Wafer appearance defect detection method, device and equipment

Abstract

The invention discloses a wafer appearance defect detection method, device and equipment. The method comprises the following steps: determining a starting column center coordinate and an ending column center coordinate of each row of crystal grains of a target wafer; traversing each row of crystal grains according to the center coordinates of the starting column and the center coordinates of the ending column, and controlling a height measurement sensor to measure the height of each crystal grain in the traversing process to obtain a height data set of each crystal grain; according to the height data set of each crystal grain, determining height compensation data of the corresponding crystal grain; performing path planning on each row of crystal grains to obtain a moving path and height compensation data, and controlling a target camera to perform relative movement, height adjustment and photographing to obtain a target image set; and performing defect identification on the target image set based on a pre-trained difference model and an AI model to obtain a first detection result. According to the invention, the accuracy and efficiency of finished wafer appearance defect detection can be improved.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and particularly to a method, device and equipment for detecting appearance defects of wafers. Background Art

[0002] In the process of semiconductor product production, the appearance defects of the finished wafers are detected to ensure the product quality. In the related art, most of the AOI (Automated Optical Inspection) equipment is used for appearance defect detection. However, the existing AOI equipment still needs to be improved in terms of detection accuracy and detection efficiency. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art. For this purpose, the present invention provides a method, device and equipment for detecting appearance defects of wafers, which can improve the accuracy and efficiency of detecting appearance defects of finished wafers.

[0004] On the one hand, an embodiment of the present invention provides a method for detecting appearance defects of wafers, including: Determining the starting column center coordinates and the ending column center coordinates of the grains in each row of the target wafer; Traversing each row of grains according to the starting column center coordinates and the ending column center coordinates, and controlling a height measurement sensor to measure the height of each grain during the traversing process to obtain a height data set of each grain; Determining height compensation data corresponding to each grain according to the height data set of each grain; Controlling a target camera to perform relative movement, height adjustment and photographing according to the movement path obtained by path planning for each row of grains and the height compensation data to obtain a target image set; Performing defect recognition on the target image set based on a pre-trained difference model and an AI model to obtain a first detection result.

[0005] According to some embodiments of the present invention, before determining the starting column center coordinates and the ending column center coordinates of the grains in each row of the target wafer, it further includes: Obtaining the Map coordinates and the theoretical mechanical coordinates of the preset marks of the target wafer; Adjusting the field of view center of the target camera to align with the center of the preset mark according to the theoretical mechanical coordinates to determine the actual mechanical coordinates of the preset mark; Determining a coordinate system mapping transformation matrix according to the Map coordinates and the actual mechanical coordinates of the preset mark.

[0006] According to some embodiments of the present invention, determining the starting column center coordinates and the ending column center coordinates of the grains in each row of the target wafer includes: Determine the first starting column center coordinate and the first ending column center coordinate of the Map graph of the grains in each row of the target wafer; The traversing of the grains in each row according to the starting column center coordinate and the ending column center coordinate includes: According to the coordinate system mapping transformation matrix, convert the first starting column center coordinate and the first ending column center coordinate into the second starting column center coordinate and the second ending column center coordinate in the mechanical coordinate system; Traverse the grains in each row according to the second starting column center coordinate and the second ending column center coordinate.

[0007] According to some embodiments of the present invention, the determining of the height compensation data corresponding to each grain according to the height data set of each grain includes: Sort the data in the height data set of each grain and evenly divide it into N parts to obtain the first height data subset, where N is a positive integer greater than 2; Remove the data of the first 1 / N and the last 1 / N of the first height data subset to obtain the second height data subset; Remove the high-frequency data of the second height data subset and calculate the average value of the remaining data to obtain the height compensation data corresponding to the grain.

[0008] According to some embodiments of the present invention, the defect recognition of the target image set based on the pre-trained difference model and the AI model to obtain the first detection result includes: Obtain the first target image of the target grain from the target image set; Based on a preset ROI region, perform region recognition on the first target image and perform Blob analysis to intercept the second target image corresponding to the region where the target grain is located; Based on a preset detection template, perform alignment adjustment and identify the region to be detected on the second target image to obtain the third target image; Based on the pre-trained difference model, perform defect recognition on the third target image to determine the first defect region; Based on the first defect region, perform a local screenshot of the third target image to obtain a defect image; Based on the pre-trained defect classification model, classify the defect image to obtain the first detection result.

[0009] According to some embodiments of the present invention, before the performing of region recognition on the first target image based on a preset ROI region and performing Blob analysis to intercept the second target image corresponding to the region where the target grain is located, it further includes: Based on a preset ROI region, perform region recognition on the first target image to obtain the fourth target image; Perform sharpness detection on the fourth target image, and output a second detection result if the sharpness detection fails.

[0010] According to some embodiments of the present invention, the third target image includes a photosensitive area image and a wire area image. The defect recognition of the third target image based on the pre-trained difference model to determine the first defect area includes: Perform defect recognition on the photosensitive area image based on the pre-trained first difference model to obtain a photosensitive defect area; Perform defect recognition on the wire area image based on the pre-trained second difference model to obtain a wire defect area; Merge the photosensitive defect area and the wire defect area to obtain the first defect area.

[0011] According to some embodiments of the present invention, the local screenshot of the third target image based on the first defect area to obtain a defect image includes: Perform regional range expansion based on the first defect area to obtain a second defect area; Perform a local screenshot of the third target image according to the second defect area to obtain a defect image.

[0012] On the other hand, an embodiment of the present invention provides a wafer appearance defect detection device, including: A first determination module for determining the starting column center coordinates and the ending column center coordinates of each row of grains of the target wafer; A height measurement module for traversing each row of grains according to the starting column center coordinates and the ending column center coordinates, and controlling a height measurement sensor to measure the height of each grain during the traversal to obtain a height data set of each grain; A second determination module for determining the height compensation data of the corresponding grain according to the height data set of each grain; An image acquisition module for controlling the target camera to perform relative movement, height adjustment and photographing according to the movement path obtained by path planning for each row of grains and the height compensation data to obtain a target image set; A defect detection module for performing defect recognition on the target image set based on a pre-trained difference model and an AI model to obtain a first detection result.

[0013] In yet another aspect, an embodiment of the present invention provides a wafer appearance defect detection device, including a processor and a memory. A computer program is stored in the memory, and when the processor runs the computer program, it is used to implement the wafer appearance defect detection method as described above.

[0014] The embodiments of the present invention have at least the following beneficial effects: By traversing the target wafer row by row, height measurement of each die and determination of height compensation data are achieved, and the movement, height adjustment, and photographing of the target camera are controlled through path planning and height compensation, which can optimize the movement path of the target camera and ensure clear images obtained by photographing, and defect identification of the target image set is performed based on a pre-trained difference model and an AI model, so that the accuracy and efficiency of appearance defect detection of the finished wafer can be improved.

[0015] The additional aspects and advantages of the present invention will be partly given in the following description, partly will become obvious from the following description, or will be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and / or additional aspects and advantages of the present invention will become obvious and easy to understand from the description of the embodiments in conjunction with the following drawings, in which: Figure 1 is a schematic diagram of a method for detecting appearance defects of a wafer according to an embodiment of the present invention; Figure 2 is a schematic structural diagram of a detection device according to an embodiment of the present invention; Figure 3 is a partial structural diagram of a target wafer according to an embodiment of the present invention; Figure 4 is a schematic diagram of path planning for a single row of dies according to an embodiment of the present invention; Figure 5 is a schematic diagram of a target wafer according to an embodiment of the present invention; Figure 6 is Figure 5 a partial enlarged view of the area where the mark M1 is located in; Figure 7 is a schematic diagram of equal division of a height measurement data set of a target die according to an embodiment of the present invention; Figure 8 is a schematic diagram of a detection template according to an embodiment of the present invention; Figure 9 is a partial enlarged view of a target wafer according to an embodiment of the present invention; Figure 10 is a schematic diagram of a photosensitive area image and a wire area image of a target wafer according to an embodiment of the present invention; Figure 11 is a template image of a first difference model and a photosensitive area image with defects according to an embodiment of the present invention; Figure 12 is a template image of a second difference model and a wire area image with defects according to an embodiment of the present invention; Figure 13 is a schematic diagram of various defect types according to an embodiment of the present invention; Figure 14 It is a principle block diagram of a wafer appearance defect detection device according to an embodiment of the present invention; Figure 15 It is a principle block diagram of a wafer appearance defect detection device according to an embodiment of the present invention.

[0017] Reference numerals: XY stage 10, Z-axis lifting mechanism 20, industrial camera 21, telecentric lens 22, point light source 23, height measurement sensor 24, target wafer 210, photosensitive area 211, wire area 212, first determination module 310, height measurement module 320, second determination module 330, image acquisition module 340, defect detection module 350, processor 410, memory 420. Detailed implementation manners

[0018] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as a limitation to the present invention.

[0019] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by terms such as up, down, front, back, left, right, etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0020] In the description of the present invention, the meaning of "several" is one or more, the meaning of "multiple" is two or more, greater than, less than, exceeding, etc. are understood as not including the number itself, and "above", "below", "within", etc. are understood as including the number itself. If there is a description of "first", "second", etc., it is only for the purpose of distinguishing technical features and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or the sequence relationship of the indicated technical features.

[0021] In the description of the present invention, unless otherwise clearly defined, terms such as "set", "installed", "connected", etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above terms in the present invention in combination with the specific content of the technical solution.

[0022] Please refer to Figure 1, this embodiment discloses a method for detecting appearance defects of a wafer, including steps S100 to S500. It should be noted that the step numbers in this embodiment are only for the convenience of review and understanding, rather than limiting the execution order of the steps. Before describing the content of each step, a brief description of the hardware structure of the detection equipment involved in this detection method is given. Among them, please refer to Figure 2 , the detection equipment includes an XY stage 10 and a Z-axis lifting mechanism 20. The XY stage 10 is used to support the target wafer 210 (i.e., the wafer to be detected), and can move bidirectionally along the X direction and the Y direction in the XY plane. The Z-axis lifting mechanism 20 is equipped with an industrial camera 21 and a height measurement sensor 24. The Z-axis lifting mechanism 20 is located above the XY stage 10 and can perform linear motion along the Z direction. The industrial camera 21 is connected with a telecentric lens 22 and a point light source 23. The industrial camera 21 is used for image acquisition. The height measurement sensor 24 uses a spectral confocal height measurement sensor 24, with a height measurement interval of 10 μm and an overlapping accuracy of 0.8 μm, and precisely measures the height data of different positions on the surface of the die of the target wafer 210.

[0023] The details of each step are described below: S100. Determine the starting column center coordinates and ending column center coordinates of each row of dies on the target wafer 210; Exemplarily, please refer to Figure 3 , Figure 3 shows a partial structure of the target wafer 210. A plurality of dies are arranged in an array distribution in the horizontal and vertical directions in the target wafer 210. As shown by the label Die in the figure, each die has a certain length and width. Taking the first complete die in each row as the starting die and the last complete die in each row as the ending die, then taking the outer edge of the starting die as the starting column and the outer edge of the ending die as the ending column, so that the starting column center coordinates and ending column center coordinates can be determined. Among them, the starting die can be arranged on the left or on the right. If the starting die is arranged on the left, the outer edge of the starting die refers to the left edge of the starting die, and the outer edge of the ending die refers to the right edge of the ending die; if the starting die is arranged on the right, the outer edge of the starting die refers to the right edge of the starting die, and the outer edge of the ending die refers to the left edge of the ending die. For example, Figure 3 point A in shows the center of the starting column, point B shows the center of the ending column, and the line L between point A and point B is the scanning path for height measurement.

[0024] S200. Traverse each row of dies according to the starting column center coordinates and ending column center coordinates, and control the height measurement sensor 24 to measure the height of each die during the traversal process to obtain the height data set of each die; Exemplarily, the relative positions of the target wafer 210 and the height sensor 24 are adjusted according to the starting column center coordinates and the ending column center coordinates, so that the height sensor 24 can move along the scanning path, thereby traversing each row of die, and measuring the height of each die during the traversal to obtain a height data set for each die, where the height data set for each die contains height data at multiple positions on the die surface. It should be noted that the target wafer 210 in this embodiment is placed on the XY stage 10, and the XY stage 10 can drive the target wafer 210 to move relative to the height sensor 24, so as to realize the traversal of each die by the height sensor 24; in some application examples, for example, the height sensor 24 is installed on the XY moving mechanism, and the height sensor 24 can move along the X direction and the Y direction under the drive of the XY moving mechanism, so as to traverse each die in each row.

[0025] S300. Determine the height compensation data for the corresponding die according to the height data set of each die; Exemplarily, during the production process, the target wafer 210 may have the problem of uneven surface. Determining the height compensation data for the corresponding die according to the height data set of each die is convenient for subsequent photo compensation for each die, solving the problem of out-of-focus photo caused by the uneven surface of the target wafer 210 and the temperature drift of the target camera lens, which is beneficial to improving the clarity of die image acquisition and further improving the accuracy of image recognition.

[0026] S400. Control the target camera to perform relative movement, height adjustment and photo taking according to the movement path obtained by path planning for each row of die and the height compensation data to obtain a target image set; Exemplarily, please refer to Figure 4 , Figure 4 shows a schematic diagram of the photo taking path and the running direction of a row of die on the target wafer 210. In the figure, the mark Die represents a die, where the photo taking path and the running direction are as shown by the arrow in the figure, and the starting point of the arrow is the photo taking position. Planning the photo taking path is beneficial to improving the photo taking speed, reducing the empty running time of the axis, and improving the photo taking efficiency. It should be noted that in this embodiment, controlling the relative movement of the target camera (i.e., the industrial camera 21 above) can be achieved by driving the target wafer 210 to move relative to the target camera in the XY plane by the XY stage 10, or, for the target camera installed on the XY moving mechanism, the target camera can move along the X direction and the Y direction under the drive of the XY moving mechanism, so as to realize the relative movement. During the movement, adjusting the height of the target camera according to the height compensation data of each die can ensure that the target camera can obtain the clearest image during photo taking, which is beneficial to improving the accuracy of image recognition.

[0027] S500. Use a pre-trained differential model and an AI model to identify defects in the target image set, and obtain the first detection result.

[0028] Exemplarily, using a pre-trained differential model and an AI model to identify defects in the target image set can significantly reduce the labor intensity, improve the detection efficiency, accuracy, and stability compared with manual detection. Among them, the AI (Artificial Intelligence) model can be trained using existing deep neural networks or fine-tuned based on large models, such as the defect classification model below.

[0029] Through the above solution, by traversing the target wafer 210 row by row, height measurement of each die and determination of height compensation data are realized, and the movement, height adjustment, and photographing of the target camera are controlled through path planning and height compensation, which can optimize the movement path of the target camera and ensure the clarity of the captured images. Additionally, using a pre-trained differential model and an AI model to identify defects in the target image set can improve the accuracy and efficiency of the appearance defect detection of the finished wafer.

[0030] In some application examples, before step S100, it also includes: Obtain the Map coordinates and theoretical mechanical coordinates of the preset marks on the target wafer 210; According to the theoretical mechanical coordinates, adjust the center of the field of view of the target camera to align with the center of the preset mark, and determine the actual mechanical coordinates of the preset mark; According to the Map coordinates and actual mechanical coordinates of the preset mark, determine the coordinate system mapping transformation matrix.

[0031] Exemplarily, the Map is a visualization graph used to represent the positions and states of each die on the wafer during the semiconductor manufacturing process. The Map coordinates record the relative positions of the dies on the target wafer 210, while the mechanical coordinates record the relative positions of the dies with respect to the coordinate origin of the detection device (such as the XY stage 10) after the target wafer 210 is placed on the detection device. To facilitate the positioning of each die on the target wafer 210, some features of the target wafer 210 are selected as alignment marks, i.e., preset marks, before production. For example, the edge of the target wafer 210 is used as the preset mark, as shown by marks M1, M2, and M3 in Figure 5 The local enlarged view of the area where mark M1 is located can be referred to in Figure 6 , Figure 6 where mark M1-C represents the center of mark M1. Through edge mark positioning, no additional alignment marks need to be set, the target wafer 210 will not be damaged, and the compatibility is strong.

[0032] During the actual production process, there may be a certain positional deviation between the actual placement position and the theoretical placement position of the target wafer 210, resulting in a reduction in the positioning accuracy of the theoretical mechanical coordinates of the target wafer 210. Therefore, it is necessary to re-calibrate the mechanical coordinates of the target wafer 210. According to the theoretical coordinates of the preset mark, adjust the center of the field of view of the target camera to align with the center of the preset mark (such as Figure 6 shown by the mark M1-C in

[0033] ). For example, according to the theoretical coordinates of the preset mark, by adjusting the relative position between the target wafer 210 and the target camera, make the preset mark enter the field of view of the target camera, and then fine-tune by detecting the distance between the preset mark and the center of the field of view of the target camera, so as to achieve the alignment of the center of the field of view of the target camera and the center of the preset mark. At this time, the mechanical coordinates of the preset mark are the actual mechanical coordinates. Among them, the relative movement between the target wafer 210 and the target camera can be realized by driving the target wafer 210 to move relative to the target camera in the XY plane by the XY stage 10. Or, for the target camera installed on the XY moving mechanism, the target camera can move along the X direction and the Y direction under the drive of the XY moving mechanism, so as to achieve relative movement. According to the Map coordinates and the actual mechanical coordinates of the preset mark, the coordinate mapping transformation matrix between the Map coordinate system and the mechanical coordinate system can be calculated. Since the Map coordinates corresponding to each die on the target wafer 210 are known, according to the coordinate mapping transformation matrix, the Map coordinates of any point on the target wafer 210 can be converted into mechanical coordinates, which is convenient for calculation and motion control.

[0034] Correspondingly, step S100 includes: determining the first starting column center coordinate and the first ending column center coordinate of the Map of each row of dies on the target wafer 210; In step S200, traversing each row of dies according to the starting column center coordinate and the ending column center coordinate includes: According to the coordinate mapping transformation matrix, convert the first starting column center coordinate and the first ending column center coordinate into the second starting column center coordinate and the second ending column center coordinate in the mechanical coordinate system; Traverse each row of dies according to the second starting column center coordinate and the second ending column center coordinate.

[0035] Exemplarily, the Map records the Map coordinates of each die on the target wafer 210. By querying the Map, it is convenient to determine the first starting column center coordinate and the first ending column center coordinate of each row of dies on the target wafer 210. At this time, the determined center coordinates are all Map coordinates. Since the Map coordinates are relative coordinates, it is not convenient for calculation and motion control. Coordinate conversion is performed according to the coordinate mapping transformation matrix, that is, the Map coordinates are converted into mechanical coordinates, so as to facilitate calculation and motion control.

[0036] The photographing height compensation of the target camera has a great influence on the clarity of the photographed image. In order to improve the clarity of the photographed image, step S300 includes: Sort the data of the height data set of each grain and divide it into N equal parts to obtain a first height data subset, where N is a positive integer greater than 2; Remove the first 1 / N and the last 1 / N of the data in the first height data subset to obtain a second height data subset; Remove the high-frequency data in the second height data subset and calculate the average value of the remaining data to obtain the height compensation data corresponding to the grain.

[0037] Exemplarily, the height measurement interval of the height measurement sensor 24 is 10 μm, which is smaller than the surface area of a single grain. When measuring the height of a single grain, the height data of different positions on the grain surface can be measured. Please refer to Figure 7 ., sort the height data of a single grain and divide it into N equal parts, for example, 6 parts, remove the first 1 / 6 and the last 1 / 6 of the height data (i.e., invalid data), and retain the middle 4 / 6 of the height data (i.e., valid data) to obtain data with higher accuracy. Then remove the high-frequency data therein to obtain the remaining data, and calculate the average value of the remaining data to obtain the height compensation data of the grain.

[0038] Automatic defect detection by means of a differential model and an AI model can greatly improve the detection efficiency and accuracy. Among them, step S500 includes: Obtain a first target image of a target grain from the target image set; Perform region recognition on the first target image based on a preset ROI region and perform Blob analysis to intercept a second target image corresponding to the region where the target grain is located; Based on a preset detection template, align and adjust the second target image and identify the region to be detected to obtain a third target image; Perform defect recognition on the third target image based on a pre-trained differential model to determine a first defect region; Perform a partial screenshot of the third target image based on the first defect region to obtain a defect image; Classify the defect image based on a pre-trained defect classification model to obtain a first detection result.

[0039] Exemplarily, the target image set contains images of all the dies on the target wafer 210, and defect identification is performed on each die one by one. The first target image of the target die is obtained from the target image set, and then the area where the target die is located is identified from the first target image, and a local image of the area is intercepted for precise detection. In computer vision, a Blob refers to a connected area in an image. Blob analysis is to perform connected component extraction and labeling on a binary image after foreground / background separation. For example, the content of Blob analysis includes preprocessing such as grayscale threshold, length, width, aspect ratio, opening operation, and closing operation. During the photographing process, there may be a certain deviation between the placement pose of the target wafer 210 and the theoretical pose, resulting in a certain deviation between the contour and angle of the image and the preset detection template. Therefore, it is necessary to perform alignment adjustment on the second target image and identify the area to be detected. Based on the pre-trained difference model, defect identification is performed on the third target image, and it can be determined whether there are defects in the third target image and the area where the defects are located. According to the area where the defects are located, a local screenshot of the third target image is taken, and a local image containing the defects can be accurately obtained, reducing the interference features in the defect image, which is beneficial for the defect classification model to accurately identify and classify the defect types, thereby obtaining an accurate first detection result.

[0040] Among them, the size of the preset detection template is the same as the size of the first target image, and it is configured with multiple detection areas. The detection template is as Figure 8 shown. The squares in the figure represent the detection areas. A positioning mark is set at the center of the detection template to facilitate obtaining the coordinate data and angle data of the image. During detection, the positioning mark at the center of the detection template is aligned with the center of the current image to obtain the position of the current detection area, and the current image is rotated and translated so that the current image completely coincides with the detection template, thereby achieving alignment adjustment. The area to be detected can be identified through Blob analysis.

[0041] In some application examples, before performing region identification on the first target image based on the preset ROI region and performing Blob analysis to intercept the second target image corresponding to the area where the target die is located, it also includes: Performing region identification on the first target image based on the preset ROI region to obtain a fourth target image; Performing clarity detection on the fourth target image, and outputting a second detection result in the case where the clarity detection fails.

[0042] For example, ROI refers to a region of interest (ROI), which selects a specific area from an image to reduce interference from irrelevant information. Regional recognition is performed on the first target image based on the ROI region, focusing on the area to be analyzed to obtain a fourth target image. The fourth target image is then subjected to clarity testing, for example, by using a Laplace image processing algorithm to calculate the clarity of the fourth target image. If the clarity test fails, a second test result is output. For example, if the target die is defective, if the clarity test passes, regional recognition is performed on the first target image based on the preset ROI region, and blob analysis is performed to capture a second target image corresponding to the area where the target die is located. This ensures that the clarity of the image to be tested meets requirements and avoids misjudgments due to unclear images.

[0043] Please refer to Figure 9 and Figure 10 The target wafer 210 includes a photosensitive area 211 and a wire area 212. The third target image includes a photosensitive area image and a wire area image. The photosensitive area image is as follows: Figure 10 As shown in (a), the wire area image is as follows Figure 10 As shown in (b), defect recognition is performed on the third target image based on the pre-trained differential model to determine the first defect area, including: Defect recognition is performed on the photosensitive area image based on the pre-trained first difference model to obtain a photosensitive defect area; For example, the first differential model needs to be pre-trained before detection. For example, a defect-free photosensitive area image is selected, and a differential template with the same size as the first target image and the entire detection area is prepared for model training to obtain the first differential model. Defect recognition is performed on the photosensitive area image based on the first differential model, wherein the template image of the first differential model is as follows: Figure 11 As shown in (a), the image of the defective photosensitive area is as follows Figure 11 As shown in (b), the defect recognition conditions are used for screening to obtain the length, width and area of the defect area, and then compared with the preset parameter threshold. When the parameter exceeds the standard, the corresponding area is marked as a photosensitive defect area, as shown by the mark Q1.

[0044] Defect recognition is performed on the wire area image based on the pre-trained second difference model to obtain the wire defect area; For example, the training and recognition principle of the second differential model is the same as that of the first differential model, except that the training image of the second differential model is an image of a non-defective wire area. After the defect recognition of the second differential model, the wire defect area can be obtained. For example, the template image of the second differential model is as follows: Figure 12 As shown in (a), the image of the defective wire area is as follows Figure 12As shown in (b), the wire defect area is shown as the marked Q2. By using the first difference model and the second difference model to identify defects in the photosensitive area 211 and the wire area 212 respectively, accurate identification can be carried out according to the different characteristics of the photosensitive area 211 and the wire area 212, which is beneficial to improving the accuracy of identification.

[0045] Merge the photosensitive defect area and the wire defect area to obtain the first defect area.

[0046] Exemplarily, the types of defects include various types such as scratches, color abnormalities, and dirt. In most cases, the location of the defect will not appear at a single location. For example, it will not only appear in the photosensitive area 211 or only in the wire area 212. That is to say, the defect may appear in the adjacent photosensitive area 211 and the wire area 212 at the same time, or may span multiple photosensitive areas 211 and wire areas 212 at the same time. The defects identified by the first difference model and the second difference model are discrete defect areas. For example, the photosensitive defect area marks the defects in the photosensitive area 211 and will not mark the defects in the wire area 212. If the defect spans multiple photosensitive areas 211 and wire areas 212, only the corresponding photosensitive area 211 will be marked in the photosensitive defect area. Therefore, it is necessary to merge the discrete photosensitive defect areas and wire defect areas to obtain the complete defect area, that is, the first defect area.

[0047] After obtaining the first defect area, it can be determined whether there are defects. If there are, the local image where the defect is located is intercepted for defect classification. Among them, based on the first defect area, a local screenshot of the third target image is taken to obtain a defect image, including: Based on the first defect area, the area range is expanded to obtain the second defect area; According to the second defect area, a local screenshot of the third target image is taken to obtain a defect image.

[0048] Exemplarily, the first defect area is an accurate area containing the defect. The field of view range of the image may be relatively narrow and the contained feature content is relatively small. Based on the first defect area, the area range is expanded. For example, the length and width of each area are each expanded by 20 pixels to obtain the second defect area with a wider field of view range and relatively more contained feature content. For example, the boundary of the defect is clearer, so as to facilitate the identification and classification of the defect type by the defect classification model. According to the second defect area, a local screenshot of the third target image is taken to obtain a defect image, as Figure 13 shows various types of defect images.

[0049] In this embodiment, defect detection is respectively performed on the photosensitive area 211 and the wire area 212, which can adapt to the characteristics of different areas, improve the accuracy of detection, merge the detected photosensitive defect area and wire defect area, then perform regional range expansion to obtain a second defect area with a wider field of view, and perform local screenshot according to the second defect area, so as to focus on the local image containing defect features, and then perform defect classification based on the defect classification model. In this way, the detection and classification of defects are carried out step by step, which is beneficial to taking into account the accuracy and efficiency of defect recognition.

[0050] Please refer to Figure 14 , an embodiment of the present invention also provides a wafer appearance defect detection device, including: A first determination module 310, configured to determine the starting column center coordinates and ending column center coordinates of each row of grains of the target wafer 210; A height measurement module 320, configured to traverse each row of grains according to the starting column center coordinates and ending column center coordinates, and control the height measurement sensor 24 to measure the height of each grain during the traversal process to obtain a height data set of each grain; A second determination module 330, configured to determine the height compensation data corresponding to each grain according to the height data set of each grain; An image acquisition module 340, configured to control the target camera to perform relative movement, height adjustment and photographing according to the movement path obtained by path planning for each row of grains and the height compensation data to obtain a target image set; A defect detection module 350, configured to perform defect recognition on the target image set based on a pre-trained differential model and an AI model to obtain a first detection result.

[0051] The inventive concept of this embodiment of the wafer appearance defect detection device is the same as that of the above embodiment of the wafer appearance defect detection method. The content not involved in this embodiment of the wafer appearance defect detection device can be referred to the above embodiment of the wafer appearance defect detection method, and will not be elaborated here. By traversing the target wafer 210 row by row, the height of each grain is measured and the height compensation data is determined, and the movement of the target camera, height adjustment and photographing are controlled through path planning and height compensation, which can optimize the movement path of the target camera and ensure the clarity of the photographed image, and perform defect recognition on the target image set based on a pre-trained differential model and an AI model. In this way, the accuracy and efficiency of the finished wafer appearance defect detection can be improved.

[0052] Please refer to Figure 15, an embodiment of the present invention further provides a wafer appearance defect detection device, including a processor 410 and a memory 420. A computer program is stored in the memory 420. When the processor 410 runs the computer program, it is used to implement the wafer appearance defect detection method as described above. The content of the wafer appearance defect detection method can be referred to the above text and will not be elaborated here. By traversing the target wafer 210 row by row, height measurement of each die and determination of height compensation data are achieved, and the movement, height adjustment and photographing of the target camera are controlled through path planning and height compensation, which can optimize the movement path of the target camera and ensure the clarity of the captured image, and defect recognition of the target image set is performed based on the pre-trained difference model and AI model, so as to improve the accuracy and efficiency of the finished wafer appearance defect detection.

[0053] The embodiments of the present invention have been described in detail above with reference to the drawings. However, the present invention is not limited to the above embodiments. Within the knowledge scope of those of ordinary skill in the art to which the present invention pertains, various changes can be made without departing from the purpose of the present invention.

Claims

1. A method for detecting appearance defects of a wafer, characterized in that, Including: Determine the starting column center coordinates and ending column center coordinates of the grains in each row of the target wafer; Traverse the grains in each row according to the starting column center coordinates and the ending column center coordinates, and control the height measurement sensor to measure the height of each grain during the traversal process to obtain the height data set of each grain; Determine the height compensation data of the corresponding grain according to the height data set of each grain; According to the movement path obtained by path planning for each row of grains and the height compensation data, control the target camera to perform relative movement, height adjustment and take pictures to obtain the target image set; Based on the pre-trained differential model and AI model, perform defect recognition on the target image set to obtain the first detection result.

2. The wafer appearance defect detection method according to claim 1, wherein, Before determining the starting column center coordinates and ending column center coordinates of the grains in each row of the target wafer, it further includes: Obtain the Map coordinates and theoretical mechanical coordinates of the preset mark of the target wafer; According to the theoretical mechanical coordinates, adjust the center of the field of view of the target camera to align with the center of the preset mark, and determine the actual mechanical coordinates of the preset mark; Determine the coordinate system mapping transformation matrix according to the Map coordinates and the actual mechanical coordinates of the preset mark.

3. The wafer appearance defect detection method according to claim 2, wherein Determining the starting column center coordinates and ending column center coordinates of the grains in each row of the target wafer includes: Determine the first starting column center coordinate and the first ending column center coordinate of the Map of the grains in each row of the target wafer; The traversing each row of grains according to the starting column center coordinates and the ending column center coordinates includes: According to the coordinate system mapping transformation matrix, convert the first starting column center coordinate and the first ending column center coordinate into the second starting column center coordinate and the second ending column center coordinate in the mechanical coordinate system; Traverse each row of grains according to the second starting column center coordinate and the second ending column center coordinate.

4. The wafer appearance defect detection method according to claim 1, wherein Determining the height compensation data of the corresponding grain according to the height data set of each grain includes: Sort the data in the height data set of each grain and divide it into N equal parts to obtain the first height data subset, where N is a positive integer greater than 2; Remove the data in the first 1 / N and the last 1 / N of the first height data subset to obtain the second height data subset; Remove the high-frequency data in the second height data subset, and calculate the average value of the remaining data to obtain the height compensation data of the corresponding grain.

5. The wafer appearance defect detection method according to any one of claims 1 to 4, characterized in that, The performing defect recognition on the target image set based on the pre-trained differential model and AI model to obtain the first detection result includes: Obtain the first target image of the target grain from the target image set; Perform region recognition on the first target image based on the preset ROI region, and perform Blob analysis to intercept the second target image corresponding to the region where the target grain is located; Based on the preset detection template, perform alignment adjustment and identify the area to be detected on the second target image to obtain the third target image; Perform defect recognition on the third target image based on the pre-trained differential model to determine the first defect region; Perform local screenshot on the third target image based on the first defect region to obtain the defect image; Classify the defective image based on a pre-trained defect classification model to obtain a first detection result.

6. The wafer appearance defect detection method according to claim 5, characterized in that, Before regionally identifying the first target image based on a preset ROI region and performing Blob analysis to intercept the second target image corresponding to the region where the target grain is located, it further includes: Regionally identify the first target image based on a preset ROI region to obtain a fourth target image; Perform clarity detection on the fourth target image, and output a second detection result when the clarity detection fails.

7. The wafer appearance defect detection method according to claim 5, characterized in that The third target image includes a photosensitive region image and a wire region image. The defective identification of the third target image based on a pre-trained difference model to determine a first defective region includes: Perform defective identification on the photosensitive region image based on a pre-trained first difference model to obtain a photosensitive defective region; Perform defective identification on the wire region image based on a pre-trained second difference model to obtain a wire defective region; Merge the photosensitive defective region and the wire defective region to obtain a first defective region.

8. The wafer appearance defect detection method according to claim 7, wherein, The partial screenshot of the third target image based on the first defective region to obtain a defective image includes: Expand the region range based on the first defective region to obtain a second defective region; Perform a partial screenshot of the third target image according to the second defective region to obtain a defective image.

9. A wafer appearance defect detection device, characterized in that, It includes: A first determination module for determining the starting column center coordinates and ending column center coordinates of the grains in each row of the target wafer; A height measurement module for traversing each row of grains according to the starting column center coordinates and the ending column center coordinates, and controlling a height measurement sensor to measure the height of each grain during the traversal to obtain a height data set for each grain; A second determination module for determining the height compensation data corresponding to each grain according to the height data set of each grain; An image acquisition module for controlling a target camera to perform relative movement, height adjustment, and photographing according to the movement path obtained by path planning for each row of grains and the height compensation data to obtain a set of target images; A defect detection module for performing defect identification on the set of target images based on a pre-trained difference model and an AI model to obtain a first detection result.

10. A wafer appearance defect detection device, comprising a processor and a memory, wherein a computer program is stored in the memory, and is characterized in that, When running the computer program, the processor is used to implement the wafer appearance defect detection method according to any one of claims 1 to 8.

Images