Glass substrate edge detection system and repairing method thereof

Abstract

The invention belongs to the technical field of glass production and manufacturing, and particularly discloses a glass substrate edge detection system and a repairing method thereof, and the glass substrate edge detection system comprises a transmission unit, a detection module, a processing control module, a repairing execution module and an auxiliary unit. The detection module adopts a multispectral imaging and laser displacement sensing technology: a visible light imaging unit captures surface texture defects, a near-infrared imaging unit detects internal microcracks in a penetrating manner, and an ultraviolet imaging unit excites a fluorescent substance to enhance the defect contrast; the laser displacement sensor is combined with an adaptive focusing technology to realize three-dimensional contour submicron-level measurement. And the processing control module identifies the defect type through a deep learning algorithm and generates a repair instruction based on the three-dimensional data. And the repairing execution module is used for quickly changing a laser repairing head, a glue repairing head or a grinding head through a mechanical arm and executing repairing processes such as laser cladding, nano spraying or UV glue repairing according to parameters such as the edge breakage depth and the crack length, so that the detection precision and the repairing efficiency of the edge defects of the glass substrate are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of glass production and manufacturing, and in particular relates to a glass substrate edge detection system and a repairing method thereof. Background Art

[0002] Glass substrates are the core base material for precision electronic devices such as display panels and photovoltaic cells. The edge quality of these substrates directly affects the mechanical strength, optical performance, and service life of the products. During high-speed automated production, the edges of glass substrates are prone to defects such as chipping, cracks, and burrs due to processes such as cutting and handling. If these defects are not detected and repaired in a timely manner, they will lead to a decrease in product yield or even functional failure. Traditional inspection methods rely on manual sampling or single-spectrum imaging, which suffer from high missed detection rates and inaccurate defect positioning. The repair process is inefficient and has large precision fluctuations due to factors such as fixed process parameters and delayed tool switching, making it difficult to meet the stringent requirements of high-precision electronic devices for the edge quality of glass substrates. Summary of the Invention

[0003] In view of the problem that the efficiency and accuracy of glass substrate edge repair in the existing technology need to be improved, a glass substrate edge detection system and its repair method are proposed. The present invention provides the following technical solutions:

[0004] A glass substrate edge detection system, comprising:

[0005] a transport unit for transporting the glass substrate at a stable speed;

[0006] A detection module, disposed above and on both sides of the transmission unit, for detecting edge defects of the glass substrate, comprising a multispectral imaging module and a laser displacement sensor for detecting the three-dimensional profile depth of the edge. The multispectral imaging module comprises a visible light imaging unit for capturing edge surface texture and color defects, a near-infrared imaging unit for detecting internal microcrack extension, and an ultraviolet imaging unit for stimulating fluorescent substances at edge defects and enhancing defect contrast. The visible light imaging units are disposed on both sides of the transmission unit, the near-infrared imaging unit is disposed above the transmission unit, and the ultraviolet imaging unit is disposed below the transmission unit.

[0007] A processing control module, used to analyze the inspection data and generate repair instructions, including an industrial computer connected to the linear array camera and the laser displacement sensor, a motion controller connected to the industrial computer, and a database;

[0008] The repair execution module is located behind the detection module and is used to repair defects according to the repair instructions. It includes a robotic arm located on one side of the transmission line, with a laser repair head, a glue repair head, and a grinding head installed at the end of the robotic arm through a quick-change device. The robotic arm is connected to the motion controller through a signal line;

[0009] The auxiliary unit is set in the detection and repair area to assist in the detection and repair work, including a dust removal device and a temperature control system connected to the laser repair head.

[0010] Preferably, the visible light imaging unit, the near-infrared imaging unit and the ultraviolet imaging unit are all linear array cameras.

[0011] Preferably, the visible light imaging unit, near-infrared imaging unit, and ultraviolet imaging unit of the multispectral imaging module are all equipped with narrow-band filters and coaxial parallel light sources, and the angle between the light source and the normal of the glass edge is 30-45° to eliminate mirror reflection interference.

[0012] Preferably, the processing control module has a built-in defect determination algorithm, and when any imaging unit detects an edge depth change ΔH>5μm, it generates an edge chipping defect signal and marks the three-dimensional coordinates and volume parameters of the defect.

[0013] A method for repairing an edge of a glass substrate comprises the following steps:

[0014] S1. Data acquisition: When the glass substrate is transported from the transmission unit to the detection module, the visible light imaging unit, near-infrared imaging unit, and ultraviolet imaging unit collect two-dimensional images of the glass edge under the illumination of a parallel backlight source. The laser displacement sensor simultaneously detects the three-dimensional contour data of the glass edge and transmits the image and data to the industrial computer.

[0015] S2. Edge defect identification: The industrial computer uses a deep learning algorithm to process the image, identify the defect type, and calculate the size and location coordinates of the defect based on the 3D contour data. If a defect is detected, the process proceeds to the next step; if no defect is detected, the glass substrate directly enters the next process.

[0016] S3. Repair: The industrial computer retrieves the corresponding repair process parameters from the database based on the defect type and size, and sends a repair instruction to the motion controller. The motion controller controls the robotic arm to move to the defect location and automatically replaces the laser repair head, glue repair head, or grinding head according to the defect type, and performs the repair operation according to the preset process parameters.

[0017] S4, re-inspection; after the repair is completed, the glass substrate passes through the inspection module again to re-inspect the repair effect. If the repair is qualified, it is transferred to the next process; if it is unqualified, the robot arm repairs it again through step S3. If it still fails to meet the requirements after multiple repairs, it is marked as scrapped.

[0018] Preferably, in step S2:

[0019] The industrial computer obtains the edge optical contrast C through the multispectral imaging module, directly obtains the crack length l and width w through the laser profile scanning module, and calculates the edge chipping depth Δh and the chamfer angle α:

[0020] Edge chipping depth

[0021] Chamfer angle

[0022] Where λ is the laser wavelength, Δφ is the phase difference, h1 and h2 are the heights of adjacent sampling points, and d is the laser profile scanning sampling interval.

[0023] Preferably, in step S3:

[0024] When the crack length l≥100μm, annealing treatment is performed at P0=80W and the cooling rate is T is temperature;

[0025] When the burr height is ≥10 μm, it is removed by laser micro-ablation;

[0026] When the chipping width w≥50μm, the nano-spray filling is started synchronously, the spray flow rate Q=0.5·v·w, v is the flow rate, and the spray deposition efficiency η=0.8·(1-e -0.02p ), p is the nano-spraying delivery pressure.

[0027] Preferably, in step S3, when the chipping depth Δh ≥ 10% D, the laser power is Perform laser cladding; where t is the current time, t0 is the start time of laser power attenuation, P0 is the initial laser power, k is the material absorption coefficient, and D is the substrate thickness;

[0028] Preferably, in step S3, when the chipping depth Δh<10%D, the processing control module automatically calls the glue patch repair process, controls the glue patch head to extrude UV curing glue along the defect contour, and the glue layer thickness is Δh+2μm; after the glue liquid is extruded, it is cured by a UV curing lamp with a wavelength of 365nm and a power of 1000mW / cm 2 Irradiate with high power for 5-10 seconds to complete rapid curing.

[0029] Preferably, in step S4, during re-inspection, the strength recovery rate σ is calculated by secondary inspection. r / σ0, if it is less than 90%, it is marked as defective; if C>0.3 and the laser displacement sensor detects an edge depth change ΔH>5μm, it is determined to be an edge collapse defect, and the defect coordinates and repair instructions are generated to trigger the repair process; where σ0 is the design bending strength, σ r is the bending strength after repair;

[0030] Bending strength after repair Among them, F max is the maximum load of the three-point bending test, W is the width of the substrate, L is the span of the three-point bending test, and D is the thickness of the substrate.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. Through the visible light imaging unit, near-infrared imaging unit and ultraviolet imaging unit, the integrated visible light, near-infrared and ultraviolet three-band imaging modules cover cross-scale detection of surface texture, internal cracks and fluorescence excitation defects, eliminating the blind spots of single-modal detection;

[0033] 2. The built-in defect determination algorithm can quickly and accurately identify edge depth changes and generate corresponding defect signals, 3D coordinates, and volume parameters, providing precise data support for subsequent repair work. The application of the YOLOv8 deep learning algorithm improves the accuracy and efficiency of defect identification. The defect size and location coordinates calculated in combination with 3D contour data are more accurate, providing strong support for subsequent repair work.

[0034] 3. Annealing treatment combined with infrared thermal imaging monitoring can monitor the situation in the heating area in real time, prevent the spread of thermal damage, ensure the repair effect while protecting the integrity of the glass substrate. Laser micro-ablation technology can accurately remove burrs, reduce surface damage and improve accuracy. The combination of laser cladding and glue repair processes can accurately repair edge collapse defects of different depths, improving the flexibility and effect of repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 1 is a schematic diagram of the system structure of the glass substrate edge detection system of the present invention;

[0036] Figure 2 1 is a schematic diagram of a method flow of a glass substrate edge repair method according to the present invention; DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention is clearly and completely described below in conjunction with the accompanying drawings of the present invention. The directional words mentioned in the following embodiments, such as "up", "down", "left" and "right", etc., are only referenced to the directions of the accompanying drawings. Therefore, the directional words used are used to illustrate rather than limit the invention.

[0038] A glass substrate edge detection system, comprising:

[0039] The transport unit is used to transport glass substrates at a stable speed. It integrates an active vibration reduction mechanism to suppress image acquisition blur caused by environmental vibration during the transport process, ensuring the accuracy of subsequent defect detection and improving the overall efficiency of the production line and product quality.

[0040] The detection module, located above and on both sides of the transmission unit, is used to detect defects on the edge of the glass substrate. It includes a multispectral imaging module and a laser displacement sensor for detecting the three-dimensional depth of the edge profile. The laser displacement sensor, combined with adaptive focusing technology, ensures consistent profile measurement for substrates of varying thicknesses. This allows the laser displacement sensor to adapt to glass substrates of varying thicknesses, ensuring high-precision and consistent profile measurement and providing reliable data support for subsequent repair work.

[0041] The multispectral imaging module includes:

[0042] Visible light imaging unit, with a wavelength of 400-700nm and a resolution of ≤5μm, is used to capture edge surface texture and color defects; near-infrared imaging unit, with a wavelength of 850-950nm, is used to detect the extension of internal microcracks;

[0043] UV imaging unit, with a wavelength of 320-380nm and a resolution of ≤2μm, is used to excite fluorescent substances at edge defects and enhance defect contrast;

[0044] To make full use of the characteristics of different imaging units, visible light imaging units are set on both sides of the transmission unit, the near-infrared imaging unit is set above the transmission unit, and the ultraviolet imaging unit is set on the bottom side of the transmission unit, realizing all-round detection of the edge of the glass substrate and ensuring the comprehensiveness and accuracy of the detection.

[0045] A processing control module, used to analyze the inspection data and generate repair instructions, including an industrial computer connected to the linear array camera and the laser displacement sensor, a motion controller connected to the industrial computer, and a database;

[0046] The repair execution module is located behind the detection module and is used to repair defects according to the repair instructions. It includes a robotic arm located on one side of the transmission line. The robotic arm is connected to the motion controller via a signal line. The end of the robotic arm is equipped with a quick-change device to install and replace the laser repair head, glue repair head, and grinding head. This enables rapid switching and use of multiple repair tools, improving the flexibility and efficiency of repair.

[0047] The auxiliary unit is set in the inspection and repair area to assist in the inspection and repair work. It includes a dust removal device and a temperature control system connected to the laser repair head, providing a good working environment for the inspection and repair work and ensuring the accuracy and stability of the inspection and repair.

[0048] Specifically, the visible light imaging unit, the near-infrared imaging unit, and the ultraviolet imaging unit all use line array cameras.

[0049] Specifically, the visible light imaging unit, near-infrared imaging unit, and ultraviolet imaging unit of the multispectral imaging module are all equipped with narrow-band filters and coaxial parallel light sources. The angle between the light source and the normal of the glass edge is 30-45°, which is used to eliminate mirror reflection interference, improve imaging quality, and provide clearer and more reliable images for subsequent defect identification.

[0050] Specifically, the processing control module has a built-in defect judgment algorithm. When any imaging unit detects an edge depth change ΔH>5μm, it generates a chipping defect signal and marks the defect's three-dimensional coordinates and volume parameters. It can quickly and accurately identify edge depth changes and generate corresponding defect signals, three-dimensional coordinates, and volume parameters, providing accurate data support for subsequent repair work.

[0051] A method for repairing an edge of a glass substrate comprises the following steps:

[0052] S1. Data acquisition: When the glass substrate is transported from the transmission unit to the inspection module, the visible light imaging unit, near-infrared imaging unit, and ultraviolet imaging unit collect two-dimensional images of the glass edge under the illumination of a parallel backlight source. The laser displacement sensor simultaneously detects the three-dimensional contour data of the glass edge and transmits the image and data to the industrial computer. This ensures the comprehensiveness and accuracy of the data, providing a reliable foundation for subsequent defect identification and repair.

[0053] S2. Edge defect recognition: The industrial computer uses the YOLOv8 deep learning algorithm to process the image and identify the defect type. After training, it can identify multiple defect types such as chipping, cracks, notches, burrs, etc. The size and location coordinates of the defect are calculated based on the 3D contour data. If a defect is detected, the process proceeds to the next step; if no defect is detected, the glass substrate directly enters the next process.

[0054] Specifically, in step S2:

[0055] The industrial computer obtains the edge optical contrast C through the multispectral imaging module, directly obtains the crack length l and width w through the laser profile scanning module, and calculates the edge chipping depth Δh and the chamfer angle α:

[0056] Edge chipping depth

[0057] Chamfer angle

[0058] Wherein, is the laser wavelength, is the phase difference, h1 and h2 are the heights of adjacent sampling points, and d is the laser profile scanning sampling interval.

[0059] S3. Repair: The industrial computer retrieves the corresponding repair process parameters from the database based on the defect type and size, and sends a repair instruction to the motion controller. The motion controller controls the robotic arm to move to the defect location and automatically replaces the laser repair head, glue repair head, or grinding head according to the defect type, and performs the repair operation according to the preset process parameters.

[0060] Specifically, the repair strategy in step S3 is:

[0061] When the local protrusion area is larger than 5000 μm 2 , according to the design accuracy, the grinding pressure data is automatically retrieved from the database, and the grinding is performed through the grinding head, using a progressive multi-stage grinding strategy, rough grinding to remove the main body protrusions and then fine grinding to restore the surface smoothness.

[0062] When the crack length l≥100μm, annealing treatment is performed at P0=80W. The annealing process is monitored by infrared thermal imaging, and the heating area is adjusted in real time to prevent the spread of thermal damage. The cooling rate T is temperature;

[0063] When the burr height is ≥10 μm, it is removed by laser micro-ablation;

[0064] When the chipping width w≥50μm, the nano-spray filling is started synchronously, the spray flow rate Q=0.5·v·w, v is the flow rate, and the spray deposition efficiency η=0.8·(1-e -0.02p ), p is the nano-spraying delivery pressure.

[0065] When the chipping depth Δh ≥ 10% D, it is determined to be a deep chipping defect. Laser cladding is performed, and the laser power is initialized according to P0 = 50W + 0.3Δh; where t is the current time, t0 is the start time of laser power attenuation, P0 is the initial laser power, k is the material absorption coefficient, and D is the substrate thickness; when the edge collapse depth Δh < 10% D, it is determined to be a shallow edge collapse defect, and the processing control module automatically calls the glue repair process, controls the glue repair head to extrude UV curing glue along the defect contour, and the glue layer thickness is Δh + 2μm; after the glue liquid is extruded, it is passed through a UV curing lamp with a wavelength of 365nm and 1000mW / cm 2 Irradiate with high power for 5-10 seconds to complete rapid curing.

[0066] S4, re-inspection; after the repair is completed, the glass substrate passes through the inspection module again to re-inspect the repair effect. If the repair is qualified, it is transferred to the next process; if it is unqualified, the robot arm repairs it again through step S3. If it still fails to meet the requirements after multiple repairs, it is marked as scrapped.

[0067] Specifically, in step S4, during re-testing, the strength recovery rate σ is calculated by secondary testing. r / σ0, if it is less than 90%, it is marked as defective; if C>0.3 and the laser displacement sensor detects an edge depth change ΔH>5μm, it is determined to be an edge collapse defect, and the defect coordinates and repair instructions are generated to trigger the repair process; where σ0 is the design bending strength, σ r is the bending strength after repair;

[0068] Bending strength after repair Among them, F max is the maximum load of the three-point bending test, W is the width of the substrate, L is the span of the three-point bending test, and D is the thickness of the substrate.

[0069] Specifically, if the repair fails, the robot arm will be repaired up to three times. If it still fails, it will be marked as scrapped; avoiding the cost and time waste caused by unlimited repairs while ensuring the overall efficiency of the production line.

[0070] Take the 10.5 generation glass substrate (substrate thickness D = 0.5mm) as an example:

[0071] When the multispectral imaging module detects C = 0.4, Δh = 80 μm, and 80 μm > 10% D = 50 μm, it is determined to be deep edge chipping;

[0072] Laser initial power = 50 + 0.3 × 80 = 74W, according to Cladding, simultaneous nano-spraying (w = 60 μm, v = 10 mm / s, Q = 0.5 × 10 × 60 = 300 μL / s);

[0073] Secondary detection = σ r / σ0=92%>90%, so the repair is qualified. It can be seen that the glass substrate edge detection system and repair method have good effect and reliability.

Claims

1. A glass substrate edge detection system, characterized in that: include: a transport unit for transporting the glass substrate at a stable speed; A detection module, disposed above and on both sides of the transmission unit, for detecting edge defects of the glass substrate, comprising a multispectral imaging module and a laser displacement sensor for detecting the three-dimensional profile depth of the edge. The multispectral imaging module comprises a visible light imaging unit for capturing edge surface texture and color defects, a near-infrared imaging unit for detecting internal microcrack extension, and an ultraviolet imaging unit for stimulating fluorescent substances at edge defects and enhancing defect contrast. The visible light imaging units are disposed on both sides of the transmission unit, the near-infrared imaging unit is disposed above the transmission unit, and the ultraviolet imaging unit is disposed below the transmission unit. A processing control module, used to analyze the inspection data and generate repair instructions, including an industrial computer connected to the linear array camera and the laser displacement sensor, a motion controller connected to the industrial computer, and a database; The repair execution module is located behind the detection module and is used to repair defects according to the repair instructions. It includes a robotic arm located on one side of the transmission line, with a laser repair head, a glue repair head, and a grinding head installed at the end of the robotic arm through a quick-change device. The robotic arm is connected to the motion controller through a signal line; The auxiliary unit is set in the detection and repair area to assist in the detection and repair work, including a dust removal device and a temperature control system connected to the laser repair head.

2. The glass substrate edge detection system according to claim 1, wherein: The visible light imaging unit, the near infrared imaging unit and the ultraviolet imaging unit all adopt line array cameras.

3. The glass substrate edge detection system according to claim 1, wherein: The visible light imaging unit, near-infrared imaging unit, and ultraviolet imaging unit of the multispectral imaging module are all equipped with narrow-band filters and coaxial parallel light sources. The angle between the light source and the normal of the glass edge is 30-45 degrees, which is used to eliminate mirror reflection interference.

4. The glass substrate edge detection system according to claim 1, wherein: The processing control module has a built-in defect judgment algorithm. When any imaging unit detects an edge depth change ΔH>5μm, it generates an edge collapse defect signal and marks the three-dimensional coordinates and volume parameters of the defect.

5. A glass substrate edge repair method, characterized in that: The following steps are involved: S1. Data acquisition: When the glass substrate is transported from the transmission unit to the detection module, the visible light imaging unit, near-infrared imaging unit, and ultraviolet imaging unit collect two-dimensional images of the glass edge under the illumination of a parallel backlight source. The laser displacement sensor simultaneously detects the three-dimensional contour data of the glass edge and transmits the image and data to the industrial computer. S2. Edge defect identification: The industrial computer uses a deep learning algorithm to process the image, identify the defect type, and calculate the size and location coordinates of the defect based on the 3D contour data. If a defect is detected, the process proceeds to the next step. If no defects are detected, the glass substrate directly enters the next process; S3. Repair: The industrial computer retrieves the corresponding repair process parameters from the database based on the defect type and size, and sends a repair instruction to the motion controller. The motion controller controls the robotic arm to move to the defect location and automatically replaces the laser repair head, glue repair head, or grinding head according to the defect type, and performs the repair operation according to the preset process parameters. S4, re-examination; After the repair is completed, the glass substrate passes through the inspection module again to re-inspect the repair effect. If the repair is qualified, it is transferred to the next process; if it is unqualified, the robot arm repairs it again through step S3. If it still fails after multiple repairs, it is marked as scrapped.

6. A glass substrate edge repair method according to claim 4, characterized in that: In step S2: The industrial computer obtains the edge optical contrast C through the multispectral imaging module, directly obtains the crack length l and width w through the laser profile scanning module, and calculates the edge chipping depth Δh and the chamfer angle α: Edge chipping depth Chamfer angle Where λ is the laser wavelength, Δφ is the phase difference, h1 and h2 are the heights of adjacent sampling points, and d is the laser profile scanning sampling interval.

7. A glass substrate edge repair method according to claim 4, characterized in that: In step S3: When the crack length l≥100μm, annealing treatment is performed at P0=80W and the cooling rate is T is temperature; When the burr height is ≥10 μm, it is removed by laser micro-ablation; When the chipping width w≥50μm, the nano-spray filling is started synchronously, the spray flow rate Q=0.5·v·w, v is the flow rate, and the spray deposition efficiency η=0.8·(1-e -0.02p ), p is the nano-spraying delivery pressure.

8. The glass substrate edge repair method according to claim 7, wherein: In step S3, when the chipping depth Δh ≥ 10% D, the laser power is adjusted according to the real-time Laser cladding is performed; where t is the current moment, t0 is the start time of laser power attenuation, P0 is the initial laser power, k is the material absorption coefficient, and D is the substrate thickness.

9. The glass substrate edge repair method according to claim 7, wherein: In step S3, when the chipping depth Δh < 10% D, the processing control module automatically calls the glue patch repair process, controls the glue patch head to extrude UV curing glue along the defect contour, and the glue layer thickness is Δh + 2μm; after the glue liquid is extruded, it is passed through a UV curing lamp with a wavelength of 365nm and 1000mW / cm 2 Irradiate with high power for 5-10 seconds to complete rapid curing.

10. The glass substrate edge repair method according to claim 4, characterized in that: In step S4, during re-test, the strength recovery rate σ is calculated by secondary testing. r / σ0, if it is less than 90%, it is marked as defective; if C>0.3 and the laser displacement sensor detects an edge depth change ΔH>5μm, it is determined to be an edge collapse defect, and the defect coordinates and repair instructions are generated to trigger the repair process; where σ0 is the design bending strength, σ r is the bending strength after repair; Bending strength after repair Among them, F max is the maximum load of the three-point bending test, W is the width of the substrate, L is the span of the three-point bending test, and D is the thickness of the substrate.

Images