Intelligent method for removing half-hole burrs of AIOT module board based on laser ablation

Abstract

The application relates to the field of electronic component manufacturing, and particularly discloses an AIOT module plate semi-hole burr intelligent removal method based on laser ablation, aiming to solve the problems of low efficiency, easy damage to the substrate, uneven removal, lack of intelligent identification and accurate adaptation, low control precision and real-time feedback mechanism and the like of the existing AIOT module plate semi-hole burr removal method. The method intelligently identifies, accurately positions and three-dimensionally quantizes micron-level burrs through high-precision visual detection technology combined with deep learning; utilizes an artificial intelligence algorithm to intelligently generate customized laser ablation parameters according to burr characteristics and material physical properties; and performs non-contact removal through a precise laser ablation system in cooperation with an optimized scanning path. The application can realize non-contact, high-efficiency, high-precision and lossless intelligent removal of burrs by adopting the above technical solution, and greatly improves the product yield, electrical performance and manufacturing level.

Description

Technical Field

[0001] This invention belongs to the field of laser processing technology, specifically relating to an intelligent method for removing burrs from half-holes of AIOT module boards based on laser ablation. Background Technology

[0002] With the rapid development of Artificial Intelligence of Things (AIOT) technology, the manufacturing precision and reliability of AIoT module boards, as core components of smart hardware, are becoming increasingly critical. During the manufacturing process, especially in the half-hole machining stage that enables device interconnection, microburrs are inevitably generated. These burrs severely affect the electrical performance, signal integrity, and subsequent assembly precision of the module board.

[0003] However, existing burr removal methods, such as traditional mechanical grinding, chemical cleaning, or manual finishing, generally suffer from low efficiency, easy damage to the substrate, uneven processing, or environmental pollution, making it difficult to meet the increasingly stringent precision manufacturing requirements of AIoT module boards. While laser ablation technology, a non-contact processing method, shows promise for burr removal, current technologies typically rely on preset parameters and experience-based adjustments, lacking the ability to intelligently identify and precisely adapt to burrs of different shapes and sizes. This results in low precision control of the ablation process, easily leading to defects such as over-burning, under-burning, or incomplete removal, thus affecting product yield and consistency. Furthermore, the lack of a real-time intelligent feedback mechanism makes it difficult to address the challenges of the randomness and diversity of burrs during production.

[0004] Therefore, a smart method for removing burrs from half-holes in AIOT module boards based on laser ablation is desired. Summary of the Invention

[0005] To address the challenges of inefficient burr removal during the manufacturing of AIoT module boards with half-holes, which can damage the substrate, result in uneven processing, or pollute the environment, and to overcome the shortcomings of existing laser ablation technologies such as lack of intelligent recognition and precise adaptation to burrs of different shapes and sizes, low control precision, and susceptibility to over-burning, under-burning, or incomplete removal, as well as the lack of a real-time intelligent feedback mechanism to handle the randomness and diversity of burrs, this application is proposed. The embodiments of this application provide an intelligent burr removal method for half-holes of AIoT module boards based on laser ablation. This method utilizes high-precision visual inspection technology to achieve intelligent recognition and precise positioning of burrs, combines this with artificial intelligence algorithms to intelligently generate customized laser ablation parameters, and executes the removal operation through a precision laser ablation system. Simultaneously, a real-time feedback mechanism is introduced for ablation effect evaluation and parameter adaptive adjustment, thereby effectively removing burrs from half-holes of AIoT module boards, improving product yield, electrical performance, and assembly accuracy. This achieves non-contact, high-efficiency, high-precision, and non-destructive intelligent burr removal, significantly improving the manufacturing level and consistency of AIoT module boards and meeting the requirements of precision manufacturing.

[0006] According to one aspect of this application, a method for intelligent removal of burrs in half-holes of an AIOT module board based on laser ablation is provided, comprising:

[0007] Acquire high-resolution image data of the half-hole area of ​​the AIOT module board to be processed;

[0008] Intelligent analysis of high-resolution image data to identify burrs and quantify burr characteristic parameters;

[0009] Laser ablation parameters are intelligently generated based on burr characteristic parameters and a preset material ablation model.

[0010] The burrs in the semi-hole area are removed by laser ablation using a laser ablation system.

[0011] The semi-hole area after laser ablation is inspected to evaluate the burr removal effect and generate a feedback signal;

[0012] The laser ablation parameters are adaptively adjusted based on the feedback signal until the burr removal meets the preset standard.

[0013] According to one embodiment of this application, obtaining high-resolution image data of the half-hole region of the AIOT module board to be processed specifically includes:

[0014] Place the AIoT module board to be processed on the high-precision positioning platform;

[0015] Start a high-resolution industrial camera array to acquire images of the half-hole area of ​​the AIOT module board from multiple angles and focal planes;

[0016] During image acquisition, a synchronously controlled adjustable light source system provides uniform and shadowless illumination to ensure image quality and feature visibility.

[0017] The acquired images are transmitted to the image processing unit for initial integration and correction.

[0018] According to another embodiment of this application, intelligent analysis of high-resolution image data to identify burrs and quantify burr feature parameters specifically includes:

[0019] Perform image preprocessing operations on image data, including noise reduction filtering, brightness and contrast enhancement, geometric correction, and distortion correction, to optimize image quality;

[0020] A deep learning-based image segmentation network model is used to perform pixel-level segmentation of burrs in the half-hole region, thereby accurately distinguishing burrs from the substrate body.

[0021] Connectivity analysis is performed on the segmented burr regions to extract the geometric features of each independent burr, including but not limited to the area, perimeter, maximum length, average width, and shape factor of the burr.

[0022] By combining 3D reconstruction algorithms, the 3D morphology of burrs is reconstructed based on multi-angle image data, thereby quantifying the 3D feature parameters of burrs such as height, volume and surface roughness.

[0023] Based on the geometric features and three-dimensional feature parameters of the burr, a set of burr feature parameters is generated, which includes the burr location coordinates, size, shape, and estimated volume.

[0024] According to another embodiment of this application, the intelligent generation of laser ablation parameters based on burr characteristic parameters and a preset material ablation model specifically includes:

[0025] Physical properties of the AIOT module board substrate material and burr material were obtained from the material database, including thermal conductivity, specific heat capacity, density, melting point, vaporization point and laser absorption rate.

[0026] Input the burr feature parameter set and material physical properties into a laser ablation model based on finite element analysis or neural network training;

[0027] The laser ablation model calculates initial ablation parameters such as laser output power, laser pulse width, laser pulse frequency, laser scanning speed, laser spot size, and focal length based on the input data.

[0028] Using a multi-objective optimization algorithm, with the optimization objectives of maximizing burr removal efficiency, minimizing substrate damage, and minimizing ablation residue, the initial ablation parameters are iteratively optimized to generate a set of optimal laser ablation parameter combinations.

[0029] Generate customized laser ablation path planning data for each identified burr. The path planning data includes the laser scanning sequence, scanning direction and scanning trajectory to ensure that the laser accurately acts on the burr area and avoids the non-burr area.

[0030] According to another embodiment of this application, the laser ablation removal of burrs in the half-hole region using a laser ablation system specifically includes:

[0031] The optimal combination of laser ablation parameters and laser ablation path planning data are loaded into the control unit of the laser ablation system;

[0032] A high-precision motion control platform drives the AIOT module board or laser head to move to the initial ablation position in the target half-hole area;

[0033] The laser is activated, and based on the loaded laser ablation parameter combination and path planning data, the laser beam is precisely focused and applied to the burr area through the galvanometer scanning system and dynamic focusing system.

[0034] During the ablation process, the laser beam performs high-precision scanning along a preset path, vaporizing or melting the burr material through photothermal or photochemical effects;

[0035] The laser ablation system features high repeatability and high energy stability, ensuring the accuracy and consistency of the ablation process.

[0036] According to a further embodiment of this application, detecting the semi-hole region after laser ablation to evaluate the burr removal effect and generating a feedback signal specifically includes:

[0037] After the laser ablation operation is completed, the high-resolution industrial camera array is restarted to acquire images of the ablated half-hole area.

[0038] Intelligent analysis is performed on the image data acquired in the second phase to determine whether there are residual burrs and to quantify the characteristic parameters of the residual burrs.

[0039] The characteristic parameters of the residual burrs are compared with the preset acceptance standards to determine whether the burr removal meets the requirements.

[0040] If residual burrs exist or the ablation effect is not ideal, a feedback signal containing the location and size of the residual burrs and the degree of insufficient ablation will be generated.

[0041] If the burrs have been completely removed and the substrate is undamaged, a completion signal indicating that the substrate is qualified is generated.

[0042] According to a further embodiment of this application, adaptively adjusting the laser ablation parameters based on feedback signals until burr removal meets a preset standard specifically includes:

[0043] When a feedback signal indicating residual burrs or unsatisfactory ablation effect is received, it is input to the intelligent parameter adjustment module.

[0044] The intelligent parameter adjustment module uses an adaptive control algorithm or reinforcement learning model to recalculate and optimize ablation parameters such as laser output power, laser pulse width, and laser scanning speed based on the characteristic parameters of residual burrs in the feedback signal and the degree of insufficient ablation, so as to improve the removal effect of the next ablation.

[0045] Repeat the steps of laser ablation to remove burrs in the half-hole area using a laser ablation system, and then inspect the half-hole area after laser ablation to evaluate the burr removal effect and generate a feedback signal.

[0046] The ablation operation stops only when a completion signal indicating that the burr removal meets the preset standard is received, and the next half of the hole area is processed.

[0047] According to one embodiment of this application, the method further includes: constructing and maintaining a database containing various AIOT module board material properties, different burr morphology characteristics, and corresponding optimal laser ablation parameters; when generating laser ablation parameters, relevant parameters are preferentially matched and obtained from this database as initial values, thereby accelerating the parameter optimization process. The database is continuously updated and learned through historical production data and experimental data to improve the accuracy and robustness of parameter generation.

[0048] Compared with existing technologies, the intelligent removal method for half-hole burrs in AIOT module boards based on laser ablation provided in this application achieves accurate burr identification, customized ablation, non-destructive removal, and adaptive effect by introducing high-precision visual detection, deep learning algorithms, intelligent parameter optimization, and real-time closed-loop feedback control mechanism.

[0049] Specifically, the beneficial effects of this application include, but are not limited to:

[0050] First, by acquiring high-resolution image data and using deep learning-based intelligent analysis, micron-level burrs can be accurately identified, located, and quantified in three dimensions, overcoming the problem of insufficient accuracy in burr identification by traditional methods and providing a data foundation for subsequent precise ablation.

[0051] Secondly, by combining material physical properties and intelligent optimization algorithms to generate customized laser ablation parameters, the laser energy is precisely applied to the burrs rather than the substrate, which greatly reduces the risk of overburning, underburning or substrate damage and improves the level of fine control of the ablation process.

[0052] Furthermore, the use of a high-precision laser ablation system, combined with optimized scanning path planning, enables non-contact, efficient, and clean removal of burrs, avoiding the inefficiencies, environmental pollution, and secondary damage associated with traditional methods such as mechanical grinding and chemical cleaning.

[0053] In addition, the introduction of a real-time detection and feedback mechanism enables immediate evaluation of the ablation effect and adaptive adjustment of laser parameters based on feedback signals, ensuring thorough burr removal that meets preset standards. This effectively addresses the challenges of the randomness and diversity of burrs during the production process, significantly improving product yield and consistency.

[0054] Ultimately, the method provided by this invention can achieve a high degree of automation and intelligence in the process of removing burrs from half-holes of AIoT module boards, significantly reducing the need for manual intervention, improving production efficiency, and reducing manufacturing costs. It provides an innovative solution for precision manufacturing in the AIoT industry and has significant technological progress and economic and social benefits. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention;

[0056] Figure 2 This is a schematic diagram of the architecture for adaptive adjustment of laser ablation parameters by feedback signal in this invention. Detailed Implementation

[0057] In the following description, specific terminology and structures are used to illustrate embodiments of the invention for ease of understanding. However, these descriptions are not intended to limit the invention, and those skilled in the art will understand that various modifications and substitutions can be made without departing from the spirit and scope of the invention. The accompanying drawings are for illustrative purposes only, are not drawn to scale, and should not be construed as limiting the invention.

[0058] Please refer to Figure 1 and Figure 2 This application provides an intelligent method for removing burrs from half-holes in an AIoT (Artificial Intelligence of Things) module board based on laser ablation. The overall process includes the following main steps: First, acquiring high-resolution image data of the half-hole area of ​​the AIoT module board to be processed; second, intelligently analyzing the high-resolution image data to identify burrs and quantify burr feature parameters; third, intelligently generating laser ablation parameters based on the burr feature parameters and a preset material ablation model; subsequently, removing the burrs in the half-hole area using a laser ablation system; next, detecting the half-hole area after laser ablation to evaluate the burr removal effect and generate a feedback signal; finally, adaptively adjusting the laser ablation parameters based on the feedback signal until the burr removal meets the preset standard.

[0059] In the aforementioned intelligent removal method for half-hole burrs on AIoT module boards based on laser ablation, acquiring high-resolution image data of the half-hole area of ​​the AIoT module board to be processed specifically includes the following sub-steps: placing the AIoT module board to be processed on a high-precision positioning platform; activating a high-resolution industrial camera array to acquire images of the half-hole area of ​​the AIoT module board from multiple angles and focal planes; during image acquisition, simultaneously controlling an adjustable light source system to provide uniform and shadowless illumination to ensure image quality and feature visibility; and transmitting the acquired multiple image data to an image processing unit for preliminary integration and correction. It should be understood that this step aims to provide high-quality, high-precision visual input data for subsequent burr identification and removal. The half-hole structure of the AIoT module board is complex and its size is tiny; any minute burr can affect its electrical performance and assembly accuracy. Therefore, acquiring sufficiently detailed and comprehensive image data is the foundation of the entire intelligent removal process. Specifically, the AIoT module board to be processed is placed on a high-precision positioning platform. This platform boasts micron-level repeatability and sub-micron-level motion resolution. It typically combines an XY-axis motion stage driven by a precision linear motor with a Z-axis lifting stage, and uses a closed-loop optical or magnetic scale for position feedback to ensure the module board can be precisely moved and stably fixed in the center of the camera array's field of view in three-dimensional space. The positioning platform also integrates vacuum adsorption or mechanical clamping devices to prevent displacement or vibration of the module board during image acquisition or laser ablation. A high-resolution industrial camera array is then activated to acquire images of the semi-aperture area of ​​the AIoT module board from multiple angles and focal planes. The camera array typically consists of two or more industrial-grade cameras with high resolution (e.g., 5 million to 20 million pixels) and high frame rate (e.g., 50 to 200 frames per second), equipped with apochromatic microscope objectives or telecentric lenses to eliminate image distortion and ensure measurement accuracy. The camera array works collaboratively through a synchronous trigger controller to achieve synchronized image acquisition among multiple cameras, avoiding image inconsistencies caused by time differences. Multi-angle acquisition can be achieved by adjusting the camera's tilt angle relative to the module board or by rotating the positioning platform. This aims to capture burr information inside and at the edges of the semi-hole from different perspectives, solving the occlusion problem under a single viewpoint. Multi-focal plane acquisition, on the other hand, involves fine-tuning the camera's focal length or the module board height along the Z-axis to acquire a series of images with different focal planes. This allows for subsequent 3D information reconstruction, ensuring that the details of the burrs in the depth direction are also clearly recorded.During image acquisition, a synchronously controlled adjustable light source system provides uniform and shadowless illumination, ensuring image quality and feature visibility. This adjustable light source system typically includes multiple illumination modes such as ring lights, coaxial lights, and backlights. The light source intensity and angle can be programmed and controlled according to the module board material characteristics and burr reflection conditions. By optimizing the illumination, the contrast between burrs and the substrate is maximized, specular reflection and shadow interference are suppressed, and the texture, edges, and morphological features of the burrs are clearly presented in the image, providing high-quality image data for subsequent intelligent analysis. Multiple acquired image data are then transmitted to an image processing unit for initial integration and correction. The image processing unit is typically an industrial computer equipped with a high-performance graphics processing unit (GPU). Image data is transmitted in real time via a high-speed data interface (e.g., gigabit Ethernet or camera link). The initial integration includes timestamp alignment of multiple images, image stitching and fusion (for data acquired from multiple angles), and preliminary image correction, such as lens distortion correction, brightness uniformity correction, and white balance adjustment based on preset correction parameters, thereby forming a unified, high-quality raw image dataset, laying a solid foundation for subsequent intelligent spur recognition and quantization.

[0060] In the aforementioned intelligent removal method for half-hole burrs in AIoT module boards based on laser ablation, the intelligent analysis of high-resolution image data to identify burrs and quantify burr feature parameters specifically includes: performing image preprocessing operations on the image data, including noise reduction filtering, brightness and contrast enhancement, geometric correction, and distortion correction, to optimize image quality; using a deep learning-based image segmentation network model to perform pixel-level segmentation of burrs in the half-hole region, thereby accurately distinguishing burrs from the substrate body; performing connected component analysis on the segmented burr region to extract the geometric features of each independent burr, including but not limited to the burr area, perimeter, maximum length, average width, and shape factor; combining a 3D reconstruction algorithm to reconstruct the 3D morphology of the burrs based on multi-angle image data, thereby quantifying the burr's height, volume, surface roughness, and other 3D feature parameters; and generating a burr feature parameter set containing the burr's location coordinates, size, shape, and estimated volume based on the burr's geometric features and 3D feature parameters. It should be understood that this step is crucial for achieving precise laser ablation, as it transforms the original image information into quantified burr features that can be understood and processed by the intelligent system. Traditional image processing methods often struggle to achieve high-precision and robust recognition and quantization when faced with complex burr morphologies, minute dimensions, and variable backgrounds. Specifically, image preprocessing operations are performed on the image data, including denoising filtering, brightness and contrast enhancement, geometric correction, and distortion correction. Denoising filtering employs algorithms such as Gaussian filtering, median filtering, or bilateral filtering to effectively remove random noise introduced during image acquisition, smoothing the image while preserving edge details of the burrs as much as possible. Brightness and contrast enhancement uses methods such as histogram equalization, gamma correction, or local contrast enhancement (e.g., contrast-limited adaptive histogram equalization) to improve the distinction between burrs and the background, making even subtle burr features more prominent. Geometric correction corrects non-perpendicular projections caused by camera mounting angles or module placement deviations, correcting the image to a standard view through perspective transformation and other methods. Distortion correction utilizes pre-calibrated camera parameters to accurately compensate for radial and tangential distortions, ensuring that the geometric dimensions in the image are highly consistent with the actual physical dimensions, laying the foundation for subsequent accurate measurements. A deep learning-based image segmentation network model is used to perform pixel-level segmentation of burrs in the semi-aperture region, thereby accurately distinguishing burrs from the substrate body. The deep learning model can employ advanced semantic segmentation or instance segmentation network architectures such as U-Net, DeepLabV3+, or MaskR-CNN. The model is trained on a large dataset of labeled burr images, learning the complex visual patterns of burrs and outputting a probability map of each pixel belonging to either a burr or the substrate, or directly generating a pixel-level mask. Pixel-level segmentation enables precise capture of burr contours, accurately identifying even extremely small or irregularly shaped burrs. Connectivity analysis is performed on the segmented burr regions to extract the geometric features of each individual burr.Connectivity analysis algorithms (e.g., based on breadth-first search or depth-first search) can identify adjacent spur regions in an image as independent spur individuals. For each identified independent spur, the system extracts a series of geometric features, including but not limited to area (the number of pixels occupied by the spur multiplied by the actual area per unit pixel), perimeter (the length of the spur outline), maximum length (the maximum projected length of the spur in any direction, also known as the Fretter diameter), average width (an approximation of the area divided by the maximum length), and shape factor (four π multiplied by the area and then divided by the square of the perimeter, used to quantify the shape regularity of the spur; for example, the shape factor of a circular spur is close to one). These geometric features together constitute a two-dimensional morphological description of the spur. Combined with a three-dimensional reconstruction algorithm, the spur is reconstructed in three dimensions based on multi-angle image data, thereby quantifying three-dimensional feature parameters such as the spur's height, volume, and surface roughness. Three-dimensional reconstruction can employ methods such as stereo vision technology (based on parallax calculation from multiple images), structured light scanning (projecting a known pattern and analyzing its deformation), or multi-view photogrammetry. These methods can generate 3D point cloud data or mesh models of the burr surface, allowing for precise calculation of the burr's average height, maximum height, total volume, and surface roughness (e.g., by calculating changes in surface normals or local height fluctuations). Based on the burr's geometric features and 3D characteristic parameters, a burr feature parameter set is generated, containing the burr's location coordinates, size, shape, and estimated volume. This parameter set is stored in a structured data format (e.g., JSON or XML) and includes a unique identifier for each identified burr, its precise 3D position (X, Y, Z coordinates) in the module board coordinate system, length, width, height dimensions, area, volume, perimeter, shape factor, and all other quantified features. This parameter set serves as the direct input for subsequent intelligent generation of laser ablation parameters.

[0061] In the aforementioned intelligent removal method for half-hole burrs in AIoT module boards based on laser ablation, the intelligent generation of laser ablation parameters according to burr characteristic parameters and a preset material ablation model specifically includes: obtaining the physical properties of the substrate material and burr material of the AIoT module board from a material database, including thermal conductivity, specific heat capacity, density, melting point, vaporization point, and laser absorptivity; inputting the burr characteristic parameter set and material physical properties into a laser ablation model based on finite element analysis or neural network training; the laser ablation model calculating initial ablation parameters such as laser output power, laser pulse width, laser pulse frequency, laser scanning speed, laser spot size, and focal length based on the input data; using a multi-objective optimization algorithm, with the optimization objectives of maximizing burr removal efficiency, minimizing substrate damage, and minimizing ablation residue, iteratively optimizing the initial ablation parameters to generate a set of optimal laser ablation parameter combinations; and generating customized laser ablation path planning data for each identified burr, including the laser scanning sequence, scanning direction, and scanning trajectory, to ensure that the laser accurately acts on the burr area and avoids non-burr areas. As can be understood, this step is the core of achieving customized, high-precision laser ablation. It combines the characteristic information of burrs with the physical properties of the material, and uses intelligent models to predict and optimize laser action parameters to remove burrs to the greatest extent while protecting the substrate. Traditional methods often use fixed or empirical laser parameters, which are difficult to adapt to the diversity of burrs. Specifically, the physical properties of the substrate material and burr material of the AIoT module board are obtained from a material database. The material database is a structured digital storage system that contains detailed physical parameters of various commonly used substrate materials (e.g., FR-4, polyimide, ceramic substrates) and common burr materials (e.g., copper, tin, nickel, polymer residues). These parameters include thermal conductivity (a physical quantity that measures a material's ability to conduct heat, measured in watts per meter Kelvin), specific heat capacity (a physical quantity that measures a material's ability to absorb heat, measured in joules per kilogram Kelvin), density (the mass-to-volume ratio of a material, measured in kilograms per cubic meter), melting point and vaporization point (the temperatures at which a material undergoes a phase transition, measured in Kelvin or degrees Celsius), and laser absorptivity (the proportion of laser energy absorbed by a material at a specific wavelength, dimensionless). These physical properties are crucial for predicting the interaction between lasers and materials; for example, materials with high absorptivity are more easily ablated at the same laser energy. The set of burr feature parameters (from the burr geometry and 3D features obtained in the previous step) and the material's physical properties are input into a laser ablation model trained based on finite element analysis or a neural network. The laser ablation model can be a physics-based finite element analysis (FEA) model, which numerically simulates the heat transfer, phase transition, and material removal processes of materials under laser irradiation by discretizing the physical space and time.The FEA model can predict the temperature field distribution, the formation and expansion of melting and vaporization zones within a material based on laser parameters (such as power, pulse width, and spot size) and material properties, thereby assessing the ablation depth, width, and thermal impact on the surrounding substrate. Another model is a machine learning model trained on neural networks, such as multilayer perceptrons, convolutional neural networks, or recurrent neural networks. This model establishes a nonlinear mapping relationship between input and output by learning from a large amount of historical ablation experimental data (inputs are burr characteristics, material properties, and laser parameters; output is the ablation effect). The laser ablation model calculates initial ablation parameters such as laser output power, laser pulse width, laser pulse frequency, laser scanning speed, laser spot size, and focal length based on the input data. Based on the specific size, shape, and material of the burr, combined with the physical properties of the substrate material, the model initially calculates a set of laser parameters that can initially achieve burr removal. For example, for large metal burrs, higher laser output power and longer pulse widths may be required; for small polymer burrs close to a sensitive substrate, lower power, ultrashort pulses, and smaller spot sizes may be needed. A multi-objective optimization algorithm is employed to iteratively optimize the initial ablation parameters, aiming to maximize burr removal efficiency, minimize substrate damage, and minimize ablation residue. This generates a set of optimal laser ablation parameter combinations. Multi-objective optimization algorithms, such as Non-Dominated Sorting Genetic Algorithm-II (NSGA-II), Particle Swarm Optimization (PSO), or Differential Evolutionary Algorithm, are designed to resolve multiple conflicting optimization objectives. For example, increasing removal efficiency may increase the risk of substrate damage, while reducing damage may decrease efficiency. The optimization algorithm iteratively searches the parameter space to generate a series of Pareto optimal solutions, meaning solutions that are not inferior to other solutions on any given objective and are superior to other solutions on at least one objective. The system selects one of the optimal parameter combinations based on preset weights or priorities, achieving the best balance between burr removal efficiency, substrate damage control, and ablation residue. Burr removal efficiency is typically measured by the volume of burrs removed per unit time (cubic micrometers per second), substrate damage minimization is quantified by limiting the etching depth or heat-affected zone size on the substrate surface, and ablation residue minimization is evaluated by the volume or area percentage of residual burrs after ablation. Customized laser ablation path planning data is generated for each identified burr. This data includes the laser scanning sequence, direction, and trajectory, ensuring the laser precisely targets the burr area and avoids non-burr regions. The path planning module generates a refined scanning path based on the precise location, 3D morphology, and optimal laser parameters for each burr. For example, a single scan along its axis might be used for long, thin burrs, while a spiral or zigzag scan might be used for clustered burrs. The path planning algorithm considers the laser spot size, ablation depth, and thermal diffusion range to ensure the laser energy is precisely applied to the burr body while minimizing accidental damage to the surrounding substrate.Path planning data is typically output in G-code or other machine-readable formats to guide the motion control unit of the laser ablation system.

[0062] In the aforementioned intelligent burr removal method for half-holes in AIoT module boards based on laser ablation, the laser ablation system removes burrs from the half-hole area by: loading the optimal combination of laser ablation parameters and laser ablation path planning data into the control unit of the laser ablation system; a high-precision motion control platform drives the AIoT module board or laser head to move to the initial ablation position in the target half-hole area; the laser is activated, and according to the loaded combination of laser ablation parameters and path planning data, the laser beam is precisely focused and applied to the burr area through a galvanometer scanning system and a dynamic focusing system; during the ablation process, the laser beam performs high-precision scanning along a preset path, vaporizing or melting the burr material through photothermal or photochemical effects; the laser ablation system possesses high repeatability and high energy stability, ensuring the accuracy and consistency of the ablation process. It can be understood that this step transforms the results of the preceding intelligent analysis and parameter generation into actual physical operations, efficiently and accurately removing burrs. The non-contact nature of laser ablation avoids secondary damage that may be caused by mechanical contact. Specifically, the optimal combination of laser ablation parameters and laser ablation path planning data are loaded into the control unit of the laser ablation system. The control unit is typically an integrated industrial controller, containing a high-performance processor, a digital signal processor (DSP), and a dedicated programmable logic controller (PLC), responsible for parsing ablation instructions from the intelligent parameter generation module. This data is transmitted to the control unit's internal memory via standard industrial communication protocols (e.g., EtherCAT or Profinet) or a dedicated high-speed interface, awaiting execution. A high-precision motion control platform drives the AIoT module board or laser head to move to the initial ablation position of the target half-hole area. The motion control platform can be the same system as the platform in the image acquisition stage. It receives instructions from the control unit and precisely moves the half-hole area to be ablated or the laser head into the focal plane of the laser, aligning it with the starting ablation point of the first burr. Its repeatability is typically less than or equal to one micrometer, ensuring that each ablation operation begins at a preset, precise position. The laser is activated, and based on the loaded laser ablation parameter combination and path planning data, the laser beam is precisely focused and applied to the burr area through a galvanometer scanning system and a dynamic focusing system. The laser is typically a high-power, short-pulse (e.g., picosecond or femtosecond) or ultrafast ultraviolet laser to achieve a "cold ablation" effect and minimize the heat-affected zone. After beam expansion and collimation, the laser beam enters the galvanometer scanning system, which includes two high-speed, high-precision galvanometer mirrors. This system can deflect the laser beam with extremely high speed and accuracy (e.g., scanning speed up to several meters per second, positioning accuracy down to microradians), enabling rapid scanning within a very small area.The dynamic focusing system adjusts the laser focal length in real time based on the three-dimensional morphology data of the burr, ensuring that the laser beam remains focused on the burr surface throughout the ablation process, compensating for changes in burr height and maintaining optimal energy density. During ablation, the laser beam performs a high-precision scan along a preset path, vaporizing or melting the burr material through photothermal or photochemical effects. The photothermal effect refers to the rapid temperature rise of the material after absorbing laser energy, reaching its melting or vaporization point, resulting in melting, evaporation, or sublimation removal. The photochemical effect (often occurring with ultraviolet or ultrafast lasers) refers to high-energy photons directly breaking the molecular bonds of the material, decomposing it into smaller molecules or atomic clusters, achieving a "cold-cutting" effect and minimizing thermal damage. The laser beam, with a pre-planned scanning sequence, direction, and trajectory, precisely targets the burr area, achieving layer-by-layer burr removal. The laser ablation system possesses high repeatability and high energy stability, ensuring the accuracy and consistency of the ablation process. The laser's output power stability is typically controlled within two percent, with minimal pulse energy fluctuations to guarantee uniform energy for each laser action. The repeatability of the galvanometer and motion platform ensures that the laser beam can accurately return to the same position for multiple ablation operations or seamlessly switch between adjacent positions. These high-precision and high-stability characteristics are the fundamental guarantee for achieving non-destructive removal of micron-level burrs.

[0063] In the aforementioned intelligent burr removal method for half-holes in artificial IoT module boards based on laser ablation, the specific steps for detecting the half-hole area after laser ablation to evaluate the burr removal effect and generate feedback signals include: after the laser ablation operation is completed, a high-resolution industrial camera array is restarted to acquire images of the ablated half-hole area; the acquired image data is intelligently analyzed to determine whether residual burrs exist and to quantify the characteristic parameters of the residual burrs; the characteristic parameters of the residual burrs are compared with preset acceptance standards to determine whether the burr removal meets the requirements; if residual burrs exist or the ablation effect is unsatisfactory, a feedback signal containing the location, size, and degree of insufficient ablation of the residual burrs is generated; if the burrs have been completely removed and the substrate is undamaged, a completion signal indicating acceptance is generated. It can be understood that this step constitutes the feedback loop of the intelligent removal method and is the core link for achieving adaptive adjustment and ensuring the quality of the final product. It ensures the thoroughness of burr removal and avoids excessive damage to the substrate. Specifically, after the laser ablation operation is completed, a high-resolution industrial camera array is restarted to acquire images of the ablated half-hole area. The camera array configuration, light source control, and image acquisition parameters are kept consistent with the initial detection stage to ensure the comparability of the detection results. Image acquisition is usually performed after the ablated area has cooled sufficiently to avoid image artifacts caused by thermal effects. The second-acquired image data is also transmitted to the image processing unit for preliminary integration and correction. Intelligent analysis is performed on the second-acquired image data to determine whether there are residual burrs and to quantify the characteristic parameters of the residual burrs. This analysis process is similar to the initial burr identification (the aforementioned S2 step), using image preprocessing techniques (denoising, enhancement, etc.) and a deep learning-based image segmentation network model to comprehensively scan and detect the ablated half-hole area. The model identifies any remaining or newly formed burr-like structures and extracts their geometric features (area, perimeter, maximum length, etc.) and three-dimensional features (height, volume, surface roughness). These features are precisely quantified to form a set of residual burr feature parameters. The feature parameters of the residual burrs are compared with preset acceptance standards to determine whether the burr removal meets the requirements. The acceptance criteria are a predefined dataset containing quantitative indicators of burr removal effectiveness. For example, the maximum height of residual burrs must not exceed X micrometers, the total volume of residual burrs must not exceed Y cubic micrometers, the area of ​​any single residual burr must not exceed Z square micrometers, and the substrate surface must not show obvious overburning marks, color changes, or new damage. The comparison process checks each parameter of each residual burr individually and performs a comprehensive evaluation. If residual burrs are present or the ablation effect is unsatisfactory, a feedback signal is generated containing the location, size, and degree of insufficient ablation of the residual burrs.Feedback signals are generated in the form of structured data packets, containing the precise location coordinates of each incompletely removed burr, its current dimensions (e.g., remaining height, volume), and the difference between these dimensions and the pass / fail criteria (e.g., percentage of insufficient ablation or required additional ablation depth). The degree of insufficient ablation can be calculated based on the ratio of the burr volume difference before and after ablation to the initial burr volume. This feedback signal is a key input guiding subsequent adaptive parameter adjustments. If the burr has been completely removed and the substrate is undamaged, a completion signal indicating pass / fail is generated. This completion signal indicates that the burr removal task for the current half-hole area has been successfully completed, and the system can proceed to the processing flow for the next half-hole area or end the current processing task.

[0064] In the aforementioned intelligent burr removal method for half-holes in AIoT module boards based on laser ablation, the adaptive adjustment of laser ablation parameters based on feedback signals until the burr removal meets the preset standards specifically includes: when a feedback signal indicating the presence of residual burrs or unsatisfactory ablation effect is received, it is input to the intelligent parameter adjustment module; the intelligent parameter adjustment module, based on the characteristic parameters of the residual burrs and the degree of insufficient ablation in the feedback signal, uses an adaptive control algorithm or reinforcement learning model to recalculate and optimize ablation parameters such as laser output power, laser pulse width, and laser scanning speed to improve the removal effect of the next ablation; the steps of laser ablation removal of burrs in the half-hole area through the laser ablation system and detection of the half-hole area after laser ablation to evaluate the burr removal effect and generate feedback signals are repeatedly executed; until a completion signal indicating that the burr removal meets the preset standards is received, the ablation operation is stopped, and the next half-hole area is processed. This step essentially establishes a closed-loop control system, enabling the laser ablation process to dynamically adjust based on real-time detection results. This effectively addresses the randomness and diversity of burrs, ensuring the final removal effect meets expectations and avoiding manual intervention and repeated trial and error. Specifically, when a feedback signal indicating the presence of residual burrs or unsatisfactory ablation results is received, the signal is transmitted to the intelligent parameter adjustment module via a high-speed data interface. The intelligent parameter adjustment module is a software and hardware integrated intelligent decision-making unit. Its core consists of an embedded processor and an adaptive control or reinforcement learning algorithm running within it. Based on the characteristic parameters of the residual burrs (such as remaining size and morphology) and the degree of insufficient ablation in the feedback signal, the intelligent parameter adjustment module uses adaptive control algorithms or reinforcement learning models to recalculate and optimize ablation parameters such as laser output power, laser pulse width, and laser scanning speed to improve the removal effect of the next ablation.

[0065] Adaptive control algorithms can employ adaptive variants of the proportional-integral-derivative (PID) controller, dynamically adjusting PID parameters to optimize laser output based on the error between the current state and the target state of the residual burrs. Another approach is model predictive control (MPC), which utilizes an internal laser ablation prediction model, combined with current feedback, to predict future ablation effects and optimize laser parameters to minimize prediction errors.

[0066] Reinforcement learning models treat the laser ablation process as an interaction between an agent and its environment (the laser ablation system and the module board). The agent observes the environmental state (residual burr characteristics, degree of ablation insufficiency), selects actions (adjusting laser parameters), and learns the optimal strategy based on environmental reward signals (burr removal effect, substrate damage). Reinforcement learning models (e.g., based on deep Q-networks (DQN) or Actor-Critic architectures) are trained in offline simulations or with limited real-world trial and error, enabling them to respond quickly and intelligently to complex feedback situations. For example, if the feedback signal indicates insufficient burr volume ablation, the intelligent parameter adjustment module might increase the laser output power or extend the pulse width; if it indicates a slight risk of overburning, it might decrease the power or increase the scanning speed. These adjustments are finely calculated based on precise 3D information of the burrs and the degree of ablation insufficiency, aiming to achieve millimeter-level or even micrometer-level parameter adjustment accuracy. The process involves repeatedly performing laser ablation to remove burrs from the half-hole region using the laser ablation system and then inspecting the ablated half-hole region to evaluate the burr removal effect and generate feedback signals. This cyclical operation forms a closed-loop feedback loop. Based on the detection results after each ablation, the system continuously adjusts the laser parameters and performs ablation again until the preset pass standard is met. Each iteration updates the burr characteristic parameters and ablation effect, and the intelligent parameter adjustment module uses the latest information to make decisions. The ablation operation stops only when a completion signal indicating that the burr removal meets the preset standard is received, and the process moves to the next half-hole area. When it is detected that all burrs have been completely removed and the substrate is intact, the system issues a completion signal, terminating the ablation task for the current half-hole area. The entire process achieves a high degree of automation and intelligence, significantly improving the efficiency and reliability of burr removal.

[0067] According to one embodiment of this application, the method further includes: constructing and maintaining a database containing various AIoT module board material properties, different burr morphology characteristics, and corresponding optimal laser ablation parameters; when generating laser ablation parameters, relevant parameters are preferentially matched and obtained from this database as initial values, thereby accelerating the parameter optimization process. The database is continuously updated and learned through historical production data and experimental data to improve the accuracy and robustness of parameter generation. It should be understood that the introduction of this database aims to improve the efficiency and accuracy of the intelligent parameter generation module. By accumulating and utilizing historical data, the system can find suitable initial parameters more quickly, reducing the time required for iterative optimization. Specifically, the database is a centralized and scalable digital information repository that stores historical records of various commonly used AIoT module board materials (such as FR-4, polyimide, and ceramic substrates of different thicknesses and layers), different types of burrs (such as copper shavings, solder dross, carbonization residue, and resin overflow), and the optimal laser ablation parameters for these materials and burrs. These parameters include laser output power, pulse width, pulse frequency, scanning speed, spot size, and focal length, etc. The database also contains evaluation data on burr removal effectiveness, substrate damage assessment data, and other relevant process parameters and environmental conditions for each ablation operation. When generating laser ablation parameters, the system first performs fuzzy or exact matching in the database based on the currently identified burr feature parameter set (including burr type, size, location, etc.) and module board material properties. The system prioritizes searching historical records most similar to the current processing situation and extracts the corresponding optimal laser ablation parameters as initial values. This empirical knowledge-based initialization strategy significantly reduces the search space of the parameter optimization algorithm, thereby accelerating the convergence speed of the optimal parameter combination and reducing the number of iterations and computational resource consumption. The database is continuously updated and learned through historical production data and experimental data. Each successful burr removal operation (i.e., meeting the qualification standard and recording the final laser parameters and burr features used) sends its data back to the database as new training samples or empirical knowledge for storage. In addition, through regular experimental verification, the optimal ablation parameters under new material combinations, burr morphologies, or laser technologies are explored, and this new knowledge is injected into the database. By employing machine learning algorithms, such as supervised or semi-supervised learning, continuous analysis and model training are performed on data in the database to discover deeper correlations between parameters and results, thereby continuously improving the accuracy, robustness, and adaptability of parameter generation. This dynamic update and learning mechanism enables the system to continuously enhance its intelligence and processing capabilities as time and application scenarios expand.

[0068] In summary, this application provides a method for intelligent removal of burrs in half-holes of artificial Internet of Things (IoT) module boards based on laser ablation. By introducing high-precision visual inspection, deep learning algorithms, intelligent parameter optimization, and a real-time closed-loop feedback control mechanism, it achieves accurate burr identification, customized ablation, non-destructive removal, and adaptive effect.

[0069] Specifically, the beneficial effects of this application include, but are not limited to:

[0070] First, by acquiring high-resolution image data and using deep learning-based intelligent analysis, micron-level burrs can be accurately identified, located, and their three-dimensional morphology quantified. This overcomes the insufficient accuracy of traditional methods in burr identification, providing a data foundation for subsequent precise ablation. Precise pixel-level segmentation and three-dimensional reconstruction techniques ensure a comprehensive understanding of the burr size, shape, and location information, significantly improving the reliability of the initial data.

[0071] Secondly, by combining material physical properties and intelligent optimization algorithms to generate customized laser ablation parameters, the laser energy is precisely applied to the burrs rather than the substrate, significantly reducing the risk of over-burning, under-burning, or substrate damage, and improving the level of precision control in the ablation process. The multi-objective optimization algorithm finds the optimal balance between removal efficiency, substrate damage, and residue, enabling each burr to receive a personalized and optimal removal solution.

[0072] Furthermore, the use of a high-precision laser ablation system, coupled with optimized scanning path planning, enables non-contact, efficient, and clean burr removal, avoiding the inefficiencies, environmental pollution, and secondary damage associated with traditional methods such as mechanical grinding and chemical cleaning. The coordinated operation of the laser selection, galvanometer scanning, and dynamic focusing system ensures the precise action and efficient removal of the laser beam at the micrometer scale.

[0073] Furthermore, the introduction of a real-time detection and feedback mechanism enables immediate evaluation of the ablation effect and adaptive adjustment of laser parameters based on feedback signals. This ensures thorough burr removal that meets preset standards, effectively addressing the challenges of the randomness and diversity of burrs during production and significantly improving product yield and consistency. The closed-loop control strategy allows the system to self-correct and continuously optimize until the ideal removal effect is achieved.

[0074] Ultimately, the method provided by this invention can achieve a high degree of automation and intelligence in the process of removing burrs from half-holes of AIoT module boards, significantly reducing the need for manual intervention, improving production efficiency, and reducing manufacturing costs. It provides an innovative solution for precision manufacturing in the AIoT industry and has significant technological progress and economic and social benefits.

Claims

1. A method for intelligent removal of burrs in half-holes of AIOT module boards based on laser ablation, characterized in that, Includes the following steps: Acquire high-resolution image data of the half-hole area of ​​the AIOT module board to be processed; Intelligent analysis is performed on the high-resolution image data to identify burrs in the semi-aperture region and quantify the geometric and three-dimensional feature parameters of the burrs. Based on the burr geometric feature parameters, three-dimensional feature parameters and preset material ablation model, the system intelligently generates a combination of laser ablation parameters, including laser output power, laser pulse width, laser pulse frequency, laser scanning speed, laser spot size and focal length, as well as customized laser ablation path planning data for the burr. Using a laser ablation system, based on the laser ablation parameter combination and laser ablation path planning data, burrs in the half-hole area are removed by laser ablation. The semi-hole area after laser ablation is inspected to evaluate the burr removal effect and generate a feedback signal, which includes the location, size and degree of incomplete ablation of the residual burrs. Based on the feedback signal, the laser ablation parameter combination is adaptively adjusted by the intelligent parameter adjustment module using an adaptive control algorithm or reinforcement learning model, and the detection and ablation steps are repeated until the burr removal meets the preset standard. The intelligent generation of laser ablation parameters based on the burr geometric feature parameters, three-dimensional feature parameters, and preset material ablation model specifically includes: The physical properties of the AIOT module board substrate material and burr material are obtained from the material database. The physical properties include thermal conductivity, specific heat capacity, density, melting point, vaporization point and laser absorption rate. The set of burr geometric feature parameters and material physical properties are input into a laser ablation model based on finite element analysis or neural network training; The laser ablation model calculates the laser output power, laser pulse width, laser pulse frequency, laser scanning speed, laser spot size, and initial ablation parameters based on the input data. The process of detecting the semi-hole area after laser ablation to evaluate the burr removal effect and generate a feedback signal specifically includes: After the laser ablation operation is completed, the high-resolution industrial camera array is restarted to acquire images of the ablated half-hole area. Intelligent analysis is performed on the image data acquired in the second phase to determine whether there are residual burrs, and the geometric and three-dimensional feature parameters of the residual burrs are quantified. The characteristic parameters of the residual burrs are compared with the preset acceptance criteria, which include the maximum height of the residual burrs, the total volume, the area of ​​a single burr, and the absence of overheating marks or new damage on the substrate surface. If residual burrs exist or the ablation effect is not ideal, a feedback signal containing the location and size of the residual burrs and the degree of insufficient ablation will be generated. If the burrs have been completely removed and the substrate is undamaged, a completion signal indicating that the substrate is qualified is generated. The adaptive adjustment of laser ablation parameters based on the feedback signal specifically includes: When a feedback signal indicating residual burrs or unsatisfactory ablation effect is received, it is input to the intelligent parameter adjustment module. The intelligent parameter adjustment module uses an adaptive control algorithm or reinforcement learning model to recalculate and optimize the laser output power, laser pulse width, and laser scanning speed ablation parameters based on the characteristic parameters of residual burrs in the feedback signal and the degree of insufficient ablation, so as to improve the removal effect of the next ablation. The adaptive control algorithm employs an adaptive variant of the proportional-integral-derivative controller or model predictive control; the reinforcement learning model is based on a deep Q-network or Actor-Critic architecture, which learns the optimal policy by observing the environmental state, selecting actions, and based on reward signals.

2. The intelligent removal method for half-hole burrs on AIoT module boards based on laser ablation according to claim 1, characterized in that, The acquisition of high-resolution image data of the half-hole region of the AIOT module board to be processed specifically includes: The AIOT module board to be processed is placed on a high-precision positioning platform, which integrates a vacuum adsorption or mechanical clamping device. A high-resolution industrial camera array consisting of two or more industrial-grade cameras is activated to acquire images of the half-hole area of ​​the AIOT module board from multiple angles and focal planes. The high-resolution industrial camera array is equipped with an apochromatic microscope objective or a telecentric lens and works collaboratively through a synchronous trigger controller. During image acquisition, a synchronously controlled adjustable light source system provides uniform and shadowless illumination. The adjustable light source system includes multiple illumination modes such as a ring shadowless lamp, a coaxial light source, and a backlight. The acquired multiple image data are transmitted to the image processing unit for preliminary integration and correction. The preliminary integration includes timestamp alignment of multiple images, image stitching and fusion, and preliminary image correction.

3. The intelligent removal method for half-hole burrs on AIoT module boards based on laser ablation according to claim 2, characterized in that, The high-precision positioning platform has micron-level repeatability and submicron-level motion resolution. It is composed of an XY-axis motion stage driven by a precision linear motor and a Z-axis lifting stage, and uses a closed-loop optical or magnetic grating ruler for position feedback. The high-resolution industrial camera array has industrial-grade cameras with a high resolution of five to twenty million pixels and a high frame rate of fifty to two hundred frames per second.

4. The intelligent removal method for half-hole burrs on AIoT module boards based on laser ablation according to claim 1, characterized in that, The intelligent analysis of the high-resolution image data to identify burrs and quantify their geometric and three-dimensional feature parameters specifically includes: The image data is subjected to image preprocessing operations, including noise reduction filtering using Gaussian filtering, median filtering or bilateral filtering algorithms, brightness and contrast enhancement using histogram equalization or gamma correction methods, geometric correction using perspective transformation methods, and distortion correction using pre-calibrated camera parameters. A deep learning-based image segmentation network model is used to perform pixel-level segmentation of burrs in the half-hole region, thereby accurately distinguishing burrs from the substrate body, thus forming a deep learning model. The deep learning model adopts U-Net, DeepLabV3+ or MaskR-CNN network architecture. Connectivity analysis is performed on the segmented burr regions to extract the geometric features of each individual burr.

5. The intelligent removal method for half-hole burrs on AIOT module boards based on laser ablation according to claim 4, characterized in that, The geometric features include the area, perimeter, maximum length, average width, and shape factor of the burr; the intelligent analysis also includes: By combining stereo vision technology, structured light scanning or multi-view photogrammetry 3D reconstruction algorithms, the 3D morphology of burrs is reconstructed based on multi-angle image data, thereby quantifying the 3D feature parameters of burr height, volume and surface roughness. Based on the geometric features and three-dimensional feature parameters of the burr, a set of burr feature parameters is generated, which includes the burr position coordinates, size, shape, and estimated volume.

6. The intelligent removal method for half-hole burrs on AIOT module boards based on laser ablation according to claim 1, characterized in that, The intelligently generated laser ablation parameters also include: Using a multi-objective optimization algorithm such as non-dominated sorting genetic algorithm, particle swarm optimization, or differential evolution algorithm, with the optimization objectives of maximizing burr removal efficiency, minimizing substrate damage, and minimizing ablation residue, the initial ablation parameters are iteratively optimized to generate a set of optimal laser ablation parameter combinations. A customized laser ablation path planning data is generated for each identified burr. The customized laser ablation path planning data includes the laser scanning sequence, scanning direction and scanning trajectory to ensure that the laser accurately acts on the burr area and avoids the non-burr area.

7. The intelligent removal method for half-hole burrs on AIOT module boards based on laser ablation according to claim 1, characterized in that, The laser ablation removal of burrs in the semi-hole region using a laser ablation system specifically includes: The optimal combination of laser ablation parameters and laser ablation path planning data are loaded into the control unit of the laser ablation system. The control unit includes a processor, a digital signal processor and a dedicated programmable logic controller. A high-precision motion control platform drives the AIOT module board or laser head to move to the initial ablation position in the target half-hole area; The laser is activated. The laser is a high-power, short-pulse or ultrafast ultraviolet laser. The laser beam is precisely focused and applied to the burr area by a galvanometer scanning system and a dynamic focusing system. During the ablation process, the laser beam performs high-precision scanning along a preset path, vaporizing or melting the burr material through photothermal or photochemical effects. The laser ablation system has a repeatability accuracy of less than or equal to one micrometer and a laser output power stability of less than two percent.

Images