Silicon wafer position compensation method and related equipment

Abstract

The embodiment of the present application discloses a silicon wafer position compensation method, an image processing device, and a computer-readable storage medium, which are used to provide a silicon wafer position compensation method, so that the position of the silicon wafer can be corrected to an ideal position through a compensation vector value. The method of the embodiment of the present application includes: obtaining an original silicon wafer image, inputting the original silicon wafer image into a pre-trained silicon wafer position compensation model, determining the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determining the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determining the ideal silicon wafer center point position of the original silicon wafer image, determining the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and converting the compensation vector value from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of silicon wafer position compensation, and more specifically, to a silicon wafer position compensation method, an image processing device, and a computer-readable storage medium. Background Art

[0002] In the semiconductor manufacturing process, semiconductor measurement and inspection machines are widely used in the wafer testing stage. Specifically, the detection device can be used to detect the wafer surface, the computer equipment can process the detected image and extract the defects, and the processed image data can be stored in the computer equipment. The wafer defect detection image can be displayed in the computer equipment, and the distribution of defects on the wafer can be determined based on the detection image.

[0003] However, in semiconductor measurement machine imaging, there are often problems such as the position of the silicon wafer deviating from the correct position center, the silicon wafer rotating to a certain extent due to mechanical vibration, and unclear edges of the silicon wafer. Therefore, in order to better detect wafer defects on the detected image and improve the accuracy of defect location recognition, a silicon wafer position compensation method is urgently needed to correct the silicon wafer to the ideal position. Summary of the invention

[0004] The embodiments of the present application provide a silicon wafer position compensation method, an image processing device, and a computer-readable storage medium, which are used to provide a silicon wafer position compensation method so that the position of the silicon wafer can be corrected to an ideal position through a compensation vector value.

[0005] In a first aspect, an embodiment of the present application provides a silicon wafer position compensation method, comprising:

[0006] Obtaining original silicon wafer images;

[0007] The original silicon wafer image is input into a pre-trained silicon wafer position compensation model, the silicon wafer position compensation model determines the silicon wafer contour information of the original silicon wafer image, the silicon wafer center point position of the original silicon wafer image is determined based on the silicon wafer contour information, the ideal silicon wafer center point position of the original silicon wafer image is determined, the compensation vector value of the silicon wafer is determined based on the silicon wafer center point position and the ideal silicon wafer center point position, and the compensation vector value is converted from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

[0008] Optionally, determining the position of the center point of the silicon wafer of the original silicon wafer image based on the silicon wafer contour information includes:

[0009] Performing arc fitting on the silicon wafer contour information based on the least square method to obtain a first arc corresponding to the silicon wafer contour information;

[0010] A first center position corresponding to the first arc is determined, and the first center position is used as the center point position of the silicon wafer.

[0011] Optionally, the original silicon wafer image describes information of grooves and / or information of tentacles; wherein the grooves and / or the tentacles are structures in a wafer suction cup for fixing the wafer to be inspected;

[0012] The step of determining the ideal silicon wafer center point position of the original silicon wafer image comprises:

[0013] Determining the boundary information of the groove and / or the boundary information of the tentacle in the original silicon wafer image by an edge-finding algorithm;

[0014] Based on the least square method, arc fitting is performed on the boundary information of the groove and / or the boundary information of the tentacle to obtain a second arc;

[0015] A second center position corresponding to the second arc is determined, and the second center position is used as the ideal silicon wafer center point position.

[0016] Optionally, before inputting the original silicon wafer image into a pre-trained silicon wafer position compensation model, the method further includes:

[0017] Obtaining an original silicon wafer image sample; wherein the original silicon wafer image sample is marked with a target compensation vector value;

[0018] Input the original silicon wafer image samples into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information samples of the original silicon wafer image samples by the silicon wafer position compensation model, determine the silicon wafer center point position samples of the original silicon wafer image samples based on the silicon wafer contour information samples, determine the ideal silicon wafer center point position samples of the original silicon wafer image samples, determine the compensation vector value samples of the silicon wafer based on the silicon wafer center point position samples and the ideal silicon wafer center point position samples, and transform the compensation vector value samples from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the predicted target compensation vector value output by the silicon wafer position compensation model; wherein the predicted target compensation vector value is used to compensate the position of the silicon wafer in the image coordinate system;

[0019] The loss between the predicted target compensation vector value and the marked target compensation vector value is calculated according to the regression loss function. When the loss meets the convergence condition, a trained silicon wafer position compensation model is obtained.

[0020] Optionally, after obtaining the original silicon wafer image, the method further includes:

[0021] Converting the grayscale value of the pixel in the original silicon wafer image into a binary domain to obtain the grayscale value of the pixel in the original silicon wafer image after the conversion;

[0022] The step of inputting the original silicon wafer image into a pre-trained silicon wafer position compensation model comprises:

[0023] The grayscale values ​​of the converted representation of the pixels in the original silicon wafer image are input into a pre-trained silicon wafer position compensation model.

[0024] Optionally, after converting the grayscale value of the pixel point in the original silicon wafer image into a binary domain, the method further includes:

[0025] The grayscale values ​​of the converted pixel points in the original silicon wafer image are saved in the memory, and the grayscale values ​​of the converted pixel points in the original silicon wafer image are locked by byte lock;

[0026] After determining the compensation vector value of the silicon wafer based on the center point position of the silicon wafer and the ideal center point position of the silicon wafer, the method further includes:

[0027] The byte lock is closed, and the step of converting the compensation vector value from the image coordinate system to the world coordinate system based on the conversion matrix is ​​performed to obtain the target compensation vector value output by the silicon wafer position compensation model.

[0028] Optionally, determining the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determining the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determining the ideal silicon wafer center point position of the original silicon wafer image, determining the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and converting the compensation vector value from the image coordinate system to the world coordinate system based on a conversion matrix to obtain a target compensation vector value output by the silicon wafer position compensation model, includes:

[0029] Execute a first task through a first thread; wherein the first task is a task corresponding to the silicon wafer contour information of the original silicon wafer image determined by the silicon wafer position compensation model;

[0030] Execute a second task through a second thread; wherein the second task is a task corresponding to the position of the center point of the silicon wafer of the original silicon wafer image determined by the silicon wafer position compensation model based on the silicon wafer contour information;

[0031] Executing a third task through a third thread; wherein the third task is a task corresponding to the ideal silicon wafer center point position of the original silicon wafer image determined by the silicon wafer position compensation model;

[0032] Executing a fourth task through a fourth thread; wherein the fourth task is to determine a compensation vector value of a silicon wafer based on the center point position of the silicon wafer and the center point position of the ideal silicon wafer by the silicon wafer position compensation model, so as to obtain a task corresponding to the compensation vector value output by the silicon wafer position compensation model;

[0033] The fifth task is executed through the fifth thread; wherein the fifth task is a task corresponding to the target compensation vector value output by the silicon wafer position compensation model by converting the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix; and the first thread, the second thread, the third thread, the fourth thread and the fifth thread are executed in parallel.

[0034] In a second aspect, an embodiment of the present application provides an image processing device, including:

[0035] An acquisition unit, used for acquiring an original silicon wafer image;

[0036] A silicon wafer position compensation unit is used to input the original silicon wafer image into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determine the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determine the ideal silicon wafer center point position of the original silicon wafer image, determine the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and convert the compensation vector value from the image coordinate system to the world coordinate system based on a conversion matrix to obtain a target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

[0037] In a third aspect, an embodiment of the present application provides an image processing device, including:

[0038] CPU, memory, input and output interfaces, wired or wireless network interfaces, and power supply;

[0039] The memory is a short-term storage memory or a persistent storage memory;

[0040] The central processing unit is configured to communicate with the memory and execute instructions in the memory to perform the aforementioned silicon wafer position compensation method.

[0041] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which includes instructions. When the instructions are executed on a computer, the computer executes the aforementioned silicon wafer position compensation method.

[0042] In a fifth aspect, an embodiment of the present application provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the aforementioned silicon wafer position compensation method.

[0043] It can be seen from the above technical scheme that the embodiment of the present application has the following advantages: the original silicon wafer image can be obtained, the original silicon wafer image is input into the pre-trained silicon wafer position compensation model, the silicon wafer contour information of the original silicon wafer image is determined by the silicon wafer position compensation model, the silicon wafer center point position of the original silicon wafer image is determined based on the silicon wafer contour information, the ideal silicon wafer center point position of the original silicon wafer image is determined, the compensation vector value of the silicon wafer is determined based on the silicon wafer center point position and the ideal silicon wafer center point position, and the compensation vector value is converted from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model, wherein the target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system. A silicon wafer position compensation method is provided, which can accurately determine the position deviation of the silicon wafer and calculate the compensation vector value, and the position of the silicon wafer can be corrected to the ideal position through the compensation vector value. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of the architecture of a silicon wafer position compensation system disclosed in an embodiment of the present application;

[0045] Figure 2 A schematic flow chart of a silicon wafer position compensation method disclosed in an embodiment of the present application;

[0046] Figure 3 A schematic flow chart of another silicon wafer position compensation method disclosed in an embodiment of the present application;

[0047] Figure 4 A schematic diagram of the structure of an image processing device disclosed in an embodiment of the present application;

[0048] Figure 5 A schematic diagram of the structure of another image processing device disclosed in an embodiment of the present application;

[0049] Figure 6 This is a schematic diagram of the structure of another image processing device disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0050] The embodiments of the present application provide a silicon wafer position compensation method, an image processing device, and a computer-readable storage medium, which are used to provide a silicon wafer position compensation method so that the position of the silicon wafer can be corrected to an ideal position through a compensation vector value.

[0051] See also Figure 1 , the architecture of the silicon wafer position compensation system in the embodiment of the present application includes:

[0052] Image processing device 101 and semiconductor measurement and testing machine 102. When performing silicon wafer position compensation, the image processing device 101 can be connected to the semiconductor measurement and testing machine 102. The semiconductor measurement and testing machine 102 can send the original silicon wafer image to the image processing device 101, and the image processing device 101 can input the original silicon wafer image into a pre-trained silicon wafer position compensation model to obtain a target compensation vector value output by the silicon wafer position compensation model, wherein the target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system.

[0053] based on Figure 1 See the silicon wafer position compensation system shown in Figure 2 , Figure 2 The present invention is a flow chart of a silicon wafer position compensation method disclosed in an embodiment of the present application, the method comprising:

[0054] 201. Obtain an original silicon wafer image.

[0055] In this embodiment, when the silicon wafer position compensation is performed, the original silicon wafer image can be obtained.

[0056] 202. Input the original silicon wafer image into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determine the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determine the ideal silicon wafer center point position of the original silicon wafer image, determine the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and convert the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix to obtain a target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

[0057] After obtaining the original silicon wafer image, the original silicon wafer image can be input into a pre-trained silicon wafer position compensation model, and the silicon wafer position compensation model determines the silicon wafer contour information of the original silicon wafer image, determines the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determines the ideal silicon wafer center point position of the original silicon wafer image, determines the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and converts the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

[0058] It should be understood that the position information of the silicon wafer in the world coordinate system usually includes the position information of the center point and the angle information. The target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system. In addition to compensating the center point position of the silicon wafer, the rotation angle of the silicon wafer around its center point can also be compensated. The compensation vector value represents the offset and rotation of the silicon wafer. By adjusting the value of the compensation vector, the precise positioning and rotation control of the silicon wafer in the world coordinate system can be achieved.

[0059] Among them, the method for determining the position of the center point of the silicon wafer of the original silicon wafer image based on the silicon wafer contour information may be to first perform arc fitting on the silicon wafer contour information based on the least squares method to obtain a first arc corresponding to the silicon wafer contour information, and then determine the first center position corresponding to the first arc, and use the first center position as the position of the center point of the silicon wafer. It can be understood that in addition to the method for determining the position of the center point of the silicon wafer of the original silicon wafer image based on the silicon wafer contour information described above, other reasonable methods may also be used, which are not specifically limited here.

[0060] It is understandable that in addition to arc fitting of silicon wafer contour information based on least squares method, other reasonable fitting algorithms can also be used, including but not limited to RANSAC algorithm (random sampling consistency), nonlinear optimization method, deep learning method, weighted least squares method, overall least squares method, etc.

[0061] In an embodiment of the present application, an original silicon wafer image can be obtained, and the original silicon wafer image can be input into a pre-trained silicon wafer position compensation model, and the silicon wafer contour information of the original silicon wafer image can be determined by the silicon wafer position compensation model, and the silicon wafer center point position of the original silicon wafer image can be determined based on the silicon wafer contour information, and the ideal silicon wafer center point position of the original silicon wafer image can be determined, and the compensation vector value of the silicon wafer can be determined based on the silicon wafer center point position and the ideal silicon wafer center point position, and the compensation vector value can be converted from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model, wherein the target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system. A silicon wafer position compensation method is provided, which can accurately determine the position deviation of the silicon wafer and calculate the compensation vector value, and the position of the silicon wafer can be corrected to the ideal position through the compensation vector value.

[0062] In the embodiment of the present application, there are multiple methods for silicon wafer position compensation, based on Figure 2 The silicon wafer position compensation method shown is described below.

[0063] In this embodiment, when the silicon wafer position compensation is performed, the original silicon wafer image can be obtained.

[0064] After obtaining the original silicon wafer image, the original silicon wafer image can be input into a pre-trained silicon wafer position compensation model, and the silicon wafer position compensation model determines the silicon wafer contour information of the original silicon wafer image, determines the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determines the ideal silicon wafer center point position of the original silicon wafer image, determines the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and converts the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the image coordinate system.

[0065] Among them, the method for determining the center point position of the silicon wafer of the original silicon wafer image based on the silicon wafer contour information can be, first, performing arc fitting on the silicon wafer contour information based on the least squares method to obtain the first arc corresponding to the silicon wafer contour information, and then determining the first center position corresponding to the first arc, and using the first center position as the center point position of the silicon wafer.

[0066] It is worth mentioning that arc fitting of the silicon wafer contour information based on the least squares method can obtain an approximate arc, which can reflect the overall shape of the silicon wafer. Secondly, determining the center position of the circle corresponding to the first arc can locate the center point of the silicon wafer, which is of great significance for subsequent process steps (such as cutting, testing, measurement, etc.) and accuracy requirements. Furthermore, determining the center position of the circle corresponding to the first arc can determine the rotation angle of the silicon wafer, which is critical for alignment, calibration and assembly with other components. Finally, determining the center point position of the silicon wafer can also help better arrange the production line and improve production efficiency. For example, silicon wafers closer to the center point can be processed first to reduce the time for movement and adjustment.

[0067] Among them, the method for determining the ideal silicon wafer center point position of the original silicon wafer image can be, first, determining the boundary information of the groove and / or the boundary information of the tentacle in the original silicon wafer image through an edge finding algorithm, and then based on the least squares method, performing arc fitting on the boundary information of the groove and / or the boundary information of the tentacle to obtain a second arc, and finally determining the second center position corresponding to the second arc, and using the second center position as the ideal silicon wafer center point position.

[0068] It is worth mentioning that the boundary information of the grooves and tentacles in the original silicon wafer image can be determined by an edge-finding algorithm. The edge-finding algorithm includes but is not limited to the Canny edge detection algorithm, the Sobel operator edge detection algorithm, the Laplacian operator edge detection algorithm, the Scharr operator edge detection algorithm, the Roberts operator edge detection algorithm or the Prewitt operator edge detection algorithm. Secondly, based on the least squares method, arc fitting can be performed on the boundary information of the grooves and tentacles. The least squares method is a commonly used mathematical optimization method that determines the fitting curve or function by minimizing the square difference between the actual observed value and the fitted value. The arc fitting can obtain a second arc, and the center position corresponding to the second arc can be used as an estimate of the center point position of the ideal silicon wafer. The advantage of this is that based on the boundary information of the grooves and tentacles, the center position of the silicon wafer can be accurately obtained through a mathematical model, so as to further analyze and process it.

[0069] Wherein, before the original silicon wafer image is input into the pre-trained silicon wafer position compensation model, the silicon wafer position compensation model may be pre-trained.

[0070] Among them, the method for training the silicon wafer position compensation model can be, first obtaining the original silicon wafer image sample, wherein the original silicon wafer image sample is marked with a target compensation vector value, and then inputting the original silicon wafer image sample into the pre-trained silicon wafer position compensation model, determining the silicon wafer contour information sample of the original silicon wafer image sample by the silicon wafer position compensation model, determining the silicon wafer center point position sample of the original silicon wafer image sample based on the silicon wafer contour information sample, determining the ideal silicon wafer center point position sample of the original silicon wafer image sample, determining the compensation vector value sample of the silicon wafer based on the silicon wafer center point position sample and the ideal silicon wafer center point position sample, and converting the compensation vector value sample from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the predicted target compensation vector value output by the silicon wafer position compensation model, wherein the predicted target compensation vector value is used to compensate for the position of the silicon wafer in the image coordinate system, and finally calculating the loss between the predicted target compensation vector value and the marked target compensation vector value according to the regression loss function, and when the loss meets the convergence condition, a trained silicon wafer position compensation model is obtained.

[0071] Among them, after obtaining the original silicon wafer image, the grayscale value representation of the pixel points in the original silicon wafer image can be converted into a binary domain to obtain the grayscale value of the converted pixel points in the original silicon wafer image; and wherein, the method of inputting the original silicon wafer image into a pre-trained silicon wafer position compensation model can be to input the grayscale value of the converted pixel points in the original silicon wafer image into the pre-trained silicon wafer position compensation model.

[0072] It is worth mentioning that converting the grayscale value of the pixel points in the original silicon wafer image into a binary domain and performing silicon wafer position compensation processing based on the grayscale value of the binary domain of the pixel points in the original silicon wafer image can provide higher storage efficiency, transmission efficiency and processing efficiency. This conversion can make image data easier to store, transmit and process, and provides more convenience and flexibility for subsequent operations. Therefore, compared with the spatial domain and frequency domain, the efficiency of silicon wafer position compensation is improved.

[0073] Among them, after converting the grayscale value of the pixel in the original silicon wafer image into a binary domain, the grayscale value of the converted pixel in the original silicon wafer image can be saved to the memory, and the grayscale value of the converted pixel in the original silicon wafer image can be locked by a byte lock; and after determining the compensation vector value of the silicon wafer based on the center point position of the silicon wafer and the ideal center point position of the silicon wafer, the byte lock can be closed, and the compensation vector value can be converted from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model.

[0074] It is worth mentioning that the use of byte locks can avoid memory leaks and illegal operations, and improve the security and accuracy of silicon wafer position compensation.

[0075] Among them, the method of determining the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determining the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determining the ideal silicon wafer center point position of the original silicon wafer image, determining the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and converting the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model can be, performing a first task through a first thread, wherein the first task is a task corresponding to the silicon wafer contour information of the original silicon wafer image determined by the silicon wafer position compensation model, and performing a second task through a second thread, wherein the second task is a task corresponding to the silicon wafer contour information of the original silicon wafer image determined by the silicon wafer position compensation model. The first thread, the second thread, the third thread, the third thread, the fourth ...

[0076] Specifically, by utilizing multi-threaded parallel algorithms, multi-memory data can be processed in batches efficiently and quickly, with a speed of up to 100ms, almost real-time processing, and real-time mapping.

[0077] It is worth mentioning that the image enhancement processing can be performed in a multi-threaded parallel manner, which improves the efficiency of silicon chip position compensation.

[0078] Specifically, to facilitate understanding of the embodiments of the present application, a specific example is given below. Figure 3 , Figure 3 FIG. 1 is a flow chart of another silicon wafer position compensation method disclosed in an embodiment of the present application, comprising Figure 3 It can be seen that the original silicon wafer image can be obtained, the original silicon wafer image can be transferred to the binary domain and a byte lock can be added, the original silicon wafer image transferred to the binary domain can be input into the silicon wafer position compensation model, and various tasks of silicon wafer position compensation processing can be performed separately in a multi-threaded parallel manner. The tasks include using the Robert operator to calculate the maximum gradient, extracting the contour information in each image (that is, extracting the corresponding contour information in each original silicon wafer image), fitting the arc using the least squares of the contour information, calculating the center radius information based on the fitted arc, and calculating the rotation angle using the vector between multiple circle centers (compensation vector value). The byte lock can be closed and transferred to the imaging domain to obtain the position information and rotation angle of the silicon wafer.

[0079] It can be understood that the embodiments of the present application can efficiently, accurately and quickly perform silicon wafer center positioning and rotation angle identification, and is not affected by noise. The calculation time is relatively short, about 100ms for 5 pictures, and real-time processing of silicon wafer position angle correction can be achieved.

[0080] It is worth mentioning that the position information of the silicon wafer in the world coordinate system usually includes the position information of the center point and the angle information. The target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system. In addition to compensating the center point position of the silicon wafer, the rotation angle of the silicon wafer around its center point can also be compensated. The compensation vector value represents the offset and rotation of the silicon wafer. By adjusting the value of the compensation vector, the precise positioning and rotation control of the silicon wafer in the world coordinate system can be achieved.

[0081] It can be understood that in addition to the method described above of determining the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information; in addition to the method described above of determining the ideal silicon wafer center point position of the original silicon wafer image; in addition to the method described above of training the silicon wafer position compensation model; in addition to the method described above of inputting the original silicon wafer image into the pre-trained silicon wafer position compensation model; there may also be other methods, which are not limited here.

[0082] In this embodiment, the original silicon wafer image can be obtained, the original silicon wafer image is input into the pre-trained silicon wafer position compensation model, the silicon wafer contour information of the original silicon wafer image is determined by the silicon wafer position compensation model, the silicon wafer center point position of the original silicon wafer image is determined based on the silicon wafer contour information, the ideal silicon wafer center point position of the original silicon wafer image is determined, the compensation vector value of the silicon wafer is determined based on the silicon wafer center point position and the ideal silicon wafer center point position, and the compensation vector value is converted from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model, wherein the target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system. A silicon wafer position compensation method is provided, which can accurately determine the position deviation of the silicon wafer and calculate the compensation vector value, and the position of the silicon wafer can be corrected to the ideal position through the compensation vector value. Secondly, the compensation vector value represents the offset and rotation of the silicon wafer, and by adjusting the value of the compensation vector, the precise positioning and rotation control of the silicon wafer in the world coordinate system can be achieved. Next, the arc fitting of the silicon wafer contour information can be performed based on the least squares method to obtain an approximate arc, which can reflect the overall shape of the silicon wafer and improve the accuracy of determining the center point position of the silicon wafer. Furthermore, based on the boundary information of the groove and the tentacle, the ideal center position of the silicon wafer can be accurately obtained through a mathematical model, so that the compensation vector value can be determined, which improves the accuracy of determining the compensation vector value. Furthermore, the grayscale value of the pixel point in the original silicon wafer image can be converted into a binary domain, and the silicon wafer position compensation processing can be performed based on the grayscale value of the binary domain of the pixel point in the original silicon wafer image, which can provide higher storage efficiency, transmission efficiency and processing efficiency. This conversion can make image data easier to store, transmit and process, and provide more convenience and flexibility for subsequent operations. Therefore, compared with the spatial domain and frequency domain, the efficiency of silicon wafer position compensation is improved. Furthermore, byte locks can be used to avoid memory leaks and illegal operations, and improve the security and accuracy of silicon wafer position compensation. Furthermore, image enhancement processing can be performed in a multi-threaded parallel manner, which improves the efficiency of silicon wafer position compensation. Finally, the problems of compensating the deviation of the silicon wafer from the correct position center, compensating the rotation of the silicon wafer due to mechanical vibration, positioning the unclear edge of the silicon wafer, and the long positioning time of the silicon wafer were solved.

[0083] The above describes the silicon wafer position compensation method in the embodiment of the present application. The following describes the image processing device in the embodiment of the present application. Figure 4 , an embodiment of the image processing device in the embodiment of the present application includes:

[0084] An obtaining unit 401 is used to obtain an original silicon wafer image;

[0085] The silicon wafer position compensation unit 402 is used to input the original silicon wafer image into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determine the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determine the ideal silicon wafer center point position of the original silicon wafer image, determine the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and convert the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

[0086] In an embodiment of the present application, an original silicon wafer image can be obtained, and the original silicon wafer image can be input into a pre-trained silicon wafer position compensation model, and the silicon wafer contour information of the original silicon wafer image can be determined by the silicon wafer position compensation model, and the silicon wafer center point position of the original silicon wafer image can be determined based on the silicon wafer contour information, and the ideal silicon wafer center point position of the original silicon wafer image can be determined, and the compensation vector value of the silicon wafer can be determined based on the silicon wafer center point position and the ideal silicon wafer center point position, and the compensation vector value can be converted from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model, wherein the target compensation vector value is used to compensate the position of the silicon wafer in the world coordinate system. A silicon wafer position compensation method is provided, which can accurately determine the position deviation of the silicon wafer and calculate the compensation vector value, and the position of the silicon wafer can be corrected to the ideal position through the compensation vector value.

[0087] The following is a detailed description of the image processing device in the embodiment of the present application. Figure 5 Another embodiment of the image processing device in the embodiment of the present application includes:

[0088] An obtaining unit 501 is used to obtain an original silicon wafer image;

[0089] The silicon wafer position compensation unit 502 is used to input the original silicon wafer image into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determine the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determine the ideal silicon wafer center point position of the original silicon wafer image, determine the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and convert the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the image coordinate system.

[0090] The silicon wafer position compensation unit 502 is specifically used to perform arc fitting on the silicon wafer contour information based on the least squares method to obtain a first arc corresponding to the silicon wafer contour information, determine a first center position corresponding to the first arc, and use the first center position as the center point position of the silicon wafer.

[0091] The silicon wafer position compensation unit 502 is specifically used to determine the boundary information of the groove and / or the boundary information of the tentacle in the original silicon wafer image through an edge-finding algorithm, and based on the least squares method, perform arc fitting on the boundary information of the groove and / or the boundary information of the tentacle to obtain a second arc, determine the second center position corresponding to the second arc, and use the second center position as the ideal silicon wafer center point position, wherein the original silicon wafer image describes the information of the groove and / or the tentacle, and the groove and / or the tentacle are structures in the wafer suction cup for fixing the wafer to be inspected.

[0092] The image processing device further includes: an input unit 503, a calculation unit 504;

[0093] The obtaining unit 501 is further used to obtain an original silicon wafer image sample; wherein the original silicon wafer image sample is marked with a target compensation vector value;

[0094] The input unit 503 is used to input the original silicon wafer image sample into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information sample of the original silicon wafer image sample by the silicon wafer position compensation model, determine the silicon wafer center point position sample of the original silicon wafer image sample based on the silicon wafer contour information sample, determine the ideal silicon wafer center point position sample of the original silicon wafer image sample, determine the compensation vector value sample of the silicon wafer based on the silicon wafer center point position sample and the ideal silicon wafer center point position sample, and convert the compensation vector value sample from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the predicted target compensation vector value output by the silicon wafer position compensation model; wherein the predicted target compensation vector value is used to compensate the position of the silicon wafer in the image coordinate system;

[0095] The calculation unit 504 is used to calculate the loss between the predicted target compensation vector value and the marked target compensation vector value according to the regression loss function, and when the loss meets the convergence condition, a trained silicon wafer position compensation model is obtained.

[0096] The image processing device further comprises:

[0097] A conversion unit 505 is used to convert the grayscale value of the pixel point in the original silicon wafer image into a binary domain to obtain the grayscale value of the pixel point in the original silicon wafer image after the conversion;

[0098] The silicon wafer position compensation unit 502 is specifically configured to input the grayscale values ​​of the converted pixel points in the original silicon wafer image into a pre-trained silicon wafer position compensation model.

[0099] The image processing device further comprises:

[0100] A locking unit 506, used for saving the grayscale value of the converted pixel point in the original silicon wafer image to the memory, and locking the grayscale value of the converted pixel point in the original silicon wafer image through a byte lock;

[0101] The closing unit 507 is used to close the byte lock and execute the step of converting the compensation vector value from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model.

[0102] The silicon wafer position compensation unit 402 is specifically configured to execute a first task through a first thread, wherein the first task is a task corresponding to the silicon wafer contour information of the original silicon wafer image determined by the silicon wafer position compensation model; execute a second task through a second thread, wherein the second task is a task corresponding to the position of the silicon wafer center point of the original silicon wafer image determined by the silicon wafer position compensation model based on the silicon wafer contour information; execute a third task through a third thread, wherein the third task is a task corresponding to the position of the ideal silicon wafer center point of the original silicon wafer image determined by the silicon wafer position compensation model; and execute a fourth task through a fourth thread. Among them, the fourth task is a task corresponding to the compensation vector value of the silicon wafer determined by the silicon wafer position compensation model based on the silicon wafer center point position and the ideal silicon wafer center point position, so as to obtain the compensation vector value output by the silicon wafer position compensation model, and the fifth task is executed by the fifth thread, wherein the fifth task is a task corresponding to the target compensation vector value output by the silicon wafer position compensation model by converting the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix by the silicon wafer position compensation model, and the first thread, the second thread, the third thread, the fourth thread and the fifth thread are executed in parallel.

[0103] In this embodiment, each unit in the image processing device performs the above Figure 2 The operation of the image processing device in the illustrated embodiment will not be described in detail herein.

[0104] See below Figure 6 Another embodiment of the image processing device 600 in the embodiment of the present application includes:

[0105] CPU 601, memory 605, input / output interface 604, wired or wireless network interface 603 and power supply 602;

[0106] The memory 605 is a temporary storage memory or a permanent storage memory;

[0107] The CPU 601 is configured to communicate with the memory 605 and execute the instructions in the memory 605 to perform the aforementioned Figure 1 The method in the illustrated embodiment.

[0108] The embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium includes instructions, when the instructions are executed on a computer, the computer executes the aforementioned Figure 2 The method in the illustrated embodiment.

[0109] The present application also provides a computer program product including instructions, which, when executed on a computer, enables the computer to execute the aforementioned Figure 2 The method in the illustrated embodiment.

[0110] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0111] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0112] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0113] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0114] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0115] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk and other media that can store program code.

Claims

1. A silicon wafer position compensation method, characterized in that: include: Obtaining original silicon wafer images; The original silicon wafer image is input into a pre-trained silicon wafer position compensation model, the silicon wafer position compensation model determines the silicon wafer contour information of the original silicon wafer image, the silicon wafer center point position of the original silicon wafer image is determined based on the silicon wafer contour information, the ideal silicon wafer center point position of the original silicon wafer image is determined, the compensation vector value of the silicon wafer is determined based on the silicon wafer center point position and the ideal silicon wafer center point position, and the compensation vector value is converted from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

2. The method according to claim 1, characterized in that The step of determining the center point position of the silicon wafer of the original silicon wafer image based on the silicon wafer contour information includes: Performing arc fitting on the silicon wafer contour information based on the least square method to obtain a first arc corresponding to the silicon wafer contour information; A first center position corresponding to the first arc is determined, and the first center position is used as the center point position of the silicon wafer.

3. The method according to claim 1, characterized in that The original silicon wafer image describes information about grooves and / or information about tentacles; wherein the grooves and / or the tentacles are structures in a wafer suction cup for fixing the wafer to be inspected; The step of determining the ideal silicon wafer center point position of the original silicon wafer image comprises: Determining the boundary information of the groove and / or the boundary information of the tentacle in the original silicon wafer image by an edge-finding algorithm; Based on the least square method, arc fitting is performed on the boundary information of the groove and / or the boundary information of the tentacle to obtain a second arc; A second center position corresponding to the second arc is determined, and the second center position is used as the ideal silicon wafer center point position.

4. The method according to claim 1, characterized in that: Before inputting the original silicon wafer image into the pre-trained silicon wafer position compensation model, the method further includes: Obtaining an original silicon wafer image sample; wherein the original silicon wafer image sample is marked with a target compensation vector value; Input the original silicon wafer image samples into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information samples of the original silicon wafer image samples by the silicon wafer position compensation model, determine the silicon wafer center point position samples of the original silicon wafer image samples based on the silicon wafer contour information samples, determine the ideal silicon wafer center point position samples of the original silicon wafer image samples, determine the compensation vector value samples of the silicon wafer based on the silicon wafer center point position samples and the ideal silicon wafer center point position samples, and transform the compensation vector value samples from the image coordinate system to the world coordinate system based on the transformation matrix to obtain the predicted target compensation vector value output by the silicon wafer position compensation model; wherein the predicted target compensation vector value is used to compensate the position of the silicon wafer in the image coordinate system; The loss between the predicted target compensation vector value and the marked target compensation vector value is calculated according to the regression loss function. When the loss meets the convergence condition, a trained silicon wafer position compensation model is obtained.

5. The method according to claim 1, characterized in that After obtaining the original silicon wafer image, the method further comprises: Converting the grayscale value of the pixel in the original silicon wafer image into a binary domain to obtain the grayscale value of the pixel in the original silicon wafer image after the conversion; The step of inputting the original silicon wafer image into a pre-trained silicon wafer position compensation model comprises: The grayscale values ​​of the converted representation of the pixels in the original silicon wafer image are input into a pre-trained silicon wafer position compensation model.

6. The method according to claim 5, characterized in that After converting the grayscale value of the pixel in the original silicon wafer image into a binary domain, the method further includes: The grayscale values ​​of the converted pixel points in the original silicon wafer image are saved in the memory, and the grayscale values ​​of the converted pixel points in the original silicon wafer image are locked by byte lock; After determining the compensation vector value of the silicon wafer based on the center point position of the silicon wafer and the ideal center point position of the silicon wafer, the method further includes: The byte lock is closed, and the step of converting the compensation vector value from the image coordinate system to the world coordinate system based on the conversion matrix is ​​performed to obtain the target compensation vector value output by the silicon wafer position compensation model.

7. The method according to claim 1, characterized in that The method of determining the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determining the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determining the ideal silicon wafer center point position of the original silicon wafer image, determining the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and converting the compensation vector value from the image coordinate system to the world coordinate system based on the conversion matrix to obtain the target compensation vector value output by the silicon wafer position compensation model includes: Execute a first task through a first thread; wherein the first task is a task corresponding to the silicon wafer contour information of the original silicon wafer image determined by the silicon wafer position compensation model; Execute a second task through a second thread; wherein the second task is a task corresponding to the position of the center point of the silicon wafer of the original silicon wafer image determined by the silicon wafer position compensation model based on the silicon wafer contour information; Executing a third task through a third thread; wherein the third task is a task corresponding to the ideal silicon wafer center point position of the original silicon wafer image determined by the silicon wafer position compensation model; Executing a fourth task through a fourth thread; wherein the fourth task is to determine a compensation vector value of a silicon wafer based on the center point position of the silicon wafer and the center point position of the ideal silicon wafer by the silicon wafer position compensation model, so as to obtain a task corresponding to the compensation vector value output by the silicon wafer position compensation model; The fifth task is executed through the fifth thread; wherein the fifth task is a task corresponding to the target compensation vector value output by the silicon wafer position compensation model by converting the compensation vector value from the image coordinate system to the world coordinate system based on the transformation matrix; and the first thread, the second thread, the third thread, the fourth thread and the fifth thread are executed in parallel.

8. An image processing device, characterized in that: include: An acquisition unit, used for acquiring an original silicon wafer image; A silicon wafer position compensation unit is used to input the original silicon wafer image into a pre-trained silicon wafer position compensation model, determine the silicon wafer contour information of the original silicon wafer image by the silicon wafer position compensation model, determine the silicon wafer center point position of the original silicon wafer image based on the silicon wafer contour information, determine the ideal silicon wafer center point position of the original silicon wafer image, determine the compensation vector value of the silicon wafer based on the silicon wafer center point position and the ideal silicon wafer center point position, and convert the compensation vector value from the image coordinate system to the world coordinate system based on a conversion matrix to obtain a target compensation vector value output by the silicon wafer position compensation model; wherein the target compensation vector value is used to compensate for the position of the silicon wafer in the world coordinate system.

9. An image processing device, characterized in that: include: CPU and memory; The memory is a short-term storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute instructions in the memory to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium comprises instructions, and when the instructions are executed on a computer, the computer is caused to perform the method according to any one of claims 1 to 7.

Images