Glass-based adapter plate glass via femtosecond laser processing device

Abstract

This invention discloses a femtosecond laser processing device for through-holes in glass-based adapter plates, belonging to the field of laser processing technology. It includes a fixed base, a workpiece fixing assembly, a femtosecond laser emitting assembly, a motion control assembly, a cleaning and inspection assembly, and a central control module. All components are integrated into the fixed base for collaborative operation. The motion control assembly, through a three-dimensional precision micro-motion platform and a vacuum adsorption stage, ensures processing accuracy. The femtosecond laser emitting assembly supports power adjustment to adapt to different processing requirements. The cleaning and inspection assembly removes debris using dry gas and, combined with laser speckle detection and a 30° angle acquisition by a high-definition industrial camera, can identify nanoscale hidden defects. The central control module achieves fully automated closed-loop control, and data integration, backup, and preprocessing ensure reliable traceability. This device improves the accuracy and comprehensiveness of through-hole processing, reduces the scrap rate, and is suitable for industrial mass production needs.

Description

Technical Field

[0001] This invention relates to the field of laser processing technology, specifically to a femtosecond laser processing device for glass through-holes in a glass-based adapter plate. Background Technology

[0002] In the field of laser processing technology, glass-based adapters are widely used in high-end fields such as semiconductors and electronic packaging due to their excellent insulation, thermal conductivity, and dimensional stability. The processing accuracy of their glass through-holes directly affects the assembly accuracy and lifespan of the devices. However, existing laser processing equipment for glass through-holes still has many technical challenges: First, the laser reflection signals and image data collected during the testing process are easily affected by ambient light interference and equipment vibration. The lack of effective data integration, backup and preprocessing mechanisms leads to poor data validity and affects the accuracy of defect identification. At the same time, the lack of traceability of test data from different batches is not conducive to quality control in mass production.

[0003] To address the shortcomings of existing technologies, this invention provides a femtosecond laser processing apparatus for glass-based adapter plates with glass through-holes, thereby solving one or more of the problems mentioned in the background art. Summary of the Invention

[0004] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a femtosecond laser processing device for glass through-holes in glass-based adapter plates, thereby solving the problems mentioned in the background section.

[0005] (II) Technical Solution To achieve the above objectives, the present invention provides a femtosecond laser processing device for glass through-holes in a glass-based adapter plate. The device includes a fixed base, a workpiece fixing assembly, a femtosecond laser emitting assembly, a motion control assembly, a cleaning and inspection assembly, and a central control module. The fixed base is equipped with the workpiece fixing assembly, femtosecond laser emitting assembly, motion control assembly, and cleaning and inspection assembly. The workpiece fixing assembly supports and fixes the glass-based adapter plate, preventing displacement or damage during processing. The femtosecond laser emitting assembly generates a femtosecond-level pulsed laser source. The motion control assembly is connected to both the workpiece fixing assembly and the femtosecond laser emitting assembly, driving either assembly to make precise micro-movements along the thickness direction of the glass-based adapter plate. The cleaning and inspection assembly includes a robotic arm, an air blowing unit, and a vision inspection unit. The air blowing unit blows dry gas into the processed glass through-holes to remove debris. The vision inspection unit captures images of the glass through-holes and inspects their diameter, perpendicularity, and inner wall smoothness.

[0006] Preferably, the femtosecond laser emitting assembly includes a femtosecond laser body, which contains a pulse generator, a gain dielectric cavity and a resonant cavity, a laser output head and a power adjustment knob. The laser output head is installed at the bottom of the femtosecond laser body, and the laser output head is equipped with a power adjustment knob.

[0007] Preferably, the motion control assembly includes an X-axis precision micro-motion platform, a Z-axis precision micro-motion platform, and a Y-axis precision micro-motion platform. The X-axis precision micro-motion platform is mounted on a fixed base, and a vacuum adsorption stage is embedded on the top surface of the X-axis precision micro-motion platform. The Y-axis precision micro-motion platform is mounted on the side wall of the fixed base, and the Z-axis precision micro-motion platform is mounted on the Y-axis precision micro-motion platform.

[0008] Preferably, the central control module is electrically connected to the workpiece fixing component, the femtosecond laser emitting component, the motion control component, and the cleaning and detection component, respectively, to coordinate the collaborative work of each component and realize the automated processing of glass through holes.

[0009] Preferably, the robotic arm is mounted on a fixed base, and the air blowing unit includes an air pump and an adjustable nozzle. The air pump is mounted on the robotic arm, and the adjustable nozzle is mounted on the actuator end of the robotic arm.

[0010] Preferably, the visual inspection unit includes a laser speckle emission unit, a reflection signal acquisition unit, a signal preprocessing unit, a defect feature recognition unit, and a defect information output unit; Each unit is integrated on the robotic arm and arranged coaxially with the high-definition industrial camera. The lens axis of the high-definition industrial camera is set at a 30° angle with the laser optical path to synchronously acquire images of the end face and inner wall of the glass through hole. The laser speckle emitting unit is used to receive the detection trigger signal sent by the central control module, and generate a pulsed laser speckle signal adapted to the glass substrate material through the built-in laser modulation module, and project it onto the inner wall of the through hole and the near hole area according to the preset emission angle. The reflection signal acquisition unit is used to start synchronously with the laser speckle emission unit. It captures the laser speckle reflection signal after reflection through the inner wall of the through hole and the near hole area through a high-sensitivity photodetector. It synchronously records the phase shift, intensity attenuation and speckle pattern distortion data of the reflection signal, forms the original speckle signal dataset and transmits it to the signal preprocessing unit. The signal preprocessing unit receives the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera. It performs filtering, baseline calibration and signal enhancement on the raw speckle signal, noise reduction and edge sharpening on the image data, removes invalid data caused by environmental interference and equipment vibration, and generates effective speckle feature data and effective image data. The defect feature recognition unit is used to call the pre-stored speckle feature reference library of defect-free glass-based materials, compare the effective speckle feature data output by the signal preprocessing unit with the reference library data, and combine the size calibration information of the effective image data to identify hidden defects such as nanoscale microcracks and stress concentration areas on the inner wall of the through hole and near the hole, and determine the specific location and extension parameters of the defects. The defect information output unit is used to convert the latent defect location, extension parameters and feature data obtained by the defect feature identification unit into standardized detection data according to the preset data encoding rules, establish a two-way signal transmission link with the central control module, synchronously upload detection data and receive instruction feedback from the central control module.

[0011] Preferably, the reflected signal acquisition unit includes a band adaptive calibration module, a dynamic timing synchronization module, a phase detection module, an intensity detection module, an array acquisition module, and a pixel-level pattern comparison module; The band adaptive calibration module is used to adapt to the pulsed laser speckle signal band, emission frequency and dynamic spot parameters output by the laser speckle emitting unit, complete the gain initialization and signal acquisition channel self-test, and after confirming that there are no faults in each channel, it aligns with the reflection path of the inner wall of the through hole and the near hole area with the help of the posture positioning results of the robotic arm. The dynamic timing synchronization module, based on the projection parameters of the laser speckle emitting unit, matches the preset acquisition frequency and captures the laser speckle reflection signal after reflection through the inner wall of the through hole and the near hole area in real time. It synchronously locks the acquisition timing and corresponding detection position coordinates of each reflection signal to form an initial reflection signal stream with timing marks. The phase detection module is used to call the original phase parameters pre-stored by the laser speckle emission unit, extract the phase of each frame of the signal waveform in the initial reflected signal stream, compare it with the original phase parameters to obtain the phase offset of each frame of reflected signal, and generate a subset of phase offset data by associating it with the corresponding capture time sequence and detection position coordinates through the multi-dimensional data binding module. The intensity detection module is used to synchronously read the peak intensity of each frame of the initial reflected signal stream, and calculate the intensity attenuation value of the reflected signal by combining the original signal intensity parameters fed back in real time by the laser speckle emission unit. It relies on the multi-dimensional data binding module to accurately bind the temporal and positional information of the phase offset data subset to form an intensity attenuation data subset. An array-type acquisition module and a pixel-level pattern comparison module are used to synchronously acquire the reflected laser speckle pattern. The standard speckle pattern pre-stored in the laser speckle emitting unit is called, and the shape distortion, distribution offset and gray-scale difference feature parameters of the pattern are extracted as speckle pattern distortion data. The corresponding time and position phase offset and intensity attenuation data are associated through a multi-dimensional data binding module to form a subset of distortion data.

[0012] Preferably, the reflected signal acquisition unit also includes a built-in data integration module and a local backup module; The built-in data integration module is used to perform time-series alignment and redundant data removal on three major data subsets: phase shift, intensity attenuation, and speckle pattern distortion. It also labels the signal acquisition position coordinates and detector gain parameters according to a preset encoding format and integrates them to form a complete original speckle signal dataset. The local backup module is used to back up the generated raw speckle signal dataset in real time. The backup data is associated with the corresponding detection batch number. At the same time, the raw speckle signal dataset is synchronously sent to the signal preprocessing unit via the data transmission interface according to the transmission protocol agreed with the signal preprocessing unit.

[0013] Preferably, the signal preprocessing unit includes an adaptive parameter matching module, a cross-data dimension verification module, a baseline calibration module, and an effective data filtering module, which are used to receive the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera, and first complete the synchronous calibration of the dual data receiving channels to ensure the stability of the data transmission link; The adaptive parameter matching module, based on the detector gain and signal band parameters associated with the original speckle signal dataset and the image resolution parameters fed back by the high-definition industrial camera, calls existing adaptation algorithms to perform frame-by-frame verification of the format standardization and transmission integrity of the two types of data. The cross-data dimension verification module is used to accurately bind the detection location coordinates and capture time sequence markers of similar data to complete data time sequence alignment; for the original speckle signal, the existing frequency band filtering algorithm is called to perform targeted frequency band filtering to remove ambient light interference, noise signals caused by equipment vibration and circuit noise; The baseline calibration module retrieves preset baseline parameters and, combined with the original signal reference value fed back in real time by the laser speckle emission unit, dynamically corrects the baseline of the filtered speckle signal to eliminate parameter deviations caused by signal drift. Based on the baseline-calibrated speckle signal, an existing feature-preserving signal enhancement algorithm is used to moderately amplify the weak speckle feature signal, while simultaneously preserving the original phase shift, intensity attenuation, and speckle distortion feature parameters to avoid feature loss, resulting in a preprocessed speckle signal. Simultaneously, image data acquired by a high-definition industrial camera is processed in parallel, and an existing adaptive noise reduction algorithm is called to remove salt-and-pepper noise, environmental interference spots, and reflection artifacts from the image. Then, the edge sharpening module enhances the contour features of the through-hole end face and inner wall, preserving key information for image size calibration to ensure the accuracy of subsequent data association, and generating preprocessed image data. The effective data filtering module, based on preset signal amplitude thresholds and image blur thresholds, judges the validity of preprocessed speckle signals and image data, eliminating speckle signals with abnormal amplitudes, image regions with excessive blur, and invalid pixels one by one; simultaneously, it classifies and organizes the filtered speckle signals according to preset feature dimensions, extracts core parameters such as phase, intensity, and distortion, and forms structured effective speckle feature data; simultaneously, it deeply binds the filtered image data with corresponding time sequence and detection location information, and supplements and annotates basic parameters such as image resolution and processing time to form effective image data; The signal preprocessing unit uses a built-in data formatting module to standardize the effective speckle feature data and effective image data, establish a one-to-one correspondence between the two types of data, complete the entire preprocessing process, and provide standardized input data for the subsequent defect feature recognition unit.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: This device achieves micron-level precision machining of through holes in glass-based adapter plates through the linkage control of precision micro-motion platforms on the X, Y, and Z axes and the negative pressure fixation of the vacuum adsorption stage, combined with the adjustable power design of the femtosecond laser emission component. This effectively improves the perpendicularity of the through holes, the smoothness of the inner wall, and the dimensional consistency, avoiding workpiece displacement and breakage. Furthermore, it innovatively adopts a collaborative acquisition structure of laser speckle detection and a high-definition industrial camera at a 30° angle. Combined with multi-dimensional data integration, preprocessing, and comparison with a defect-free benchmark library, it can accurately identify hidden defects such as nanoscale microcracks and stress concentration areas, filling existing detection gaps. The data integration and local backup mechanism of the reflected signal acquisition unit, combined with the filtering, calibration, and screening processes of the signal preprocessing unit, ensures that the detection data is complete, standardized, traceable, and accurate. The electrical connection design between the central control module and each component constructs an automated closed loop for the entire "processing-cleaning-detection" process, reducing manual intervention, improving processing efficiency and batch consistency, and reducing the scrap rate. The modular design and adjustable parameters of each component adapt to the processing and testing needs of glass-based adapter plates of different specifications and materials, expanding the versatility of the device, reducing equipment adaptation costs, and fully meeting the high-precision quality control requirements of industrial mass production. Attached Figure Description

[0015] Figure 1 This is a three-dimensional structural diagram of the present invention; Figure 2 This is a schematic diagram of the main structure of the present invention; Figure 3 This is a flowchart illustrating the central control module of the present invention. Figure 4 This is a logic diagram of the visual detection unit of the present invention.

[0016] In the diagram: 1. Fixed base; 2. X-axis precision micro-motion platform; 3. Vacuum adsorption stage; 4. Femtosecond laser body; 5. Laser output head; 6. Power adjustment knob; 7. Robotic arm; 8. Z-axis precision micro-motion platform; 9. Y-axis precision micro-motion platform; 10. Air pump; 11. Adjustable nozzle; 12. High-definition industrial camera; 13. Central control module. Detailed Implementation

[0017] In this invention, the terms "first," "second," etc., are used for descriptive purposes only and do not specifically refer to any order or sequence, nor are they intended to limit the invention. They are merely used to distinguish protective components or operations described using the same technical terms, and should not be construed as indicating or implying their relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Furthermore, the technical solutions and features of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed by this invention.

[0018] Example 1, please refer to Figures 1-4 The present invention provides a technical means comprising a fixed base 1, a workpiece fixing component, a femtosecond laser emitting component, a motion control component, a cleaning and inspection component, and a central control module 13. The fixed base 1 is equipped with the workpiece fixing component, the femtosecond laser emitting component, the motion control component, and the cleaning and inspection component. The workpiece fixing component is used to support and fix the glass substrate adapter plate, preventing displacement or damage to the glass substrate adapter plate during processing. The femtosecond laser emitting component is used to generate a femtosecond-level pulsed laser source. The motion control component is connected to the workpiece fixing component and the femtosecond laser emitting component respectively, and is used to drive the workpiece fixing component or the femtosecond laser emitting component to make precise micro-movements along the thickness direction of the glass substrate adapter plate. The cleaning and inspection component includes a robotic arm 7, an air blowing unit, and a vision inspection unit. The air blowing unit is used to blow dry gas into the processed glass through-hole to remove debris from the hole. The vision inspection unit is used to capture images of the glass through-hole and detect the diameter, perpendicularity, and inner wall smoothness of the through-hole.

[0019] The working principle and beneficial effects of the above scheme are as follows: During operation, the glass-based adapter plate is first placed on the workpiece fixing component, which firmly clamps it to prevent displacement or damage caused by laser impact or platform micro-movements during processing, thus laying the foundation for subsequent precise processing. The fixed base 1 provides a stable mounting carrier for the workpiece fixing component, femtosecond laser emitting component, motion control component, and cleaning and detection component, ensuring the accurate installation position of each component and preventing processing accuracy from being affected by carrier shaking. After the central control module 13 is started, it synchronously triggers the femtosecond laser emitting component and the motion control component. The femtosecond laser emitting component generates a femtosecond-level pulsed laser source, and the motion control component drives the workpiece fixing component or the femtosecond laser emitting component to make precise micro-movements along the thickness direction of the glass-based adapter plate according to the preset processing parameters, so that the laser can act on the glass-based adapter plate layer by layer to complete the through-hole processing. After processing, the cleaning and inspection component is activated. The robotic arm 7 moves the air blowing unit to the through-hole position, and the air blowing unit blows dry gas into the hole to thoroughly remove the processing debris, preventing debris residue from affecting the quality of the through-hole and the accuracy of subsequent inspections. Subsequently, the vision inspection unit captures images of the through-hole and simultaneously inspects the diameter, perpendicularity, and inner wall smoothness of the through-hole, forming basic inspection data. Among these features, the stable workpiece fixing component of the fixed base 1 effectively avoids the risk of processing displacement, the precise micro-motion capability of the motion control component ensures the dimensional accuracy of the through-hole processing, and the cleaning and inspection component enables immediate processing and initial quality inspection after processing. The overall structure is compact and provides reliable hardware support for subsequent optimized processing and automated control.

[0020] Example 2: Please refer to Figures 1-4 Based on Embodiment 1, the femtosecond laser emitting assembly includes a femtosecond laser body 4, which contains a pulse generator, a gain medium cavity and a resonant cavity, a laser output head 5 and a power adjustment knob 6. The laser output head 5 is installed at the bottom of the femtosecond laser body 4, and the power adjustment knob 6 is provided on the laser output head 5.

[0021] The working principle and beneficial effects of the above scheme are as follows: This embodiment is based on the overall architecture of Embodiment 1, and refines the structure and working logic of the femtosecond laser emitting component. Its working principle is compatible with the overall process of Embodiment 1. After startup, the pulse generator inside the femtosecond laser body 4 first generates a basic pulse signal. This signal is transmitted to the gain medium cavity, which amplifies the energy of the pulse signal. Then, the amplified pulse signal enters the resonant cavity. Through the oscillation and screening of the resonant cavity, a femtosecond-level pulse laser that meets the processing requirements is formed. Finally, it is directionally emitted to the glass substrate adapter plate processing position through the laser output head 5 at the bottom of the femtosecond laser body 4. The operator can use the power adjustment knob 6 on the laser output head 5 to precisely adjust the laser output power according to the thickness, material and through-hole processing requirements of the glass substrate adapter plate. This avoids the problems of excessive power causing the glass substrate adapter plate to crack or the inner wall to crack, or insufficient power to penetrate the glass substrate adapter plate and low processing efficiency. The synergistic effect of the pulse generator, gain dielectric cavity, and resonant cavity within the femtosecond laser body 4 ensures the pulse stability and energy concentration of the output laser, guaranteeing consistency when the laser acts on the glass-based adapter plate and reducing processing errors. The directional emission design of the laser output head 5 avoids laser scattering and improves energy utilization. The power adjustment knob 6 provides flexibility for processing glass-based adapter plates of different specifications, meeting diverse processing needs without replacing the laser components and reducing equipment adaptation costs. Furthermore, the component's rational structural layout, integrated into the fixed base 1, precisely coordinates with the micro-adjustment of the motion control component, further improving the accuracy and stability of through-hole processing and providing a high-quality processing foundation for the accurate detection of subsequent inspection components.

[0022] Example 3: Please refer to Figures 1-4 Based on any one of Embodiments 1-2, the motion control component includes an X-axis precision micro-motion platform 2, a Z-axis precision micro-motion platform 8, and a Y-axis precision micro-motion platform 9. The X-axis precision micro-motion platform 2 is mounted on a fixed base 1, and a vacuum adsorption stage 3 is embedded on the top surface of the X-axis precision micro-motion platform. The Y-axis precision micro-motion platform 9 is mounted on the side wall of the fixed base 1, and the Z-axis precision micro-motion platform 8 is mounted on the Y-axis precision micro-motion platform 9.

[0023] The working principle and beneficial effects of the above solution are as follows: During operation, the glass-based adapter plate is placed on the vacuum adsorption table 3 on the top surface of the X-axis precision micro-motion platform 2. After the vacuum adsorption table 3 is started, negative pressure is generated to firmly adsorb and fix the glass-based adapter plate. In cooperation with the workpiece fixing component, the limiting effect is further strengthened, completely eliminating the micro-displacement of the workpiece during the processing. The X-axis precision micro-motion platform 2 is installed on the fixed base 1 and can drive the vacuum adsorption table 3 and the glass-based adapter plate to move along the X-axis direction to achieve horizontal adjustment of the through-hole processing position; the Y-axis precision micro-motion platform 9 is installed on the side wall of the fixed base 1, the Z-axis precision micro-motion platform 8 is installed on the Y-axis precision micro-motion platform 9, and the femtosecond laser emission component is correspondingly installed on the Z-axis precision micro-motion platform 8. The Y-axis precision micro-motion platform 9 can drive the Z-axis precision micro-motion platform 8 and the laser component to move along the Y-axis direction, and the Z-axis precision micro-motion platform 8带动 the laser component to move in the vertical direction. The three cooperate with each other to achieve all-round adjustment of the laser component and the workpiece along the X, Y, and Z axes, accurately position the through-hole processing position, and complete layer-by-layer micro-motion processing along the thickness direction of the glass-based adapter plate. The independent adjustment and linkage control of the X-axis, Y-axis, and Z-axis precision micro-motion platforms can ensure that the movement accuracy reaches the micron level, meeting the high-precision processing requirements of the through-holes of the glass-based adapter plate; the adsorption and fixing method of the vacuum adsorption table 3 can avoid damage to the glass-based adapter plate caused by excessive clamping force compared with traditional clamping, and at the same time adapt to glass-based adapter plates of different sizes and shapes, improving the universality of the device. Each platform is respectively installed on the fixed base 1 and the side wall, with a reasonable layout and no interference with each other, ensuring the stability during the movement process, reducing the vibration interference during the platform linkage, and cooperating with the stable laser output of the femtosecond laser emission component, effectively improving the perpendicularity of the through-hole and the smoothness of the inner wall, reducing the difficulty of subsequent cleaning and detection, and further optimizing the overall processing quality.

[0024] Embodiment 4: Please refer to Figures 1-4 , based on any one of Embodiments 1-3, the central control module 13 is electrically connected to the workpiece fixing component, the femtosecond laser emission component, the motion control component, and the cleaning and detection component respectively, and is used to coordinate the cooperation of each component to achieve the automatic processing of glass through-holes.

[0025] The working principle and beneficial effects of the above scheme are as follows: Before operation, the operator inputs the processing parameters and inspection standards into the central control module 13. After startup, the central control module 13 first sends instructions to the workpiece fixing component and the vacuum adsorption stage 3 to complete the fixing of the glass-based adapter plate. Subsequently, it synchronously sends signals to the femtosecond laser emitting component and the motion control component to control the femtosecond laser body 4 to start and output laser with a set power. At the same time, it adjusts the X-axis, Y-axis, and Z-axis precision micro-motion platforms to drive the workpiece to accurately align with the laser component and process through holes along the thickness direction of the glass-based adapter plate according to the preset path. During the processing, the central control module 13 receives feedback signals from each component in real time and dynamically adjusts the laser output power and the platform micro-motion speed to ensure the stability of the processing process. After the through hole processing is completed, the central control module 13 triggers the cleaning and inspection components in sequence, controls the robotic arm 7 to move the air blowing unit to the designated position, and after the debris is cleaned, the vision inspection unit is started to perform quality inspection. The inspection data is transmitted back to the central control module 13 in real time, and the module makes a preliminary judgment on the data. The electrical connection design of the central control module 13 breaks the independent working mode of each component, realizing the coordinated control of the entire process from workpiece fixing, laser processing, platform adjustment to cleaning and inspection, significantly reducing manual intervention and improving processing efficiency. Its dynamic adjustment function can promptly avoid abnormal situations during processing, reducing the scrap rate; the coordinated consistency of each component also further ensures processing accuracy, avoiding processing defects caused by lag or error in manual operation. At the same time, the automated control mode enables the device to continuously process glass-based adapter plates in batches, adapting to the needs of industrial production and enhancing the practicality and production value of the device.

[0026] Example 5: Please refer to Figures 1-4 Based on any one of embodiments 1-4, the robotic arm 7 is mounted on the fixed base 1, and the air blowing unit includes an air pump 10 and an adjustable nozzle 11. The air pump 10 is mounted on the robotic arm 7, and the adjustable nozzle 11 is mounted on the execution end of the robotic arm 7.

[0027] The working principle and beneficial effects of the above scheme are as follows: During operation, triggered by the command of the central control module 13, the robotic arm 7 mounted on the fixed base 1 starts. Based on the coordinate feedback of the through-hole processing position, it drives the adjustable nozzle 11 at the execution end to move precisely to the top or corresponding side of the through-hole. At the same time, the air pump 10 mounted on the robotic arm 7 starts, and the air pump 10 generates dry gas and delivers it to the adjustable nozzle 11, which blows it into the through-hole at high speed. The operator can adjust the air outlet angle and air flow rate of the adjustable nozzle 11 according to the diameter and depth of the through-hole to perform targeted cleaning of the debris remaining on the inner wall and bottom of the through-hole, avoiding the scattering of dry gas that prevents the debris from being completely removed, or the excessive air flow impacting the inner wall of the through-hole and causing micro-damage. The robotic arm 7 has multi-degree-of-freedom adjustment capability, which can adapt to the cleaning needs of through-holes of different positions and specifications, ensuring precise alignment between the nozzle and the through-hole; the integrated installation of the air pump 10 and the robotic arm 7 can avoid pipe entanglement, ensure the flexible movement of the robotic arm 7, and reduce pressure loss during gas delivery, thereby improving the cleaning effect. The use of dry gas prevents moisture from remaining on the inner wall of the through-hole, thus preventing the glass-based adapter plate from becoming damp and affecting its performance. It also prevents moisture and debris from mixing and adhering to the inner wall, improving the thoroughness of cleaning. After cleaning, the robotic arm 7 can quickly move the vision inspection unit to the inspection position, shortening the process interval and improving overall efficiency. This solution, through structural optimization, makes the cleaning process more targeted and flexible, effectively ensuring the cleanliness of the inner wall of the through-hole. This provides a guarantee for the subsequent vision inspection unit to accurately detect the diameter, perpendicularity, and inner wall smoothness, indirectly improving the reliability of the overall processing and inspection.

[0028] Example 6: Please refer to Figures 1-4 Based on any one of Examples 1-5, the visual inspection unit includes a laser speckle emission unit, a reflection signal acquisition unit, a signal preprocessing unit, a defect feature recognition unit, and a defect information output unit. Each unit is integrated on the robotic arm 7 and arranged coaxially with the high-definition industrial camera 12. The lens axis of the high-definition industrial camera 12 is set at a 30° angle with the laser optical path to synchronously acquire images of the end face and inner wall of the glass through hole. The laser speckle emitting unit is used to receive the detection trigger signal sent by the central control module 13, and generate a pulsed laser speckle signal adapted to the glass substrate material through the built-in laser modulation module, and project it onto the inner wall of the through hole and the near hole area according to the preset emission angle. The reflection signal acquisition unit is used to start synchronously with the laser speckle emission unit. It captures the laser speckle reflection signal after reflection through the inner wall of the through hole and the near hole area through a high-sensitivity photodetector. It synchronously records the phase shift, intensity attenuation and speckle pattern distortion data of the reflection signal, forms the original speckle signal dataset and transmits it to the signal preprocessing unit. The signal preprocessing unit is used to receive the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera 12, to filter, calibrate the baseline and enhance the signal of the raw speckle signal, to reduce noise and sharpen the edge of the image data, to remove invalid data caused by environmental interference and equipment vibration, and to generate effective speckle feature data and effective image data. The defect feature recognition unit is used to call the pre-stored speckle feature reference library of defect-free glass-based materials, compare the effective speckle feature data output by the signal preprocessing unit with the reference library data, and combine the size calibration information of the effective image data to identify hidden defects such as nanoscale microcracks and stress concentration areas on the inner wall of the through hole and near the hole, and determine the specific location and extension parameters of the defects. The defect information output unit is used to convert the latent defect location, extension parameters and feature data obtained by the defect feature identification unit into standardized detection data according to the preset data encoding rules, establish a two-way signal transmission link with the central control module 13, synchronously upload detection data and receive instruction feedback from the central control module 13.

[0029] In this embodiment, the pulsed laser speckle signal refers to the laser signal generated by the built-in laser modulation module after receiving the trigger signal detected by the central control module 13, which is adapted to the glass substrate material and projected onto the inner wall of the through hole and the near hole area at a preset emission angle.

[0030] In this embodiment, the original speckle signal dataset refers to the data set that is integrated and transmitted to the subsequent processing structure after the high-sensitivity photodetector captures the reflected signal, synchronously records the phase shift, intensity attenuation and speckle pattern distortion data.

[0031] In this embodiment, effective speckle feature data refers to speckle feature data that can be used for defect identification after the original speckle signal has been filtered, baseline calibrated and signal enhanced to remove invalid data caused by environmental interference and equipment vibration.

[0032] In this embodiment, the defect-free glass-based material speckle feature benchmark library refers to a pre-stored set of speckle feature data of glass-based materials in a defect-free state, used for comparison with the processed speckle feature data.

[0033] In this embodiment, standardized detection data refers to the standardized data that can be transmitted to the central control module 13 via a bidirectional link after the defect location, extension parameters, and feature data have been converted according to preset coding rules.

[0034] The working principle and beneficial effects of the above scheme are as follows: This scheme relies on the robotic arm 7 to achieve flexible movement of the detection structure, forming a collaborative detection structure with the high-definition industrial camera 12. The 30° angle setting can comprehensively capture image information of the end face and inner wall of the through hole. By receiving instructions from the central control module 13, a laser signal adapted to the glass substrate material is generated. After projection, the reflected information is captured and relevant data is recorded. After multi-layer signal processing to eliminate interference, and combined with the comparison of preset benchmark data, hidden defects and location parameters are accurately identified. The detection data is converted according to rules and uploaded to the central control module 13 through a bidirectional link, and instructions are received synchronously, forming a closed loop with the cleaning and processing components mentioned above. This effectively improves the comprehensiveness of glass through hole detection, can identify minute defects missed by conventional detection, and the data transmission and processing are adapted to the needs of automated processing. It provides a reliable basis for the central control module 13 to adjust the processing parameters, ensuring the stability of the processing quality of the glass substrate adapter plate.

[0035] Example 7: Please refer to Figures 1-4 Based on any one of Examples 1-6, the visual inspection unit includes a laser speckle emission unit, a reflection signal acquisition unit, a signal preprocessing unit, a defect feature recognition unit, and a defect information output unit; Each unit is integrated on the robotic arm 7 and arranged coaxially with the high-definition industrial camera 12. The lens axis of the high-definition industrial camera 12 is set at a 30° angle with the laser optical path to synchronously acquire images of the end face and inner wall of the glass through hole. The laser speckle emitting unit is used to receive the detection trigger signal sent by the central control module 13, and generate a pulsed laser speckle signal adapted to the glass substrate material through the built-in laser modulation module, and project it onto the inner wall of the through hole and the near hole area according to the preset emission angle. The reflection signal acquisition unit is used to start synchronously with the laser speckle emission unit. It captures the laser speckle reflection signal after reflection through the inner wall of the through hole and the near hole area through a high-sensitivity photodetector. It synchronously records the phase shift, intensity attenuation and speckle pattern distortion data of the reflection signal, forms the original speckle signal dataset and transmits it to the signal preprocessing unit. The signal preprocessing unit is used to receive the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera 12, to filter, calibrate the baseline and enhance the signal of the raw speckle signal, to reduce noise and sharpen the edge of the image data, to remove invalid data caused by environmental interference and equipment vibration, and to generate effective speckle feature data and effective image data. The defect feature recognition unit is used to call the pre-stored speckle feature reference library of defect-free glass-based materials, compare the effective speckle feature data output by the signal preprocessing unit with the reference library data, and combine the size calibration information of the effective image data to identify hidden defects such as nanoscale microcracks and stress concentration areas on the inner wall of the through hole and near the hole, and determine the specific location and extension parameters of the defects. The defect information output unit is used to convert the latent defect location, extension parameters and feature data obtained by the defect feature identification unit into standardized detection data according to the preset data encoding rules, establish a two-way signal transmission link with the central control module 13, synchronously upload detection data and receive instruction feedback from the central control module 13.

[0036] In this embodiment, the pulsed laser speckle signal refers to the laser signal generated by the built-in laser modulation module after receiving the trigger signal detected by the central control module 13, which is adapted to the glass substrate material and projected onto the inner wall of the through hole and the near hole area at a preset emission angle.

[0037] In this embodiment, the original speckle signal dataset refers to the data set that, after the high-sensitivity photodetector and the laser projection structure are started synchronously, capture the reflected signals of the inner wall of the through hole and the near-hole area, synchronously record the phase shift, intensity attenuation and speckle pattern distortion data, integrate them to form a data set and transmit it to the subsequent signal processing structure.

[0038] In this embodiment, effective speckle feature data refers to speckle feature data that can be used for latent defect identification after the original speckle signal has been filtered, baseline calibrated and signal enhanced to remove invalid data caused by environmental interference and equipment vibration.

[0039] In this embodiment, the effective image data refers to the images of the through-hole end face and inner wall captured by the high-definition industrial camera 12. After noise reduction and edge sharpening, the interfering light spots and blurred areas are removed, and the image data with clear outlines and size calibration information is retained.

[0040] In this embodiment, the defect-free glass-based material speckle feature benchmark library refers to a pre-stored set of speckle feature data of glass-based materials in a state without latent defects, used for comparison with the processed speckle feature data.

[0041] In this embodiment, standardized detection data refers to the location, extension parameters, and characteristic data of latent defects. After being converted according to preset data encoding rules, the standardized data can be interactively transmitted with the central control module 13 through a bidirectional transmission link.

[0042] The working principle and beneficial effects of the above scheme are as follows: This scheme relies on the robotic arm 7 to achieve integrated layout and flexible posture adjustment of the detection structure. It adopts a coaxial arrangement design with the high-definition industrial camera 12, and with a 30° angled optical path setting, it can simultaneously capture clear images of the glass through-hole end face and inner wall, while ensuring the consistency of laser signal and image acquisition, avoiding blind spots. By receiving instructions from the central control module 13, it can generate a laser signal adapted to the glass substrate material, accurately projecting it onto the inner wall of the through-hole and the near-hole area at a preset angle, simultaneously capturing the reflected signal and recording multi-dimensional data, achieving a comprehensive scan of the detection area. Through a multi-layer signal processing flow, interference data caused by ambient light and equipment vibration is eliminated, while image quality is optimized, significantly improving data validity and providing accurate basis for subsequent defect identification. By comparing with pre-stored defect-free benchmark data and combining image size calibration information, it can accurately identify hidden defects that are difficult to detect by conventional inspection, such as nanoscale microcracks and stress concentration areas, clarifying the specific location, extension range, and related parameters of the defects, filling the gap in the detection of minute defects. After standardization and conversion, the test data interacts in real time with the central control module 13 via a two-way transmission link. This allows for both synchronous uploading of test results and adjustment of test parameters in response to commands, forming a complete closed-loop control system with the preceding processing, cleaning, and motion adjustment processes. This not only adapts to the testing needs of glass-based adapter plates of different specifications, improving the versatility and flexibility of the testing, but also provides reliable data support for the central control module 13 to dynamically optimize processing parameters. This helps to promptly avoid subsequent processing defects, reduce scrap rates, enhance the automation and precision of the overall processing flow, and ensure the processing quality and structural stability of the glass-based adapter plates.

[0043] Example 8: Please refer to Figures 1-4 Based on any one of embodiments 1-7, the reflected signal acquisition unit further includes a built-in data integration module and a local backup module; The built-in data integration module is used to perform time-series alignment and redundant data removal on three major data subsets: phase shift, intensity attenuation, and speckle pattern distortion. It also labels the signal acquisition position coordinates and detector gain parameters according to a preset encoding format and integrates them to form a complete original speckle signal dataset. The local backup module is used to back up the generated raw speckle signal dataset in real time. The backup data is associated with the corresponding detection batch number. At the same time, the raw speckle signal dataset is synchronously sent to the signal preprocessing unit via the data transmission interface according to the transmission protocol agreed with the signal preprocessing unit.

[0044] In this embodiment, the data subset refers to the three independent data sets recorded by the high-sensitivity photodetector after capturing the reflected signal: phase shift, intensity attenuation, and speckle pattern distortion.

[0045] In this embodiment, the preset encoding format refers to the preset encoding rules used to standardize and label the signal capture position coordinates and detector gain parameters, thereby unifying and standardizing the data.

[0046] In this embodiment, the detector gain parameter refers to the gain value parameter set by the high-sensitivity photodetector when capturing reflected signals, which is used to characterize the detector's signal capture sensitivity.

[0047] In this embodiment, the batch number refers to a unique identifier set to distinguish different batches of glass substrate adapter plate testing tasks, and is associated with and bound to the backup of the original speckle signal dataset of the corresponding batch.

[0048] In this embodiment, the agreed transmission protocol refers to the pre-agreed data transmission rules between the reflected signal acquisition unit and the signal preprocessing unit to ensure stable data transmission and format compatibility between the two.

[0049] In this embodiment, the data transmission interface refers to the hardware interface used to transmit the raw speckle signal dataset from the reflection signal acquisition unit to the signal preprocessing unit.

[0050] The working principle and beneficial effects of the above scheme are as follows: Based on the original detection data acquisition, this scheme further optimizes the processing and transmission flow of raw data. Relying on a corresponding structure, it specifically integrates the captured multi-dimensional data. First, it precisely aligns the phase shift, intensity attenuation, and speckle pattern distortion-related data according to time sequence, eliminating duplicate and invalid redundant information. Simultaneously, it labels the signal acquisition position coordinates and detector gain parameters, ensuring that the resulting raw data set is complete, standardized, and traceable. Simultaneously, the integrated raw data is backed up in real-time through the corresponding structure. The backup data is associated with the detection batch number, enabling the classification, retention, and rapid traceability of different batches of detection data. This facilitates verification of raw data and troubleshooting when detection anomalies occur. While completing the backup, the data is synchronously sent to the subsequent processing structure via the data transmission interface according to a preset protocol, ensuring the stability and compatibility of data transmission and avoiding data loss or format errors during transmission. This solution not only improves the integrity and accuracy of the original test data, providing a high-quality data foundation for subsequent data preprocessing and defect identification, but also ensures secure data retention through local backup. The batch association design adapts to the needs of batch processing and testing, and the standardized data transmission method ensures the continuity of the testing process. It also deeply collaborates with the aforementioned processing, testing, and control structures, further enhancing the reliability and practicality of the overall device and providing strong support for quality control in automated batch production.

[0051] Example 9: Please refer to Figures 1-4Based on any one of Examples 1-8, the signal preprocessing unit includes an adaptive parameter matching module, a cross-data dimension verification module, a baseline calibration module, and an effective data filtering module. It is used to receive the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera 12. It first completes the synchronous calibration of the dual data receiving channels to ensure the stability of the data transmission link. The adaptive parameter matching module, based on the detector gain and signal band parameters associated with the original speckle signal dataset and the image resolution parameters fed back by the high-definition industrial camera, calls existing adaptation algorithms to perform frame-by-frame verification of the format standardization and transmission integrity of the two types of data. The cross-data dimension verification module is used to accurately bind the detection location coordinates and capture time sequence markers of similar data to complete data time sequence alignment; for the original speckle signal, the existing frequency band filtering algorithm is called to perform targeted frequency band filtering to remove ambient light interference, noise signals caused by equipment vibration and circuit noise; The baseline calibration module retrieves preset baseline parameters and, combined with the original signal reference value fed back in real time by the laser speckle emission unit, dynamically corrects the baseline of the filtered speckle signal to eliminate parameter deviations caused by signal drift. Based on the baseline-calibrated speckle signal, an existing feature-preserving signal enhancement algorithm is used to moderately amplify the weak speckle feature signal, while simultaneously preserving the original phase shift, intensity attenuation, and speckle distortion feature parameters to avoid feature loss, resulting in a preprocessed speckle signal. Simultaneously, image data acquired by a high-definition industrial camera is processed in parallel, and an existing adaptive noise reduction algorithm is called to remove salt-and-pepper noise, environmental interference spots, and reflection artifacts from the image. Then, the edge sharpening module enhances the contour features of the through-hole end face and inner wall, preserving key information for image size calibration to ensure the accuracy of subsequent data association, and generating preprocessed image data. The effective data filtering module, based on preset signal amplitude thresholds and image blur thresholds, judges the validity of preprocessed speckle signals and image data, eliminating speckle signals with abnormal amplitudes, image regions with excessive blur, and invalid pixels one by one; simultaneously, it classifies and organizes the filtered speckle signals according to preset feature dimensions, extracts core parameters such as phase, intensity, and distortion, and forms structured effective speckle feature data; simultaneously, it deeply binds the filtered image data with corresponding time sequence and detection location information, and supplements and annotates basic parameters such as image resolution and processing time to form effective image data; The signal preprocessing unit uses a built-in data formatting module to standardize the effective speckle feature data and effective image data, establish a one-to-one correspondence between the two types of data, complete the entire preprocessing process, and provide standardized input data for the subsequent defect feature recognition unit.

[0052] In this embodiment, the dual data receiving channels refer to two independent data receiving channels for the signal preprocessing unit to receive the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera 12, respectively.

[0053] In this embodiment, the signal band parameter refers to the band-related numerical parameters of the laser beam projected by the laser speckle emitting unit, which are associated with the original speckle signal dataset.

[0054] In this embodiment, the preset baseline parameter refers to the preset standard parameter built into the signal preprocessing unit, which is used to correct the baseline of the filtered speckle signal.

[0055] In this embodiment, the feature-preserving signal enhancement algorithm refers to the existing technology that can amplify weak speckle feature signals while completely preserving the original phase shift, intensity attenuation, and speckle distortion feature parameters of the signal.

[0056] In this embodiment, salt-and-pepper noise refers to isolated white and black dots that appear in the images captured by the high-definition industrial camera 12, which need to be removed by an adaptive noise reduction algorithm.

[0057] In this embodiment, the preset signal amplitude threshold refers to the critical value preset by the signal preprocessing unit to determine whether the amplitude of the speckle signal is normal, and to remove speckle signals with abnormal amplitude.

[0058] In this embodiment, the image blur threshold refers to the critical standard preset by the signal preprocessing unit for judging the clarity of image regions, and is used to remove image regions with excessive blur.

[0059] In this embodiment, the preset feature dimension refers to the feature direction preset by the signal preprocessing unit for classifying and organizing the filtered speckle signal, which is used to extract the core parameters of phase, intensity, and distortion.

[0060] In this embodiment, the data formatting module refers to the functional module built into the signal preprocessing unit, which is used to perform unified format standardization processing on effective speckle feature data and effective image data.

[0061] The working principle and beneficial effects of the above scheme are as follows: Based on the original detection data transmission, this scheme further optimizes the dual-channel preprocessing process of signals and images. First, the received raw speckle data and the image data acquired by the high-definition industrial camera 12 are simultaneously calibrated in both channels to ensure the stability of the data transmission link and avoid data loss or misalignment due to link fluctuations. Combining the detector parameters associated with the raw data and the image resolution parameters of the high-definition industrial camera 12, the existing algorithm is used to verify the format standardization and transmission integrity of the two types of data frame by frame. At the same time, the detection position coordinates and acquisition time sequence of the same type of data are precisely bound to achieve data time sequence alignment, providing a regular data foundation for subsequent processing. Targeted frequency band filtering is performed on the raw speckle signal to effectively eliminate interference caused by ambient light, equipment vibration, and circuit noise. Then, combined with the laser projection of the original reference value, the signal baseline is dynamically corrected to eliminate signal drift deviation, synchronously amplify weak feature signals, and completely retain core parameters, thereby improving signal clarity and avoiding the loss of feature information. Image data is processed synchronously and in parallel to remove salt-and-pepper noise, interference spots, and reflection artifacts, enhance the through-hole contour features, and retain key size calibration information to ensure the usability of image data. The preprocessed signal and image data are filtered for validity using preset thresholds to remove abnormal data and invalid areas. The data is then organized according to standardized dimensions, and core parameters are extracted. Finally, the two types of valid data are standardized in a unified format to establish a one-to-one correspondence. The entire process progressively optimizes data quality, ensuring processing reliability by adapting to existing algorithms, and significantly improving data accuracy and standardization through multi-dimensional calibration, filtering, screening, and formatting. This provides high-quality input for subsequent defect identification and deeply collaborates with the preceding data acquisition, backup, control, and defect identification structures, further strengthening the coherence and accuracy of the detection process. It adapts to the high-precision detection requirements of automated batch processing, helping to improve the overall quality control level of the device.

[0062] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A femtosecond laser processing device for glass through-holes in a glass-based adapter plate, characterized in that, The system includes a fixed base (1), a workpiece fixing component, a femtosecond laser emitting component, a motion control component, a cleaning and inspection component, and a central control module (13). The fixed base (1) is equipped with the workpiece fixing component, the femtosecond laser emitting component, the motion control component, and the cleaning and inspection component. The workpiece fixing component is used to support and fix the glass substrate adapter plate, which can prevent the glass substrate adapter plate from being displaced or damaged during processing. The femtosecond laser emitting component is used to generate a femtosecond-level pulsed laser source. The motion control component is connected to the workpiece fixing component and the femtosecond laser emitting component respectively, and is used to drive the workpiece fixing component or the femtosecond laser emitting component to make precise micro-movements along the thickness direction of the glass substrate adapter plate. The cleaning and inspection component includes a robotic arm (7), an air blowing unit, and a vision inspection unit. The air blowing unit is used to blow dry gas into the processed glass through hole to remove debris in the hole. The vision inspection unit is used to take images of the glass through hole and detect the diameter, verticality, and inner wall smoothness of the through hole.

2. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 1, characterized in that, The femtosecond laser emitting assembly includes a femtosecond laser body (4), which contains a pulse generator, a gain medium cavity and a resonant cavity, a laser output head (5) and a power adjustment knob (6). The laser output head (5) is installed at the bottom of the femtosecond laser body (4), and the power adjustment knob (6) is set on the laser output head (5).

3. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 2, characterized in that, The motion control assembly includes an X-axis precision micro-motion platform (2), a Z-axis precision micro-motion platform (8), and a Y-axis precision micro-motion platform (9). The X-axis precision micro-motion platform (2) is mounted on a fixed base (1), and a vacuum adsorption stage (3) is embedded on the top surface of the X-axis precision micro-motion platform. The Y-axis precision micro-motion platform (9) is mounted on the side wall of the fixed base (1), and the Z-axis precision micro-motion platform (8) is mounted on the Y-axis precision micro-motion platform (9).

4. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 1, characterized in that, The central control module (13) is electrically connected to the workpiece fixing component, the femtosecond laser emitting component, the motion control component and the cleaning and detection component respectively, and is used to coordinate the collaborative work of each component to realize the automated processing of glass through holes.

5. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 1, characterized in that, The robotic arm (7) is mounted on a fixed base (1). The air blowing unit includes an air pump (10) and an adjustable nozzle (11). The air pump (10) is mounted on the robotic arm (7), and the adjustable nozzle (11) is mounted on the execution end of the robotic arm (7).

6. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 5, characterized in that, The visual inspection unit includes a laser speckle emission unit, a reflection signal acquisition unit, a signal preprocessing unit, a defect feature recognition unit, and a defect information output unit; Each unit is integrated on the robotic arm (7) and arranged coaxially with the high-definition industrial camera (12). The lens axis of the high-definition industrial camera (12) is set at a 30° angle with the laser optical path to synchronously acquire images of the end face and inner wall of the glass through hole. The laser speckle emission unit is used to receive the detection trigger signal sent by the central control module (13), and generate a pulsed laser speckle signal adapted to the glass substrate material through the built-in laser modulation module, and project it onto the inner wall of the through hole and the near hole area according to the preset emission angle. The reflection signal acquisition unit is used to start synchronously with the laser speckle emission unit. It captures the laser speckle reflection signal after reflection through the inner wall of the through hole and the near hole area through a high-sensitivity photodetector. It synchronously records the phase shift, intensity attenuation and speckle pattern distortion data of the reflection signal, forms the original speckle signal dataset and transmits it to the signal preprocessing unit. The signal preprocessing unit is used to receive the original speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera (12), to filter, calibrate the baseline and enhance the signal of the original speckle signal, to reduce noise and sharpen the edge of the image data, to remove invalid data caused by environmental interference and equipment vibration, and to generate effective speckle feature data and effective image data. The defect feature recognition unit is used to call the pre-stored speckle feature reference library of defect-free glass-based materials, compare the effective speckle feature data output by the signal preprocessing unit with the reference library data, and combine the size calibration information of the effective image data to identify hidden defects such as nanoscale microcracks and stress concentration areas on the inner wall of the through hole and near the hole, and determine the specific location and extension parameters of the defects. The defect information output unit is used to convert the latent defect location, extension parameters and feature data obtained by the defect feature identification unit into standardized detection data according to the preset data encoding rules, establish a two-way signal transmission link with the central control module (13), synchronously upload detection data and receive instruction feedback from the central control module (13).

7. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 6, characterized in that, The reflected signal acquisition unit includes a band adaptive calibration module, a dynamic timing synchronization module, a phase detection module, an intensity detection module, an array acquisition module, and a pixel-level pattern comparison module; The band adaptive calibration module is used to adapt to the pulsed laser speckle signal band, emission frequency and dynamic spot parameters output by the laser speckle emitting unit, complete the gain initialization and signal acquisition channel self-test, and after confirming that there are no faults in each channel, it aligns with the reflection path of the inner wall of the through hole and the near hole area with the help of the posture positioning results of the robotic arm. The dynamic timing synchronization module, based on the projection parameters of the laser speckle emitting unit, matches the preset acquisition frequency and captures the laser speckle reflection signal after reflection through the inner wall of the through hole and the near hole area in real time. It synchronously locks the acquisition timing and corresponding detection position coordinates of each reflection signal to form an initial reflection signal stream with timing marks. The phase detection module is used to call the original phase parameters pre-stored by the laser speckle emission unit, extract the phase of each frame of the signal waveform in the initial reflected signal stream, compare it with the original phase parameters to obtain the phase offset of each frame of reflected signal, and generate a subset of phase offset data by associating it with the corresponding capture time sequence and detection position coordinates through the multi-dimensional data binding module. The intensity detection module is used to synchronously read the peak intensity of each frame of the initial reflected signal stream, and calculate the intensity attenuation value of the reflected signal by combining the original signal intensity parameters fed back in real time by the laser speckle emission unit. It relies on the multi-dimensional data binding module to accurately bind the temporal and positional information of the phase offset data subset to form an intensity attenuation data subset. An array-type acquisition module and a pixel-level pattern comparison module are used to synchronously acquire the reflected laser speckle pattern. The standard speckle pattern pre-stored in the laser speckle emitting unit is called, and the shape distortion, distribution offset and gray-scale difference feature parameters of the pattern are extracted as speckle pattern distortion data. The corresponding time and position phase offset and intensity attenuation data are associated through a multi-dimensional data binding module to form a subset of distortion data.

8. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 7, characterized in that, The reflected signal acquisition unit also includes a built-in data integration module and a local backup module; The built-in data integration module is used to perform time-series alignment and redundant data removal on three major data subsets: phase shift, intensity attenuation, and speckle pattern distortion. It also labels the signal acquisition position coordinates and detector gain parameters according to a preset encoding format and integrates them to form a complete original speckle signal dataset. The local backup module is used to back up the generated raw speckle signal dataset in real time. The backup data is associated with the corresponding detection batch number. At the same time, the raw speckle signal dataset is synchronously sent to the signal preprocessing unit via the data transmission interface according to the transmission protocol agreed with the signal preprocessing unit.

9. The femtosecond laser processing device for glass through-holes in a glass-based adapter plate according to claim 6, characterized in that, The signal preprocessing unit includes an adaptive parameter matching module, a cross-data dimension verification module, a baseline calibration module, and an effective data filtering module. It is used to receive the raw speckle signal dataset output by the reflection signal acquisition unit and the image data acquired by the high-definition industrial camera (12). First, it completes the synchronous calibration of the dual data receiving channels to ensure the stability of the data transmission link. The adaptive parameter matching module, based on the detector gain and signal band parameters associated with the original speckle signal dataset and the image resolution parameters fed back by the high-definition industrial camera, calls existing adaptation algorithms to perform frame-by-frame verification of the format standardization and transmission integrity of the two types of data. The cross-data dimension verification module is used to accurately bind the detection location coordinates and capture time sequence markers of similar data to complete data time sequence alignment; for the original speckle signal, the existing frequency band filtering algorithm is called to perform targeted frequency band filtering to remove ambient light interference, noise signals caused by equipment vibration and circuit noise; The baseline calibration module retrieves preset baseline parameters and, combined with the original signal reference value fed back in real time by the laser speckle emission unit, dynamically corrects the baseline of the filtered speckle signal to eliminate parameter deviations caused by signal drift. Based on the baseline-calibrated speckle signal, an existing feature-preserving signal enhancement algorithm is used to moderately amplify the weak speckle feature signal, while simultaneously preserving the original phase shift, intensity attenuation, and speckle distortion feature parameters to avoid feature loss, resulting in a preprocessed speckle signal. Simultaneously, image data acquired by a high-definition industrial camera is processed in parallel, and an existing adaptive noise reduction algorithm is called to remove salt-and-pepper noise, environmental interference spots, and reflection artifacts from the image. Then, the edge sharpening module enhances the contour features of the through-hole end face and inner wall, preserving key information for image size calibration to ensure the accuracy of subsequent data association, and generating preprocessed image data. The effective data filtering module, based on preset signal amplitude thresholds and image blur thresholds, judges the validity of preprocessed speckle signals and image data, and removes speckle signals with abnormal amplitudes, image areas with excessive blur, and invalid pixels one by one; at the same time, the filtered speckle signals are classified and organized according to preset feature dimensions, and the core parameters of phase, intensity, and distortion are extracted to form structured effective speckle feature data. Simultaneously, the filtered image data is deeply bound with the corresponding time sequence and detection location information, and basic parameters such as image resolution and processing time are supplemented and annotated to form effective image data; The signal preprocessing unit uses a built-in data formatting module to standardize the effective speckle feature data and effective image data in a unified format, establish a one-to-one correspondence between the two types of data, complete the entire preprocessing process, and provide standardized input data for the subsequent defect feature recognition unit.

Images