A nozzle clogging state monitoring method, device, equipment, medium and product

Abstract

The application provides a nozzle blockage state monitoring method, device, equipment, medium and product, and relates to the fields of optical detection and image processing. The method comprises the following steps: acquiring a plurality of reflection light patterns corresponding to nozzle outlet sections respectively; performing decoupling processing on a target reflection light pattern to obtain an equidistant development diagram in a two-dimensional plane coordinate system; wherein the target reflection light pattern is a reflection light pattern corresponding to a target nozzle outlet section, and the target nozzle outlet section is any one of the plurality of nozzle outlet sections; obtaining a target imaging area according to the equidistant development diagram; performing adaptive binarization processing on the target imaging area to obtain a laser feature pattern; obtaining a gray-scale image corresponding to the target nozzle outlet section according to the laser feature pattern; and inputting the gray-scale image corresponding to the target nozzle outlet section and a pre-stored normal state gray-scale image into a pre-trained blockage state recognition model to obtain a blockage state judgment result and a blockage level corresponding to the target nozzle. The accuracy of nozzle blockage state monitoring is effectively improved.

Description

Technical Field

[0001] This application relates to the field of optical inspection and image processing, and in particular to a method, apparatus, equipment, medium and product for monitoring nozzle blockage status. Background Technology

[0002] In semiconductor and organic light-emitting diode (OLED) manufacturing processes, vapor deposition is a core technology used to deposit organic or metallic materials onto a substrate with high precision, forming light-emitting layers, electrode layers, or circuit structures. As the source of material evaporation, the unobstructed flow of the vapor deposition source's nozzle directly determines the uniform distribution and deposition quality of the material. However, in actual production, vapor deposition source nozzles are prone to clogging due to material condensation, impurity deposition, or fluctuations in process parameters. This clogging not only leads to uneven distribution of the vapor deposition material, causing defects such as substrate surface thickness deviation, uneven brightness, or color shift, but can also cause large particles to splash due to sudden increases in local pressure, resulting in short circuits or open circuits and directly rendering the device unusable. Furthermore, frequent downtime for maintenance significantly reduces equipment uptime and increases production costs. Currently, the need for real-time monitoring of nozzle status in vapor deposition processes is particularly urgent.

[0003] Currently, in the vapor deposition process of semiconductors and organic light-emitting diodes, temperature sensor-based monitoring schemes are used: by monitoring nozzle temperature and heating current in real time, early warning of material condensation trends and identification of abnormal operating conditions are achieved. However, this type of method mainly relies on signals such as temperature, current, and heating rate to indirectly judge the risk of blockage. The initial response to blockage is delayed, and it is easily affected by factors such as power fluctuations and changes in thermal radiation, resulting in misjudgments. At the same time, it can only provide limited process parameters and cannot accurately reflect the location, degree, and morphology of the blockage in the nozzle orifice. Therefore, the accuracy of nozzle blockage monitoring is low. Summary of the Invention

[0004] This application provides a method, apparatus, device, medium, and product for monitoring nozzle clogging status, in order to solve the technical problem of low accuracy in nozzle clogging status monitoring caused by the prior art.

[0005] In a first aspect, embodiments of this application provide a method for monitoring nozzle blockage, including:

[0006] Obtain the reflected light patterns corresponding to the exit cross sections of multiple nozzles;

[0007] The target reflected light pattern is decoupled to obtain an isometric unfolded diagram in a two-dimensional plane coordinate system; wherein, the target reflected light pattern is the reflected light pattern corresponding to the target nozzle exit section, and the target nozzle exit section is any one of multiple nozzle exit sections;

[0008] The target imaging area is obtained based on the isometric unfolded diagram.

[0009] The target imaging area is subjected to adaptive binarization processing to obtain a laser feature pattern;

[0010] Based on the laser feature pattern, a grayscale image corresponding to the exit cross-section of the target nozzle is obtained;

[0011] The grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image are input into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle.

[0012] In one possible implementation, the step of inputting the grayscale image corresponding to the outlet cross-section of the target nozzle and the pre-stored normal state grayscale image into a pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle includes:

[0013] The grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image are input into the pre-trained blockage state recognition model to obtain the blockage feature data corresponding to the target nozzle.

[0014] Based on the clogging characteristic data corresponding to the target nozzle, the clogging status judgment result and clogging level of the target nozzle are obtained.

[0015] In one possible implementation, the clogging feature data includes central dark area data, annular bright band data, and edge grayscale gradient data;

[0016] Accordingly, obtaining the blockage status judgment result and blockage level of the target nozzle based on the blockage feature data of the target nozzle includes:

[0017] Based on the central dark area data, the annular bright band data, and the edge grayscale gradient data corresponding to the target nozzle, the clogging status judgment result and clogging level of the target nozzle are obtained.

[0018] In one possible implementation, obtaining the grayscale image corresponding to the target nozzle exit cross-section based on the laser feature pattern includes:

[0019] Based on the laser feature pattern, the actual light transmission profile and surface deposition profile of the aperture corresponding to the exit section of the target nozzle are obtained;

[0020] Based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section, a grayscale image corresponding to the target nozzle exit section is obtained.

[0021] In one possible implementation, obtaining the grayscale image corresponding to the target nozzle exit cross-section based on the actual light transmission profile and surface deposition profile corresponding to the target nozzle exit cross-section includes:

[0022] Based on the actual light transmission profile of the aperture corresponding to the exit cross-section of the target nozzle, the flow area attenuation rate and edge distortion degree corresponding to the exit cross-section of the target nozzle are obtained;

[0023] The consistency deviation of the multi-hole array is obtained based on the flow area attenuation rate and edge distortion degree corresponding to the outlet cross section of the target nozzle.

[0024] When the consistency deviation of the multi-hole array is detected to be less than a preset deviation threshold, a grayscale image corresponding to the target nozzle exit section is obtained based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section.

[0025] In one possible implementation, the reflected light pattern is obtained by the diffractive optical element lens guiding multiple annular laser spots to multiple nozzle exit sections, and then reflecting them from the multiple nozzle exit sections. The multiple annular laser spots are obtained by the diffractive optical element lens converting an initial rectangular laser spot.

[0026] Secondly, embodiments of this application provide a nozzle clogging status monitoring device, comprising:

[0027] The reflected light pattern acquisition module is used to acquire the reflected light patterns corresponding to the exit sections of multiple nozzles.

[0028] A decoupling module is used to decouple the target reflected light pattern to obtain an isometric unfolded diagram in a two-dimensional plane coordinate system; wherein, the target reflected light pattern is the reflected light pattern corresponding to the target nozzle exit section, and the target nozzle exit section is any one of multiple nozzle exit sections;

[0029] The target imaging area acquisition module is used to obtain the target imaging area based on the isometric unfolded diagram.

[0030] The binarization processing module is used to perform adaptive binarization processing on the target imaging area to obtain a laser feature pattern;

[0031] The grayscale image acquisition module is used to obtain the grayscale image corresponding to the exit section of the target nozzle based on the laser feature pattern.

[0032] The prediction module is used to input the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle.

[0033] Thirdly, embodiments of this application provide an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0034] The memory stores computer-executed instructions;

[0035] The processor executes computer execution instructions stored in the memory to implement the nozzle clogging status monitoring method provided in the first aspect of this application.

[0036] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the nozzle clogging status monitoring method provided in the first aspect of this application.

[0037] Fifthly, embodiments of this application provide a computer program product, including a computer program, which, when executed by a processor, is used to implement the nozzle clogging status monitoring method provided in the first aspect of this application.

[0038] This application provides a method, apparatus, device, medium, and product for monitoring nozzle blockage status, including: acquiring reflected light patterns corresponding to multiple nozzle outlet cross-sections; decoupling the target reflected light patterns to obtain an isometric unfolded diagram in a two-dimensional plane coordinate system; obtaining the target imaging area based on the isometric unfolded diagram; performing adaptive binarization processing on the target imaging area to obtain a laser feature pattern; obtaining a grayscale image corresponding to the target nozzle outlet cross-section based on the laser feature pattern; and inputting the grayscale image corresponding to the target nozzle outlet cross-section and a pre-stored normal state grayscale image into a pre-trained blockage status recognition model to obtain the blockage status judgment result and blockage level corresponding to the target nozzle. The above method achieves the following technical effects: By decoupling the reflected light pattern into a two-dimensional isometric unfolded image, the geometric nonlinearity and radial stretching distortion caused by polar coordinate projection are fundamentally eliminated, thus significantly reducing the complexity of subsequent feature extraction and the risk of mismatch. Based on this, adaptive binarization processing of the target imaging area dynamically adapts to fluctuations in overall light intensity at different deposition stages, ensuring accurate and continuous segmentation boundaries between normal reflection areas and defect areas in the laser feature pattern. Furthermore, the binarized feature pattern is converted into a standardized grayscale image and input into a pre-trained blockage state recognition model along with a pre-stored normal state grayscale image. This allows the model to directly compare the pixel-by-pixel differences between the current state and the health baseline, quantitatively capturing minute grayscale attenuation, area loss, and edge distortion. Through deep extraction and weighted fusion of feature data by the model, accurate acquisition of blockage state judgment results and blockage levels is achieved, effectively solving the problems of low monitoring accuracy, easy missed detections, and false detections. Attached Figure Description

[0039] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0040] Figure 1This is a schematic diagram illustrating an application scenario of a nozzle clogging status monitoring method provided in an embodiment of this application.

[0041] Figure 2 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 1 ;

[0042] Figure 3 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 2 ;

[0043] Figure 4 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 3 ;

[0044] Figure 5 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 4 ;

[0045] Figure 6 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 5 ;

[0046] Figure 7 This is a schematic diagram of the structure of a monitoring system provided in an embodiment of this application;

[0047] Figure 8 This is a schematic diagram of the structure of a nozzle clogging status monitoring device provided in an embodiment of this application;

[0048] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0049] Explanation of reference numerals in the attached figures:

[0050] 110 - Terminal; 120 - Server; 701 - Processor; 702 - Memory; 703 - Communication Components; 704 - Bus.

[0051] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0052] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0053] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" do not necessarily imply difference. It should be noted that in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner. In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more.

[0054] It should be noted that the phrase "at...time" in the embodiments of this application can refer to the instant at which a certain situation occurs, or to a period of time after the occurrence of a certain situation; the embodiments of this application do not specifically limit this. Furthermore, the nozzle clogging status monitoring method provided in the embodiments of this application is merely an example, and the nozzle clogging status monitoring method may include more or fewer elements.

[0055] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0056] To clearly understand the technical solution of this application, the solutions of the prior art will be described in detail first.

[0057] In the manufacturing processes of semiconductors and organic light-emitting diodes, vapor deposition technology is one of the core processes, and the need for real-time monitoring of nozzle status in the vapor deposition process is particularly urgent.

[0058] Currently, in the vapor deposition process of semiconductors and organic light-emitting diodes, temperature sensor-based monitoring schemes are used: by monitoring nozzle temperature and heating current in real time, early warning of material condensation trends and identification of abnormal operating conditions are achieved. However, this type of method mainly relies on signals such as temperature, current, and heating rate to indirectly judge the risk of blockage. The initial response to blockage is delayed, and it is easily affected by factors such as power fluctuations and changes in thermal radiation, resulting in misjudgments. At the same time, it can only provide limited process parameters and cannot accurately reflect the location, degree, and morphology of the blockage in the nozzle orifice. Therefore, the accuracy of nozzle blockage monitoring is low.

[0059] In summary, how to design a technical solution to the problem of low accuracy in nozzle clogging status monitoring caused by existing technologies is the problem that this application urgently needs to solve.

[0060] Therefore, in view of the above-mentioned technical problems existing in the prior art, the embodiments of this application provide a nozzle clogging status monitoring method, device, equipment, medium and product, aiming to effectively improve the accuracy of nozzle clogging status monitoring.

[0061] The following describes the application scenarios of a nozzle clogging status monitoring method, apparatus, device, medium, and product provided in the embodiments of this application. These application scenarios are merely examples, intended to help those skilled in the art understand the technical content of this application, but do not imply that the embodiments of this application cannot be used in other devices, systems, environments, or scenarios.

[0062] Manufacturing of semiconductors or organic light-emitting diodes: The nozzle clogging status monitoring method, apparatus, equipment, medium and product provided in the embodiments of this application can be used to detect in real time whether a single nozzle in a multi-hole array is clogged, avoid defects such as broken wires and insufficient adhesive, and ensure production consistency and yield.

[0063] Figure 1 This is a schematic diagram illustrating an application scenario of a nozzle clogging status monitoring method provided in an embodiment of this application, such as... Figure 1 As shown, it includes: terminal 110 and server 120.

[0064] Terminal 110 can be a client that initiates the nozzle blockage status monitoring process. Server 120 is used to control the acquisition of the blockage status judgment result and blockage level corresponding to the target nozzle.

[0065] Figure 2 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 1 The execution entity in this embodiment can be Figure 1 The server 120 in this embodiment can also be other computer-related devices; no particular limitation is made to this embodiment. The nozzle clogging status monitoring method provided in this embodiment includes the following steps:

[0066] S101. Obtain the reflected light patterns corresponding to the exit sections of multiple nozzles.

[0067] In this embodiment, the nozzle outlet section, i.e. the outlet end face of the nozzle channel, is an annular plane with a central circular hole. The outer circle of the annular plane is the outer diameter end face of the nozzle, which is the outer boundary of the nozzle tube wall; the inner circle of the annular plane is the boundary of the nozzle channel, which is the opening edge of the channel; the annular region is the end face region between the inner and outer circles, which is the main area for sediment accumulation and also the main target of laser irradiation.

[0068] When there is no blockage, the annular area corresponding to the nozzle outlet section is a flat reflective surface. When the laser irradiates the nozzle outlet section, it will produce bright and uniform specular reflection or diffuse reflection. As a result, the reflected light pattern captured by the camera appears as a clear and complete bright ring.

[0069] When blockage occurs, the evaporated coating material first condenses on the lower-temperature annular area, gradually forming deposits and causing edge buildup. These deposits on the annular area gradually grow inward, slowly encroaching on the boundary of the nozzle channel. The deposits cross the inner circle boundary from the annular area and hang into the nozzle channel like a curtain, causing partial blockage.

[0070] Therefore, the clogging status of the nozzle orifice can be inferred from changes in the annular region. When the reflected light pattern captured by the camera is a clear and complete bright ring, the nozzle orifice is considered to be in a healthy state and unobstructed. When local dark spots or protrusions appear on the bright ring corresponding to the reflected light pattern captured by the camera, it is believed that there is accumulation at the edge of the nozzle orifice due to the accumulation absorbing light or changing the direction of reflection. When the bright ring corresponding to the reflected light pattern captured by the camera is broken or gapped, and the dark area in the center of the nozzle orifice expands, it is believed that there is an object blocking the nozzle orifice, causing partial blockage, because the accumulation intruding from the edge casts a shadow.

[0071] In a vacuum evaporation equipment, multiple nozzles are mounted in a known fixed array on the inner wall of a cylindrical cavity. Upon installation or first use, a calibration process is performed: In a clean, unobstructed environment, a laser is projected and an image of the overall reflected light pattern corresponding to all nozzle exit sections is captured. This ensures that all bright rings corresponding to the nozzle exit sections are clear and complete. Image processing algorithms, such as speckle analysis, are then used to automatically identify all bright connected regions in the overall reflected light pattern. If two bright ring connected regions are adhered, they can be separated through morphological segmentation or based on prior knowledge of the array spacing. If the adhesion is severe and automatic separation is impossible, manual segmentation by the operator is also possible. Each bright connected region represents a reflected light pattern corresponding to one nozzle exit section. For cases where automatic identification is inaccurate, such as low contrast in edge areas, the operator can manually select or click on the position of each nozzle exit section in the image. Based on the identification results of the bright connected regions, a coordinate position record is generated for each nozzle exit section and written into a coordinate mapping table. The coordinate mapping table includes the physical index rows and columns corresponding to all nozzle exit sections, the pixel coordinates in the overall reflected light pattern, and the center coordinates. Optionally, the coordinate mapping table can be recalibrated periodically, for example, after each batch of vapor deposition, to eliminate drift.

[0072] In subsequent use, multiple annular laser spots are guided and projected onto multiple nozzle exit sections via a projection lens. A light-collecting lens then collects the reflected light from these illuminated nozzle exit sections, forming a clear overall reflected light pattern corresponding to all nozzle exit sections on the camera's sensor. Based on the newly captured overall reflected light pattern corresponding to all nozzle exit sections and a coordinate mapping table, the reflected light patterns corresponding to each nozzle exit section can be obtained.

[0073] For the overall reflected light pattern corresponding to all nozzle exit sections captured, only the elliptical effective imaging area is extracted and processed. This area strictly corresponds to the physical area of ​​the nozzle array covered by the laser projection, eliminating stray reflections from the inner wall of the cavity and interference from structural obstruction.

[0074] A direct projection scheme is adopted, in which the laser beam directly illuminates the nozzle exit section without passing through any reflectors. This design avoids aberrations introduced by reflectors, reduces the space occupied by the vacuum cavity, and ensures optical path stability and pattern positioning accuracy.

[0075] S102. Decouple the target reflected light pattern to obtain an isometric unfolded diagram in a two-dimensional plane coordinate system.

[0076] In this embodiment, the target reflected light pattern is the reflected light pattern corresponding to the target nozzle exit cross-section, and the target nozzle exit cross-section is any one of multiple nozzle exit cross-sections.

[0077] In this embodiment, the reflected light pattern corresponding to a certain nozzle outlet cross section is analyzed to determine whether the nozzle is blocked and the level of blockage after blockage occurs.

[0078] For cylindrical cavities equipped with nozzle arrays, the target surface is a non-planar curved surface. Conventional direct illumination causes radial stretching distortion of the circular illumination pattern, directly affecting the geometric measurement accuracy of the reflected light pattern corresponding to the nozzle exit section. Therefore, a single-frame unfolding imaging technique is adopted. Through precise optical path calibration and geometric mapping algorithms, the reflected light pattern is decoupled in real time into an isometric unfolded image in a two-dimensional planar coordinate system. That is, through mathematical mapping, the reflected light pattern is converted into an image in a Cartesian coordinate system to eliminate the spatial nonlinear distortion introduced by curvature.

[0079] As an optional implementation, in order to transform the reflected light pattern from the polar coordinate system to the rectangular coordinate system, firstly, a radial and circumferential sampling grid is established in the polar coordinate system with the center of the nozzle channel as the origin. Secondly, the annular bright ring corresponding to the reflected light pattern is sampled at equal angular intervals along the circumferential direction, and the brightness distribution at each angle is extracted along the radial direction to obtain a discrete sampling point set. Then, the discrete sampling point set is interpolated to generate a brightness matrix on a regular grid, and the brightness matrix is ​​smoothed and filtered to suppress optical noise and sampling discontinuities. The filtered matrix is ​​used as an isometric unfolded diagram reflecting the complete brightness information of the annular region of the nozzle exit section.

[0080] S103. Obtain the target imaging area based on the isometric unfolded diagram.

[0081] In this embodiment, the target imaging area is the effective imaging area (Region of Interest, ROI). ROI refers to a rectangular area in the equidistant unfolded diagram that contains complete annular bright ring brightness information. This rectangular area corresponds to the effective optical signal range of the annular reflective surface of the nozzle exit section.

[0082] As an optional implementation method, the specific steps for extracting the ROI from the isometric unfolded diagram are as follows: First, determine the inner and outer diameter boundaries corresponding to the annular bright rings along the vertical (radial) direction in the isometric unfolded diagram, that is, find the two row indices where the brightness values ​​change significantly, which correspond to the inner diameter r in the original polar coordinates. min and outer diameter r maxSecondly, determine the starting and ending positions of the angle corresponding to the complete annular bright ring along the horizontal axis (angle), which is usually from 0° to 360°. If the sampling starting angle is not 0°, angle translation or cyclic shift is required. Then, based on the inner diameter, outer diameter and angle range mentioned above, cut out the corresponding rectangular sub-region in the equidistant unfolded diagram. Take the range between the inner diameter boundary and the outer diameter boundary as the row range, and the range between the starting angle and the ending angle as the column range. Take this rectangular sub-region as the effective imaging area.

[0083] S104. Perform adaptive binarization processing on the target imaging area to obtain the laser feature pattern.

[0084] In this embodiment, the laser feature pattern is a high-brightness target, and the background noise is the low-grayscale cavity surface. Based on the local Otsu thresholding method, the laser feature pattern and background noise are dynamically separated to ensure that the pattern segmentation accuracy is maintained at different deposition stages, such as the initial clean state, the intermediate uniform deposition state, or the final locally blocked state.

[0085] As an optional implementation, firstly, the effective imaging area can be divided into multiple local regions according to a fixed size. Within each local region, the optimal segmentation threshold is dynamically calculated. For example, the Otsu method can be used to calculate the optimal threshold for that local region, or the average brightness or Gaussian weighted average of each local region can be used as the simplified binarization threshold. Secondly, the grayscale value of each pixel is compared with the corresponding local region threshold. If it is greater than the threshold, the pixel is assigned a value of 255 (white); otherwise, it is assigned a value of 0 (black), thus generating a preliminary binary image. Then, morphological filtering is applied to the preliminary binary image to remove isolated noise and fill small holes. Finally, the processed binary image is output, which is a laser feature pattern containing features such as the edge of a bright ring, notches, or dark spots. Pixels assigned a value of 255 (white) represent the normal reflection area of ​​the bright ring, while pixels assigned a value of 0 (black) represent the background, the center of the aperture, dark spots, or notches. These two types of pixels together constitute a complete laser feature pattern for subsequent AI image recognition.

[0086] S105. Based on the laser feature pattern, obtain the grayscale image corresponding to the exit section of the target nozzle.

[0087] In this embodiment, the grayscale image is a standardized grayscale image that can be used for AI image recognition.

[0088] As an optional implementation, firstly, the pixel values ​​in the binarized laser feature pattern are converted from integer to floating-point type and divided by 255.0 to map the pixel value range to [0.0, 1.0]. Then, the floating-point grayscale matrix is ​​normalized by interpolation to a preset fixed size, such as 224×224 or 128×128 pixels, to meet the input requirements of the AI ​​model. The scaled grayscale image is then used as a standardized grayscale image for AI image recognition.

[0089] S106. Input the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle.

[0090] In this embodiment, before the vacuum evaporation equipment is installed or used for the first time, and after confirming that all nozzle outlet sections are in a clean, unblocked, and healthy state, a calibration process is performed: a ring laser spot array is projected to illuminate all nozzle outlet sections, and the overall reflected light pattern is captured by a camera; the reflected light pattern corresponding to each nozzle outlet section is extracted from the overall reflected light pattern according to the coordinate mapping table; the reflected light pattern corresponding to each nozzle outlet section is sequentially subjected to polar coordinate expansion, ROI extraction, binarization, and grayscale processing to obtain a standard grayscale image of each nozzle outlet section in a healthy state, that is, a pre-stored normal state grayscale image of each nozzle outlet section is obtained, which is associated with the physical row index and physical column index of the nozzle and stored in the database for subsequent actual testing.

[0091] The pre-trained congestion state recognition model can be obtained by training a neural network model used for congestion state and congestion level judgment. For example, the pre-trained congestion state recognition model is a trained residual neural network (RNN).

[0092] An initial residual neural network is constructed using the labeled grayscale image of the nozzle outlet cross-section, a pre-stored grayscale image of the normal state, and corresponding clogging state and clogging level labels as training samples. During training, the grayscale image of the nozzle outlet cross-section and the pre-stored grayscale image of the normal state are used as input, and the corresponding clogging state and clogging level are used as supervision signals. The cross-entropy loss function is used to calculate the error between the predicted value and the true label, and the network weights are updated through backpropagation and an optimizer. The training continues until a preset number of iterations is reached or the loss function value converges to a preset threshold (e.g., the change in loss value is less than a certain threshold). The model stops training, and the trained residual neural network is obtained.

[0093] In actual testing, the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored grayscale image of the normal state are input together into the pre-trained blockage state recognition model, and the blockage state judgment result and blockage level of the target nozzle can be obtained from the output of the model.

[0094] Structured light laser patterns are precisely projected onto the nozzle array plane, and the shape of the reflected / transmitted light spots is analyzed in real time by a deep learning-based laser contour semantic segmentation algorithm. This algorithm, together with a highly robust real-time image processing engine, constitutes the core technology.

[0095] By decoupling the reflected light pattern into a two-dimensional isometric unfolded image, the geometric nonlinearity and radial stretching distortion caused by polar coordinate projection are fundamentally eliminated, significantly reducing the complexity of subsequent feature extraction and the risk of mismatch. Based on this, adaptive binarization processing of the target imaging area dynamically adapts to fluctuations in overall light intensity at different deposition stages, ensuring accurate and continuous segmentation boundaries between normal and defective areas in the laser feature pattern. Furthermore, the binarized feature pattern is converted into a standardized grayscale image and input into a pre-trained blockage state recognition model along with a pre-stored normal state grayscale image. This allows the model to directly compare pixel-by-pixel differences between the current state and a healthy baseline, quantitatively capturing minute grayscale attenuation, area loss, and edge distortion. Through deep extraction and weighted fusion of feature data, the model achieves accurate acquisition of blockage state judgment results and blockage levels, effectively solving the problems of low monitoring accuracy, easy missed detections, and false detections.

[0096] This application provides a method for monitoring nozzle blockage status, comprising: acquiring reflected light patterns corresponding to multiple nozzle outlet cross sections respectively; decoupling the target reflected light patterns to obtain an isometric unfolded diagram in a two-dimensional plane coordinate system; obtaining the target imaging area based on the isometric unfolded diagram; performing adaptive binarization processing on the target imaging area to obtain a laser feature pattern; obtaining a grayscale image corresponding to the target nozzle outlet cross section based on the laser feature pattern; and inputting the grayscale image corresponding to the target nozzle outlet cross section and a pre-stored normal state grayscale image into a pre-trained blockage status recognition model to obtain the blockage status judgment result and blockage level corresponding to the target nozzle. The above method achieves the following technical effects: By decoupling the reflected light pattern into a two-dimensional isometric unfolded image, the geometric nonlinearity and radial stretching distortion caused by polar coordinate projection are fundamentally eliminated, thus significantly reducing the complexity of subsequent feature extraction and the risk of mismatch. Based on this, adaptive binarization processing of the target imaging area dynamically adapts to fluctuations in overall light intensity at different deposition stages, ensuring accurate and continuous segmentation boundaries between normal reflection areas and defect areas in the laser feature pattern. Furthermore, the binarized feature pattern is converted into a standardized grayscale image and input into a pre-trained blockage state recognition model along with a pre-stored normal state grayscale image. This allows the model to directly compare the pixel-by-pixel differences between the current state and the health baseline, quantitatively capturing minute grayscale attenuation, area loss, and edge distortion. Through deep extraction and weighted fusion of feature data by the model, accurate acquisition of blockage state judgment results and blockage levels is achieved, effectively solving the problems of low monitoring accuracy, easy missed detections, and false detections.

[0097] Figure 3 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 2 In the nozzle clogging status monitoring method provided in this embodiment, S106 includes the following steps:

[0098] S201. Input the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image into the pre-trained blockage state recognition model to obtain the blockage feature data corresponding to the target nozzle.

[0099] In this embodiment, the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored grayscale image of the normal state are input into the pre-trained blockage state recognition model. This model can obtain the blockage feature data corresponding to the target nozzle by extracting depth features.

[0100] S202. Based on the clogging characteristic data corresponding to the target nozzle, obtain the clogging status judgment result and clogging level corresponding to the target nozzle.

[0101] In this embodiment, by comparing the clogging feature data corresponding to the target nozzle with the corresponding threshold, the clogging status judgment result and clogging level of the target nozzle can be obtained.

[0102] Using feature data to determine congestion status and classify levels has stronger robustness and interpretability. It also facilitates the determination of the weight coefficients and classification thresholds of each feature through experimental calibration, making the judgment results more objective and stable. Furthermore, it can quantitatively output specific levels, providing clear quantitative basis for process traceability and closed-loop control.

[0103] Figure 4 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 3 In the nozzle clogging status monitoring method provided in this embodiment, the clogging feature data includes central dark area data, annular bright band data, and edge grayscale gradient data. S202 includes the following steps:

[0104] S301. Based on the central dark area data, annular bright band data and edge grayscale gradient data corresponding to the target nozzle, obtain the clogging status judgment result and clogging level of the target nozzle.

[0105] In this embodiment, the grayscale image corresponding to the target nozzle outlet cross-section and the pre-stored normal state grayscale image are input into a pre-trained blockage state recognition model. This model can extract depth features and output feature vectors of three branches: central dark area data, annular bright band data, and edge grayscale gradient data. The central dark area data characterizes the degree to which the grayscale value of the nozzle center region is lower than the normal threshold, the proportion of the dark area, and the grayscale attenuation rate; the annular bright band data characterizes the grayscale uniformity, integrity, circumferential continuity, and brightness attenuation degree of the annular bright ring; and the edge grayscale gradient data characterizes the grayscale gradient amplitude, gradient direction consistency, and edge roughness of the light-transmitting aperture edge.

[0106] After laser irradiation, the reflected light intensity distribution image is acquired. By comparing it with a standard normal state image at the pixel level, blockage characteristics such as expansion of the central dark area, breakage of the annular bright band, and abnormal gray-scale gradient at the edge can be quantitatively identified, thereby directly determining whether a blockage has occurred and its severity level.

[0107] A difference image is generated by subtracting the grayscale image corresponding to the exit section of the target nozzle from the pre-stored grayscale image of the normal state pixel by pixel. The grayscale value of each pixel in the difference image represents the brightness difference at that location between the current state of the target nozzle and the normal state of the target nozzle. Based on the geometry of the nozzle, the difference image is divided into three feature regions: the central dark area corresponding to the center region of the nozzle, the annular bright band area corresponding to the annular bright ring area, and the edge transition area corresponding to the edge region of the light aperture.

[0108] Within the central dark area, the positive differences in the difference image are calculated, i.e., pixels whose gray values ​​corresponding to the current state of the target nozzle are less than those corresponding to the normal state of the target nozzle are obtained. Then, the mean gray value of these positively differing pixels, the area ratio of these positively differing pixels to the total number of pixels in the central dark area, and the ratio of the maximum gray value of these positively differing pixels to the mean gray value of the normal state are calculated. The central dark area data is obtained by weighted fusion of the above three indicators. The weight coefficients corresponding to the above three indicators can be determined by experimental calibration, and the sum of the three weight coefficients is 1. For example, the weight coefficients corresponding to the above three indicators are 0.4, 0.4, and 0.2, respectively.

[0109] Within the annular bright band region, the positive difference of the difference image is calculated, i.e., the pixels in the difference image whose gray value corresponding to the current state of the target nozzle is less than the gray value corresponding to the normal state of the target nozzle are obtained; then, the mean gray value of the positive difference pixels is calculated, and the ratio of the mean gray value of the positive difference pixels to the mean gray value of the normal state is calculated, which is used as the normalized brightness decay rate; the annular bright band region is divided into 16 or 32 fan-shaped sub-regions along the circumference, and the area ratio of the positive difference pixels in each sub-region is calculated, and the standard deviation of the area ratio of each sub-region is calculated as the circumferential non-uniformity; the area ratio of the non-difference pixels in the annular bright band region to the total number of pixels in the entire annular bright band region is calculated as the brightness band integrity, and the non-difference pixels are the pixels whose |gray value corresponding to the current state of the target nozzle in the difference image - gray value corresponding to the normal state of the target nozzle| ≤ ε, for example, ε = 3~5 gray levels; the above three indicators are weighted and fused to obtain the annular bright band data. The weight coefficients corresponding to the above three indicators can be determined by experimental calibration, and the sum of the three weight coefficients is 1. For example, the weight coefficients corresponding to the above three indicators are 0.3, 0.5 and 0.2 respectively.

[0110] Within the edge transition zone, firstly, based on the actual light-transmitting contour of the aperture, a strip-shaped region extending by a preset pixel width on both sides of the contour line is designated as the edge transition zone, for example, a preset pixel width of 3 to 5 pixels. Then, the Sobel operator is used to calculate the gradient magnitude of each pixel in the edge transition zone of the difference image, and the mean of the gradient magnitude is calculated as the steepness of the gray-level difference between the two sides of the edge. Next, the gradient direction angle of each pixel within the edge transition zone is calculated, and the circular standard deviation of the gradient direction angles of all pixels is statistically analyzed as a measure of gradient direction consistency. Then, the local variance of the gradient magnitude within the edge transition zone is calculated as the edge roughness. The above three indicators are then weighted and fused to obtain the edge gray-level gradient data. The weight coefficients corresponding to the above three indicators can be determined through experimental calibration, and the sum of the three weight coefficients is 1. For example, the weight coefficients corresponding to the above three indicators are 0.4, 0.3, and 0.3, respectively.

[0111] After obtaining the central dark area data, the annular bright band data and the edge gray gradient data, the central dark area data is compared with a first data threshold, the annular bright band data is compared with a second data threshold, and the edge gray gradient data is compared with a third data threshold. If all of the central dark area data, the annular bright band data and the edge gray gradient data are less than the corresponding data thresholds, the judgment result of the clogging state corresponding to the target nozzle is that no clogging occurs; if at least one of the central dark area data, the annular bright band data and the edge gray gradient data is greater than or equal to the corresponding data threshold, the judgment result of the clogging state corresponding to the target nozzle is that clogging occurs. Optionally, the first data threshold, the second data threshold and the third data threshold may be determined by a statistical quantile method, that is, a large number of samples of nozzles in an unclogged state are collected, the central dark area data, the annular bright band data and the edge gray gradient data of each sample are calculated respectively, for each type of data, the 95% quantile or 99% quantile of its distribution is calculated, and the quantile is taken as the corresponding data threshold. For example, each of the first data threshold, the second data threshold and the third data threshold may be 0.25.

[0112] When it is detected that the judgment result of the clogging state corresponding to the target nozzle is that clogging occurs, the central dark area data, the annular bright band data and the edge gray gradient data are respectively normalized to the interval [0, 1], and then weighted and fused to calculate a comprehensive clogging score: S=α×D dark +β×D ring +γ×D gradient , where α, β and γ are preset weight coefficients, and α+β+γ=1. α, β and γ may be determined by an experimental calibration method, for example, α=0.4, β=0.3, γ=0.3; D dark is the central dark area data, D ring is the annular bright band data, D gradient is the edge gray gradient data, and S is the comprehensive clogging score.

[0113] The comprehensive clogging score S is mapped to a corresponding clogging grade according to a preset score interval where the S is located: if S<T1, the clogging grade corresponding to the target nozzle is slight clogging; if T1≤S<T2, the clogging grade corresponding to the target nozzle is moderate clogging; if S≥T2, the clogging grade corresponding to the target nozzle is severe clogging. T1 and T2 are preset grade thresholds, which may be determined by the experimental calibration method, for example, T1=0.5, T2=0.7.

[0114] By integrating features from three dimensions—central dark area data, annular bright band data, and edge grayscale gradient data—the degradation degree of nozzle clogging in different regions can be comprehensively characterized, avoiding misjudgment based on a single feature and significantly improving the robustness and accuracy of clogging status assessment. At the same time, by obtaining a comprehensive clogging score through weighted fusion, it is possible to finely classify mild, moderate, and severe clogging levels based on preset thresholds, providing process engineers with quantitative and traceable clogging assessment results, which facilitates the implementation of graded maintenance and closed-loop control.

[0115] Figure 5 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 4 In the nozzle clogging status monitoring method provided in this embodiment, S105 includes the following steps:

[0116] S401. Based on the laser feature pattern, obtain the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section.

[0117] In this embodiment, edge detection is performed on the laser feature pattern. For example, the Canny operator is used to extract the boundary line between the white pixel area and the black pixel area. This boundary line is the actual light-transmitting profile of the aperture. The surface deposition profile corresponds to the area where the reflection is weakened due to the deposit. Morphological closing operation is performed on the laser feature pattern to fill the isolated black pixels inside the white normal reflection area. Then, the difference area between the boundary of the filled white area and the boundary of the original white area is calculated, and this difference area is used as the surface deposition profile.

[0118] The edge sharpness corresponding to the exit section of the target nozzle can also be obtained from the laser feature pattern. A circularity fit is performed on the actual aperture light-transmitting profile, and the radial standard deviation of each pixel on the actual aperture light-transmitting profile to the fitted circle is calculated as an indirect measure of the edge sharpness of each pixel. The actual aperture light-transmitting profile, edge sharpness, and surface deposition profile are micrometer-level features.

[0119] S402. Based on the actual light transmission profile and surface deposition profile of the orifice corresponding to the target nozzle outlet section, obtain the grayscale image corresponding to the target nozzle outlet section.

[0120] In this embodiment, firstly, a blank grayscale image with the same size as the laser feature pattern is generated, and all pixels are initialized to 0 (black). Secondly, based on the position information of the actual light-transmitting contour of the aperture, the closed contour is drawn on the blank grayscale image, and the area inside the contour is filled with white (grayscale value 255) to represent the normal light-transmitting area. Then, based on the position information of the surface deposition contour, the pixel area corresponding to the surface deposition contour is set to medium gray, for example, grayscale value 128, within the filled white area to distinguish it from the normal light-transmitting area. Next, the drawn image is subjected to size normalization processing, for example, scaling to 224×224 pixels, and grayscale value normalization processing is performed, for example, converting the pixel values ​​of the scaled image from integer type to floating point type and dividing by 255.0, so that the pixel value range is mapped to [0.0, 1.0], and finally a standardized grayscale image that can be used for AI image recognition is obtained.

[0121] Based on the actual light transmission profile and surface deposition profile of the orifice corresponding to the target nozzle outlet section, a grayscale image corresponding to the target nozzle outlet section is obtained. This can improve the characterization ability of the grayscale image of the target nozzle outlet section for orifice shrinkage, edge accumulation and surface deposition, reduce the probability of misjudgment caused by simple brightness changes, and enhance the sensitivity to blockage morphology, thereby improving the accuracy and stability of nozzle blockage monitoring.

[0122] Figure 6 A flowchart illustrating a nozzle clogging status monitoring method provided in this application embodiment. Figure 5 In the nozzle clogging status monitoring method provided in this embodiment, S402 includes the following steps:

[0123] S501. Based on the actual light transmission profile of the orifice corresponding to the target nozzle outlet section, obtain the flow area attenuation rate and edge distortion degree corresponding to the target nozzle outlet section.

[0124] In this embodiment, the light transmission area A of the nozzle under normal reference conditions is obtained. reference For example, the normal reference state can be the factory standard or the initial clean state; calculate the actual area A enclosed by the actual light transmission profile of the aperture. actual The area can be obtained by counting pixels within the outline or by using the polygon area formula; calculate the flow area attenuation rate: , where A reference A represents the light transmission area of ​​the nozzle under normal reference conditions. actual A is the actual area enclosed by the actual light transmission profile of the aperture. reaAttenuation This represents the rate of decrease in circulation area.

[0125] The specific steps for obtaining edge distortion are as follows: Perform roundness fitting on the actual light transmission profile of the aperture to obtain the center and radius of the fitted circle. ; Calculate the distance from each pixel on the actual light transmission profile of the aperture to the center of the fitted circle. ; Calculate the relative deviations between each distance and the fitted radius: ,in, Let be the relative deviation between the distance from the i-th pixel to the center of the fitted circle and the fitted radius. Let be the distance from the i-th pixel to the center of the fitted circle. The fitted radius is used to calculate the edge distortion. , where V is the edge distortion degree.

[0126] The flow area attenuation rate is used to indicate how much the channel is blocked, i.e., the change in the channel size; the edge distortion is used to indicate the change in the channel shape.

[0127] S502. Based on the flow area attenuation rate and edge distortion degree corresponding to the target nozzle outlet section, the consistency deviation of the multi-hole array is obtained.

[0128] In this embodiment, the standard deviation is calculated based on the flow area attenuation rate of all nozzles in the nozzle array. The standard deviation obtained for the flow area attenuation rate is taken as the first standard deviation. The standard deviation is calculated based on the edge distortion degree of all nozzles in the nozzle array. The standard deviation obtained for the edge distortion degree is taken as the second standard deviation. The consistency deviation of the multi-hole array is calculated as w1×σ_area+w2×σ_distortion, where w1 and w2 are preset weighting coefficients, which can be determined by experimental calibration, for example, w1=0.5, w2=0.5; σ_area is the first standard deviation; and σ_distortion is the second standard deviation.

[0129] It can directly quantify the nozzle clogging status. The actual light transmission profile of the nozzle orifice is directly obtained through laser imaging, and the effective flow area attenuation rate, edge distortion degree and multi-hole array consistency deviation are calculated with sub-pixel accuracy. The output is a traceable, calibrable raw process data stream that can be incorporated into statistical process control (SPC).

[0130] S503. When the detected consistency deviation of the multi-hole array is less than the preset deviation threshold, the grayscale image corresponding to the target nozzle exit section is obtained based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section.

[0131] In this embodiment, if the detected consistency deviation of the multi-hole array is less than a preset deviation threshold, the nozzle array is deemed to be consistent, a pass flag is output, and the process is allowed to proceed to the next step. The method for obtaining the grayscale image corresponding to the target nozzle exit cross-section based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit cross-section has been mentioned in the previous embodiment and will not be repeated here. If the detected consistency deviation of the multi-hole array is not less than the preset deviation threshold, the nozzle array is deemed to be inconsistent, a fail flag is output, and an alarm is triggered. Optionally, the preset deviation threshold can be determined by experimental calibration, for example, a preset deviation threshold of 3%.

[0132] When the consistency deviation of the porous array is detected to be less than the preset deviation threshold, subsequent steps are performed. This ensures that the consistency of the porous array meets the requirements and avoids distortion or error accumulation in subsequent steps due to excessive deviation, thereby improving the reliability and repeatability of detection or process.

[0133] Based on the above embodiments, this application provides a nozzle clogging status monitoring method. In the nozzle clogging status monitoring method provided in this embodiment, the reflected light pattern is obtained by the reflection of multiple annular laser spots from multiple nozzle exit sections after the diffractive optical element lens guides multiple annular laser spots to multiple nozzle exit sections. The multiple annular laser spots are obtained by converting the initial rectangular laser spot by the diffractive optical element lens.

[0134] In this embodiment, the initial rectangular laser spot can be obtained through a laser module, and the cross-section of the initial rectangular laser spot is rectangularly distributed. The long axis of the initial rectangular laser spot extends along the nozzle array direction, and can cover the multi-hole area in one go along the nozzle array axis; the short axis precisely matches the diameter of a single circular nozzle, realizing multi-hole synchronous imaging in a single projection, which greatly improves detection efficiency and signal-to-noise ratio. The laser beam can reach the target observation point, i.e., the nozzle exit cross-section, without attenuation and with high fidelity. Lasers possess irreplaceable optical characteristics such as monochromaticity, high directionality, and ultra-high brightness, and are the only light source that can stably project a controllable beam into narrow, deep hole structures.

[0135] The laser module parameters are as follows: wavelength is 650nm red, 520nm green, or 450nm blue; output power is 100mW; lens angle is 40.63°; in terms of rectangular dimensions, the X-axis angle is 40.63°, corresponding to a length of 370mm at a distance of 0.5m; the Y-axis angle is 21.63°, corresponding to a length of 191mm at a distance of 0.5m. For example, the actual laser module is 70mm long and 16mm wide.

[0136] To achieve high-fidelity illumination from a circular nozzle orifice, a diffractive optical element (DOE) lens is integrated into the light path of the initial rectangular laser spot. The DOE lens is designed as an array of microstructures. Through phase modulation, the DOE lens shapes the initial rectangular laser spot into a ring-shaped array corresponding to the nozzle array. Each ring-shaped laser spot possesses sharp edges and a uniform center, precisely matching the geometric features of the nozzle exit cross-section, thus providing a high-quality optical foundation for subsequent acquisition of reflected light patterns.

[0137] To achieve a one-to-one correspondence between each annular laser spot and each nozzle exit section, a calibration procedure is performed upon completion of the vacuum evaporation equipment installation or its first use: Based on the row spacing and column spacing of the nozzle array, the row direction magnification of the DOE Lens surface projected onto the plane of the nozzle array by the projection lens, and the column direction magnification of the DOE Lens surface projected onto the plane of the nozzle array by the projection lens, the row direction period and column direction period of the DOE Lens are obtained. This ensures that the geometric distribution of the annular spot array guided by the DOE Lens matches the theoretical position of the nozzle array. The row direction period of the DOE Lens = D_row / β_row, and the column direction period of the DOE Lens = D_col / β_col, where D_row is the row spacing of the nozzle array, β_row is the row direction magnification of the DOE Lens surface projected onto the plane of the nozzle array by the projection lens, D_col is the column spacing of the nozzle array, and β_col is the column direction magnification of the DOE Lens surface projected onto the plane of the nozzle array by the projection lens. The row direction period of the DOE Lens and the DOE Lens... Using the column direction period of the Lens as design input, an array-type microstructure DOELens is fabricated. The position and orientation of the DOE Lens are adjusted using a five-dimensional precision adjustment frame to initially align the annular laser spot array with the exit sections of multiple nozzles. A camera is used to acquire superimposed images of the annular laser spot array and the exit sections of multiple nozzles. Image processing algorithms are used to identify the center of each annular laser spot and the center of each nozzle exit section, calculating the offset between them to generate an offset mapping table. During actual testing, coordinate compensation is performed on the imaging position of each annular laser spot according to the offset mapping table to ensure that the annular laser spot correctly corresponds to the nozzle exit section.

[0138] The structured light patterned laser module supports dynamic modulation of the rectangular spot size and intensity distribution.

[0139] The straight-line distance between the laser emission source or laser outlet and the target nozzle outlet section is the standard projection distance, which is 0.5m.

[0140] Optionally, the DOE Lens can also reshape the initial rectangular laser spot into a quasi-circular distribution through phase modulation.

[0141] The laser module emits an initial rectangular laser spot, which is shaped by the DOE Lens into multiple annular laser spots that match the exit cross-section of each nozzle. The annular laser spots are then projected onto the corresponding nozzle exit cross-sections. The light is focused by the lens and the exposure is controlled by the camera shutter. Finally, the camera obtains the reflected light patterns corresponding to the multiple nozzle exit cross-sections, which are then analyzed by the server.

[0142] Optionally, in the direct projection scheme, the laser module and the camera are located on opposite sides of the nozzle exit section. The laser beam is incident at a certain angle, typically 15° to 45°, onto the nozzle exit section. The camera is positioned along the specular reflection direction or a specific scattering direction of the laser beam to capture the reflected light pattern. Specifically, the angle between the optical axis of the laser module and the normal to the nozzle surface is the angle of incidence θ, and the angle between the optical axis of the camera and the normal to the nozzle surface is the angle of reflection θ. The angle of incidence must equal the angle of reflection to ensure that the camera receives the strongest reflected light signal.

[0143] The camera system is configured as follows: it uses a charge-coupled device (CCD) camera with more than 5 million pixels and a frame rate of 24fps; it is equipped with a low-distortion telecentric lens with a principal ray angle deviation of less than 0.1°, which can ensure no perspective scaling error in the entire field of view, and a bandpass filter that matches the laser wavelength is installed.

[0144] Figure 7 This is a schematic diagram of the structure of a monitoring system provided in an embodiment of this application. Figure 7 As shown, the monitoring system includes a laser module and a camera, as well as a lens, a camera shutter, a linear guide rail, and a monitoring device. The lens is used to converge the laser light reflected from the nozzle exit section, forming a clear reflected light pattern on the camera sensor. The camera shutter controls the exposure time of the camera sensor, ensuring an appropriate amount of light intake and avoiding overexposure or underexposure, thereby capturing a high-contrast reflected light pattern. The lens is mounted in front of the camera body, located in the optical path of the reflected light, i.e., the lens optical axis is aligned with the center of the nozzle exit section. The camera shutter is integrated inside the camera or between the lens and the sensor, and is triggered synchronously with the camera. The monitoring device is an apparatus capable of implementing the nozzle clogging status monitoring method provided in the above embodiments.

[0145] Linear guides are used to support the camera and lens assembly, driving them to move precisely along a preset direction to achieve automatic focusing, multi-position scanning, or adapt to the detection needs of nozzle arrays of different specifications, ensuring that the camera can be aligned with the nozzle outlet cross-section at different positions. The linear guides are mounted on an external or internal support of the vacuum chamber, and the camera and lens assembly are fixed to the guides by sliders. The direction of the guides is parallel to the row or column direction of the nozzle array, or perpendicular to the nozzle surface (for focusing).

[0146] The thermal shutter is used to automatically close during non-monitoring periods, i.e., during the evaporation process, physically isolating the laser emission path from the camera's receiving optical path. It only opens momentarily during the preset monitoring window, achieving in-situ monitoring with zero interference, zero pollution, and zero heat load, avoiding local temperature rise in the cavity, abnormal sublimation of the evaporation source material, and contamination of the optical component coating caused by continuous laser irradiation.

[0147] By directly projecting a laser beam into the nozzle orifice, the spatial distribution characteristics of reflected and scattered light are collected to reconstruct the three-dimensional morphology of the nozzle outlet cross-section and generate a standardized grayscale image for AI image recognition. Based on this image, three types of core data can be output: first, a normal state baseline map to establish a reference standard for the nozzle's health status; second, an abnormal state identification map, such as typical defect patterns like local blockage, edge deposition, or orifice deformation; and third, a quantitative analysis report that can be directly used for design iteration, including key process parameters such as blockage area ratio, orifice diameter change, and deformation vector diagram.

[0148] The process involves acquiring the original image and performing preprocessing, then setting the ROI, logging in, performing differential analysis and foreign object detection, and finally making a judgment or issuing an alarm.

[0149] Achieving precise matching between laser energy and nozzle array improves the parallel efficiency and uniformity of processing or inspection, while reducing the heat-affected zone and energy waste, ensuring consistent energy distribution and clear edges at the exit of each nozzle.

[0150] Figure 8 This is a schematic diagram of a nozzle clogging status monitoring device provided in an embodiment of this application. Figure 8 As shown, in this embodiment, the nozzle clogging status monitoring device includes:

[0151] The reflected light pattern acquisition module 601 is used to acquire the reflected light patterns corresponding to the exit sections of multiple nozzles.

[0152] The decoupling module 602 is used to decouple the target reflected light pattern to obtain an isometric unfolded diagram in a two-dimensional plane coordinate system; wherein, the target reflected light pattern is the reflected light pattern corresponding to the target nozzle outlet section, and the target nozzle outlet section is any one of multiple nozzle outlet sections.

[0153] The target imaging area acquisition module 603 is used to obtain the target imaging area based on the isometric unfolded diagram.

[0154] The binarization processing module 604 is used to perform adaptive binarization processing on the target imaging area to obtain the laser feature pattern.

[0155] The grayscale image acquisition module 605 is used to obtain the grayscale image corresponding to the exit section of the target nozzle based on the laser feature pattern.

[0156] The prediction module 606 is used to input the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle.

[0157] The nozzle clogging status monitoring device provided in this embodiment can perform... Figure 2 The technical solution of the nozzle clogging status monitoring method embodiment shown herein, its implementation principle and technical effect are similar to Figure 2 The embodiments of the nozzle clogging status monitoring method shown are similar and will not be described in detail here.

[0158] Meanwhile, the nozzle clogging status monitoring device provided in this application embodiment is a further refinement of the nozzle clogging status monitoring device provided in the previous embodiment.

[0159] Optionally, in this embodiment, the prediction module 606 is further configured to:

[0160] The grayscale image corresponding to the outlet section of the target nozzle and the pre-stored grayscale image of the normal state are input into the pre-trained blockage state recognition model to obtain the blockage feature data corresponding to the target nozzle; based on the blockage feature data corresponding to the target nozzle, the blockage state judgment result and blockage level corresponding to the target nozzle are obtained.

[0161] Optionally, in this embodiment, the clogging feature data includes central dark area data, annular bright band data, and edge grayscale gradient data. The prediction module 606 is further used for:

[0162] Based on the data of the central dark area, the annular bright band, and the edge grayscale gradient corresponding to the target nozzle, the clogging status judgment result and clogging level of the target nozzle are obtained.

[0163] Optionally, in this embodiment, the grayscale image acquisition module 605 is further configured to:

[0164] Based on the laser feature pattern, the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section are obtained; based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section, the grayscale image corresponding to the target nozzle exit section is obtained.

[0165] Optionally, in this embodiment, the grayscale image acquisition module 605 is further configured to:

[0166] Based on the actual light transmission profile of the aperture corresponding to the target nozzle exit section, the flow area attenuation rate and edge distortion degree corresponding to the target nozzle exit section are obtained; based on the flow area attenuation rate and edge distortion degree corresponding to the target nozzle exit section, the consistency deviation of the multi-hole array is obtained; when the consistency deviation of the multi-hole array is detected to be less than the preset deviation threshold, the grayscale image corresponding to the target nozzle exit section is obtained based on the actual light transmission profile of the aperture and the surface deposition profile corresponding to the target nozzle exit section.

[0167] Optionally, in this embodiment, the reflected light pattern in the reflected light pattern acquisition module 601 is obtained by the reflection of multiple annular laser spots after the diffractive optical element lens guides multiple nozzle exit sections, and the multiple annular laser spots are obtained by the diffractive optical element lens converting the initial rectangular laser spot.

[0168] Figure 9 This is a schematic diagram of an electronic device provided for an embodiment of this application. The electronic device is intended for various electronic devices that can be used to perform nozzle clogging status monitoring methods, such as microcomputers, single-chip microcomputers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the embodiments of this application described and / or claimed herein.

[0169] like Figure 9 As shown, the electronic device includes at least one processor 701 and a memory 702. The electronic device also includes a communication component 703. The processor 701, memory 702, and communication component 703 are connected via a bus 704.

[0170] In the specific implementation process, at least one processor 701 executes computer execution instructions stored in memory 702, causing at least one processor 701 to execute the nozzle blockage status monitoring method as executed on the electronic device side as described above.

[0171] The specific implementation process of processor 701 can be found in the above embodiment of nozzle blockage status monitoring method. Its implementation principle and technical effect are similar, and will not be repeated here.

[0172] In the above embodiments, it should be understood that the processor 701 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor 701 can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or being executed by a combination of hardware and software modules in the processor.

[0173] The memory 702 may include high-speed random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0174] Bus 704 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Bus 704 can be divided into an address bus, a data bus, and a control bus. For ease of illustration, the bus 704 in the accompanying drawings of this application is not limited to only one bus or one type of bus.

[0175] The above description addresses the functions implemented by electronic devices and main control devices, and introduces the solutions provided in the embodiments of this application. It is understood that, in order to achieve the above functions, the electronic device or main control device includes hardware structures and / or software modules corresponding to the execution of each function. By combining the units and algorithm steps of the various examples described in the embodiments disclosed in this application, the embodiments of this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the technical solutions of the embodiments of this application.

[0176] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-mentioned nozzle blockage status monitoring method.

[0177] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0178] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. The readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in application-specific integrated circuits (ASICs). Alternatively, the processor and the readable storage medium can exist as discrete components in an electronic device or a host device.

[0179] The memory 702 is the non-transitory computer-readable storage medium provided in the embodiments of this application. The non-transitory computer-readable storage medium in the embodiments of this application stores a computer pointer, used to enable the computer to execute the nozzle clogging status monitoring method provided in the embodiments of this application.

[0180] The memory 702, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules. The processor 701 executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory 702, thereby implementing the nozzle blockage status monitoring method in the above method embodiments.

[0181] In addition, this embodiment also provides a computer program product, including a computer program, which, when executed by a processor, is used to implement the nozzle clogging status monitoring method of the above embodiment.

[0182] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0183] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0184] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0185] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0186] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0187] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0188] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for monitoring nozzle blockage status, characterized in that, include: Obtain the reflected light patterns corresponding to the exit cross sections of multiple nozzles; The target reflected light pattern is decoupled by sampling the polar coordinate annular pattern at equal angular intervals along the circumference and extracting the brightness distribution at each angle along the radial direction to obtain a discrete sampling point set. This discrete sampling point set is then interpolated to generate a brightness matrix on a regular grid. The brightness matrix is ​​then smoothed and filtered, and the filtered matrix is ​​used as an isometric unfolded map in a two-dimensional plane coordinate system. The target reflected light pattern is the reflected light pattern corresponding to the exit section of the target nozzle, and the exit section of the target nozzle can be any one of multiple nozzle exit sections. The target imaging area is obtained based on the isometric unfolded diagram. Adaptive binarization processing is performed on the target imaging area to obtain a laser feature pattern containing annular bright ring edges, notches, or dark spots; Based on the laser feature pattern, contour information is obtained, and based on the contour information, a grayscale image corresponding to the target nozzle outlet cross-section is obtained. The grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image are input into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle.

2. The method according to claim 1, characterized in that, The step of inputting the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle includes: The grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image are input into the pre-trained blockage state recognition model to obtain the blockage feature data corresponding to the target nozzle. Based on the clogging characteristic data corresponding to the target nozzle, the clogging status judgment result and clogging level of the target nozzle are obtained.

3. The method according to claim 2, characterized in that, The blockage feature data includes central dark area data, annular bright band data, and edge grayscale gradient data; Accordingly, obtaining the blockage status judgment result and blockage level of the target nozzle based on the blockage feature data of the target nozzle includes: Based on the central dark area data, the annular bright band data, and the edge grayscale gradient data corresponding to the target nozzle, the clogging status judgment result and clogging level of the target nozzle are obtained.

4. The method according to claim 1, characterized in that, The step of obtaining contour information based on the laser feature pattern and obtaining a grayscale image corresponding to the target nozzle exit cross-section based on the contour information includes: Based on the laser feature pattern, the actual aperture light transmission profile and surface deposition profile corresponding to the target nozzle exit section are obtained; wherein, the actual aperture light transmission profile is the boundary line obtained by edge detection of the laser feature pattern; the surface deposition profile is the area where the difference between the front and back boundaries is filled after performing morphological closing operation on the laser feature pattern, corresponding to the area where the reflection of the nozzle exit section is weakened due to the deposition. Based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section, a grayscale image corresponding to the target nozzle exit section is obtained.

5. The method according to claim 4, characterized in that, The step of obtaining the grayscale image corresponding to the target nozzle exit cross-section based on the actual light transmission profile and surface deposition profile corresponding to the aperture of the target nozzle exit cross-section includes: Based on the actual light transmission profile of the aperture corresponding to the exit cross-section of the target nozzle, the flow area attenuation rate and edge distortion degree corresponding to the exit cross-section of the target nozzle are obtained; The flow area attenuation rate and edge distortion degree of each of the other nozzles in the multi-hole array are obtained. Combined with the flow area attenuation rate and edge distortion degree of the target nozzle, the consistency deviation of the multi-hole array is obtained. When the consistency deviation of the multi-hole array is detected to be less than a preset deviation threshold, a grayscale image corresponding to the target nozzle exit section is obtained based on the actual light transmission profile and surface deposition profile of the aperture corresponding to the target nozzle exit section.

6. The method according to claim 1, characterized in that, The reflected light pattern is obtained by the diffractive optical element lens guiding multiple annular laser spots to multiple nozzle exit sections, and then reflecting them from the multiple nozzle exit sections. The multiple annular laser spots are obtained by the diffractive optical element lens converting the initial rectangular laser spot.

7. A nozzle clogging status monitoring device, characterized in that, include: The reflected light pattern acquisition module is used to acquire the reflected light patterns corresponding to the exit sections of multiple nozzles. The decoupling module is used to decouple the target reflected light pattern. It samples the target reflected light pattern from a polar coordinate annular pattern along the circumference at equal angular intervals, and extracts the brightness distribution at each angle along the radial direction to obtain a discrete sampling point set. The discrete sampling point set is then interpolated to generate a brightness matrix on a regular grid. This brightness matrix is ​​then smoothed and filtered, and the filtered matrix is ​​used as an isometric unfolded map in a two-dimensional plane coordinate system. The target reflected light pattern is the reflected light pattern corresponding to the exit section of the target nozzle, and the target nozzle exit section can be any one of multiple nozzle exit sections. The target imaging area acquisition module is used to obtain the target imaging area based on the isometric unfolded diagram. The binarization processing module is used to perform adaptive binarization processing on the target imaging area to obtain a laser feature pattern containing annular bright ring edges, notches, or dark spots. The grayscale image acquisition module is used to obtain contour information based on the laser feature pattern, and to obtain a grayscale image corresponding to the target nozzle outlet cross section based on the contour information. The prediction module is used to input the grayscale image corresponding to the outlet section of the target nozzle and the pre-stored normal state grayscale image into the pre-trained blockage state recognition model to obtain the blockage state judgment result and blockage level corresponding to the target nozzle.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the nozzle clogging status monitoring method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the nozzle clogging status monitoring method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, is used to implement the nozzle clogging status monitoring method as described in any one of claims 1 to 6.

Images