Method for detecting edge breakage of silicon carbide polished wafer and related product

Abstract

The invention discloses a silicon carbide polished wafer edge breakage detection method and related products, which can be applied to the technical field of semiconductors, and the method comprises the following steps: placing a polished wafer to be detected on an objective table with a first plane facing upwards; performing image acquisition on the first plane based on a light source to obtain a first to-be-detected image; and based on an edge breakage detection model, performing feature comparison on the first to-be-detected image and an edge breakage image in an edge breakage database, and determining, capturing and recording a first real edge breakage on the to-be-detected polished wafer. Therefore, full-automatic edge breakage detection is performed on the silicon carbide polished wafer based on the edge breakage detection model, so that the problem of harm to human eyes during manual visual inspection is solved, and the detection efficiency and the detection precision are improved.

Description

Technical Field

[0001] The present application relates to the field of semiconductor technology, and in particular to a method for detecting edge chipping of a silicon carbide polishing sheet and related products. Background Art

[0002] Silicon carbide is a semiconductor material with high breakdown electric field strength, high thermal conductivity, high electron saturation rate and strong radiation resistance. It is widely used in the manufacture of high-temperature, high-frequency and high-power electronic devices. Silicon carbide polishing sheets are produced through multiple processes such as cutting, grinding and polishing of silicon carbide crystal rods. The process itself has a long production cycle and high cost, and is often accompanied by edge collapse.

[0003] The existing inspection methods for silicon carbide polishing sheets are mostly manual visual inspection, which is performed by manually holding the sheet under a strong light. This often results in missed detection and false detection of defects due to the different angles of holding the sheet during visual inspection. In addition, the human eye is prone to visual fatigue when working under high-intensity light for a long time, which leads to low inspection efficiency and accuracy.

[0004] Therefore, how to improve the detection efficiency and detection accuracy of chipping of silicon carbide polishing sheets is a problem that technicians in this field need to solve urgently. Summary of the invention

[0005] Based on the above problems, the present application provides a method and related products for detecting edge chipping of silicon carbide polishing sheets. Based on the edge chipping detection model, the silicon carbide polishing sheets are fully automatically detected for edge chipping, which solves the harm to the human eye during manual visual inspection and improves the detection efficiency and accuracy.

[0006] In a first aspect, an embodiment of the present application provides a method for detecting edge chipping of a silicon carbide polishing sheet, comprising:

[0007] Place the polishing sheet to be tested on a sheet stage with the first plane facing upward;

[0008] Capturing an image of the first plane based on a light source to obtain a first image to be measured;

[0009] Based on the edge chipping detection model, feature comparison is performed between the first image to be tested and the edge chipping images in the edge chipping database, and the first real edge chipping on the polishing sheet to be tested is determined and captured for recording.

[0010] Optionally, the performing image acquisition on the first plane based on the light source to obtain a first image to be measured includes:

[0011] Controlling the bright field light source and the transmission light source for irradiating the inspection polishing sheet to perform stroboscopic lighting;

[0012] According to the preset algorithm sampling step, a high-precision camera is used to collect images of different areas of the first plane during each exposure, and the collected image set is used as the first image to be tested.

[0013] Optionally, the edge chipping detection model is obtained by training with edge chipping images through artificial intelligence;

[0014] The edge chipping database records edge chipping images of different real edge chips and edge chipping features corresponding to each real edge chipping.

[0015] Optionally, based on the chipping detection model, feature comparison is performed between the first image to be tested and the chipping images in the chipping database, and the first real chipping on the polishing sheet to be tested is determined and captured for recording, including:

[0016] Selecting an image that is located at the periphery and includes the edge of the polishing sheet to be detected from the first images to be detected as a target image to be detected;

[0017] Extracting feature information of the target detection image based on a chipping detection model;

[0018] Compare the feature information with the edge chipping features of the edge chipping images pre-stored in the edge chipping database, and obtain a comparison result;

[0019] Based on the comparison result, the first real chipped edge on the polishing sheet to be detected is determined and captured, and recorded.

[0020] Optionally, the method further includes:

[0021] Determine the position of the first real chipped edge by bright field and transmission channel imaging;

[0022] The position includes a first plane and a second plane; the second plane is the other side of the first plane.

[0023] Optionally, the method further includes:

[0024] The size of the first real chipping edge is calculated based on a pixel number method and the size is recorded.

[0025] Optionally, the method further includes:

[0026] Determine the number of chipped edges corresponding to the polishing sheet to be tested;

[0027] Based on the number of chipped edges and the size of the first real chipped edges, the polishing sheet to be detected is graded in combination with preset size and number thresholds.

[0028] In a second aspect, an embodiment of the present application provides a device for detecting chipping of a silicon carbide polishing sheet, comprising:

[0029] A transport module, used for placing the polishing sheet to be inspected on a sheet carrier with the first plane facing upwards;

[0030] An acquisition module, used for acquiring an image of the first plane based on a light source to obtain a first image to be measured;

[0031] The detection module is used to compare the features of the first image to be detected with the edge chipping images in the edge chipping database based on the edge chipping detection model, determine and capture the first real edge chipping on the polishing sheet to be detected for recording.

[0032] In a third aspect, an embodiment of the present application provides a device for detecting chipping of a silicon carbide polishing sheet, comprising:

[0033] Memory for storing computer programs;

[0034] The processor is used to implement the steps of the method for detecting edge chipping of the silicon carbide polishing sheet as described above when executing the computer program.

[0035] In a fourth aspect, an embodiment of the present application provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for detecting edge chipping of a silicon carbide polishing sheet as described above are implemented.

[0036] It can be seen from the above technical solutions that compared with the prior art, the present application has the following advantages:

[0037] The present application first places the first plane of the polishing sheet to be detected on the wafer stage with the first plane facing upward. Then, based on the light source, an image of the first plane is captured to obtain a first image to be tested. Finally, based on the edge collapse detection model, the first image to be tested is feature-matched with the edge collapse images in the edge collapse database, and the first real edge collapse on the polishing sheet to be detected is determined and captured for recording. In this way, the edge collapse detection model is used to perform fully automatic edge collapse detection on the silicon carbide polishing sheet, which solves the harm to the human eye during manual visual inspection and improves the detection efficiency and detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of the structure of a chipping detection machine provided in an embodiment of the present application;

[0039] Figure 2 A top view of a chipping detection machine provided in an embodiment of the present application;

[0040] Figure 3 A flow chart of a method for detecting edge chipping of a silicon carbide polishing sheet provided in an embodiment of the present application;

[0041] Figure 4 A schematic diagram of a chipped edge morphology provided in an embodiment of the present application;

[0042] Figure 5 A schematic diagram of a chipping map provided in an embodiment of the present application;

[0043] Figure 6 A schematic structural diagram of a device for detecting edge chipping of a silicon carbide polishing sheet provided in an embodiment of the present application. DETAILED DESCRIPTION

[0044] As mentioned above, the existing detection methods for silicon carbide polishing sheets have the problem of low detection efficiency and detection accuracy. Specifically, manual visual inspection will cause defects to be missed or misdetected due to the different angles at which the person holds the sheet, and the different brightness and morphology of the broken edge under different lighting angles. In addition, the human eye is prone to visual fatigue when working under high-intensity light for a long time, which will also reduce work efficiency and observation accuracy.

[0045] In order to solve the above problems, the embodiment of the present application provides a method for detecting chipping of a silicon carbide polishing sheet, comprising: firstly, placing the first plane of the polishing sheet to be detected upward on a wafer stage. Then, based on a light source, an image of the first plane is captured to obtain a first image to be detected. Finally, based on a chipping detection model, a feature comparison is performed between the first image to be detected and the chipping images in the chipping database, and the first real chipping on the polishing sheet to be detected is determined and captured for recording.

[0046] In this way, the chipping edge detection of silicon carbide polishing sheets is fully automatic based on the chipping edge detection model, which solves the harm to human eyes during manual visual inspection and improves the detection efficiency and accuracy.

[0047] It should be noted that the method for detecting chipping of silicon carbide polishing sheets and related products provided by the present application can be applied to the field of semiconductor technology. The above is only an example and does not limit the application field of the method for detecting chipping of silicon carbide polishing sheets and related products provided by the present application.

[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] Figure 1 A schematic diagram of the structure of a chipping detection machine provided in an embodiment of the present application. Figure 1As shown, it includes: 1-Z axis; 2-X axis; 3-Y axis; 4-camera; 5-stage; 6-motor; 7-bright field light source and 8-transmitted light source. Before testing, the Z axis is adjusted to the appropriate position and no change is required afterwards. The stage is set on the X axis and the Y axis. The motor controls the movement of the X axis and the Y axis, and drives the stage to move synchronously. The camera and the bright field light source are fixed on the Z axis, and the transmitted light source is set under the stage. The stage is used to load silicon carbide polishing sheets. Figure 2 A top view of a chipping detection machine provided in an embodiment of the present application. Figure 2 As shown, the chipping detection machine also includes: 9-safe loading and unloading area, when the wafer carrier is in this position, the silicon carbide polishing wafer can be loaded or unloaded.

[0050] Figure 3 A flow chart of a method for detecting chipped edges of a silicon carbide polishing sheet provided in an embodiment of the present application. Figure 3 As shown, a method for detecting chipped edges of a silicon carbide polishing sheet provided in an embodiment of the present application may include:

[0051] S301: Place the polishing wafer to be inspected on a wafer stage with the first plane facing upward.

[0052] In actual application, the technician opens the chipping detection software installed on the machine, and the machine automatically resets. The motor controls the X-axis and Y-axis to synchronously drive the wafer stage to the safe loading and unloading area to wait for the wafer to be loaded and perform chipping detection. The polished wafer to be tested can be manually placed on the wafer stage by the technician using a suction pen with the Si (first plane) of the silicon carbide single crystal polished wafer facing up, or the program controls the robotic arm to take the polished wafer to be tested from the wafer box where the polished wafer is placed and place it on the wafer stage with the Si facing up. The technician then enters the lettering number of the polished wafer to be tested on the test computer to perform chipping detection.

[0053] S302: Capturing an image of the first plane based on a light source to obtain a first image to be measured.

[0054] In actual application, the technician clicks start on the test computer or calls the program control to start after the machine detects that a polished sheet to be tested is placed on the wafer stage, the light source is exposed, and the camera collects images of the Si surface of the silicon carbide single crystal polished sheet, and saves the collected images in the same folder as the first image to be tested. It can be understood that a complete light film image is composed of multiple first images to be tested.

[0055] In addition, since the methods of acquiring the first image to be measured are not the same, the embodiment of the present application can illustrate a possible acquisition method.

[0056] In one case, S302: performing image acquisition on the first plane based on a light source to obtain a first image to be measured may specifically include:

[0057] Controlling the bright field light source and the transmission light source for irradiating the inspection polishing sheet to perform stroboscopic lighting;

[0058] According to the preset algorithm sampling step, a high-precision camera is used to collect images of different areas of the first plane during each exposure, and the collected image set is used as the first image to be tested.

[0059] In actual application, after the test starts, the machine calls the program to control the X-axis and Y-axis to drive the wafer stage loaded with the polished sheet to be tested to move directly below the camera. Then the sampling step size is moved according to the preset algorithm, and the polished sheet to be tested is driven to move synchronously. Among them, the bright field light source illuminates the first plane of the polished sheet to be tested, and the transmitted light source illuminates the second plane of the polished sheet to be tested, and the second plane is the other side of the first plane. The bright field light source and the transmitted light source are strobed. During each exposure, the camera takes a shot, scans the polished sheet to be tested in an S-shaped manner from top to bottom and from left to right, and collects a set of images containing multiple first images to be tested for storage. Taking a four-inch polished sheet to be tested as an example, assuming that the image collected by the camera each time is a square with a side length of 3.5 cm, then a square area with a side length of 14 cm can be set, which is divided into 4×4 small blocks, and the center of the polished sheet to be tested is aligned with the center of the square area. The camera corresponds to the small block and collects multiple first images to be tested in 16 areas to obtain a complete picture of the polished sheet to be tested. In order to ensure the quality of detection, the embodiment of the present application can use a high-precision camera to detect small chipped edges that are undetectable to the human eye without adding additional costs.

[0060] S303: Based on the chipping detection model, feature comparison is performed between the first image to be tested and the chipping images in the chipping database, and the first real chipping on the polishing sheet to be tested is determined and captured for recording.

[0061] In practical applications, there may or may not be real edge collapse in the collected multiple first images to be tested, and it is necessary to use the edge collapse detection model provided in the embodiment of the present application for detection. Specifically, after the first image to be tested is input into the edge collapse detection model, the model extracts the image features, and then compares the extracted features with the features of the recorded real edge collapse image to determine whether there is real edge collapse in the first image to be tested. If so, the real edge collapse is mapped to the polishing sheet to be tested based on the first detection image for capture, and recorded in the test result folder corresponding to the engraving number. The recorded content includes the edge collapse morphology map, edge collapse size and distribution map.

[0062] Figure 4A schematic diagram of a chipping morphology provided in an embodiment of the present application, combined with Figure 4 As shown in the picture, the technicians can understand the morphology of the chipped edge and that the arc length of the chipped edge is 0.53, the radial length of the chipped edge is 0.14, etc.

[0063] Figure 5 A schematic diagram of a broken edge map provided in an embodiment of the present application, combined with Figure 5 As shown, the real edge chipping of the first detection image with real edge chipping is mapped onto the polished sheet to be detected, and then the position of the real edge chipping on the polished sheet to be detected is determined, thereby forming an edge chipping distribution map. Among them, points of different colors can indicate whether the real edge chipping is specifically on the first plane or the second plane.

[0064] In addition, the edge collapse detection model is trained by artificial intelligence using edge collapse images;

[0065] The edge chipping database records edge chipping images of different real edge chips and edge chipping features corresponding to each real edge chipping.

[0066] Specifically, the embodiment of the present application adopts artificial intelligence to train the edge chipping detection model with a large number of edge chipping images with real edge chipping as samples, so that the detection accuracy of the edge chipping detection model is very high. In addition, a edge chipping database is constructed to store these various edge chipping images with real edge chipping and their corresponding edge chipping features for feature comparison by the edge chipping detection model.

[0067] In addition, since the methods of feature comparison are not the same, the embodiments of the present application can illustrate a possible comparison method.

[0068] In one case, S303: based on the chipping detection model, feature comparison is performed between the first image to be tested and the chipping images in the chipping database, and the first real chipping on the polishing sheet to be tested is determined and captured for recording, which may specifically include:

[0069] Selecting an image that is located at the periphery and includes the edge of the polishing sheet to be detected from the first images to be detected as a target image to be detected;

[0070] Extracting feature information of the target detection image based on a chipping detection model;

[0071] Compare the feature information with the edge chipping features of the edge chipping images pre-stored in the edge chipping database, and obtain a comparison result;

[0072] Based on the comparison result, the first real chipped edge on the polishing sheet to be detected is determined and captured, and recorded.

[0073] In practical applications, edge collapse refers to defects on the edge of the silicon carbide polishing sheet. In order to reduce the amount of calculation, the embodiment of the present application can analyze only the first image to be tested that contains the edge of the polishing sheet to be detected in the periphery through program control. Continuing with the example of the above-mentioned four-inch polishing sheet, if a square area with a side length of 14 cm is still set to divide it into 4×4 small blocks, then the first image to be tested corresponding to the four small blocks closest to the center of the multiple first images to be tested collected can be deleted (because there is no edge of the polishing sheet in these four small blocks), and then the remaining images in the periphery and containing the edge of the polishing sheet to be detected are used as target images to be tested. Then, these target detection images are input into the trained edge collapse detection model to extract feature information, and the extracted feature information is compared with the edge collapse features of the edge collapse images pre-stored in the edge collapse database to obtain the comparison results. If the comparison result determines that there is edge collapse on the target detection image, all the edge collapses are recorded as the first real edge collapse, and their positions are mapped to the polishing sheet to be detected, and their number, position and size are recorded.

[0074] In addition, since the methods for determining the edge chipping position are different, the embodiments of the present application can illustrate a possible determination method.

[0075] In one embodiment, the method further comprises:

[0076] Determine the position of the first real chipped edge by bright field and transmission channel imaging;

[0077] The position includes a first plane and a second plane; the second plane is the other side of the first plane.

[0078] In practical applications, there are two types of edge chipping: edge chipping on the upper surface (first plane) and edge chipping on the lower surface (second surface). In the embodiment of the present application, a bright field light source and a transmitted light source are respectively arranged to collect upper and lower layer images by bright field reflection and transmission. Then, the two channel images are cross-validated through an artificial intelligence algorithm to accurately distinguish the edge chipping from dirty morphologies similar to edge chipping, determine whether the first real edge chipping is on the first plane or the second plane, and record it.

[0079] In addition, since different imaging methods are used depending on the surface where the chipping occurs, the embodiment of the present application can illustrate a possible imaging method.

[0080] In one embodiment, the method further comprises:

[0081] When the chipped edge is on the upper surface (first plane), the bright field reflection channel is used for imaging;

[0082] When the chipped edge is on the lower surface (the second plane), the output transmission channel is imaged;

[0083] When chipping exists on both the upper and lower surfaces, the algorithm is used to fit the imaging of the reflection channel and the transmission channel to output a complete chipping image.

[0084] Among them, algorithm fitting refers to the process of finding a mathematical function that can well describe a set of observation data. In this application, the edge collapse situation in the upper surface and the edge collapse situation in the lower surface are integrated through the fitting algorithm, and then a complete edge collapse image is output. In addition, since the methods for determining the edge collapse size are different, the embodiments of this application can illustrate a possible determination method.

[0085] In one embodiment, the method further comprises:

[0086] The size of the first real chipping edge is calculated based on a pixel number method and the size is recorded.

[0087] In practical applications, the number of pixels is an important parameter of a camera, which refers to the number of tiny units on the camera's photosensitive element that are used to capture light and form an image. Generally speaking, the higher the number of pixels, the more details the camera can capture, and the higher the resolution of the photo. Therefore, the size of the first real chipping edge can be determined by combining the resolution of the image and the number of pixels occupied by the first real chipping edge, and then the size is recorded in correspondence with the first real chipping edge.

[0088] In addition, since the level division methods are not the same, the embodiment of the present application can illustrate a possible division method.

[0089] In one embodiment, the method further comprises:

[0090] Determine the number of chipped edges corresponding to the polishing sheet to be tested;

[0091] Based on the number of chipped edges and the size of the first real chipped edges, the polishing sheet to be detected is graded in combination with preset size and number thresholds.

[0092] In practical applications, there may be multiple first real chipping edges on a polishing sheet to be detected, and the sizes of each chipping edge are also different. It is understandable that the number of chipping edges and the size of chipping edges on each polishing sheet to be detected are different. To this end, the embodiment of the present application introduces the concept of preset size and quantity thresholds to grade different polishing sheets to be detected. For example, the preset quantity thresholds 0 and 3 are set, and when the number of the first real chipping edges on the polishing sheet to be detected is 0, the polishing sheet grade is A-level, when the number is 1, 2 or 3, it is B-level, and when the number is 4 or more, it is NG-level. For another example, the preset size threshold is set to a width of 1mm, then when the first real chipping edge size on the polishing sheet to be detected is detected to be less than 1mm, the polishing sheet grade is B-level, and when the size is 1mm or more, the polishing sheet grade is NG-level.

[0093] In summary, the present application first places the first plane of the polishing sheet to be detected on the wafer stage with the first plane facing upward. Then, based on the light source, an image of the first plane is captured to obtain a first image to be tested. Finally, based on the edge collapse detection model, a feature comparison is performed between the first image to be tested and the edge collapse image in the edge collapse database, and the first real edge collapse on the polishing sheet to be detected is determined and captured for recording. In this way, the edge collapse detection model is used to perform fully automatic edge collapse detection on the silicon carbide polishing sheet, which solves the harm to the human eye during manual visual inspection and improves the detection efficiency and detection accuracy.

[0094] Figure 6 A schematic diagram of the structure of a device for detecting chipping of a silicon carbide polishing sheet provided in an embodiment of the present application. Figure 6 As shown, the device 600 for detecting chipped edges of a silicon carbide polishing sheet provided in an embodiment of the present application comprises:

[0095] A transport module 601 is used to place the polishing wafer to be inspected on a wafer stage with the first plane facing upwards;

[0096] An acquisition module 602 is used to acquire an image of the first plane based on a light source to obtain a first image to be measured;

[0097] The detection module 603 is used to perform feature comparison between the first image to be detected and the edge chipping images in the edge chipping database based on the edge chipping detection model, and determine and capture the first real edge chipping on the polishing sheet to be detected for recording.

[0098] As an implementation manner, regarding how to acquire the first image to be tested, the acquisition module 602 is specifically used to:

[0099] Controlling the bright field light source and the transmission light source for irradiating the inspection polishing sheet to perform stroboscopic lighting;

[0100] According to the preset algorithm sampling step, a high-precision camera is used to collect images of different areas of the first plane during each exposure, and the collected image set is used as the first image to be tested.

[0101] As an implementation manner, regarding how to determine the real edge collapse through feature comparison, the detection module 603 is specifically used to:

[0102] Selecting an image that is located at the periphery and includes the edge of the polishing sheet to be detected from the first images to be detected as a target image to be detected;

[0103] Extracting feature information of the target detection image based on a chipping detection model;

[0104] Compare the feature information with the edge chipping features of the edge chipping images pre-stored in the edge chipping database, and obtain a comparison result;

[0105] Based on the comparison result, the first real chipped edge on the polishing sheet to be detected is determined and captured, and recorded.

[0106] Among them, the edge collapse detection model is obtained through artificial intelligence and training using edge collapse images;

[0107] The edge chipping database records edge chipping images of different real edge chips and edge chipping features corresponding to each real edge chipping.

[0108] As an implementation method, with respect to how to determine the edge chipping position, the above-mentioned device 600 for detecting edge chipping of the silicon carbide polishing sheet further includes: a judgment module;

[0109] A judgment module, used for determining the position of the first real chipped edge through bright field and transmission channel imaging;

[0110] The position includes a first plane and a second plane; the second plane is the other side of the first plane.

[0111] As an implementation manner, with respect to how to determine the chipping size, the chipping detection device 600 of the silicon carbide polishing sheet further includes: a determination module;

[0112] The determination module is used to calculate the size of the first real chipped edge based on a pixel count method and record the size.

[0113] As an implementation method, with respect to how to classify polishing sheets, the above-mentioned silicon carbide polishing sheet edge chipping detection device 600 further includes: a classifying module;

[0114] A grading module, used to determine the number of chipped edges corresponding to the polishing sheet to be tested;

[0115] Based on the number of chipped edges and the size of the first real chipped edges, the polishing sheet to be detected is graded in combination with preset size and number thresholds.

[0116] In summary, the present application first places the first plane of the polishing sheet to be detected on the wafer stage with the first plane facing upward. Then, based on the light source, an image of the first plane is captured to obtain a first image to be tested. Finally, based on the edge collapse detection model, a feature comparison is performed between the first image to be tested and the edge collapse image in the edge collapse database, and the first real edge collapse on the polishing sheet to be detected is determined and captured for recording. In this way, the edge collapse detection model is used to perform fully automatic edge collapse detection on the silicon carbide polishing sheet, which solves the harm to the human eye during manual visual inspection and improves the detection efficiency and detection accuracy.

[0117] In addition, the present application also provides a device for detecting chipping of a silicon carbide polishing sheet, comprising:

[0118] Memory for storing computer programs;

[0119] A processor is used to implement the steps of the method for detecting edge chipping of a silicon carbide polishing sheet as described in any one of the above items when executing the computer program.

[0120] In addition, the present application also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for detecting edge chipping of a silicon carbide polishing sheet as described in any one of the above items are implemented.

[0121] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting edge collapse of a silicon carbide polishing sheet, characterized in that: The method comprises: Place the polishing sheet to be tested on a sheet stage with the first plane facing upward; Capturing an image of the first plane based on a light source to obtain a first image to be measured; Based on the edge chipping detection model, feature comparison is performed between the first image to be tested and the edge chipping images in the edge chipping database, and the first real edge chipping on the polishing sheet to be tested is determined and captured for recording.

2. The method according to claim 1, characterized in that The step of performing image acquisition on the first plane based on the light source to obtain a first image to be measured includes: Controlling the bright field light source and the transmission light source for irradiating the inspection polishing sheet to perform stroboscopic lighting; According to the preset algorithm sampling step, a high-precision camera is used to collect images of different areas of the first plane during each exposure, and the collected image set is used as the first image to be tested.

3. The method according to claim 1, characterized in that The edge collapse detection model is obtained by training with edge collapse images through artificial intelligence; The edge chipping database records edge chipping images of different real edge chips and edge chipping features corresponding to each real edge chipping.

4. The method according to claim 2, characterized in that: Based on the edge chipping detection model, feature comparison is performed between the first image to be tested and the edge chipping images in the edge chipping database, and the first real edge chipping on the polishing sheet to be tested is determined and captured for recording, including: Selecting an image that is located at the periphery and includes the edge of the polishing sheet to be detected from the first images to be detected as a target image to be detected; Extracting feature information of the target detection image based on a chipping detection model; Compare the feature information with the edge chipping features of the edge chipping images pre-stored in the edge chipping database, and obtain a comparison result; Based on the comparison result, the first real chipped edge on the polishing sheet to be detected is determined and captured, and recorded.

5. The method according to claim 4, characterized in that The method further comprises: Determine the position of the first real chipped edge by bright field and transmission channel imaging; The position includes a first plane and a second plane; the second plane is the other side of the first plane.

6. The method according to claim 1, characterized in that The method further comprises: The size of the first real chipping edge is calculated based on a pixel number method and the size is recorded.

7. The method according to claim 6, characterized in that The method further comprises: Determine the number of chipped edges corresponding to the polishing sheet to be tested; Based on the number of chipped edges and the size of the first real chipped edges, the polishing sheet to be detected is graded in combination with preset size and number thresholds.

8. A device for detecting chipped edges of silicon carbide polishing sheets, characterized in that: include: A transport module, used for placing the polishing sheet to be inspected on a sheet carrier with the first plane facing upwards; An acquisition module, used for acquiring an image of the first plane based on a light source to obtain a first image to be measured; The detection module is used to compare the features of the first image to be tested with the edge chipping images in the edge chipping database based on the edge chipping detection model, determine and capture the first real edge chipping on the polishing sheet to be tested for recording.

9. A device for detecting chipping of silicon carbide polishing sheets, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the method for detecting edge chipping of a silicon carbide polishing sheet as claimed in any one of claims 1 to 7 when executing the computer program.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for detecting edge chipping of a silicon carbide polishing sheet according to any one of claims 1 to 7 are implemented.

Images