Method and equipment for detecting cracks of silicon carbide polished wafer

Abstract

The invention discloses a silicon carbide polished section crack detection method and device, and relates to the technical field of detection, and the method comprises the steps: firstly obtaining a dark field image of a silicon carbide polished section, extracting an image region satisfying a gray value abnormal condition from the dark field image to determine a polished section region corresponding to the image region, i.e., a defect region, and then obtaining a polarization image of the region, and according to the associated residual stress around the crack defect, calling the pre-trained defect identification model to carry out image processing to obtain a defect type identification result of the polished wafer, so that the purpose of distinguishing other defects such as scratches and smudginess from the crack defect is achieved, and finally, under the condition of determining that the crack defect exists in the silicon carbide polished wafer according to the identification result, the defect type of the polished wafer is identified. The crack size is calculated according to the polished section image, the crack defect detection result is generated, the crack defect detection task is achieved, and the crack detection precision and accuracy are improved.

Description

Technical Field

[0001] The present application relates to the field of detection technology, and in particular to a method and device for detecting cracks in a silicon carbide polishing sheet. Background Art

[0002] With the development of power electronics technology and the rapid growth of the power electronics market, the third-generation wide bandgap semiconductor silicon carbide (SiC) has gradually become a strong competitor for the next generation of power devices for electric vehicles and power converters. Compared with traditional devices made of silicon (Si), silicon carbide power electronic devices have advantages such as lower on-resistance, higher electron mobility, wider band gap and higher breakdown voltage, which helps to reduce power consumption, increase the switching speed of devices, improve the high temperature resistance of devices, and provide a larger operating range and power range.

[0003] However, during the growth or annealing process of silicon carbide crystals, if thermal stress introduced into the crystal or internal defects of the crystal exceed the tolerance threshold of the crystal, it will cause the crystal to crack, resulting in through-type or partial penetration cracks, reducing the electron mobility and mechanical lightness of the crystal, causing serious impact on the overall quality and performance of the crystal, and thus affecting the quality of the silicon carbide polishing sheet obtained.

[0004] Therefore, before manufacturing power devices based on silicon carbide polishing wafers, crack detection of the silicon carbide polishing wafers is required. Summary of the invention

[0005] In view of the above problems, the present application provides a method and device for detecting cracks in a silicon carbide polishing sheet to achieve the crack detection task of the silicon carbide polishing sheet and improve the accuracy and precision of the detection.

[0006] The specific plan is as follows:

[0007] The first aspect of the present application provides a method for detecting cracks in a silicon carbide polishing sheet, comprising:

[0008] Acquire a dark field image of the silicon carbide polishing sheet, wherein the dark field image of the silicon carbide polishing sheet is obtained based on dark field reflection imaging;

[0009] Extracting an image region satisfying a grayscale value abnormality condition from a dark field image of the silicon carbide polishing sheet, and determining a polishing sheet region corresponding to the extracted image region from the silicon carbide polishing sheet;

[0010] Acquire a polarization image of the polishing sheet area, wherein the polarization image of the polishing sheet area is obtained based on polarization imaging;

[0011] Calling a pre-trained defect recognition model to process the polarization image of the polishing sheet area to obtain a defect type recognition result of the silicon carbide polishing sheet; wherein the defect recognition model is trained using a polarization sample image of the polishing sheet annotated with defect type information;

[0012] When it is determined that the silicon carbide polishing sheet has a crack defect based on the defect type identification result, the crack size of the silicon carbide polishing sheet is calculated based on the image of the silicon carbide polishing sheet, and a crack defect detection result of the silicon carbide polishing sheet is generated; the image of the silicon carbide polishing sheet includes a dark field image of the silicon carbide polishing sheet or a polarization image of the silicon carbide polishing sheet.

[0013] A second aspect of the present application provides a silicon carbide polishing sheet crack detection device, the device comprising: a dark field light source, a first CCD camera, a polarized light source, a second CCD camera, a wafer stage, a vacuum adsorption component and a moving component;

[0014] The dark field light source is used to provide dark field light for the image acquisition range of the first CCD camera;

[0015] The first CCD camera is used to collect dark field images;

[0016] The polarized light source is used to provide polarized light for the image acquisition range of the second CCD camera;

[0017] The second CCD camera is used to collect polarization images;

[0018] The wafer carrier is used for carrying the silicon carbide polishing wafer;

[0019] The vacuum adsorption component is arranged on the wafer carrier and connected to a vacuum pump, and the vacuum adsorption component is used to adsorb the silicon carbide polishing sheet placed on the wafer carrier;

[0020] The moving component is used to drive the wafer stage to move so that each area of ​​the silicon carbide polishing sheet placed on the wafer stage is located within the image acquisition range of the first CCD camera or within the image acquisition range of the second CCD camera;

[0021] The device also includes: a silicon carbide polishing sheet crack detection controller;

[0022] The silicon carbide polishing sheet crack detection controller is connected to the first CCD camera and the second CCD camera through a signal line, and obtains a dark field image and a polarization image from the first CCD camera and the second CCD camera respectively; the silicon carbide polishing sheet crack detection controller includes at least one processor and a memory connected to the processor, wherein the memory is used to store a computer program; the processor is used to execute the computer program so that the silicon carbide polishing sheet crack detection controller can implement the above-mentioned silicon carbide polishing sheet crack detection method.

[0023] By means of the above technical scheme, the present application first obtains a dark field image of a silicon carbide polishing sheet, extracts the image area that satisfies the grayscale value abnormality condition, determines the polishing sheet area corresponding to the image area, obtains the polarization image of the domain, and calls the pre-trained defect recognition model for image processing to obtain the defect type recognition result of the polishing sheet, and finally, when it is determined that there is a crack defect in the silicon carbide polishing sheet based on the recognition result, the crack size is calculated based on the polishing sheet image to generate a crack defect detection result.

[0024] The present application determines the abnormal grayscale value area in the dark field image through dark field reflection imaging, and then determines the defective area on the polishing sheet. On this basis, according to the residual stress associated with the crack defect, the grayscale value change caused by the stress lines around the defect is determined through polarization imaging observation, thereby achieving the purpose of distinguishing scratches, dirt and other defects from crack defects, and realizing the accurate capture and classification of crack defects.

[0025] Furthermore, the present application processes the dark field image and polarization image of the silicon carbide polished sheet through a terminal with data processing capabilities, eliminating the need for inspectors to conduct long-term observations under strong light or manually mark dimensions through a microscope, thereby reducing the labor and time costs of defect detection and improving the accuracy and efficiency of crack detection to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0027] Figure 1 A schematic diagram of a process for detecting cracks in a silicon carbide polishing sheet provided in the present application;

[0028] Figure 2 A schematic diagram of the structure of a silicon carbide polishing sheet crack detection device provided in this application;

[0029] Figure 3 A schematic diagram of the structure of a silicon carbide polishing sheet crack detection controller provided in this application;

[0030] Figure 4 An example of a structural schematic diagram of a crack detection device for a silicon carbide polishing sheet is provided;

[0031] Figure 5 A schematic top view of a crack detection device for a silicon carbide polishing sheet is illustrated;

[0032] Figure 6 The structural schematic diagram of the slide stage is illustrated;

[0033] Figure 7 The schematic diagram of the defect map is shown as an example;

[0034] Figure 8 A schematic diagram of a dark field image showing a crack defect;

[0035] Fig. 9 Schematic diagram of a polarization image showing a crack defect.

[0036] Reference numerals:

[0037] 41-dark field light source; 42-first CCD camera; 43-polarized light source; 44-second CCD camera; 45-film stage; 46-vacuum adsorption component; 460-vacuum pump; 461-first group of sub-components; 462-second group of sub-components; 47-moving component; 471-screw; 472-motor. DETAILED DESCRIPTION

[0038] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation mode of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application. It is known to those skilled in the art that with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0039] The applicant in this case has found through research that currently, manual inspection is usually adopted to achieve the crack detection task of silicon carbide polishing sheets. Specifically, the silicon carbide polishing sheets can be comprehensively inspected by manually operating a microscope, or the silicon carbide polishing sheets can be first sucked up by a suction pen and inspected for cracks under a strong light. After the cracks are detected, the crack size can be determined by manually operating a microscope.

[0040] The above-mentioned manual inspection method requires manual use of tools such as suction pens and tweezers to clamp and move the sheet, which is easy to leave clamping marks on the silicon carbide polishing sheet, and the suction pen may accidentally suck the Si surface of the polishing sheet, increasing the risk of polishing sheet loss caused by manual operation. In addition, since the brightness and morphology of cracks are different under different lighting angles, when inspectors visually inspect the silicon carbide polishing sheet under strong light, defects may be missed or misdetected due to different angles of holding the sheet when inspectors visually inspect; since the morphology of some cracks and scratch defects is similar, when inspectors perform crack inspection through a microscope, they cannot distinguish between crack defects and other defects such as scratches on the polishing sheet. In addition, the manual inspection method manually determines whether there are cracks, and it is difficult to retain the evidence of the inspection results, resulting in the inability to trace back when quality problems occur later, which is not conducive to problem analysis; manually intercepting images under a microscope and manually marking the size with a ruler is prone to human error. Therefore, the current crack detection method is time-consuming and labor-intensive, and it is difficult to achieve accurate, fast, and high-precision crack detection tasks.

[0041] To solve the above problems, the present application provides a method and device for detecting cracks in a silicon carbide polishing sheet, so as to realize the crack detection task of the silicon carbide polishing sheet and improve the accuracy and precision of the detection.

[0042] Figure 1 1 is a flow chart of a method for detecting cracks in a silicon carbide polishing sheet according to an embodiment of the present application. Figure 1 As shown, the method may include the following steps:

[0043] Step S101, obtaining a dark field image of a silicon carbide polishing wafer.

[0044] Among them, the dark field image of the silicon carbide polishing sheet is obtained based on dark field reflection imaging. It should be noted that under a dark field light source, if there is no defect, the light will be reflected symmetrically, the image acquisition device will not receive the reflected light, the field of view of the image acquisition device tends to be completely black, and the grayscale value of the image collected is high; if there are cracks or other defects, the light reflection will change, the image acquisition device will receive part of the reflected light, and in the field of view of the image acquisition device, the defect is brighter, and other positions are darker, and there is a difference in grayscale value. Based on this, through light reflection imaging under a dark field light source, a dark field image basis can be provided for detecting cracks or other defects on silicon carbide polishing sheets.

[0045] Step S102: extracting an image region satisfying a grayscale value abnormality condition from the dark field image of the silicon carbide polishing wafer, and determining a polishing wafer region corresponding to the extracted image region from the silicon carbide polishing wafer.

[0046] For example, the grayscale value abnormal condition may be that the grayscale value is different from the reference grayscale value by more than a preset limit, and the reference grayscale value refers to the grayscale value of a normal area without cracks and other defects. If the grayscale value changes by more than 10 relative to the background grayscale, it indicates that the grayscale value is abnormal.

[0047] In one possible implementation, this step can be implemented by a visual model, that is, the above-mentioned extraction of the image area that meets the grayscale value abnormality condition from the dark field image of the silicon carbide polishing sheet can include: calling a pre-trained dark field image processing model to process the dark field image of the silicon carbide polishing sheet, and determining the grayscale value abnormal area in the image, that is, the above-mentioned image area that meets the grayscale value abnormality condition; wherein the dark field image processing model can be trained using a dark field sample image of a silicon carbide polishing sheet with an annotated grayscale value abnormal area.

[0048] Step S103, obtaining a polarization image of the polishing sheet area.

[0049] The polarization image of the polished sheet area is obtained based on polarization imaging. It should be noted that in the polarization scene, there will be stress lines with obvious light and dark changes around the crack defect. Based on this, if stress lines exist, it can be determined that the defect is a crack. If stress lines do not exist, it can be determined that the defect is other defects such as scratches or dirt that are different from cracks.

[0050] Step S104: calling a pre-trained defect recognition model to process the polarization image of the polishing sheet area to obtain a defect type recognition result of the silicon carbide polishing sheet.

[0051] The defect recognition model is trained using polarized sample images of polished sheets annotated with defect type information. The defect type information may at least include information characterizing crack defects.

[0052] Step S105: when it is determined that the silicon carbide polishing sheet has a crack defect according to the defect type identification result, the crack size of the silicon carbide polishing sheet is calculated according to the image of the silicon carbide polishing sheet, and a crack defect detection result of the silicon carbide polishing sheet is generated.

[0053] Wherein, the image of the silicon carbide polishing sheet may include a dark field image of the silicon carbide polishing sheet or a polarization image of the silicon carbide polishing sheet, and the polarization image of the polishing sheet area belongs to the polarization image of the silicon carbide polishing sheet. The above crack defect detection result corresponds to the calculated crack size, and exemplarily, the crack defect detection result may include the calculated crack size.

[0054] Based on the above content, the present application first obtains a dark field image of a silicon carbide polishing sheet, extracts an image area that satisfies the grayscale value abnormality condition, determines the polishing sheet area corresponding to the image area, obtains a polarization image of the domain, and calls a pre-trained defect recognition model for image processing to obtain a defect type recognition result of the polishing sheet, and finally, when it is determined that there is a crack defect in the silicon carbide polishing sheet according to the recognition result, the crack size is calculated according to the polishing sheet image to generate a crack defect detection result. This embodiment determines the grayscale value abnormal area in the dark field image through dark field reflection imaging, and then determines the defect area on the polishing sheet. On this basis, based on the associated residual stress around the crack defect, the grayscale value change caused by the stress lines around the defect is determined through polarization imaging observation, thereby achieving the purpose of distinguishing other defects such as scratches and dirt from crack defects, and achieving accurate capture and classification of crack defects.

[0055] Furthermore, this embodiment processes the dark field image and polarization image of the silicon carbide polishing sheet by a terminal with data processing capabilities, eliminating the need for inspectors to conduct long-term observations under strong light or manually mark dimensions through a microscope, thereby reducing the labor and time costs of defect detection and improving the accuracy and efficiency of crack detection to a certain extent.

[0056] In a possible implementation, the dark field image and polarization image described in the present application can be acquired by using a CCD camera, that is, the above-mentioned image acquisition device can be a CCD camera.

[0057] Among them, CCD camera refers to any camera with a charge coupled device (CCD) image sensor. By collecting polishing sheet images with a high-resolution CCD camera, the accuracy of crack detection can be improved to a certain extent.

[0058] In one or more embodiments provided in the present application, the dark field image of the silicon carbide polishing sheet may include dark field block images of various regions of the silicon carbide polishing sheet; correspondingly, the polarization image of the silicon carbide polishing sheet may include polarization block images of various regions of the silicon carbide polishing sheet.

[0059] Based on this, the above-mentioned process of processing the image of the silicon carbide polishing sheet may include: processing the block images of each area respectively. Exemplarily, the above-mentioned process of extracting the image area that meets the gray value abnormality condition from the dark field image of the silicon carbide polishing sheet may include: extracting the sub-areas that meet the gray value abnormality condition from the dark field block images of each area of ​​the silicon carbide polishing sheet respectively, and obtaining the image area that meets the gray value abnormality condition, which includes each sub-area that meets the gray value abnormality condition.

[0060] In one or more embodiments provided in the present application, the process of acquiring the block image of each area of ​​the silicon carbide polishing sheet may include:

[0061] Step A: after determining that the silicon carbide polishing sheet is placed on the wafer stage, adsorbing the silicon carbide polishing sheet by a vacuum adsorption component arranged on the wafer stage.

[0062] By using the vacuum adsorption component to adsorb the silicon carbide polishing sheet on the stage, the silicon carbide polishing sheet can be made stationary relative to the stage. On this basis, the stage can be moved to achieve relative movement between the silicon carbide polishing sheet and the acquisition range of the image acquisition device. In addition, clamping the polishing sheet by vacuum adsorption can eliminate the warping of the polishing sheet to avoid the image defocusing phenomenon caused by the warping of the polishing sheet; and this method can minimize the contact with the surface of the polishing sheet, which helps to ensure the cleanliness of the polishing sheet.

[0063] Optionally, the vacuum adsorption component can be a liftable vacuum adsorption component. By means of the liftable vacuum adsorption component, the transparent polishing sheet on the stage can be lifted up, and then the image can be collected. Compared with the solution of adsorbing the polishing sheet on the stage for image collection, the image collected by this solution can show the defects on the back of the polishing sheet, thereby achieving a more comprehensive crack defect detection task. On this basis, the crack position can also be determined by imaging resolution, such as determining whether the crack is located on the upper surface or the lower surface.

[0064] Step B: turning on a preset light source that provides preset light for an image acquisition range of a preset image acquisition device.

[0065] Step C, move the wafer stage to a preset initial position, move the wafer stage along a preset trajectory according to a preset step length, and obtain block images of various areas of the silicon carbide polishing sheet collected by the preset image acquisition device during the movement of the wafer stage.

[0066] Wherein, in the process that the wafer stage moves along the preset trajectory starting from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device. It should be noted that the preset step size can be set according to the range of the image acquisition field of view of the image acquisition device to ensure that the image acquisition device can acquire images of each area of ​​the polishing sheet. In the case where the acquired block image is a dark field block image, the preset image acquisition device is an image acquisition device for acquiring a dark field image, the preset light is a dark field light, and the preset light source is a dark field light source; in the case where the acquired block image is a polarized block image, the preset image acquisition device is an image acquisition device for acquiring a polarized image, the preset light is a polarized light, and the preset light source is a polarized light source.

[0067] In one possible implementation, the preset initial position may be a position that satisfies a preset position condition, wherein the preset position condition may be that the upper left corner of the silicon carbide polishing sheet is located within the image acquisition range of a preset image acquisition device; on this basis, the preset trajectory may be an S-shaped trajectory, that is, the preset image acquisition device may acquire the images of each block of the first row of the polishing sheet in a sequence from left to right, and then acquire the images of each block of the second row of the polishing sheet in a sequence from right to left, and so on. The motion trajectory of the stage is S-shaped or serpentine, so as to save image acquisition time.

[0068] In one or more embodiments provided in the present application, the vacuum adsorption component may include two groups of sub-components, each group of sub-components includes a preset number of vacuum tubes connected to a vacuum pump and capable of being raised and lowered, and each group of sub-components is used to adsorb the polishing sheet placed on the wafer carrier and to raise and lower the polishing sheet.

[0069] Optionally, the above-mentioned preset number may be at least 3 to achieve stable clamping of the polishing sheet.

[0070] On the basis of the above, the process of acquiring the block image of each area of ​​the silicon carbide polishing sheet may include:

[0071] Step D: after determining that the silicon carbide polishing sheet is placed on the wafer stage, the silicon carbide polishing sheet is adsorbed by a first group of subcomponents arranged on the wafer stage, and the silicon carbide polishing sheet is lifted to a preset height.

[0072] The first group of subcomponents mentioned above is one of the two groups of subcomponents constituting the vacuum adsorption component, and the other group of the two groups of subcomponents constituting the vacuum adsorption component is the second group of subcomponents.

[0073] Step E: Turn on a preset light source that provides preset light for an image acquisition range of a preset image acquisition device.

[0074] Step F, move the wafer stage to a preset initial position, move the wafer stage along a preset trajectory according to a preset step length, and use the preset image acquisition device to capture block images of various areas of the silicon carbide polishing sheet adsorbed by the first group of sub-components during the movement of the wafer stage.

[0075] Step G, switching to adsorbing the silicon carbide polishing sheet through a second group of subcomponents arranged on the wafer stage, and lifting the silicon carbide polishing sheet to a preset height.

[0076] Step H: acquiring a block image of a designated area of ​​the silicon carbide polishing sheet adsorbed by the second group of subcomponents through the preset image acquisition device.

[0077] The designated area includes an area corresponding to the adsorption position of the first group of subcomponents. It should be noted that the existence of the liftable vacuum tube will cause the acquired block image to contain the image area corresponding to the vacuum tube. Steps F and H can be used to complete the data, providing a basis for obtaining a complete polishing sheet image.

[0078] Optionally, the above-mentioned preset number can be 3 to reduce the interference of the vacuum adsorption component on the polishing wafer image acquisition, thereby reducing the polishing wafer area that requires repeated acquisition; in addition, the three vacuum tubes in the same group of sub-components can be arranged in a triangular shape on the wafer carrier to improve the adsorption stability of the silicon carbide polishing wafer.

[0079] Step I: obtaining block images of various regions of the silicon carbide polishing wafer.

[0080] The block images of each area of ​​the silicon carbide polishing sheet may include: the block images of the designated area of ​​the silicon carbide polishing sheet adsorbed by the second group of subcomponents, and the block images of the non-designated area of ​​the silicon carbide polishing sheet adsorbed by the first group of subcomponents. That is, the above step I may include: using the block images of the designated area acquired in step H to replace the images of the corresponding areas in the images acquired in step F, and finally obtaining the block images of each area.

[0081] Wherein, in the process of the stage moving along the preset trajectory from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; in the case where the acquired block image is a dark field block image, the preset image acquisition device is an image acquisition device for acquiring a dark field image, the preset light is a dark field light, and the preset light source is a dark field light source; in the case where the acquired block image is a polarized block image, the preset image acquisition device is an image acquisition device for acquiring a polarized image, the preset light is a polarized light, and the preset light source is a polarized light source. In addition, other descriptions of the above step DI can refer to the above description and will not be repeated here.

[0082] On the basis of the above, in the case where it is determined that the silicon carbide polishing sheet has a crack defect according to the defect type identification result, the method may further include:

[0083] The position of the crack in the silicon carbide polishing sheet is determined according to an imaging resolution corresponding to the crack in the silicon carbide polishing sheet, wherein the position includes an upper surface or a lower surface.

[0084] It should be noted that an image acquisition device, such as a CCD camera, can return two clarity values, namely, the clarity result of the upper surface focus and the clarity result of the lower surface focus. If the upper surface is higher, it indicates that the crack defect is located on the upper surface, otherwise, it indicates that the crack defect is located on the lower surface.

[0085] In one or more embodiments provided in the present application, the above-mentioned step S105, calculating the crack size of the silicon carbide polishing sheet according to the image of the silicon carbide polishing sheet, may include:

[0086] Step J: Detecting the number of pixels occupied by the cracks in the silicon carbide polishing sheet in the image of the silicon carbide polishing sheet.

[0087] Step K: calculating the crack size of the silicon carbide polishing sheet according to the image acquisition accuracy of the image acquisition device for acquiring the image of the silicon carbide polishing sheet and the determined number of pixels.

[0088] Exemplarily, taking a single crack as an example, the crack size of the crack can be expressed as: pixel accuracy of the image acquisition device×the number of pixels occupied by the crack.

[0089] In one or more embodiments provided in the present application, the above-mentioned step S105, generating the crack defect detection result of the silicon carbide polishing sheet, may include:

[0090] Step L, generating a crack map of the silicon carbide polishing sheet.

[0091] Step M: determining the grading result of the silicon carbide polishing sheet according to the preset corresponding relationship between the crack defect situation and the polishing sheet grade.

[0092] Wherein, the crack defect condition includes the number of cracks and the crack size; exemplarily, the above-mentioned crack size can be the length of a single crack or the cumulative length of the cracks. Exemplarily, the corresponding relationship between the crack defect condition and the polishing sheet grade can be expressed as: Grade A corresponding to the first condition, Grade B corresponding to the second request, and Grade NG corresponding to the third condition; wherein, the first condition is that the number of cracks is zero, the second condition is that the number of cracks is not greater than 2 and the cumulative length of cracks is not greater than 10 millimeters (mm), and the third condition is that the number of cracks is greater than 2 and / or the cumulative length of cracks is greater than 10 mm.

[0093] Step N, generating a crack defect detection result of the silicon carbide polishing sheet.

[0094] The crack defect detection result may include: the crack map, the grading result, and the crack size, dark field image and polarization image of each crack in the silicon carbide polishing sheet.

[0095] In one or more embodiments provided in the present application, before extracting the image area satisfying the gray value abnormality condition from the dark field image of the silicon carbide polishing sheet, the method may further include:

[0096] The dark field image of the silicon carbide polishing sheet is subjected to image preprocessing.

[0097] Correspondingly, before calling the pre-trained defect recognition model to process the polarization image of the polishing sheet area, the following steps may also be included:

[0098] The polarization image of the polishing sheet area is subjected to image preprocessing.

[0099] The image preprocessing at least includes denoising or enhancement. Exemplarily, the enhancement may include contrast enhancement, sharpening, and using factors that interfere with subsequent image processing as background.

[0100] On the basis of the above, in a possible implementation, the process of processing the dark field image of the silicon carbide polishing sheet or the polarization image of the polishing sheet area may include image feature extraction and feature data processing.

[0101] The silicon carbide polishing sheet crack detection device provided in an embodiment of the present application is described below. The silicon carbide polishing sheet crack detection device described below and the silicon carbide polishing sheet crack detection method described above can be referred to each other.

[0102] Figure 2 Schematic diagram of the structure of a silicon carbide polishing sheet crack detection device disclosed in the embodiment of the present application. Figure 2 As shown, the device may include:

[0103] An image acquisition unit 21 is used to acquire a dark field image of the silicon carbide polishing sheet, wherein the dark field image of the silicon carbide polishing sheet is obtained based on dark field reflection imaging;

[0104] A data processing unit 22 is used to extract an image area that meets the gray value abnormality condition from the dark field image of the silicon carbide polishing sheet, and determine a polishing sheet area corresponding to the extracted image area from the silicon carbide polishing sheet;

[0105] The image acquisition unit 21 is further used to acquire a polarization image of the polishing sheet area, wherein the polarization image of the polishing sheet area is obtained based on polarization imaging;

[0106] The data processing unit 22 is also used to call a pre-trained defect recognition model to process the polarization image of the polishing sheet area to obtain a defect type recognition result of the silicon carbide polishing sheet; wherein the defect recognition model is trained using a polarization sample image of the polishing sheet annotated with defect type information; when it is determined that the silicon carbide polishing sheet has a crack defect based on the defect type recognition result, the crack size of the silicon carbide polishing sheet is calculated based on the image of the silicon carbide polishing sheet, and a crack defect detection result of the silicon carbide polishing sheet is generated; the image of the silicon carbide polishing sheet includes a dark field image of the silicon carbide polishing sheet or a polarization image of the silicon carbide polishing sheet.

[0107] In one or more embodiments provided in the present application, the process in which the data processing unit 22 calculates the crack size of the silicon carbide polishing sheet according to the image of the silicon carbide polishing sheet may include:

[0108] The number of pixels occupied by the crack in the silicon carbide polishing sheet in the image of the silicon carbide polishing sheet is detected, and the crack size of the silicon carbide polishing sheet is calculated based on the image acquisition accuracy of the image acquisition device that acquires the image of the silicon carbide polishing sheet and the determined number of pixels.

[0109] In one or more embodiments provided in the present application, the process of the data processing unit 22 generating the crack defect detection result of the silicon carbide polishing wafer may include:

[0110] Generating a crack map of the silicon carbide polishing sheet;

[0111] Determining the grading result of the silicon carbide polishing sheet according to the correspondence between the pre-set crack defect situation and the polishing sheet grade; the crack defect situation includes the number of cracks and the crack size;

[0112] Generate a crack defect detection result of the silicon carbide polishing sheet, wherein the crack defect detection result includes: the crack map, the grading result, and the crack size, dark field image and polarization image of each crack in the silicon carbide polishing sheet.

[0113] In one or more embodiments provided in the present application, the data processing unit 22 is also used to perform image preprocessing on the dark field image of the silicon carbide polishing sheet before extracting the image area that meets the grayscale value abnormality condition from the dark field image of the silicon carbide polishing sheet; and, before calling a pre-trained defect recognition model to process the polarization image of the polishing sheet area, perform image preprocessing on the polarization image of the polishing sheet area; wherein the image preprocessing includes at least denoising processing or enhancement processing.

[0114] In one or more embodiments provided in the present application, the dark field image of the silicon carbide polishing sheet includes dark field block images of each region of the silicon carbide polishing sheet; the polarization image of the silicon carbide polishing sheet includes polarization block images of each region of the silicon carbide polishing sheet.

[0115] In one or more embodiments provided in the present application, the process of the image acquisition unit 21 acquiring the block image of each area of ​​the silicon carbide polishing wafer may include:

[0116] After determining that the silicon carbide polishing sheet is placed on the wafer stage, the silicon carbide polishing sheet is adsorbed by a vacuum adsorption component provided on the wafer stage, a preset light source that provides preset light for an image acquisition range of a preset image acquisition device is turned on, the wafer stage is moved to a preset initial position, the wafer stage is moved along a preset trajectory according to a preset step length, and block images of various regions of the silicon carbide polishing sheet acquired by the preset image acquisition device during the movement of the wafer stage are acquired;

[0117] Wherein, in the process that the wafer stage moves along the preset trajectory starting from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; when the acquired block image is a dark field block image, the preset image acquisition device is an image acquisition device for acquiring a dark field image, the preset light is a dark field light, and the preset light source is a dark field light source; when the acquired block image is a polarized block image, the preset image acquisition device is an image acquisition device for acquiring a polarized image, the preset light is polarized light, and the preset light source is a polarized light source.

[0118] In one or more embodiments provided in the present application, the process of the image acquisition unit 21 acquiring the block image of each area of ​​the silicon carbide polishing wafer may include:

[0119] After determining that the silicon carbide polishing sheet is placed on the wafer stage, the silicon carbide polishing sheet is adsorbed by a first group of subcomponents arranged on the wafer stage, and the silicon carbide polishing sheet is lifted to a preset height;

[0120] Turning on a preset light source that provides preset light for an image acquisition range of a preset image acquisition device;

[0121] Move the wafer stage to a preset initial position, move the wafer stage along a preset trajectory according to a preset step length, and collect block images of various areas of the silicon carbide polishing sheet adsorbed by the first group of sub-components during the movement of the wafer stage by the preset image acquisition device;

[0122] Switching to adsorbing the silicon carbide polishing sheet through a second group of subcomponents arranged on the wafer stage, and lifting the silicon carbide polishing sheet to a preset height;

[0123] The block image of the designated area of ​​the silicon carbide polishing sheet adsorbed by the second group of sub-components is collected by the preset image acquisition device; the designated area includes an area corresponding to the adsorption position of the first group of sub-components;

[0124] Obtaining block images of various regions of the silicon carbide polishing sheet, wherein the block images of various regions of the silicon carbide polishing sheet include: block images of the designated regions of the silicon carbide polishing sheet adsorbed by the second group of subcomponents, and block images of non-designated regions of the silicon carbide polishing sheet adsorbed by the first group of subcomponents;

[0125] Wherein, each of the first group of sub-components and the second group of sub-components includes a preset number of vacuum tubes connected to a vacuum pump and capable of being raised and lowered, and each group of sub-components is used to absorb the polishing sheet placed on the wafer stage and raise and lower the polishing sheet; in the process of the wafer stage moving along the preset trajectory starting from a preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; in the case where the acquired block image is a dark field block image, the preset image acquisition device is an image acquisition device for acquiring a dark field image, the preset light is a dark field light, and the preset light source is a dark field light source; in the case where the acquired block image is a polarized block image, the preset image acquisition device is an image acquisition device for acquiring a polarized image, the preset light is polarized light, and the preset light source is a polarized light source.

[0126] In one or more embodiments provided in the present application, the data processing unit 22 can also be used to determine the position of the crack in the silicon carbide polishing sheet based on the imaging resolution corresponding to the crack in the silicon carbide polishing sheet when it is determined that the silicon carbide polishing sheet has a crack defect based on the defect type identification result, and the position includes the upper surface or the lower surface.

[0127] The silicon carbide polishing sheet crack detection device provided in the embodiment of the present application can be applied to a silicon carbide polishing sheet crack detection controller, such as a terminal with data processing capability: a computer, etc. Optionally, Figure 3 The hardware structure diagram of the silicon carbide polishing sheet crack detection controller is shown in FIG. Figure 3 , the hardware structure of the silicon carbide polishing sheet crack detection controller may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;

[0128] In the embodiment of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 communicate with each other through the communication bus 4;

[0129] The processor 1 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention, etc.;

[0130] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), etc., such as at least one disk memory;

[0131] The memory is used to store computer programs, and the processor is used to execute the computer programs, so that the silicon carbide polishing sheet crack detection controller can implement any of the above-mentioned silicon carbide polishing sheet crack detection methods.

[0132] The embodiment of the present application also provides a silicon carbide polishing sheet crack detection device, which may include: a dark field light source 41, a first CCD camera 42, a polarized light source 43, a second CCD camera 44, a wafer stage 45, a vacuum adsorption component 46 and a moving component 47.

[0133] Among them, the dark field light source 41 can be used to provide dark field light for the image acquisition range of the first CCD camera 42; the first CCD camera 42 is used to acquire dark field images; the polarized light source 43 is used to provide polarized light for the image acquisition range of the second CCD camera; the second CCD camera 44 is used to acquire polarized images; the stage 45 is used to carry the silicon carbide polishing sheet; the vacuum adsorption component 46 is arranged on the stage 45, and the vacuum adsorption component 46 is connected to the vacuum pump 460, and is used to adsorb the silicon carbide polishing sheet placed on the stage 45; the moving component 47 is used to drive the stage 45 to move, so that each area of ​​the silicon carbide polishing sheet placed on the stage 45 is located within the image acquisition range of the first CCD camera 42 or within the image acquisition range of the second CCD camera 44.

[0134] In one possible implementation, the above-mentioned stage 45 can be an X / Y / Z three-axis movable stage. During the detection process, the Z axis can be configured at a suitable position, and there is no need to adjust the Z axis subsequently, and the XOY plane is a horizontal plane.

[0135] In one possible implementation, the moving component 47 may include a lead screw 471 and a motor 472. The stage 45 is disposed on the lead screw 471 and driven by the motor 472. The step length can be precisely set, and the X-axis and Y-axis can be moved to a specified position by means of the moving component 47.

[0136] Based on the above, Figure 4The schematic diagram of the structure of a crack detection device for a silicon carbide polishing sheet is illustrated.

[0137] On the basis of the above, the silicon carbide polishing sheet crack detection device may further include a silicon carbide polishing sheet crack detection controller (referred to as the controller for short), the silicon carbide polishing sheet crack detection controller is connected to the first CCD camera 42 and the second CCD camera 44 through a signal line, and obtains a dark field image and a polarization image from the first CCD camera 42 and the second CCD camera 44 respectively; the silicon carbide polishing sheet crack detection controller may include at least one processor and a memory connected to the processor, wherein the memory is used to store a computer program; the processor is used to execute the computer program, and the computer program is used to:

[0138] Acquiring a dark field image of the silicon carbide polishing sheet, wherein the dark field image of the silicon carbide polishing sheet is obtained based on dark field reflection imaging;

[0139] Extracting an image region satisfying a grayscale value abnormality condition from a dark field image of the silicon carbide polishing sheet, and determining a polishing sheet region corresponding to the extracted image region from the silicon carbide polishing sheet;

[0140] Acquire a polarization image of the polishing sheet area, wherein the polarization image of the polishing sheet area is obtained based on polarization imaging;

[0141] Calling a pre-trained defect recognition model to process the polarization image of the polishing sheet area to obtain a defect type recognition result of the silicon carbide polishing sheet; wherein the defect recognition model is trained using a polarization sample image of the polishing sheet annotated with defect type information;

[0142] When it is determined that the silicon carbide polishing sheet has a crack defect based on the defect type identification result, the crack size of the silicon carbide polishing sheet is calculated based on the image of the silicon carbide polishing sheet, and a crack defect detection result of the silicon carbide polishing sheet is generated; the image of the silicon carbide polishing sheet includes a dark field image of the silicon carbide polishing sheet or a polarization image of the silicon carbide polishing sheet.

[0143] In one or more embodiments provided in the present application, the process of the silicon carbide polishing sheet crack detection controller calculating the crack size of the silicon carbide polishing sheet according to the image of the silicon carbide polishing sheet may include:

[0144] The number of pixels occupied by the crack in the silicon carbide polishing sheet in the image of the silicon carbide polishing sheet is detected, and the crack size of the silicon carbide polishing sheet is calculated based on the image acquisition accuracy of the image acquisition device that acquires the image of the silicon carbide polishing sheet and the determined number of pixels.

[0145] In one or more embodiments provided in the present application, the process of the silicon carbide polishing sheet crack detection controller generating the crack defect detection result of the silicon carbide polishing sheet may include:

[0146] Generating a crack map of the silicon carbide polishing sheet;

[0147] Determining the grading result of the silicon carbide polishing sheet according to the correspondence between the pre-set crack defect situation and the polishing sheet grade; the crack defect situation includes the number of cracks and the crack size;

[0148] Generate a crack defect detection result of the silicon carbide polishing sheet, wherein the crack defect detection result includes: the crack map, the grading result, and the crack size, dark field image and polarization image of each crack in the silicon carbide polishing sheet.

[0149] In one or more embodiments provided in the present application, the silicon carbide polishing sheet crack detection controller can also be used to perform image preprocessing on the dark field image of the silicon carbide polishing sheet before extracting the image area that meets the grayscale value abnormality condition from the dark field image of the silicon carbide polishing sheet; and, before calling a pre-trained defect recognition model to process the polarization image of the polishing sheet area, perform image preprocessing on the polarization image of the polishing sheet area; wherein the image preprocessing includes at least denoising processing or enhancement processing.

[0150] In one or more embodiments provided in the present application, the dark field image of the silicon carbide polishing sheet includes dark field block images of each region of the silicon carbide polishing sheet; the polarization image of the silicon carbide polishing sheet includes polarization block images of each region of the silicon carbide polishing sheet.

[0151] In one or more embodiments provided in the present application, the silicon carbide polishing sheet crack detection controller is respectively connected to the dark field light source 41, the polarized light source 43, the vacuum adsorption component 46 and the moving component 47 through signal lines. On this basis, the process of the silicon carbide polishing sheet crack detection controller acquiring the block image of each area of ​​the silicon carbide polishing sheet may include:

[0152] After determining that the silicon carbide polishing sheet is placed on the wafer stage 45, the silicon carbide polishing sheet is adsorbed by the vacuum adsorption component 46, a preset light source that provides preset light for an image acquisition range of a preset image acquisition device is turned on, the wafer stage 45 is moved to a preset initial position, the wafer stage 45 is moved along a preset trajectory according to a preset step length, and block images of various regions of the silicon carbide polishing sheet acquired by the preset image acquisition device during the movement of the wafer stage 45 are acquired;

[0153] Wherein, in the process that the wafer stage 45 moves along the preset trajectory starting from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; when the acquired block image is a dark field block image, the preset image acquisition device is the first CCD camera 42, and the preset light source is the dark field light source 41; when the acquired block image is a polarized block image, the preset image acquisition device is the second CCD camera 44, and the preset light source is the polarized light source 43.

[0154] In one or more embodiments provided in the present application, the vacuum adsorption component 46 includes two groups of sub-components, which can be represented as a first group of sub-components 461 and a second group of sub-components 462. Each group of sub-components includes a preset number of vacuum tubes connected to a vacuum pump and capable of being raised and lowered. Each group of sub-components is used to adsorb the polishing sheet placed on the wafer carrier and to lift and lower the polishing sheet.

[0155] The silicon carbide polishing sheet crack detection controller is connected to the dark field light source 41, the polarized light source 43, the first group of subcomponents 461, the second group of subcomponents 462 and the moving component 47 through signal lines. On this basis, the process of the silicon carbide polishing sheet crack detection controller acquiring the block image of each area of ​​the silicon carbide polishing sheet may include:

[0156] After determining that the silicon carbide polishing sheet is placed on the wafer stage 45, the silicon carbide polishing sheet is adsorbed by the first group of subcomponents 461, and the silicon carbide polishing sheet is lifted to a preset height;

[0157] Turn on the preset light source;

[0158] Move the wafer stage 45 to a preset initial position, move the wafer stage 45 along a preset trajectory according to a preset step length, and collect block images of various areas of the silicon carbide polishing sheet adsorbed by the first group of subcomponents 461 during the movement of the wafer stage 45 by the preset image acquisition device;

[0159] Switching to the second group of subcomponents 462 to absorb the silicon carbide polishing sheet, and lifting the silicon carbide polishing sheet to the preset height;

[0160] The block image of the designated area of ​​the silicon carbide polishing sheet adsorbed by the second group of sub-components 462 is collected by the preset image acquisition device; the designated area includes the area corresponding to the adsorption position of the first group of sub-components 461;

[0161] Obtaining block images of various regions of the silicon carbide polishing sheet, wherein the block images of various regions of the silicon carbide polishing sheet include: block images of the designated regions of the silicon carbide polishing sheet adsorbed by the second group of subcomponents, and block images of non-designated regions of the silicon carbide polishing sheet adsorbed by the first group of subcomponents;

[0162] Wherein, in the process that the wafer stage 45 moves along the preset trajectory starting from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; when the acquired block image is a dark field block image, the preset image acquisition device is the first CCD camera 42, and the preset light source is the dark field light source 41; when the acquired block image is a polarized block image, the preset image acquisition device is the second CCD camera 44, and the preset light source is the polarized light source 43.

[0163] In one or more embodiments provided in the present application, the silicon carbide polishing pad crack detection controller can also be used to determine the position of the crack in the silicon carbide polishing pad based on the imaging resolution corresponding to the crack in the silicon carbide polishing pad when it is determined that the silicon carbide polishing pad has a crack defect based on the defect type identification result, and the position includes the upper surface or the lower surface.

[0164] In addition, the silicon carbide polishing sheet crack detection controller can also be used to: before starting the detection, the wafer stage 45 can be reset, and the X-axis and the Y-axis can be moved to the preset upper and lower wafer positions to wait for the placement of the silicon carbide polishing sheet to be detected. Optionally, the silicon carbide polishing sheet can be placed on the wafer stage 45 with the Si surface facing up by a suction pen or a manipulator, and then the silicon carbide polishing sheet crack detection controller determines that the silicon carbide polishing sheet has been placed on the wafer stage 45 after receiving the polishing sheet engraving number input by the user.

[0165] For example, Figure 5 A schematic top view of a crack detection apparatus for a silicon carbide polishing pad is illustrated, wherein upper and lower pad positions 48 are shown.

[0166] For example, Figure 6 The schematic diagram of the structure of the slide stage is shown in the figure. Figure 6As shown, the vacuum adsorption component 46 may include two groups of subcomponents, the first group of subcomponents 461 is composed of three outer ring feet, and the second group of subcomponents 462 is composed of three inner ring feet, and the three outer ring feet and the three inner ring feet are alternately raised and adsorb the polishing sheet. On this basis, the controller can start to execute the detection process after determining that the silicon carbide polishing sheet has been placed on the wafer stage 45, and the process may include: dark field mode image acquisition process, polarization mode image acquisition process, image processing process. The image process flow of any mode may include: controlling the three outer ring feet to rise and vacuum absorb the polishing sheet, controlling the upper left corner of the wafer on the wafer stage to automatically move to the bottom of the CCD camera (42 or 44), and collecting images of the polishing sheet according to the preset step length. The collection sequence is to start from the upper left and move in an S shape from top to bottom. Each time the wafer stage moves, the CCD camera is controlled to take a picture. After the S shape is completed, the three inner ring feet are controlled to rise and absorb the wafer, and the three outer ring feet fall. The wafer stage is controlled to move through the moving parts and the CCD camera is controlled to take pictures of the absorption area of ​​the three outer ring feet to complete the data, and finally a complete polishing sheet image is obtained, which is composed of multiple high-resolution block images. The image processing process may include: processing the images collected in the dark field mode and the polarization mode, which may specifically include pre-processing such as denoising and enhancement, feature extraction, and data analysis processing. The data analysis and processing may include: adding traditional algorithm card control to identify and capture defects, and using the artificial intelligence (AI) large model trained through sample data to accurately identify, capture and classify defects to achieve the purpose of distinguishing cracks from other defects; after determining the existence of crack defects, calculate the crack length based on the accuracy of the CCD camera and the number of pixels occupied by the crack defect, and then grade the silicon carbide polishing sheet based on the individual crack length or the cumulative crack length, such as dividing it into A, B, and NG grades, and finally output the crack map, crack size data, crack original channel map (dark field map or polarization map) and polishing sheet grading results. For example, Figure 7 The defect map diagram is illustrated, which shows the distribution location and crack morphology of crack defects and can be used as a crack map diagram; in addition, Figure 7 The distribution location and defect morphology of other defects such as C scratch defects are also shown. Figure 8 Example of a dark field image of a crack defect, combined with Figure 8 As shown, the crack area has a lower gray value; Fig. 9 Example of a polarization image of a crack defect, combined with Fig. 9 As shown, there is a gray value change area around the crack caused by stress lines.

[0167] A storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any one of the silicon carbide polishing sheet crack detection methods provided in the embodiment of the present application.

[0168] An embodiment of the present application also provides a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any one of the silicon carbide polishing sheet crack detection methods provided in the embodiments of the present application.

[0169] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0170] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.

[0171] 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 cracks in a silicon carbide polishing sheet, characterized in that: include: Acquire a dark field image of the silicon carbide polishing sheet, wherein the dark field image of the silicon carbide polishing sheet is obtained based on dark field reflection imaging; Extracting an image region satisfying a grayscale value abnormality condition from a dark field image of the silicon carbide polishing sheet, and determining a polishing sheet region corresponding to the extracted image region from the silicon carbide polishing sheet; Acquire a polarization image of the polishing sheet area, wherein the polarization image of the polishing sheet area is obtained based on polarization imaging; Calling a pre-trained defect recognition model to process the polarization image of the polishing sheet area to obtain a defect type recognition result of the silicon carbide polishing sheet; wherein the defect recognition model is trained using a polarization sample image of the polishing sheet annotated with defect type information; When it is determined that the silicon carbide polishing sheet has a crack defect based on the defect type identification result, the crack size of the silicon carbide polishing sheet is calculated based on the image of the silicon carbide polishing sheet, and a crack defect detection result of the silicon carbide polishing sheet is generated; the image of the silicon carbide polishing sheet includes a dark field image of the silicon carbide polishing sheet or a polarization image of the silicon carbide polishing sheet.

2. The method for detecting cracks in a silicon carbide polishing sheet according to claim 1, wherein: Calculating the crack size of the silicon carbide polishing sheet according to the image of the silicon carbide polishing sheet comprises: The number of pixels occupied by the crack in the silicon carbide polishing sheet in the image of the silicon carbide polishing sheet is detected, and the crack size of the silicon carbide polishing sheet is calculated based on the image acquisition accuracy of the image acquisition device that acquires the image of the silicon carbide polishing sheet and the determined number of pixels.

3. The method for detecting cracks in a silicon carbide polishing sheet according to claim 2, wherein: Generating a crack defect detection result of the silicon carbide polishing sheet, comprising: Generating a crack map of the silicon carbide polishing sheet; Determining the grading result of the silicon carbide polishing sheet according to the correspondence between the pre-set crack defect situation and the polishing sheet grade; the crack defect situation includes the number of cracks and the crack size; Generate a crack defect detection result of the silicon carbide polishing sheet, wherein the crack defect detection result includes: the crack map, the grading result, and the crack size, dark field image and polarization image of each crack in the silicon carbide polishing sheet.

4. The method for detecting cracks in a silicon carbide polishing sheet according to any one of claims 1 to 3, characterized in that: Before extracting the image area satisfying the gray value abnormality condition from the dark field image of the silicon carbide polishing sheet, the method further comprises: Performing image preprocessing on the dark field image of the silicon carbide polishing sheet; Before calling the pre-trained defect recognition model to process the polarization image of the polishing sheet area, the method further includes: Performing image preprocessing on the polarization image of the polishing sheet area; The image preprocessing at least includes denoising or enhancement.

5. The method for detecting cracks in a silicon carbide polishing sheet according to claim 4, wherein: The dark field image of the silicon carbide polishing sheet includes dark field block images of various regions of the silicon carbide polishing sheet; and the polarization image of the silicon carbide polishing sheet includes polarization block images of various regions of the silicon carbide polishing sheet.

6. The method for detecting cracks in a silicon carbide polishing sheet according to claim 5, wherein: The process of acquiring the block images of each area of ​​the silicon carbide polishing sheet includes: After determining that the silicon carbide polishing sheet is placed on the wafer stage, the silicon carbide polishing sheet is adsorbed by a vacuum adsorption component provided on the wafer stage, a preset light source that provides preset light for an image acquisition range of a preset image acquisition device is turned on, the wafer stage is moved to a preset initial position, the wafer stage is moved along a preset trajectory according to a preset step length, and block images of various regions of the silicon carbide polishing sheet acquired by the preset image acquisition device during the movement of the wafer stage are acquired; Wherein, in the process that the wafer stage moves along the preset trajectory starting from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; when the acquired block image is a dark field block image, the preset image acquisition device is an image acquisition device for acquiring a dark field image, the preset light is a dark field light, and the preset light source is a dark field light source; when the acquired block image is a polarized block image, the preset image acquisition device is an image acquisition device for acquiring a polarized image, the preset light is polarized light, and the preset light source is a polarized light source.

7. The method for detecting cracks in a silicon carbide polishing sheet according to claim 5, wherein: The process of acquiring the block images of each area of ​​the silicon carbide polishing sheet includes: After determining that the silicon carbide polishing sheet is placed on the wafer stage, the silicon carbide polishing sheet is adsorbed by a first group of subcomponents arranged on the wafer stage, and the silicon carbide polishing sheet is lifted to a preset height; Turning on a preset light source that provides preset light for an image acquisition range of a preset image acquisition device; Move the wafer stage to a preset initial position, move the wafer stage along a preset trajectory according to a preset step length, and collect block images of various areas of the silicon carbide polishing sheet adsorbed by the first group of sub-components during the movement of the wafer stage by the preset image acquisition device; Switching to adsorbing the silicon carbide polishing sheet through a second group of subcomponents arranged on the wafer stage, and lifting the silicon carbide polishing sheet to the preset height; The block image of the designated area of ​​the silicon carbide polishing sheet adsorbed by the second group of sub-components is collected by the preset image acquisition device; the designated area includes an area corresponding to the adsorption position of the first group of sub-components; Obtaining block images of various regions of the silicon carbide polishing sheet, wherein the block images of various regions of the silicon carbide polishing sheet include: block images of the designated regions of the silicon carbide polishing sheet adsorbed by the second group of subcomponents, and block images of non-designated regions of the silicon carbide polishing sheet adsorbed by the first group of subcomponents; Wherein, each of the first group of sub-components and the second group of sub-components includes a preset number of vacuum tubes connected to a vacuum pump and capable of being raised and lowered, and each group of sub-components is used to absorb the polishing sheet placed on the wafer stage and raise and lower the polishing sheet; in the process of the wafer stage moving along the preset trajectory starting from a preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; in the case where the acquired block image is a dark field block image, the preset image acquisition device is an image acquisition device for acquiring a dark field image, the preset light is a dark field light, and the preset light source is a dark field light source; in the case where the acquired block image is a polarized block image, the preset image acquisition device is an image acquisition device for acquiring a polarized image, the preset light is polarized light, and the preset light source is a polarized light source.

8. The method for detecting cracks in a silicon carbide polishing sheet according to claim 7, wherein: In the case where it is determined that the silicon carbide polishing sheet has a crack defect according to the defect type identification result, the method further includes: The position of the crack in the silicon carbide polishing sheet is determined according to an imaging resolution corresponding to the crack in the silicon carbide polishing sheet, wherein the position includes an upper surface or a lower surface.

9. A silicon carbide polishing sheet crack detection device, characterized in that: The device comprises: a dark field light source, a first CCD camera, a polarized light source, a second CCD camera, a slide stage, a vacuum adsorption component and a moving component; The dark field light source is used to provide dark field light for the image acquisition range of the first CCD camera; The first CCD camera is used to collect dark field images; The polarized light source is used to provide polarized light for the image acquisition range of the second CCD camera; The second CCD camera is used to collect polarization images; The wafer carrier is used for carrying the silicon carbide polishing wafer; The vacuum adsorption component is arranged on the wafer carrier and connected to a vacuum pump, and the vacuum adsorption component is used to adsorb the silicon carbide polishing sheet placed on the wafer carrier; The moving component is used to drive the wafer stage to move so that each area of ​​the silicon carbide polishing sheet placed on the wafer stage is located within the image acquisition range of the first CCD camera or within the image acquisition range of the second CCD camera; The device also includes: a silicon carbide polishing sheet crack detection controller; The silicon carbide polishing sheet crack detection controller is connected to the first CCD camera and the second CCD camera through a signal line, and obtains a dark field image and a polarization image from the first CCD camera and the second CCD camera respectively; the silicon carbide polishing sheet crack detection controller includes at least one processor and a memory connected to the processor, wherein the memory is used to store a computer program; the processor is used to execute the computer program so that the silicon carbide polishing sheet crack detection controller can implement the silicon carbide polishing sheet crack detection method as described in any one of claims 1 to 5.

10. The silicon carbide polishing sheet crack detection device according to claim 9, characterized in that: The vacuum adsorption component includes a first group of sub-components and a second group of sub-components, each group of sub-components includes a preset number of vacuum tubes connected to the vacuum pump and capable of being raised and lowered, and each group of sub-components is used to adsorb the polishing sheet placed on the wafer stage and to raise and lower the polishing sheet; In the case where the dark field image of the silicon carbide polishing sheet includes dark field block images of each region of the silicon carbide polishing sheet, and the polarization image of the silicon carbide polishing sheet includes polarization block images of each region of the silicon carbide polishing sheet, the process of the silicon carbide polishing sheet crack detection controller acquiring the block images of each region of the silicon carbide polishing sheet includes: After determining that the silicon carbide polishing sheet is placed on the wafer stage, the silicon carbide polishing sheet is adsorbed by the first group of subcomponents, and the silicon carbide polishing sheet is lifted to a preset height; Turn on the preset light source; Move the wafer stage to a preset initial position, move the wafer stage along a preset trajectory according to a preset step length, and collect block images of various areas of the silicon carbide polishing sheet adsorbed by the first group of sub-components during the movement of the wafer stage by a preset image acquisition device; Switching to the second group of subcomponents to absorb the silicon carbide polishing sheet, and lifting the silicon carbide polishing sheet to the preset height; The block image of the designated area of ​​the silicon carbide polishing sheet adsorbed by the second group of sub-components is collected by the preset image acquisition device; the designated area includes an area corresponding to the adsorption position of the first group of sub-components; Obtaining block images of various regions of the silicon carbide polishing sheet, wherein the block images of various regions of the silicon carbide polishing sheet include: block images of the designated regions of the silicon carbide polishing sheet adsorbed by the second group of subcomponents, and block images of non-designated regions of the silicon carbide polishing sheet adsorbed by the first group of subcomponents; Wherein, in the process that the wafer stage moves along the preset trajectory starting from the preset initial position, each area of ​​the silicon carbide polishing sheet is sequentially located within the image acquisition range of the preset image acquisition device; when the acquired block image is a dark field block image, the preset image acquisition device is the first CCD camera, and the preset light source is the dark field light source; when the acquired block image is a polarized block image, the preset image acquisition device is the second CCD camera, and the preset light source is the polarized light source.

Images