dynamically adjusted according to the curvature of the glass surface. The grid density in areas with greater curvature is higher than that in planar areas. Based on the spatial distribution of the sub-region units and the laser energy requirements, a multi-path collaborative film removal trajectory planning scheme is generated. In the scheme, each trajectory corresponds to the angle combination of a set of mirrors in the matrix mirror group, and the overlap rate of the laser irradiation area of ​​adjacent trajectories is not less than a preset threshold to ensure the integrity of film removal. By controlling the lateral and longitudinal rotation axes of each mirror in the matrix mirror group, the parallel laser beam emitted by the laser module is sequentially guided to the top of each sub-region unit. The lateral rotation axis is used to adjust the horizontal incident angle of the laser beam to cover the width direction of the glass, and the longitudinal rotation axis is used to match the curvature of the glass to adjust the vertical incident angle of the laser beam. Based on the film thickness parameters of each sub-region unit, the laser beam energy density and irradiation time of the corresponding reflector group are dynamically adjusted. After the reflector completes the angle calibration, the laser module is controlled to emit pulsed laser, and the laser beam is focused onto the film surface of the target sub-region unit through the angle combination of the reflector group.

[0006] Furthermore, in the step of performing a three-dimensional contour scan of the area to be decocted on the automotive glass, a structured light scanner is used to acquire point cloud data of the glass surface. The point cloud data is then converted into a parameterized NURBS surface model using a surface fitting algorithm. Based on the normal vector distribution of the NURBS surface model, the spatial orientation of each sub-region unit is determined, and a correspondence between the coordinate grid and the normal vector of the glass surface is established to ensure that the angle between the incident direction of the laser beam and the normal vector of the corresponding sub-region unit does not exceed a preset angle threshold. The preset angle threshold is determined based on the energy utilization parameter of the laser decoction.

[0007] Furthermore, in the step of generating a multi-path collaborative film removal trajectory planning scheme, a tabu search algorithm is introduced to optimize the trajectory order, plan sub-region units in areas with drastic curvature changes, and at the same time avoid the mirror angle adjustment range of adjacent trajectories from exceeding the maximum angular velocity limit of the mirror drive mechanism. The trajectory planning scheme also includes a safety interval parameter to ensure that the movement path of the laser beam when switching sub-region units avoids the non-film removal area at the edge of the glass. The safety interval parameter is preset according to the installation boundary coordinates of the glass.

[0008] Furthermore, in the step of controlling the lateral and longitudinal rotation axes of each reflector in the control matrix reflector group, a dual closed-loop control strategy is adopted. The outer loop calculates the theoretical rotation angle of the reflector based on the target coordinates of the sub-region unit, while the inner loop provides real-time feedback of the actual position of the reflector through an angle encoder. When the deviation between the actual position and the theoretical angle exceeds a preset accuracy, an adaptive filtering algorithm is triggered to correct the control signal of the drive motor. During the correction process, an inertia compensation coefficient is introduced to offset the influence of mechanical vibration during the rotation of the reflector on the angle accuracy. The inertia compensation coefficient is calculated in real-time through the dynamic model of the reflector group.

[0009] Furthermore, in the step of dynamically adjusting the laser beam energy density and irradiation time of the corresponding reflector group, a film thickness-energy response database is established in advance. The database contains the correspondence between the critical film removal energy density and thickness of different film materials. When the film thickness of the sub-region unit is detected to exceed the preset threshold, a segmented energy output strategy is adopted. Pre-irradiation is performed at the base energy density to destroy the film structure. After a preset time interval, film removal is completed at the peak energy density. The preset time is determined according to the thermal diffusivity of the film material to avoid damage to the glass substrate caused by the accumulation of laser energy.

[0010] Furthermore, in the angle combination calculation step of the reflector group, a curved surface coordinate mapping compensation algorithm is introduced to perform geometric correction on the optical path of the laser beam after passing through the reflector, targeting the optical refraction effect of the double-curvature glass. The steps include: calculating the refraction angle of the laser beam when it is incident on the glass surface based on the radius of curvature of the glass, and adjusting the longitudinal rotation angle of the reflector based on Snell's law to compensate for the spot offset caused by refraction. The parameters of the curved surface coordinate mapping compensation algorithm are pre-calibrated by optical simulation software and are calibrated on a real machine before the coating removal operation.

[0011] Furthermore, before performing the film removal operation, a pre-scanning verification step is added. The planned trajectory is simulated by a low-energy laser beam, and the position image of the light spot on the glass surface is acquired by a high-speed camera. The actual position of the light spot is matched with the coordinates of the target sub-region unit. When the matching error exceeds a preset pixel threshold, the angle calibration program of the reflector group is automatically triggered. The calibration program fits the relationship model between the light spot offset and the reflector angle deviation using the least squares method to realize the real-time correction of the mapping relationship model.

[0012] Furthermore, the matrix reflector array is arranged in an arc-shaped array along the emission direction of the laser module. The lateral spacing between adjacent reflectors is determined according to the divergence angle of the laser beam and the maximum film removal distance, ensuring that the effective irradiation range of the reflector array covers automotive glass of different sizes. The radius of curvature of the arc array matches the maximum curvature radius of the standard automotive glass to reduce the angle adjustment range of the reflectors.

[0013] Furthermore, during the film removal process, the temperature distribution data of the glass surface is collected in real time. When the local temperature is detected to exceed the glass heat resistance threshold, a dynamic cooling mechanism is triggered. Low-temperature protective gas is sprayed into the laser irradiation area through the gas path system. The spray flow rate of the gas is linearly positively correlated with the laser energy density. At the same time, the irradiation of the current sub-region unit is paused until the temperature drops back to the safe range. The threshold of the safe range is dynamically adjusted according to the thermal conductivity coefficient of the glass material.

[0014] According to another aspect of this disclosure, a matrix film removal system is provided for implementing the automotive glass film removal method described above, the system comprising: The data processing unit is used to acquire the three-dimensional contour data of the automotive glass, divide the area to be removed into sub-region units and generate a coordinate grid, and generate a film removal trajectory planning scheme based on the coordinate grid and film thickness parameters. The control unit is used to calculate the lateral and longitudinal rotation angles of each mirror in the matrix mirror group according to the trajectory planning scheme, generate drive signals to control the dual-axis rotation mechanism of the mirror, and adjust the energy output parameters of the laser module according to the film thickness parameters. A matrix-type reflector assembly includes multiple reflector units arranged in an arc-shaped array. Each reflector unit is connected to an independent horizontal rotation axis and a vertical rotation axis. The horizontal rotation axis and the vertical rotation axis are equipped with an angle encoder and a servo drive motor to achieve high-precision adjustment of the reflector angle. A laser module is used to emit a pulsed laser beam. The emission direction of the laser beam is aligned with the central axis of the matrix reflector group. The laser beam is guided to the target sub-region unit by the angle combination of the reflector group. The control unit incorporates a curved surface coordinate mapping model and a reflector angle compensation algorithm to correct optical path deviations caused by the refraction of the laser beam by the curved glass surface.

[0015] The beneficial effects of this invention are: This invention constructs a dynamic coordinate grid division mechanism based on three-dimensional contour scanning, which intelligently adjusts the grid density according to the curvature characteristics of different regions of curved glass, ensuring that the processing accuracy of regions with large curvature matches the data resolution, fundamentally solving the problem that the fixed grid accuracy in traditional methods cannot adapt to the complex shape of curved surfaces.

[0016] Specifically, by introducing a multi-path collaborative trajectory planning scheme for a matrix-type reflector group, combined with a surface coordinate mapping compensation algorithm and a dual closed-loop angle control strategy, dynamic calibration of the laser beam incident direction and the surface normal vector was achieved. This effectively compensated for the spot offset caused by surface refraction, enabling the laser energy to act vertically and uniformly on the target film, thereby improving energy utilization and the consistency of the film removal edge.

[0017] Furthermore, based on the segmented energy output strategy of the film thickness-energy response database, the energy density and irradiation time can be dynamically adjusted according to the film parameters of different sub-regions. By pre-irradiating to destroy the film structure and controlling the heat diffusion process, damage to the glass substrate caused by energy accumulation is avoided, and the impact of thermal stress is minimized while ensuring the integrity of the film removal.

[0018] Furthermore, the combination of the pre-scanning verification process and the real-time temperature monitoring mechanism has constructed a closed-loop feedback system covering the entire process from trajectory planning to execution control. This enables precise quantitative control of the coating removal process on curved glass, ultimately forming a high-precision coating removal technology system that integrates curved surface adaptation, optical path calibration, and energy optimization. This comprehensively improves the processing quality and production efficiency of complex curved glass, providing technical support for the reliable installation of automotive electronic devices. Attached Figure Description

[0019] Figure 1 This is a flowchart of a method for removing film from automotive glass according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the three-dimensional curvature and adaptive coordinate grid of curved glass in one embodiment of the present invention; Figure 3 This is a schematic diagram of the optical path control of a matrix mirror assembly in one embodiment of the present invention; Figure 4 This is a schematic diagram of the curved surface mapping optical path compensation principle in one embodiment of the present invention; Figure 5 This is a schematic diagram of the film thickness and energy control strategy in one embodiment of the present invention. Detailed Implementation

[0020] 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 only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] The term "comprising" and any variations thereof in the specification and claims of this application are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product or device.

[0022] In embodiments of the present invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in embodiments of the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0023] The present invention provides the following preferred embodiments: Example 1 To address the issues of insufficient optical path control precision and uneven energy distribution in traditional laser coating removal technology for processing complex curved glass, this embodiment provides a method for removing coatings from automotive glass. Through refined processing of three-dimensional contour scanning data, multi-dimensional trajectory planning, high-precision mirror angle control, and dynamic energy adjustment, precise removal of the coating layer from the curved glass surface is achieved. Figure 1 As shown, the process of removing the film from automotive glass includes: S100: Performs a three-dimensional contour scan on the area of ​​the automotive glass to be coated, constructs a spatial coordinate grid on the glass surface based on the scan data, and divides the area to be coated into multiple sub-region units with unique coordinate identifiers. The accuracy of the coordinate grid is dynamically adjusted according to the curvature of the glass surface, and the grid density in areas with greater curvature is higher than that in planar areas.

[0024] S200: Based on the spatial distribution of sub-regional units and laser energy requirements, a multi-path collaborative film removal trajectory planning scheme is generated. In the scheme, each trajectory corresponds to the angle combination of a set of mirrors in the matrix mirror group, and the overlap rate of the laser irradiation area of ​​adjacent trajectories is not less than a preset threshold to ensure the integrity of film removal.

[0025] S300: By controlling the lateral and longitudinal rotation axes of each mirror in the matrix mirror group, the parallel laser beam emitted by the laser module is sequentially guided to the top of each sub-region unit. The lateral rotation axis is used to adjust the horizontal incident angle of the laser beam to cover the width direction of the glass, and the longitudinal rotation axis is used to match the curvature of the glass to adjust the vertical incident angle of the laser beam.

[0026] S400: Based on the film thickness parameters of each sub-region unit, dynamically adjust the laser beam energy density and irradiation time of the corresponding reflector group. After the reflector completes the angle calibration, control the laser module to emit pulsed laser, and focus the laser beam onto the film surface of the target sub-region unit through the angle combination of the reflector group.

[0027] Specifically, the process begins with a 3D contour scan and dynamic mesh construction of the area to be coated on the automotive glass. During implementation, an industrial-grade structured light scanner, such as the KEYENCE LJ-G series, is used to acquire millimeter-level point cloud data of the area to be coated, obtaining the 3D coordinates and curvature features of the glass surface. The point cloud data is then converted into a parametric model using a non-uniform rational B-spline (NURBS) surface fitting algorithm. Based on the Gaussian curvature and average curvature calculations, an adaptive spatial coordinate mesh is constructed: for hyperbolic transition regions such as glass corners, the mesh density is automatically increased to 20×20 nodes per square centimeter, while planar areas use a base density of 10×10 nodes, forming a well-spaced mesh system. Each sub-region unit is assigned a unique coordinate identifier (X, Y, Z, K), where K is the curvature feature parameter, providing a surface morphology reference for subsequent trajectory planning. Figure 2 The three-dimensional contour and mesh distribution of the curved surface shown are used to dynamically adjust the mesh accuracy through curvature, ensuring the processing resolution of complex areas.

[0028] Furthermore, a multi-path collaborative trajectory planning scheme is generated based on the spatial distribution of sub-regions. The sub-region coordinate data is imported into the trajectory generation module, and combined with the physical parameters of the matrix reflector group, such as the maximum rotation angle of the reflectors and the laser beam divergence angle, with a constraint that the overlap rate of the laser irradiation areas of adjacent trajectories is no less than a preset threshold based on the film material characteristics (e.g., 30%), a continuous trajectory sequence is generated through an improved Dijkstra algorithm. Each trajectory corresponds to the angle combination of the reflector's lateral rotation axis (X-axis) and longitudinal rotation axis (Y-axis). This combination is converted into mechanical motion parameters of the reflector through a coordinate transformation matrix: the lateral axis is adjusted within the range of 0-1500mm in the width direction of the covering glass by ±45°, and the longitudinal axis is dynamically adjusted within the range of 0°-30° to match the curvature of the surface, such as... Figure 3 The optical path control principle of the reflector group shown ensures the continuity of the trajectory and the smoothness of mechanical movement, avoiding the impact of sudden changes in angle adjustment on system stability.

[0029] Furthermore, in the reflector angle control stage, a dual-axis servo drive system, such as the Panasonic MINASA6 series servo motor paired with a 20-bit absolute encoder, is used to achieve high-precision control. The lateral rotation axis converts motor rotation into horizontal deflection via a rack and pinion mechanism, while the longitudinal axis utilizes a harmonic reducer for fine-tuning the vertical angle. After receiving the target sub-region coordinate command, the theoretical rotation angle is calculated using a coordinate mapping model, driving the motor to rotate and acquiring encoder feedback in real time. Position closed-loop control ensures an angle accuracy of ±0.01°, keeping the angle between the incident beam and the sub-region normal vector within 15°, maximizing the energy vertical incidence efficiency, and reducing beam distortion and energy loss. Figure 4 The optical path compensation principle shown achieves the adaptation of the incident angle at different positions on the curved surface through precise angle adjustment.

[0030] Furthermore, the laser energy parameters are dynamically adjusted based on the film thickness of each sub-region. Thickness data for each sub-region is pre-scanned using a spectral confocal sensor, such as the Mi-Iridium CFS series, with an accuracy of ±0.1 μm, establishing a thickness-energy mapping database. Based on this mapping, the energy density is calculated for a film of thickness t using the formula E = k·t + E0, where k is the material energy coefficient and E0 is the base threshold. The irradiation time is adjusted synchronously. After the reflector completes angle calibration, the control unit, based on a pre-calibrated surface parameter matrix, establishes a coordinate-reflector angle mapping relationship using the optical tracking software Zemax. A trigger signal is then sent to the laser module, such as the IPG YLR-500 fiber laser, to focus the laser beam onto the target position. Figure 5The energy control strategy shown ensures precise energy output for film layers of different thicknesses, avoiding thermal damage to the glass substrate.

[0031] The advantages of this embodiment are that it achieves refined modeling of surface features through the construction of a coordinate grid with curvature adaptation, solves the problem of incident angle adaptation by combining multi-path trajectory planning and dual-axis high-precision angle control, and ensures accurate energy output based on the energy dynamic adjustment mechanism of film thickness and the pre-calibrated mapping model, forming a complete closed loop from data acquisition to execution control, effectively improving the uniformity and processing accuracy of film removal on complex curved surfaces, and meeting the high-precision film removal requirements of automotive glass.

[0032] Example 2 To address the issues of insufficient matching accuracy between the laser incident angle and the surface normal vector, and limited energy utilization during the removal of coatings from complex curved glass, this embodiment optimizes the three-dimensional contour scanning step. By constructing a NURBS surface model and its correspondence with the normal vector, precise control of the laser incident direction is achieved.

[0033] Specifically, a high-precision structured light scanner, such as the German GOM ATOS III Triple Scan, is used to acquire point cloud data of the area to be decocted. This device projects a sinusoidal fringe pattern and simultaneously acquires images from a binocular camera, achieving a single-point measurement accuracy of 0.02 mm to ensure the acquisition of the microscopic curvature features of the glass surface. After noise reduction and filtering, the acquired point cloud data is imported into CAD modeling software. Parametric modeling is performed using a non-uniform rational B-spline (NURBS) surface fitting algorithm. By controlling the vertex mesh to adjust the surface smoothness, the fitting error is controlled within 0.1 mm, forming a mathematical model that accurately reflects the geometric characteristics of the glass. Figure 2 The three-dimensional contour of the surface shown provides the basic data for subsequent normal vector calculation.

[0034] Furthermore, based on the parameterized expression of the NURBS surface model, the normal vector is calculated for each mesh node: by solving the biased conductance of the surface in the parameter domain (u,v), a local tangent plane is constructed, and the unit normal vector n(u,v) is obtained through vector cross product. This normal vector represents the spatial orientation of the corresponding sub-region unit. A mapping relationship database between coordinate meshes and normal vectors is established, and the coordinate identifier (X, Y, Z) of each sub-region unit is associated and stored with its normal vector direction (θ,φ), where θ is the angle between the normal vector and the horizontal plane, and φ is the horizontal azimuth angle. Based on experimental data on laser film removal energy utilization, a preset angle threshold of 15° is determined, i.e., the angle between the incident direction of the laser beam and the normal vector is controlled to not exceed this threshold to avoid spot ellipticization and energy attenuation caused by oblique incidence. Tests show that for every 10° increase in the angle, the effective energy density decreases by approximately 8%.

[0035] Furthermore, during the coordinate mesh generation process, the geometric continuity of the NURBS surface is utilized to smooth the rate of change of the normal vector in adjacent sub-regions, avoiding oscillations in the mirror angle adjustment caused by abrupt changes in the normal vector. For example, in the hypercurvature transition region, such as... Figure 2 At locations where the curvature of the mid-surface changes abruptly, the mesh node density is increased to 25×25 units per square centimeter to ensure the normal vector calculation accuracy is within ±2°, providing precise guiding parameters for subsequent angle control of the reflector assembly. When the reflector assembly receives coordinate commands for a sub-region, it directly retrieves the corresponding normal vector data and converts the normal vector direction into the target angles of the reflector's lateral and longitudinal rotation axes through a coordinate transformation matrix, forming a direct mapping relationship of "surface coordinates - normal vector - mechanical angle," such as... Figure 3 The optical path control principle of the reflector group shown reduces intermediate conversion errors.

[0036] The advantage of this embodiment lies in achieving high-precision digital reconstruction of the geometric features of the glass surface through the combination of a structured light scanner and NURBS surface fitting technology. The coordinate grid mapping relationship established based on the normal vector distribution provides precise spatial guidance parameters for controlling the laser incident angle. By controlling the angle between the incident direction and the surface normal vector within a preset threshold, the proportion of vertical laser energy incidence is effectively increased, reducing energy loss and spot deformation caused by oblique incidence. This provides reliable geometric constraints for subsequent trajectory planning and energy control, ensuring energy utilization efficiency and processing uniformity in the film removal process on complex curved surfaces.

[0037] Example 3 To address the challenge of balancing trajectory planning efficiency and motion control safety during the removal of coatings from complex curved glass surfaces, this embodiment refines the multi-path collaborative trajectory planning process. By combining a tabu search algorithm with motion constraints, it achieves intelligent optimization and safe control of the trajectory sequence.

[0038] During implementation, based on Figure 2 The three-dimensional curvature and adaptive coordinate mesh of the surface shown are first divided into a set of sub-region cells containing curvature features. Each cell stores the coordinate position (X, Y, Z), normal vector parameters (θ, φ), and curvature value κ. For regions with drastic curvature changes, i.e., regions where κ ≥ 0.8 times the peak curvature, such as... Figure 2The hyperbolic transition zone, identified as a key planning area using density clustering, is prioritized for inclusion in the trajectory optimization sequence. A tabu search algorithm is employed to construct the trajectory optimization model: the sub-region access order serves as the search solution space, and the objective function comprehensively considers processing time (T) and the total adjustment amplitude of the reflector angle ΣΔθ, where Δθ represents the change in the lateral / vertical angle of the reflector corresponding to adjacent trajectories. During algorithm initialization, 100 randomly generated initial trajectories are used as the candidate solution set. The tabu list length is set to 50 to avoid repeatedly accessing invalid paths. An adaptive neighborhood search strategy dynamically adjusts the search step size to ensure convergence to the optimal solution within 100 iterations.

[0039] Furthermore, motion constraints of the reflector drive mechanism are embedded in the trajectory planning process. Taking a high-precision servo reflector group, such as the Newport XMS100 series, with a maximum angular velocity limited to 120° / s as an example, kinematic transformation is performed on the normal vector parameters of adjacent sub-region units. The angular changes Δθx and Δθy of the reflector's transverse axis (θx) and longitudinal axis (θy) are calculated, and the angular velocity is verified to ensure it does not exceed the equipment limit. Δt is the trajectory switching time interval, preset to 50ms. If the predicted angular velocity value of a certain trajectory exceeds the limit, the tabu list update mechanism is triggered, marking the adjacent sequence as tabu, and a penalty term is added to the objective function, forcing the algorithm to re-search for compliant paths.

[0040] Furthermore, regarding the setting of safety interval parameters, through... Figure 2 The glass mounting boundary coordinates, imported from the CAD design file and stored as a set of polygon vertex coordinates, are used to calculate the minimum distance between the boundary of each sub-region unit and the glass edge. A preset safety interval of 5mm is used. When the laser beam needs to switch sub-region units, the planned movement path must meet the following requirement: the vertical distance between the line connecting the starting point of the new trajectory and the ending point of the previous trajectory and the boundary of all non-removed areas must be ≥5mm. For sub-regions near the glass edge, such as... Figure 2 In the gently curvature area of ​​the right edge, a virtual safety point is set 0.5mm inside the boundary using a trajectory extension algorithm to ensure that the laser head always stays within the effective processing area during its movement, thus avoiding the risk of edge damage due to path deviation.

[0041] Furthermore, at the algorithm implementation level, the trajectory planning system calls optimization modules developed in Python, such as... Figure 2 The attached code shows a surface mesh data interface that directly maps NURBS surface parametric coordinates to the input space of the tabu search algorithm. The search priority for regions with severe curvature is adjusted using curvature weighting coefficients (0.6 ≤ ω ≤ 0.8). After trajectory sequence optimization, a G-code file containing the sub-region access order, mirror angle parameters, and safe movement path is generated, directly driving the algorithm. Figure 3 The matrix-type reflector group shown performs the film removal operation, realizing closed-loop control of "surface feature recognition - trajectory intelligent planning - motion constraint verification".

[0042] The advantage of this embodiment lies in its ability to effectively improve the processing efficiency of complex curved surfaces by prioritizing curvature-sensitive areas through a tabu search algorithm. Simultaneously, it keeps the adjustment range of the reflector angle within the safe operating range of the equipment, avoiding the risk of mechanical overload. The linkage setting of the safety interval parameter and boundary coordinates ensures the path safety of the laser beam when switching between regions, avoiding accidental damage to non-removal areas from the trajectory planning level, and providing a systematic solution for the reliable operation of automated film removal equipment.

[0043] Example 4 To address the challenge of synergistically optimizing the angle control accuracy and mechanical vibration interference of a matrix reflector array, this embodiment refines the control process for the reflector rotation axis. By combining a dual-closed-loop control strategy with a dynamic compensation mechanism, high-precision matching between theoretical angle planning and actual position feedback is achieved.

[0044] like Figure 3 The matrix reflector array shown is an example using a 3×3 array layout. Each reflector unit is a Newport XMS100-2D. The control process for its lateral rotation axis (θ-axis) and longitudinal rotation axis (φ-axis) is as follows: First, the outer loop control module... Figure 2 The target coordinates (X, Y, Z) and surface normal vector parameters (θ) of the sub-region unit shown are as follows. target ,φ target The theoretical rotation angle required for the reflector to reach the target position is calculated using coordinate transformation algorithms such as the Rodrigues transform. It is important to understand that this coordinate transformation needs to be combined with the laser source position, such as... Figure 3 The fixed coordinate point (-10, 0, 5) in the array is related to the initial position of the reflector as shown in the initial coordinates (3) of the reflector in the i-th row and j-th column of the array. i , 3 j-4 , 0), construct the path geometry relationship between incident light and reflected light to ensure the accuracy of the optical path calculated by theoretical angle.

[0045] Furthermore, the inner loop control module uses a 24-bit absolute angle encoder to acquire the actual rotation angle (θ) of the reflector in real time. real , φ realThe deviation is calculated between the angle and the theoretical angle output from the outer loop. When the absolute value of the angle deviation exceeds the preset accuracy threshold, an adaptive filtering algorithm is triggered to correct the control signal of the drive motor. This correction process introduces an inertial compensation coefficient K based on the dynamic model of the reflector assembly. The model is pre-established through finite element analysis and includes physical parameters such as the reflector mass m, moment of inertia J, and damping coefficient c. The real-time calculation formula is as follows: Where u is the motor input voltage. , These are the angular acceleration and angular velocity feedback values, respectively. By embedding K into the derivative element of the PID control algorithm, the impact of mechanical vibrations caused by the high-speed rotation of the reflector, such as high-frequency jitter within 100Hz, on angular accuracy is effectively offset.

[0046] Furthermore, at the hardware implementation level, the control signal processing unit is built using a DSP chip. The outer loop control cycle is set to 20ms to match the real-time requirements of trajectory planning, while the inner loop feedback frequency is increased to 1kHz to capture minute angular deviations. For Figure 3 The spatial distribution characteristics of the central reflector unit are as follows: the spacing between adjacent reflectors is 3mm. The control algorithm has a built-in redundancy fault tolerance mechanism. When the number of deviation corrections for a certain reflector exceeds 5 consecutive times, it automatically switches to the backup control channel to avoid single-point failure from affecting the overall optical path adjustment.

[0047] It should be noted that theoretical calculations require calling [the relevant API / tool / etc.]. Figure 2 The surface mesh data interface in the attached code directly maps the normal vector information of the NURBS surface parameterization to the control coordinate system of the mirror rotation axis, ensuring precise coupling between the dynamic adjustment of the normal vector in the curvature variation region and the mirror angle. During the real-time calculation of the inertia compensation coefficient, ambient temperature sensor data is simultaneously collected with an accuracy of ±0.5℃ to compensate for temperature drift in the dynamic model parameters, further improving control performance under complex operating conditions.

[0048] The advantage of this embodiment lies in its hierarchical error correction for mirror angle control achieved through a dual-loop control strategy. The outer loop ensures the geometric accuracy of the theoretical path, while the inner loop, combined with a dynamic compensation mechanism, effectively suppresses mechanical vibration interference, resulting in an angle control accuracy within ±0.005°. This control strategy, deeply coupled with the surface normal vector, provides a fundamental motion control guarantee for the subsequent precise focusing of the laser beam on complex curved surfaces. Especially in the film removal operation in areas with drastic curvature changes, it significantly improves the response speed and positioning accuracy of the mirror assembly to dynamic angle adjustments.

[0049] Example 5 To address the issue of energy control accuracy under varying film thicknesses, this embodiment further refines the technical solution for dynamically adjusting the laser beam energy density and irradiation time. Specifically, a pre-constructed film thickness-energy response database is stored in the system control module. This database, formed through experimental calibration, includes critical film removal energy density parameters for common film materials such as polyimide and metal oxides at different thicknesses. For example, when the film material is indium tin oxide (ITO), the database records a critical energy density range of 2.5-5.0 J / cm² for a thickness within the range of 0.1-1 μm. When the detection module determines, via spectral reflectance, that the film thickness of a sub-region unit exceeds a preset threshold, such as 0.3 μm, the control system initiates a segmented energy output strategy: first, pre-irradiation is performed at 60%-70% of the baseline critical energy density for 5-10 ms, disrupting the film's lattice structure through thermal stress; the interval time Δt is determined by the thermal diffusivity α of the film material, calculated using the following formula: Where k is a correction factor and d is the film thickness, for example, α = 5 × 10 -6 For materials with a density of m² / s, a Δt of 20 ms is set at d = 0.5 μm to ensure the diffusion and attenuation of the preceding irradiation energy to the glass substrate, avoiding heat accumulation during subsequent peak energy application. Film removal is then performed at 110%-120% of the critical peak energy density. This strategy utilizes... Figure 5 The diagram illustrating the film thickness and energy control strategy allows for visualized parameter configuration, enabling controlled delamination of films of varying thicknesses. This embodiment utilizes database-driven segmented energy output to ensure effective film removal while minimizing the risk of thermal damage to the glass substrate to below industry standards.

[0050] Example 6 To address the issue of light spot shift caused by refraction in hyperbolic glass, this embodiment optimizes the angle combination calculation of the reflector assembly by introducing a surface coordinate mapping compensation algorithm. The specific steps are as follows: First, the curvature data of the glass surface is acquired using a non-contact 3D scanner, based on... Figure 2 The three-dimensional curvature model of the curved glass shown is used to calculate the local curvature radius R of each sub-region.

[0051] Furthermore, for example Figure 3 As shown, according to Snell's Law Where n1 is the refractive index of air, n2 is the refractive index of glass, θ1 is the incident angle, and θ2 is the refraction angle, a refraction angle correction model for the incident laser beam at the curved interface is established, and the calculation formula for the longitudinal rotation angle Δφ of the reflector is derived: The algorithm parameters were pre-calibrated using Zemax optical simulation software. In a virtual environment, the optical path offset under different radii of curvature was simulated, generating a mapping lookup table containing over 2000 sets of sample data. Before the coating removal operation, the algorithm was verified through a real-machine calibration procedure: 100 feature points were marked on the surface of a standard hypercurvature glass sample, a laser beam was projected by driving the reflector group, and the deviation data between the actual spot position and the theoretical coordinates were collected. The angle compensation coefficient matrix was obtained by fitting using the least squares method, and the algorithm model was iteratively corrected. Figure 4 The principle of surface mapping optical path compensation shown is that the algorithm reduces the spot offset caused by refraction from ±150μm in the traditional method to within ±20μm through geometric optical correction, ensuring that the energy is accurately applied to the target film removal area.

[0052] Furthermore, the compensation algorithm is linked in real time with the servo control system of the reflector group, dynamically adjusting the angle parameters according to the real-time curvature data of the glass surface to achieve high-precision optical path calibration in the entire area.

[0053] Example 7 To improve the trajectory accuracy of the film removal operation, this embodiment adds a pre-scan verification step before performing the film removal process. The specific process is as follows: The control system first drives the laser module to output a low-energy laser beam at 10%-20% of its rated energy, simulating irradiation along the planned film removal trajectory. Simultaneously, a high-speed camera with a frame rate ≥1000fps installed at the end of the equipment captures images of the light spot on the glass surface. The image recognition module extracts the center coordinates of the light spot using a sub-pixel-level edge detection algorithm, matches them with the theoretical coordinates of the target sub-region unit, and calculates the positional deviations Δx and Δy in the X / Y axis directions. When the deviation in either direction exceeds a preset pixel threshold, the system automatically activates the reflector group angle calibration procedure: based on... Figure 3 The matrix-type reflector group optical path model shown establishes a nonlinear mapping relationship between the spot offset and the reflector pitch and deflection angles. A deviation compensation function f(Δx,Δy)=[Δα,Δβ] is obtained through least squares fitting, where Δα and Δβ are the angle correction amounts for each reflector unit. During calibration, nine checkerboard-distributed calibration points are selected for at least three iterations of fitting until the root mean square error of the deviation is less than 0.05mm. This pre-scan verification step effectively identifies positioning deviations caused by mechanical installation errors, optical path drift, and other factors. By real-time correction of the reflector group's angle combination, the actual positioning accuracy of the film removal trajectory is improved to ±50μm, providing a reliable position reference for subsequent high-energy operations. It is important to understand that the wavelength selection of the low-energy laser must be consistent with the actual film removal laser, such as a 1064nm infrared laser, to ensure that the optical path characteristics of the pre-scan are consistent with those of the actual operation, avoiding additional errors introduced by wavelength differences.

[0054] Example 8 To optimize the optical coverage performance of the matrix reflector array, this embodiment employs an arc-shaped array layout along the laser module's emission direction. The reflector array consists of 9-16 independently adjustable micromirror units, arranged along an arc-shaped trajectory with a radius of curvature R. a Matching the maximum radius of curvature of standard automotive glass, this design achieves this through... Figure 3 The optical path control diagram shown is used for layout simulation to ensure that the average angle between the principal optical axis of the reflector group and the normal direction of the curved surface is less than 15°, reducing the angle adjustment range. The lateral spacing d between adjacent reflectors is determined based on the divergence angle θ of the laser beam and the maximum film removal distance L, calculated as d ≤ L·tan(θ / 2). For example, when L = 500mm, d is set to 0.5mm to ensure that there is no significant energy blind zone in the adjacent spot area at the maximum working distance. The arc-shaped layout reduces the angle adjustment range of the reflector unit in the horizontal direction from ±30° of the traditional linear array to ±10°, significantly reducing the mechanical loss of the servo motor and extending the equipment life. Furthermore, the reflector surface is coated with a gold film layer with a reflectivity ≥98%, and is fixed to the arc-shaped bracket by a precision positioning mechanism. The bracket material is made of carbon fiber composite material with a low coefficient of thermal expansion, reducing the impact of ambient temperature changes on layout accuracy. This layout can cover more than 95% of automotive glass models on the market. Through the matching design of the curved curvature with the glass surface, it can reduce the response time of the reflector angle adjustment by more than 40% while ensuring the effective irradiation range, thereby improving the efficiency of the film removal operation.

[0055] Example 9 To address the overheating issue of the glass substrate during film removal, this embodiment integrates a temperature monitoring and dynamic cooling mechanism into the system. Specifically, an infrared thermal imager collects the temperature distribution on the glass surface in real time. When the local temperature exceeds the glass's heat resistance threshold, the control system immediately triggers the gas path system: a nozzle installed next to the laser head injects a low-temperature protective gas, typically compressed air or nitrogen, into the irradiated area, with the temperature controlled at 10-15°C. The gas injection flow rate Q is linearly positively correlated with the current laser energy density E, calculated using the formula Q=k·E, where k is the flow coefficient (unit: L / (J / cm²·s)). For example, when E=5J / cm², Q is set to 10L / s, ensuring dynamic matching between energy input and cooling capacity. Simultaneously, the system pauses laser irradiation of the current sub-region until the temperature sensor detects that the temperature in that region has dropped back to a safe range. The safe range threshold is dynamically adjusted based on the thermal conductivity λ of the glass material. For soda-lime glass with λ=1.2W / (m·K), the upper limit of safety is set at 200°C, with a safety margin of 80°C below the heat resistance threshold. During the cooling process, the gas path system employs a Venturi tube structure to achieve rapid gas cooling. The nozzle orifice diameter is designed to be 0.8mm to generate a high-speed airflow with a velocity ≥30m / s, ensuring that the local temperature drop rate is controlled above 50℃ / s within 50ms. Figure 5 The energy control strategy and collaborative mechanism shown involve real-time interaction between temperature monitoring data and the energy output module, forming a closed-loop feedback control that effectively avoids heat accumulation-induced glass thermal stress concentration or thermal deformation. This embodiment, through the coordinated control of dynamic cooling and energy output, keeps the maximum temperature fluctuation range of the glass surface within ±10% of the safe threshold, providing reliable thermal protection for high-power-density film removal operations.

[0056] Example 10 To address the challenge of systematizing automotive glass coating removal methods, this embodiment provides a matrix coating removal system compatible with the aforementioned methods. Through the coordinated design of hardware architecture and control logic, it achieves precise execution of each process step. The core components of the system are: a data processing unit employing an industrial-grade PC architecture, equipped with a quad-core processor and GPU acceleration module, and featuring a high-precision 3D vision sensor such as the Keyence VR-3200 to acquire the three-dimensional contour data of the automotive glass. Based on an STL model processing algorithm, the area to be removed is divided into sub-regions with side lengths of 0.5-1mm, generating point cloud data containing coordinate grid information. The trajectory planning module, based on the coating thickness detection results (e.g., obtained through a spectral confocal sensor), automatically densifies trajectory points in areas of uneven thickness to ensure that the energy distribution matches the coating characteristics.

[0057] Furthermore, the control unit uses a DSP+FPGA heterogeneous architecture as its core. Its built-in surface coordinate mapping model is derived from Snell's law in Example 6, and combines it with over 2000 sets of optical path parameters generated by Zemax optical simulation to form a real-time calculation lookup table. The mirror angle compensation algorithm is embedded in the FPGA using a hardware description language, achieving microsecond-level response. After receiving the trajectory coordinates from the data processing unit, this unit first calculates the incident refraction angle of each sub-region based on the curvature of the glass surface. Through matrix operations, it decomposes this angle into the lateral rotation angle α and longitudinal rotation angle β of each unit in the mirror group, outputting differential signals to drive servo drive motors, such as the Panasonic MINAS A6 series, with a positioning accuracy of ±0.001°. A 20-bit absolute angle encoder, such as the Heidenhain ERN 1387, is configured at the motor shaft end to provide real-time angle data feedback, forming a closed-loop control. For laser energy adjustment, the control unit outputs a PWM signal to the Q-switch drive circuit of the laser module based on the film thickness-energy response database in Example 5, achieving energy density adjustment with an accuracy of 0.1 J / cm².

[0058] Furthermore, the matrix-type reflector assembly consists of 12 gold-plated reflector units with a surface roughness Ra≤5nm, arranged in an arc array with a radius of curvature of 1500mm along the laser module's emission direction to match the maximum curvature of standard automotive glass. The lateral spacing between adjacent units is 0.8mm, calculated based on a 1mrad laser divergence angle and a 500mm working distance. Each reflector unit is connected to a servo motor via a biaxial flexible hinge structure. The rotation axis is supported by air bearings, with axial runout ≤1μm, ensuring high rigidity and low friction for angle adjustment. The laser module uses a 1064nm pulsed fiber laser with an adjustable pulse width of 50-200ns. The emission optical axis is aligned with the central axis of the reflector assembly via a six-dimensional adjustment frame, with a calibration accuracy ≤50μm, ensuring that the beam's pointing deviation after reflection is in the sub-millimeter range.

[0059] It is important to understand that the various modules of the system achieve data synchronization via the industrial Ethernet EtherCAT protocol, with a control cycle set to 100μs to meet the requirements of high-speed dynamic adjustment. The surface coordinate mapping relationship model and the reflector angle compensation algorithm are calibrated using a standard double-curvature glass template before leaving the factory, stored in the non-volatile memory of the control unit, and support online updates based on pre-scan data from the actual machine during field operations.

[0060] The advantage of this embodiment is that, through modular hardware design and integrated control algorithms, a complete film removal system covering data acquisition, trajectory planning, energy control, and optical path calibration has been constructed, realizing the engineering implementation of each technical solution and ensuring the automation and high-precision execution of the film removal process for complex curved glass.

[0061] Although the present invention has been specifically described above with reference to preferred embodiments, it should be understood that the present invention is not limited to the embodiments described above. Various modifications and variations can be made by those skilled in the art without departing from the spirit of the present invention, and such modifications and variations should fall within the scope defined by the appended claims and their equivalents.

Claims

1. A method for removing film from automotive glass, characterized in that, The steps of the method include: A three-dimensional contour scan is performed on the area of ​​the automotive glass to be coated. Based on the scan data, a spatial coordinate grid is constructed on the glass surface. The area to be coated is divided into multiple sub-region units with unique coordinate identifiers. The accuracy of the coordinate grid is dynamically adjusted according to the curvature of the glass surface. The grid density in areas with greater curvature is higher than that in planar areas. Based on the spatial distribution of the sub-region units and the laser energy requirements, a multi-path collaborative film removal trajectory planning scheme is generated. In the scheme, each trajectory corresponds to the angle combination of a set of mirrors in the matrix mirror group, and the overlap rate of the laser irradiation area of ​​adjacent trajectories is not less than a preset threshold to ensure the integrity of film removal. By controlling the lateral and longitudinal rotation axes of each mirror in the matrix mirror group, the parallel laser beam emitted by the laser module is sequentially guided to the top of each sub-region unit. The lateral rotation axis is used to adjust the horizontal incident angle of the laser beam to cover the width direction of the glass, and the longitudinal rotation axis is used to match the curvature of the glass to adjust the vertical incident angle of the laser beam. Based on the film thickness parameters of each sub-region unit, the laser beam energy density and irradiation time of the corresponding reflector group are dynamically adjusted. After the reflector completes the angle calibration, the laser module is controlled to emit pulsed laser, and the laser beam is focused onto the film surface of the target sub-region unit through the angle combination of the reflector group.

2. The method for removing film from automotive glass as described in claim 1, characterized in that, In the step of performing a three-dimensional contour scan of the area to be decoted on the automotive glass, a structured light scanner is used to acquire point cloud data of the glass surface. The point cloud data is then converted into a parameterized NURBS surface model using a surface fitting algorithm. Based on the normal vector distribution of the NURBS surface model, the spatial orientation of each sub-region unit is determined. In turn, the correspondence between the coordinate grid and the normal vector of the glass surface is established to ensure that the angle between the incident direction of the laser beam and the normal vector of the corresponding sub-region unit does not exceed a preset angle threshold. The preset angle threshold is determined based on the energy utilization parameter of the laser decoction.

3. The method for removing film from automotive glass as described in claim 1, characterized in that, In the step of generating a multi-path collaborative film removal trajectory planning scheme, a tabu search algorithm is introduced to optimize the trajectory order and plan sub-region units in areas with drastic curvature changes. At the same time, the adjustment range of the mirror angle of adjacent trajectories is avoided from exceeding the maximum angular velocity limit of the mirror drive mechanism. The trajectory planning scheme also includes a safety interval parameter to ensure that the movement path of the laser beam when switching sub-region units avoids the non-film removal area at the edge of the glass. The safety interval parameter is preset according to the installation boundary coordinates of the glass.

4. The method for removing film from automotive glass as described in claim 1, characterized in that, In the steps of controlling the lateral and longitudinal rotation axes of each reflector in the control matrix reflector group, a dual closed-loop control strategy is adopted. The outer loop calculates the theoretical rotation angle of the reflector based on the target coordinates of the sub-region unit, while the inner loop provides real-time feedback of the actual position of the reflector through an angle encoder. When the deviation between the actual position and the theoretical angle exceeds a preset accuracy, an adaptive filtering algorithm is triggered to correct the control signal of the drive motor. During the correction process, an inertia compensation coefficient is introduced to offset the influence of mechanical vibration during the rotation of the reflector on the angle accuracy. The inertia compensation coefficient is calculated in real-time through the dynamic model of the reflector group.

5. The method for removing film from automotive glass as described in claim 1, characterized in that, In the step of dynamically adjusting the laser beam energy density and irradiation time of the corresponding reflector group, a film thickness-energy response database is established in advance. The database contains the correspondence between the critical film removal energy density and thickness of different film materials. When the film thickness of the sub-region unit is detected to exceed the preset threshold, a segmented energy output strategy is adopted. Pre-irradiation is performed at the basic energy density to destroy the film structure. After a preset time interval, the film removal is completed at the peak energy density. The preset time is determined according to the thermal diffusivity of the film material to avoid damage to the glass substrate caused by the accumulation of laser energy.

6. The method for removing film from automotive glass as described in claim 1, characterized in that, In the angle combination calculation step of the mirror group, a curved surface coordinate mapping compensation algorithm is introduced to perform geometric correction on the optical refraction effect of the double-curvature glass after the laser beam passes through the mirror. The steps include: calculating the refraction angle of the laser beam when it is incident on the glass surface based on the radius of curvature of the glass, and adjusting the longitudinal rotation angle of the mirror based on Snell's law to compensate for the spot offset caused by refraction. The parameters of the curved surface coordinate mapping compensation algorithm are pre-calibrated by optical simulation software and are calibrated on a real machine before the film removal operation.

7. The method for removing film from automotive glass as described in claim 1, characterized in that, Before performing the film removal operation, a pre-scanning verification step is added. The planned trajectory is simulated by a low-energy laser beam, and the position image of the light spot on the glass surface is acquired by a high-speed camera. The actual position of the light spot is matched with the coordinates of the target sub-region unit. When the matching error exceeds a preset pixel threshold, the angle calibration program of the reflector group is automatically triggered. The calibration program fits the relationship model between the light spot offset and the reflector angle deviation using the least squares method to realize the real-time correction of the mapping relationship model.

8. The method for removing film from automotive glass as described in claim 1, characterized in that, The matrix reflector array is arranged in an arc-shaped array along the emission direction of the laser module. The lateral spacing between adjacent reflectors is determined according to the divergence angle of the laser beam and the maximum film removal distance, ensuring that the effective irradiation range of the reflector array covers automotive glass of different sizes. The radius of curvature of the arc array matches the maximum curvature radius of the standard automotive glass to reduce the angle adjustment range of the reflectors.

9. The method for removing film from automotive glass as described in claim 1, characterized in that, During the film removal process, the temperature distribution data of the glass surface is collected in real time. When the local temperature is detected to exceed the heat resistance threshold of the glass, a dynamic cooling mechanism is triggered. Low-temperature protective gas is sprayed into the laser irradiation area through the gas path system. The spray flow rate of the gas is linearly positively correlated with the laser energy density. At the same time, the irradiation of the current sub-region unit is paused until the temperature drops back to the safe range. The threshold of the safe range is dynamically adjusted according to the thermal conductivity coefficient of the glass material.

10. A matrix film removal system for implementing the automotive glass film removal method as described in claims 1-9, characterized in that, The system includes: The data processing unit is used to acquire the three-dimensional contour data of the automotive glass, divide the area to be removed into sub-region units and generate a coordinate grid, and generate a film removal trajectory planning scheme based on the coordinate grid and film thickness parameters. The control unit is used to calculate the lateral and longitudinal rotation angles of each mirror in the matrix mirror group according to the trajectory planning scheme, generate drive signals to control the dual-axis rotation mechanism of the mirror, and adjust the energy output parameters of the laser module according to the film thickness parameters. A matrix-type reflector assembly includes multiple reflector units arranged in an arc-shaped array. Each reflector unit is connected to an independent horizontal rotation axis and a vertical rotation axis. The horizontal rotation axis and the vertical rotation axis are equipped with an angle encoder and a servo drive motor to achieve high-precision adjustment of the reflector angle. A laser module is used to emit a pulsed laser beam. The emission direction of the laser beam is aligned with the central axis of the matrix reflector group. The laser beam is guided to the target sub-region unit by the angle combination of the reflector group. The control unit incorporates a curved surface coordinate mapping model and a reflector angle compensation algorithm to correct optical path deviations caused by the refraction of the laser beam by the curved glass surface.

Images