A wafer detection method, device, storage medium and equipment

Abstract

In a wafer detection method, device, storage medium and equipment provided in this specification, by acquiring a surface image of a wafer to be detected, selecting a grain image as a reference image from the grain images of the surface image, and using other grain images as test images, determining a first marking point in the reference image, respectively determining positioning points aligned with the first marking point in each test image through image matching, aligning each test image according to the positioning points in each test image, fusing to obtain a template image, and comparing each grain image one by one based on the template image to achieve defect detection, with low cost, high efficiency and high precision of the template image. At the same time, the template image is a defect-free grain image fused from the grain images collected in actual production, improving the accuracy of defect detection.

Description

Technical Field

[0001] This specification relates to the field of computer vision, and particularly to a wafer detection method, apparatus, storage medium, and device. Background Art

[0002] Defect detection of wafers can promptly discover and solve problems existing in the semiconductor chip production process, and is an important link to improve the yield of semiconductor chips.

[0003] Currently, wafers are generally detected by an Automated Optical Inspection (AOI) system. Defect - less grain images are manually selected from each grain image as template images, and then the grain (Die) images in the wafer to be detected are matched with the template images to determine the detection results of the grain images. However, with the update of semiconductor chip production processes, it is necessary to manually and repeatedly screen template images, which has high costs and low efficiency. Moreover, the template images are not ideal defect - free grain images, resulting in poor accuracy in determining the detection results.

[0004] Based on this, this specification provides a wafer detection method, apparatus, storage medium, and device. Summary of the Invention

[0005] This specification provides a wafer detection method, apparatus, storage medium, and device to partially solve the above - mentioned problems existing in the prior art.

[0006] This specification adopts the following technical solutions: This specification provides a wafer detection method, and the method includes:

[0007] Obtain a surface image of the wafer to be detected, where the surface image includes grain images of several grains;

[0008] Select one grain image from each grain image as a reference image, and use the other grain images as test images;

[0009] Determine a first marked point in the reference image, and respectively determine alignment points aligned with the first marked point in each test image through image matching;

[0010] Align each test image according to the alignment points in each test image, and fuse them to obtain a template image;

[0011] Compare each grain image one by one according to the template image to achieve defect detection.

[0012] Optionally, the step of respectively determining alignment points aligned with the first marked point in each test image through image matching specifically includes:

[0013] Calibrate a first window image in the reference image according to the position of the first marker point in the reference image;

[0014] For each test image, determine the position of the second marker point in the test image, and calibrate a second window image in the test image according to the position of the second marker point in the test image, wherein the second window image is larger than the first window image;

[0015] Perform image matching between the second window image and the first window image, and determine the area in the test image that matches the first window image as the target area;

[0016] In the target area, determine the positioning point in the test image that aligns with the first marker point.

[0017] Optionally, the first marker point is located at the center of the first window image, and the second marker point is located at the center of the second window image.

[0018] Optionally, the step of calibrating the second window image according to the position of the second marker point in the test image specifically includes:

[0019] Determine the second window size according to the preset size of the first window image and the deviation of the machine table movement and camera calibration;

[0020] Calibrate the second window image with the second window size according to the position of the second marker point in the test image.

[0021] Optionally, the step of determining the position of the second marker point in the test image specifically includes:

[0022] Based on the surface image of the wafer to be detected, determine the distance deviation between the test image and the reference image, and determine the pixel deviation according to the distance deviation and the pixel size of the test image;

[0023] Determine the position of the second marker point in the test image according to the pixel deviation and the position of the first marker point.

[0024] Optionally, the step of performing image matching between the second window image and the first window image, and determining the area in the test image that matches the first window image as the target area specifically includes:

[0025] Using the first window image as a window, traverse the second window image to determine each candidate area;

[0026] For each candidate area, determine the matching degree between the candidate area and the first window image;

[0027] Select a target region from the candidate regions according to the matching degree between each candidate region and the first window image.

[0028] Optionally, the step of determining the matching degree between each candidate region and the first window image specifically includes:

[0029] For each candidate region, according to the gradient values of the pixels in the candidate region and the gradient values of the pixels in the first window image, determine the gradient matching value between the candidate region and the first window image through a gradient template matching method;

[0030] Determine the matching degree between the candidate region and the first window image according to the gradient matching value.

[0031] Optionally, the step of determining the matching degree between each candidate region and the first window image specifically includes:

[0032] For each candidate region, according to the gray values of the pixels in the candidate region and the gray values of the pixels in the first window image, determine the gray matching value between the candidate region and the first window image through a mean square error method;

[0033] Determine the matching degree between the candidate region and the first window image according to the gray matching value.

[0034] Optionally, the step of obtaining the surface image of the wafer to be detected specifically includes:

[0035] Obtain the original images of the wafers to be detected scanned by each machine tool;

[0036] Call the camera calibration parameters of each machine tool, and determine the image correction matrix according to the camera calibration parameters, where the camera calibration parameters include the camera orthogonal angle and the flat field coefficient;

[0037] Call the camera parameters of each machine tool, where the camera parameters include the camera resolution of each machine tool, the lens magnification of each machine tool, the distortion coefficient of each machine tool, and the camera size of each machine tool;

[0038] Correct the original image according to the camera parameters and the image correction matrix so that the size of each pixel in the corrected original image is the same, and obtain the surface image;

[0039] Obtain at least the surface image of one wafer to be detected among the wafers to be detected.

[0040] Optionally, the step of comparing each die image with the template image one by one to implement defect detection specifically includes:

[0041] For each grain image, determine the grayscale difference image between the grain image and the template image;

[0042] According to the grayscale difference image, determine the detection results of each pixel in the grayscale difference image through a preset grayscale detection range;

[0043] Determine the defect detection result according to the detection results of each pixel in the grayscale difference image.

[0044] Optionally, the step of determining the detection results of each pixel in the grayscale difference image through a preset grayscale detection range according to the grayscale difference image specifically includes:

[0045] For each pixel in the grayscale difference image, determine whether the grayscale difference of the pixel falls within the preset grayscale detection range;

[0046] If so, determine that the pixel has no defect as the detection result of the pixel;

[0047] If not, determine the defect type of the pixel according to the size relationship between the grayscale difference of the pixel in the grayscale difference image and the grayscale detection range, and determine the detection result of the pixel according to the defect type of the pixel.

[0048] Optionally, before determining whether the grayscale difference of each pixel in the grayscale difference image falls within the preset grayscale detection range, the method further includes:

[0049] Determine the gradient image of the template image;

[0050] For each pixel in the grayscale difference image, determine the corresponding gradient value of the pixel in the gradient image according to the position of the pixel in the grayscale difference image, and determine the grayscale detection range of the pixel according to the gradient value and the preset grayscale detection range.

[0051] Optionally, the step of determining the defect type of the pixel according to the size relationship between the grayscale value of the pixel in the grayscale difference image and the grayscale detection range specifically includes:

[0052] When the grayscale difference of the pixel in the grayscale difference image is less than the grayscale detection range, the defect type of the pixel is a dark defect;

[0053] When the grayscale difference of the pixel in the grayscale difference image is greater than the grayscale detection range, the defect type of the pixel is a bright defect.

[0054] Optionally, the step of determining the defect detection result according to the detection results of each pixel in the grayscale difference image specifically includes:

[0055] According to the detection results of each pixel in the grayscale difference image, mark the defective areas in the grain image and determine the defect types of the defective areas;

[0056] Determine the image features of the defective areas;

[0057] According to the image features of the defective areas and the defect types of the defective areas, determine the process defect categories to which the defective areas belong as the defect detection results.

[0058] This specification provides a wafer detection method, including:

[0059] Obtain the surface image of the wafer to be detected, where the surface image contains the grain images of several grains;

[0060] Select one grain image from each grain image as the reference image and use the other grain images as the test images;

[0061] Determine the image matching positions aligned with the reference image in each test image, and align and fuse the test images to obtain a template image;

[0062] Compare each grain image one by one according to the template image to achieve defect detection;

[0063] Among them, determining the image matching positions includes:

[0064] First matching: Determine the first marker point in the reference image, and determine the positioning point aligned with the first marker point in the test image through image matching as the image matching position of the test image;

[0065] If the first matching fails, perform the second matching: Continuously reduce the reference image and the test image in the same proportion, perform image matching between the reduced images, and determine the image matching position in the test image aligned with the reference image.

[0066] This specification provides a wafer detection device, including:

[0067] The first acquisition module is used to acquire the surface image of the wafer to be detected, where the surface image contains the grain images of several grains;

[0068] The first selection module is used to select one grain image from each grain image as the reference image and use the other grain images as the test images;

[0069] The positioning point module is used to determine the first marker point in the reference image and determine the positioning points aligned with the first marker point in each test image through image matching;

[0070] A template image module, configured to align the test images according to the positioning points in the test images, and fuse them to obtain a template image;

[0071] A first detection module, configured to compare each grain image with the template image one by one to achieve defect detection.

[0072] This specification provides a wafer detection device, including:

[0073] A second acquisition module, configured to acquire a surface image of a wafer to be detected, where the surface image includes grain images of a plurality of grains;

[0074] A second selection module, configured to select one grain image from the grain images as a reference image, and use the other grain images as test images;

[0075] An image matching module, configured to determine an image matching position aligned with the reference image in each test image, and align and fuse the test images to obtain a template image; wherein, determining the image matching position includes: a first match: determining a first marker point in the reference image, and determining, through image matching, a positioning point aligned with the first marker point in the test image as the image matching position of the test image; if the first match fails, a second match is performed: by continuously reducing the reference image and the test image in the same proportion, performing image matching between the reduced images, and determining the image matching position in the test image aligned with the reference image;

[0076] A second defect detection module, configured to compare each grain image with the template image one by one to achieve defect detection.

[0077] This specification provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, a wafer detection method is implemented.

[0078] This specification provides an electronic device, which includes a memory, a processor, and a computer program stored on the memory and running on the processor. When the processor executes the program, a wafer detection method is implemented.

[0079] The above at least one technical solution adopted in this specification can achieve the following beneficial effects: In a wafer detection method provided in this specification, by obtaining a surface image of a wafer to be detected, selecting a grain image as a reference image from the grain images of the surface image, and using other grain images as test images, determining a first marking point in the reference image, and respectively determining positioning points aligned with the first marking point in each test image through image matching, aligning each test image according to the determined positioning points in each test image, fusing to obtain a template image, and then comparing each grain image one by one based on the template image to achieve defect detection.

[0080] As can be seen from the above method, through image matching between the test images and the reference image, the positioning points of each test image are obtained, and then each test image is aligned and fused to obtain a template image, which has low cost, high efficiency and high accuracy of the template image. At the same time, the template image is a defect-free grain image fused based on the grain images collected in actual production, which improves the accuracy of defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The drawings described herein are used to provide a further understanding of this specification, and constitute a part of this specification. The schematic embodiments and descriptions thereof of this specification are used to explain this specification, and do not constitute an improper limitation of this specification. In the drawings:

[0082] Figure 1 is a schematic flowchart of a wafer detection method provided in this specification;

[0083] Figure 2 is a schematic diagram of determining positioning points in a test image based on image matching provided in this specification;

[0084] Figure 3 is a schematic diagram of determining positioning points by combining the square difference method and the gradient template matching method provided in this specification;

[0085] Figure 4 is a schematic flowchart of a wafer detection method provided in this specification;

[0086] Figure 5 is a schematic diagram of a wafer detection device provided in this specification;

[0087] Figure 6 is a schematic diagram of a wafer detection device provided in this specification;

[0088] Figure 7 is a schematic diagram of the structure of an electronic device for implementing a wafer detection method corresponding to this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0089] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this specification will be clearly and completely described below in conjunction with specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this specification without creative efforts belong to the scope protected by this specification.

[0090] In the process of executing a wafer detection method in this specification, it involves the processing of image data. Therefore, in the embodiments of this specification, the server can execute the process of wafer detection. Of course, this specification does not limit which device executes the process of wafer detection. For example, devices such as personal computers and mobile terminals can perform wafer detection, or in actual production, the process of wafer detection can also be executed by a machine tool. For the convenience of description, the server is used as the execution entity for explanation below.

[0091] Currently, generally, AOI is used to detect defects on wafers. AOI is applied to defect detection in advanced packaging. Its working principle is to capture die images through a camera system, such as a Charge Coupled Device (CCD) camera system, and after a series of algorithmic processes through an image processing card and a computer software system, compare with a template image to find defects and generate a report. However, with the update of semiconductor chip production processes, it is necessary to manually and repeatedly screen the template images, with high costs and low efficiency, and the template image is not an ideal defect-free die image, resulting in poor accuracy in determining the detection results.

[0092] Based on this, this specification provides a wafer detection method, which can be executed by a machine tool in actual production. The machine tool includes a high-resolution CCD camera, an optical lens, and a multi-angle color light source. That is, the machine tool has different illumination modes such as bright field and dark field, and the wavelength and brightness can be adjusted. At the same time, a high-precision and high-speed motion platform cooperates with the camera to synchronously scan and obtain wafer images. While meeting the requirements for detecting various wafer surface defects, it is also possible to determine the template image through image matching based on the collected die images, so as to reduce labor costs and improve efficiency.

[0093] The following will detail the technical solutions provided by each embodiment of this specification in conjunction with the drawings.

[0094] Figure 1 The flowchart of a wafer detection method provided in this specification includes the following steps:

[0095] S100: Obtain the surface image of the wafer to be detected, where the surface image contains die images of several dies.

[0096] In one or more embodiments of this specification, a wafer is a basic material in semiconductor manufacturing, usually a circular thin slice made of silicon. A die refers to a single integrated circuit unit formed on the wafer through photolithography and other process steps. In this step, in order to determine the reference image and the test image in subsequent steps, the server needs to obtain the surface image of the wafer to be detected.

[0097] Specifically, the server can obtain the surface image of the wafer to be detected, and the surface image contains die images of several dies. This specification does not limit the method of obtaining the surface image. For example, a scanning electron microscope, an optical microscope, etc. can also capture the surface image through a camera system, and then the die images can be processed through a series of algorithms such as an image processing card and a computer software system.

[0098] S102: Select one die image from each die image as the reference image, and use the other die images as test images.

[0099] In one or more embodiments of this specification, in order to determine the positioning points for alignment in each die image through image matching in subsequent steps, in this step, a reference image needs to be selected from each die image as the target of image matching.

[0100] Specifically, the server can select one die image from each die image obtained in step S100 as the reference image, and use the other die images as test images. Among them, each die image can be segmented from the surface image through edge detection or other mature technologies in the art. This specification does not limit the specific method of selecting the reference image. For example, it can be randomly selected, or the first segmented die image can be selected, etc. Of course, in subsequent steps, the reference image is used as the target of image matching. Therefore, when selecting the reference image, a die image that meets the selection rules can be selected as the reference image. The selection rules can be set according to actual needs. For example, a die image without dirt marks, large scratches and other major defects, that is, determine the area in the die image where the gray value is lower / higher than the preset value. When the size of this area is greater than a certain value, it cannot be used as the reference image.

[0101] S104: Determine the first marking point in the reference image, and respectively determine the positioning points aligned with the first marking point in each test image through image matching.

[0102] In one or more embodiments of the present specification, the Mark in the crystal grain is a specific mark or test point on a single crystal grain for providing reference and positioning in crystal grain testing and packaging, rather than a feature point. Feature points are usually "salient points" in an image, such as edge points, corner points, etc. Defects are often "salient points" in the crystal grain image. If crystal grain image matching is implemented based on feature points, whether taking feature points in a partial area or the entire area of the crystal grain image, the accuracy of feature point matching between crystal grain images will be affected due to the "defect" feature points in the crystal grain image. Therefore, in order to improve the accuracy of alignment of each test image in subsequent steps, in this step, the server can identify and determine the Mark in each crystal grain image. However, due to possible errors in the recognition accuracy, the server can also perform "fine-tuning" on the second Mark in each test image through image matching based on the first Mark in the reference image to obtain a positioning point aligned with the first Mark.

[0103] In one or more embodiments of the present specification, the server can calibrate a first window image in the reference image according to the position of the first Mark in the reference image; then, for each test image, calibrate a second window image in the test image according to the position of the second Mark in the test image, and the second window image is larger than the first window image; finally, based on the image matching result between the second window image and the first window image, determine the positioning point in the test image that is aligned with the first Mark.

[0104] Specifically, first, the server can use the pixel points (coordinates) corresponding to the Mark of the corresponding crystal grain in the reference image as the first Mark, and calibrate a first window image in the reference image according to the position of the first Mark in the reference image.

[0105] At the same time, for each test image, use the pixel points (coordinates) corresponding to the Mark of the corresponding crystal grain in the test image as the second Mark, and calibrate a second window image in the test image according to the position of the second Mark in the test image. Among them, the second window image is larger than the first window image.

[0106] It should be noted that in the surface image obtained in step S100, due to machine movement and camera calibration, there may be deviations in the positions of the Marks in the crystal grain images, that is, there may be partial deviations in the positions of the Marks identified in each crystal grain image based on the surface image. Therefore, in order for the server to align each test image and fuse them to obtain a template image in subsequent steps. In this step, it is necessary to use the first Mark in the reference image as the alignment target, and determine the positioning point aligned with the first Mark in each test image through image matching.

[0107] In one embodiment, when calibrating the window image, the positional relationship between the first marker point and the first window image may be the same as or different from the positional relationship between the second marker point and the second window image, which is not limited in this specification.

[0108] In one embodiment, this specification does not limit the positional relationship between the marker point and the window image when calibrating the window image. For example, the marker point can be the center of the window image to calibrate the window image. In another embodiment, since the grain image includes a blank part and a grain part, and the window image is calibrated around the marker point, that is, the surrounding area of the marker point may include a blank part and a grain part. That is, for subsequent image matching, the weight of the grain part in this surrounding area is higher than that of the blank part. Therefore, the server can calibrate the first window image in the reference image according to the position of the first marker point in the reference image and the preset positional relationship between the marker point and the grain. Make the positional relationship between the first marker point in the first window image the same as the preset positional relationship between the marker point and the grain, so as to ensure that the first window image contains more grain parts, that is, more information of the "grain part" participates in the subsequent image matching between the first window image and the second window image, thereby improving the reliability and accuracy of obtaining the positioning point by image matching.

[0109] Secondly, for each test image, the server can use the first window image as a window to traverse the second window image of the test image, determine the candidate regions at each traversed position, and then for each candidate region, determine the matching degree between the candidate region and the first window image. Based on the respective matching degrees, select the target region from the candidate regions. Among them, this specification does not limit the specific method for determining the matching degree between each candidate region and the first window image, such as the gradient template matching method, the mean squared error method, the Euclidean distance based on image vectorization, etc.

[0110] Finally, the server can determine the positioning point in the target region of the test image that aligns with the first marker point based on the positional relationship between the first marker point and the first window image.

[0111] It should be noted that, based on the gradient template matching method, determining the matching degree between each candidate region and the first window image includes: for each candidate region, determining the gradient matching value between the candidate region and the first window image according to the gradient values of the pixels in the candidate region and the gradient values of the pixels in the first window image, and using it as the matching degree between the candidate region and the first window image. Based on the mean square difference method, determining the matching degree between each candidate region and the first window image includes: for each candidate region, determining the gray level matching value between the candidate region and the first window image according to the gray level values of the pixels in the candidate region and the gray level values of the pixels in the first window image, and using it as the matching degree between the candidate region and the first window image.

[0112] It should be noted that in the surface image obtained in step S100, due to the movement of the machine platform and camera calibration, the positions of the marking points in each grain image obtained by recognition are not completely consistent, that is, there are deviations in the positions of the marking points in the grain images where they are located. Therefore, in this step, by performing image matching on the area near the marking points, the positioning points aligned with the first marking point are determined in each test image, so as to ensure that the reference image or / and each test image can be accurately aligned through the positioning points determined in this step.

[0113] Specifically refer to Figure 2 As shown, the first window image is calibrated at the position of the first marking point in the reference image, and the second window image is calibrated at the position of the second marking point in the test image. Then, image matching is performed between the second window image and the first window image to determine the target region. Based on the target region, the positioning points aligned with the first marking point are determined in the test image. Among them, the rectangular patterns filled in the reference image or test image represent the grain parts, and this figure is only a schematic diagram. In this specification, it is not limited that the first window image / second window image is only rectangular when calibrated, and it can be set according to requirements.

[0114] S106: Align the test images according to the positioning points in the test images, and fuse them to obtain a template image.

[0115] In one or more embodiments of this specification, in order to perform defect detection on each grain image in subsequent steps, in this step, it is necessary to align the test images according to the positioning points in the test images determined in step S104, and fuse them to obtain a template image.

[0116] Specifically, the server can align the test images according to the positioning points in the test images, and based on the aligned test images, determine the average gray level value of each pixel, and determine the template image according to the average gray level value of each pixel.

[0117] Optionally, for each test image, the server can align the test image with the reference image according to the positioning points in the test image and the first marker points of the reference image, and determine the average gray value of each pixel based on the aligned test images and the reference image. Then, according to the average gray value of each pixel position, a template image can be determined.

[0118] It should be noted that generally, the number of grain images participating in the fusion is relatively large. For each pixel, the number of defective gray values is relatively small. Therefore, even if defective grain images participate in the fusion, the obtained template image is approximately defect-free and can be used for defect detection in subsequent steps.

[0119] At the same time, in this specification, for each pixel, the defective gray values of this pixel in each test image can be excluded, so as to further improve the accuracy of the template image. In this specification, the specific method of excluding defective gray values is not limited and can be set according to actual needs. In one or more embodiments of the specification, when fusing to obtain a template image, for each pixel, according to the gray values corresponding to this pixel in each test image respectively, a normal value detection range can be determined, and according to the gray values of this pixel within this normal value detection range, the average gray value of this pixel can be determined. Among them, the mean and standard deviation of the gray values of this pixel in each test image can be determined, and according to the mean and standard deviation, the normal value detection range can be determined. If the mean is Mean and the standard deviation is std, the normal value detection range can be (Mean - n * std) to (Mean + n * std), where n is a positive integer and can be set according to actual needs. Of course, in this specification, the specific method of determining the normal value detection range is not limited. In one or more embodiments of the specification, when fusing to obtain a template image, for each pixel, the probability distribution of the gray values of this pixel in each test image is determined, the gray values with probabilities lower than a preset threshold are removed, and according to the gray values with probabilities greater than this preset threshold in the probability distribution, the average gray value of this pixel is determined. Among them, the preset threshold can be set according to actual needs. Of course, when fusing to obtain a template image, the reference image can also participate in the alignment of each test image.

[0120] It should be noted that for the convenience of defect detection of each grain image based on the template image in subsequent steps, the size of the template image can be the same as that of the grain image. Therefore, when determining the template image based on the aligned test images, a test image can be randomly selected, and for each pixel in the selected test image, the gray values of this pixel in the reference image and other test images are determined, and then the average gray value of this pixel is obtained. According to the average gray values of each pixel, the template image is determined.

[0121] Of course, the alignment point position can also be determined based on the position of the first marker point or positioning point. According to the alignment point position and the size of the grain image, an empty template image is determined. For each pixel in the empty template image, the gray values of the pixel in the reference image and each test image are determined to obtain the average gray value of the pixel. Then, based on the average gray values of the pixels, the empty template image is filled to obtain the template image. Herein, there is no limitation on the specific manner of determining the alignment point position. For example, according to the coordinates of each positioning point or the first marker point, the average or median is taken to obtain the alignment point position.

[0122] S108: Compare each of the grain images one by one according to the template image to achieve defect detection.

[0123] In one or more embodiments of this specification, in this step, according to the template image determined in step S106, defect detection can be performed on the grain images of each grain in the wafer to be detected.

[0124] Specifically, first, the server can determine the gray difference image between the grain image of each grain and the template image.

[0125] Secondly, for each pixel in the gray difference image, it is determined whether the gray difference of the pixel falls within a preset gray detection range. If the gray difference of the pixel falls within the preset gray detection range, it is determined that the pixel has no defect, which is used as the detection result of the pixel. If the gray difference of the pixel does not fall within the preset gray detection range, the defect type of the pixel is determined, and then based on the defect type, the detection result of the pixel is determined.

[0126] Finally, according to the detection results of the pixels in the gray difference image, the defective areas in the grain image are marked, and the defect type and image features of the defective areas are determined. According to the defect type and image features of the defective areas, the process defect category to which the defective areas belong is determined. According to the process defect categories of the defective areas, the defect detection result of the grain is determined. Herein, the image features include the shape, length, width, area, etc. of the defective areas. The process defect categories can be divided into large defects, scratches, etc. For example, when the length of the defective area exceeds a preset threshold, the defective area is a large defect. According to the relationship between the gray difference of the pixel and the size of the gray detection range, the defect type of the pixel is determined as follows:

[0127] When the gray difference of the pixel in the gray difference image is less than the preset gray detection range, the defect type of the pixel is a dark defect. When the gray difference of the pixel in the gray difference image is greater than the preset gray detection range, the defect type of the pixel is a bright defect.

[0128] It should be noted that the grayscale value range is 0 to 255. When determining the grayscale difference image, in order to facilitate observing the defect conditions of each pixel, after subtracting the template image from the grain image, a fixed grayscale value (such as 128, this value is only for illustration and not a limitation on the fixed grayscale value) can be added to the grayscale value of each pixel obtained after the subtraction. If the grayscale difference after adding this fixed grayscale value to the grayscale value of each pixel exceeds the grayscale value range, the endpoint grayscale value adjacent to the grayscale value of this pixel can be selected within this grayscale value range as the grayscale difference of this pixel. Then, based on the grayscale differences of each pixel, the grayscale difference image of this grain can be obtained.

[0129] In the above method, a partial area (the first window image) around the first marked point is determined in the reference image, and a partial area (the second window image) around the second marked point is found in each test image to search for the first marked point. Then, through image matching between the first window image and the second window image, a positioning point aligned with the first marked point is found in the second window image. That is, only partial area matching is required. Therefore, the sizes of the grain images determined in the embodiments of this specification do not need to be exactly the same. That is to say, the size consistency of the grain images does not need to be considered. That is, the grain images obtained from the wafer to be detected can have different sizes, thus avoiding the workload of image processing such as "image scaling, cropping, and size matching", that is, reducing complexity. And the partial area matching reduces the workload of image matching compared with full-image matching. The server then aligns each test image based on the obtained positioning points and fuses them to obtain a template image. And when fusing to obtain this template image, defective grayscale values can be further eliminated for each pixel, so that the average grayscale value of this pixel is approximately defect-free. That is, the fused template image is a defect-free grain image. Then, when detecting defects in each grain image based on this template image, the cost is low, the efficiency is high, and the accuracy is high.

[0130] Among them, the first marked point / second marked point is a marked point (mark) in the grain image. The marked point (mark) is a physical feature that can be accurately positioned during the integrated circuit production process. The alignment between grain images is achieved through the pixel points corresponding to the marked point (mark) in the grain image, and at the same time, the deviation of the position of the marked point (Mark) in the grain image caused by machine movement and camera calibration is eliminated, making the positioning process convenient and accurate.

[0131] In the preferred embodiment, the first window image is calibrated with the first marked point as the center, and the second window image is calibrated with the second marked point as the center. In the actual application process, the calibration provided by this embodiment is convenient, and the subsequent calculation of relevant image matching is simple.

[0132] In addition, in one or more embodiments of this specification, when selecting a target region from each candidate region in step S104, the server may, for each candidate region, determine a comprehensive matching degree based on the gradient matching value between the candidate region and the first window image and the gray-scale matching value between the candidate region and the first window image, and then select the target region based on this comprehensive matching degree. Herein, the specific method for determining the comprehensive matching degree is not limited in this specification, such as taking the sum value or weighted sum of the gradient matching value and the gray-scale matching value.

[0133] It should be noted that in the combined method of the square difference method and the gradient template matching method, considering the comprehensive matching degree between each candidate region and the first window image from the two aspects of gradient and gray scale, and then selecting the target region based on each comprehensive matching degree is more accurate. However, since the gradient template matching method requires additional gradient calculations and more computing resources are needed. Therefore, this specification also provides a method of first using the square difference method to determine the gray-scale matching region and then using the gradient template matching method to determine the target region, which is as follows:

[0134] In one or more embodiments of this specification, the gray-scale matching region can be first selected from each candidate region by the square difference method, and then the third window image can be calibrated at the position of the first marked point of the reference image. Based on this third window image and the gray-scale matching region, the target region can be selected from each sub-region in the gray-scale matching region by the gradient template matching method.

[0135] Specifically, first, the server can use the first window image calibrated in step S104 as the window, traverse the second window image, determine the candidate regions at each traversed position, and for each candidate region, determine the gray-scale matching value between the candidate region and the first window image by the square difference method according to the gray-scale values of the pixels in the candidate region and the gray-scale values of the pixels in the first window image. According to each gray-scale matching value, the gray-scale matching region is selected from each candidate region.

[0136] Secondly, the server can calibrate the third window image at the position of the first marked point in the reference image, use this third window image as the window, traverse the gray-scale matching region, and determine the sub-regions corresponding to each traversed position, where the third window image is smaller than the first window image. The positional relationship between the third window image and the first marked point can refer to the "positional relationship between the first marked point and the first window image" in step S104.

[0137] Finally, for each sub-region in the gray-scale matching region, the server can determine the matching degree between the sub-region and the third window image by the gradient template matching method according to the gradient values of the pixels in the sub-region and the gradient values of the pixels in the third window image. According to the matching degrees between each sub-region and the third window image, the target region is selected from each sub-region.

[0138] It should be noted that since the gradient template matching method requires additional gradient calculation of the region, while the mean squared error method only requires subtraction and squaring of the gray values of each pixel, first use the mean squared error method to select the gray-scale matching region in each candidate region, and then, based on the third window image smaller than the first window image, through the gradient template matching method with the gray-scale matching region, the accuracy of determining the target region can be ensured from both the gray-scale and gradient perspectives while reducing the use of computing resources. Of course, in this specification, the order of using the mean squared error method and the gradient template matching method when determining the target region is not limited, and the gradient template matching method can also be used first and then the gradient template matching method.

[0139] Specifically refer to Figure 3 As shown, following the above example, the first image matching is to determine the gray-scale matching region by the mean squared error method. Secondly, at the position of the first marked point in the reference image, calibrate the third window image, and based on the gray-scale matching region and the third window image, perform the second image matching by the gradient template matching method to determine the target region. Finally, based on the target region, determine the positioning point aligned with the first marked point in the test image.

[0140] In addition, in one or more embodiments of this specification, ideally, the distance between the test image and the reference image should be an integer multiple of the grain image size. Therefore, the server can determine the distance deviation between the test image and the reference image for each test image, and based on the distance deviation and the pixel size of the test image, determine the pixel deviation. Then, based on the pixel deviation and the position of the first marked point in the reference image, determine the position of the second marked point in the test image as follows:

[0141] Specifically, the server can determine the distance deviation between the test image and the reference image based on the surface image of the wafer to be detected for each test image, and based on the distance deviation and the pixel size of the test image, determine the pixel deviation. Then, based on the pixel deviation and the position of the first marked point, determine the position of the second marked point in the test image.

[0142] Among them, the distance deviation between the test image i and the reference image 0 is determined as follows:

[0143] N x =round[(x i -x0) / pitch x

[0144] N y =round[(y i -y0) / pitch y ​​

[0145] Δx = x i - x0 - N x *pitch x

[0146] Δy = y i - y0 - N y *pitch y

[0147] Wherein, N x is the number of grain images separating the test image i and the reference image 0 in the x - dimension, and N y is the number of grain images separating the test image i and the reference image 0 in the y - dimension. (x i , y i ) are the physical coordinates of the test image i in the surface image, and (x0, y0) are the physical coordinates of the reference image 0 in the surface image. pitch x is the size of the grain image in the x - dimension, and pitch y is the size of the grain image in the y - dimension; round(*) is the rounding function. Δx is the distance deviation between the test image i and the reference image 0 in the x - dimension; Δy is the distance deviation between the test image i and the reference image 0 in the y - dimension.

[0148] Determine the pixel deviation based on the distance deviation and the pixel size of the test image, and then determine the position of the second marker point in the test image based on this pixel deviation and the position of the first marker point, as follows:

[0149] R i = R0 - Δy / pixel y

[0150] C i = C0 + Δx / pixel x

[0151] Wherein, (R i , C i ) are the pixel coordinates of the second marker point in the test image i; (R0, C0) are the pixel coordinates of the first marker point in the reference image 0. pixel x is the size of the pixel size in the x - dimension, and pixel y is the size of the pixel size in the y - dimension. The pixel size is determined during camera calibration, i.e., pixel size = (camera size / camera resolution) / lens magnification factor, and the pixel sizes of each grain image are the same.

[0152] It should be noted that in the pixel coordinates of the marker point, R i , C i, which are the coordinates in the x - dimension and y - dimension respectively. Since in the physical coordinate system, the x - dimension is positive to the right and negative to the left, and the y - dimension is positive upward and negative downward, while in the image coordinate system, the x - dimension is positive to the right and negative to the left, and the y - dimension is negative upward and positive downward. That is, there should be a sign difference in the pixel deviation from the distance deviation in the physical coordinate to the pixel coordinate in the y - dimension.

[0153] Based on the distance deviation between the above - mentioned test image and the reference image, the process of determining the position of the second marker point in the test image and then determining the positioning point in the test image is as follows:

[0154] First, the server can calibrate the first window image in the reference image according to the position of the first marker point in the reference image, and for each test image, calibrate the second window image in the test image according to the second marker point in the test image. The second window image is larger than the first window image.

[0155] Second, the server can perform image matching between the second window image and the first window image to determine the area in the test image that matches the first window image as the target area.

[0156] Finally, the server can determine the positioning point in the target area of the test image that aligns with the first marker point in the test image based on the positional relationship between the first marker point and the first window image.

[0157] In addition, in step S104, when the server calibrates the second window image in step S104, it can obtain the second window size based on the size of the first window image plus the deviation of the machine movement and camera calibration, and for each test image, calibrate the second window image with this second window size according to the position of the second marker point in the test image.

[0158] In addition, when obtaining the surface image of the wafer to be detected in step S100, the distortion of the image and the differences in camera resolution and lens magnification between each machine can be corrected based on the camera calibration parameters of each machine, so that the actual sizes represented by the images collected by each machine are the same. Then, for the wafers to be detected of the same product, that is, the wafers to be detected with the same pattern, the template image obtained from the surface image of one wafer to be detected can be used for defect detection of the die images in each wafer to be detected, thereby reducing the workload of repeatedly determining the template image.

[0159] In one or more embodiments of the present specification, the server may obtain the original images of the wafers to be detected scanned by each machine tool, and call the camera calibration parameters of each machine tool, and determine the image correction matrix according to the camera calibration parameters, where the camera calibration parameters include the camera orthogonal angle and the flat field coefficient. Call the camera parameters of each machine tool, where the camera parameters include the camera resolution of each machine tool, the lens magnification of each machine tool, the distortion coefficient of each machine tool, and the camera size of each machine tool, and then based on the camera parameters and the image correction matrix, correct the original images of the wafers to be detected so that the sizes of the pixels in the corrected original images are consistent, and obtain the surface images. At least obtain the surface images of one of the wafers to be detected for subsequent steps to determine the template image.

[0160] It should be noted that in actual production, since the machine tool includes a high-resolution CCD camera, an optical lens, and a multi-angle color light source, that is, the machine tool has different illumination modes such as bright field and dark field, and the wavelength and brightness are adjustable. Therefore, in the preferred embodiment, when obtaining the surface images of the wafers, the acquisition should be performed under the same illumination mode, the same wavelength, and the same brightness.

[0161] In addition, in step S108, when performing defect detection on each pixel in the grayscale difference image based on the preset grayscale detection range, since there is a connection between the gradient changes among the pixels in the template image. Therefore, the server can determine the grayscale detection range of each pixel based on the gradient values of the pixels in the gradient image of the template image, specifically as follows:

[0162] In one or more embodiments of the present specification, the server may determine the gradient image of the template image in step S106, and for each test image, determine the grayscale difference image between the test image and the template image. Then, for each pixel in the grayscale difference image, according to the position of the pixel in the grayscale difference image, determine the corresponding gradient value of the pixel in the gradient image of the template image, and based on the gradient value and the preset grayscale detection range, determine the grayscale detection range of the pixel.

[0163] For each pixel G in the grayscale difference image x,y , based on the preset grayscale detection range, determine the detection result of the pixel as follows:

[0164] The preset grayscale detection range is: G dark ~G bright

[0165] Determine the gradient image G of the template image grad , then determine the corresponding gradient value G grad (x,y) of the pixel in the gradient image, and based on the gradient value and the preset grayscale detection range, determine the grayscale detection range of the pixel as Gdark -k*G grad (x,y)~G bright +k*G grad (x,y)。

[0166] Among them, G dark is the preset dark defect threshold, and G bright is the preset bright defect threshold. k is the gradient coefficient, which can be set according to actual needs.

[0167] Figure 4 This is a schematic flowchart of a wafer detection method provided in this specification, including the following steps:

[0168] S400: Obtain the surface image of the wafer to be detected, where the surface image contains the grain images of several grains.

[0169] S402: Select one grain image from each grain image as the reference image, and use the other grain images as test images.

[0170] S404: Determine the image matching positions aligned with the reference image in each test image, and align and fuse the test images to obtain a template image. Among them, determining the image matching position includes: the first match: determine the first marker point in the reference image, and determine the positioning point aligned with the first marker point in the test image through image matching as the image matching position of the test image; if the first match fails, perform the second match: continuously reduce the reference image and the test image in the same proportion, perform image matching between the reduced images, and determine the image matching position in the test image aligned with the reference image.

[0171] S406: Compare each grain image one by one according to the template image to achieve defect detection.

[0172] In the above, for the relevant content of step S400, step S402, the first match in step S404, and step S406, refer to Figure 1 the content of each step in a wafer detection method provided, which will not be elaborated here.

[0173] In step S406, the server can continuously reduce the reference image and each test image in the same proportion, perform image matching between the reduced images, and determine the image matching positions in each test image aligned with the reference image. Based on the image matching positions of each test image, a template image is fused. Defect detection is performed on each grain image according to this template image. That is, by performing image matching between each test image and the reference image respectively, the image matching positions matching the reference image are found.

[0174] The above is a wafer detection method provided in one or more embodiments of this specification. Based on the same concept, this specification also provides a corresponding wafer detection device, as Figure 5 shown.

[0175] A first acquisition module 500 acquires a surface image of a wafer to be detected, where the surface image includes grain images of several grains;

[0176] A first selection module 501 selects one grain image from each grain image as a reference image, and uses the other grain images as test images;

[0177] A positioning point module 502 determines a first marking point in the reference image, and respectively determines positioning points aligned with the first marking point in each test image through image matching;

[0178] A template image module 503 aligns the test images according to the positioning points in the test images, and fuses them to obtain a template image;

[0179] A first detection module 504 compares each grain image one by one according to the template image to achieve defect detection.

[0180] Optionally, the positioning point module 502 is specifically configured to calibrate a first window image in the reference image according to the position of the first marking point in the reference image; for each test image, determine the position of a second marking point in the test image, and calibrate a second window image in the test image according to the position of the second marking point in the test image, where the second window image is larger than the first window image; perform image matching between the second window image and the first window image, determine a region in the test image that matches the first window image as a target region; and determine a positioning point in the test image that is aligned with the first marking point in the target region.

[0181] Optionally, the positioning point module 502 can also be used such that the first marking point is located at the center of the first window image, and the second marking point is located at the center of the second window image.

[0182] Optionally, the positioning point module 502 can also be used to determine a second window size according to a preset size of the first window image and deviations in machine movement and camera calibration; and calibrate a second window image with the second window size according to the position of the second marking point in the test image.

[0183] Optionally, the positioning point module 502 may also be configured to determine the distance deviation between the test image and the reference image based on the surface image of the wafer to be detected, and determine the pixel deviation according to the distance deviation and the pixel size of the test image; determine the position of the second marking point in the test image according to the pixel deviation and the position of the first marking point.

[0184] Optionally, the positioning point module 502 is specifically configured to traverse the second window image with the first window image as the window to determine each candidate region; for each candidate region, determine the matching degree between the candidate region and the first window image; select a target region from the candidate regions according to the matching degrees between the candidate regions and the first window image.

[0185] Optionally, the positioning point module 502 may also be configured to, for each candidate region, determine the gradient matching value between the candidate region and the first window image through a gradient template matching method according to the gradient values of the pixels in the candidate region and the gradient values of the pixels in the first window image; determine the matching degree between the candidate region and the first window image according to the gradient matching value.

[0186] Optionally, the positioning point module 502 may also be configured to, for each candidate region, determine the gray level matching value between the candidate region and the first window image through a mean square difference method according to the gray level values of the pixels in the candidate region and the gray level values of the pixels in the first window image; determine the matching degree between the candidate region and the first window image according to the gray level matching value.

[0187] Optionally, the first acquisition module 500 is specifically configured to acquire the original images of the wafers to be detected scanned by each machine tool; call the camera calibration parameters of each machine tool, and determine an image correction matrix according to the camera calibration parameters, where the camera calibration parameters include the camera orthogonal angle and the flat field coefficient; call the camera parameters of each machine tool, where the camera parameters include the camera resolution of each machine tool, the lens magnification of each machine tool, the distortion coefficient of each machine tool, and the camera size of each machine tool; correct the original images according to the camera parameters and the image correction matrix so that the size of each pixel in the corrected original image is the same to obtain a surface image; acquire at least the surface image of one wafer to be detected among the wafers to be detected.

[0188] Optionally, the first detection module 504 is specifically configured to, for each die image, determine the gray level difference image between the die image and the template image; determine the detection result of each pixel in the gray level difference image through a preset gray level detection range; determine the defect detection result according to the detection results of the pixels in the gray level difference image.

[0189] Optionally, the first detection module 504 is specifically configured to, for each pixel in the grayscale difference image, determine whether the grayscale difference of the pixel falls within a preset grayscale detection range; if so, determine that the pixel is defect-free as the detection result of the pixel; if not, determine the defect type of the pixel according to the relationship between the grayscale difference of the pixel in the grayscale difference image and the size of the grayscale detection range, and determine the detection result of the pixel according to the defect type of the pixel.

[0190] Optionally, the first detection module 504 is specifically configured to, when the grayscale difference of the pixel in the grayscale difference image is less than the grayscale detection range, the defect type of the pixel is a dark defect; when the grayscale difference of the pixel in the grayscale difference image is greater than the grayscale detection range, the defect type of the pixel is a bright defect.

[0191] Optionally, the first detection module 504 is specifically configured to mark the defect area in the crystal grain image according to the detection results of the pixels in the grayscale difference image, and determine the defect type of the defect area; determine the image features of the defect area; and determine the process defect category to which the defect area belongs according to the image features of the defect area and the defect type of the defect area as the defect detection result.

[0192] The above is a wafer detection method provided in one or more embodiments of this specification. Based on the same idea, this specification also provides a corresponding wafer detection device, as Figure 6 shown.

[0193] A second acquisition module 600 acquires a surface image of a wafer to be detected, where the surface image includes crystal grain images of a plurality of crystal grains;

[0194] A second selection module 601 selects one crystal grain image from the crystal grain images as a reference image, and uses the other crystal grain images as test images;

[0195] An image matching module 602 determines an image matching position aligned with the reference image in each test image, and aligns and fuses the test images to obtain a template image; where determining the image matching position includes: a first match: determining a first marker point in the reference image, and determining, through image matching, a positioning point aligned with the first marker point in the test image as the image matching position of the test image; if the first match fails, a second match is performed: by continuously reducing the reference image and the test image in the same proportion, performing image matching between the reduced images, and determining the image matching position in the test image aligned with the reference image.

[0196] The second defect detection module 603 compares each grain image one by one according to the template image to achieve defect detection.

[0197] This specification also provides a computer-readable storage medium storing a computer program that can be used to execute the above Figure 1 or Figure 4 a wafer detection method provided.

[0198] This specification also provides Figure 7 a schematic structural diagram of the electronic device shown. As Figure 7 described, at the hardware level, the device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the above Figure 1 or Figure 4 a wafer detection method described. Of course, in addition to the software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0199] In the 1990s, it was quite obvious to distinguish whether an improvement in a technology was an improvement in hardware (e.g., improvement in circuit structures such as diodes, transistors, switches, etc.) or an improvement in software (improvement in method processes). However, with the development of technology, many improvements in method processes today can be regarded as direct improvements in hardware circuit structures. Almost all designers obtain the corresponding hardware circuit structures by programming the improved method processes into the hardware circuits. Therefore, it cannot be said that an improvement in a method process cannot be implemented with a hardware entity module. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program by themselves to "integrate" a digital system on a piece of PLD without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly implemented using "logic compiler" software, which is similar to the software compiler used in program development and writing. The original code before compilation also has to be written in a specific programming language, which is called Hardware Description Language (HDL). And there is not only one type of HDL, but many types, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones currently are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also be aware that by simply performing a little logical programming on the method process with the above-mentioned several hardware description languages and programming it into the integrated circuit, it is easy to obtain the hardware circuit that implements the logical method process.

[0200] The controller can be implemented in any suitable manner. For example, the controller can take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program code (such as software or firmware) executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of the controller include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that, in addition to implementing the controller in the form of pure computer-readable program code, it is entirely possible to make the controller implement the same function in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be regarded as the structures within the hardware component. Or even, the devices for implementing various functions can be regarded as either software modules for implementing the method or the structures within the hardware component.

[0201] The systems, devices, modules, or units illustrated in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0202] For the convenience of description, when describing the above devices, they are described separately as various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0203] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0204] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, as well as the combination of flows and / or blocks in the flowchart and / or block diagram. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0205] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0206] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, thereby providing the steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0207] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0208] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0209] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0210] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0211] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0212] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0213] The above are only examples of this specification and are not intended to limit this specification. For those skilled in the art, various changes and modifications can be made to this specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this specification shall be included within the scope of the claims of this specification.

Claims

1. A wafer detection method, characterized in that, Including: Obtain a surface image of a wafer to be detected, where the surface image includes crystal grain images of a number of crystal grains; Select one crystal grain image from each of the crystal grain images as a reference image, and use the other crystal grain images as test images; According to the position of a first marking point in the reference image and a preset positional relationship between the marking point and the crystal grain, calibrate a first window image in the reference image. For each test image, based on the surface image of the wafer to be detected, determine the distance deviation between the test image and the reference image, and according to the distance deviation and the pixel size of the test image, determine the pixel deviation between the test image and the reference image. According to the pixel deviation and the position of the first marking point in the reference image, determine the position of a second marking point in the test image; according to the preset size of the first window image and the deviation of machine movement and camera calibration, determine the size of a second window image; perform image matching between the second window image and the first window image, and determine, in the test image, a region that matches the first window image as a target region. In the target region, determine a positioning point in the test image that aligns with the first marking point, where the second window image is larger than the first window image, and the positional relationship of the first marking point in the first window image is the same as the preset positional relationship between the marking point and the crystal grain, so as to ensure that more crystal grain parts are included in the first window image. The marking point in the crystal grain is a specific marking or test point that provides reference and positioning on a single crystal grain, rather than a feature point; Align the test images according to the positioning points in the test images, and fuse them to obtain a template image; Perform comparison on each of the crystal grain images according to the template image to achieve defect detection.

2. The method according to claim 1, characterized in that, The first marking point is located at the center of the first window image, and the second marking point is located at the center of the second window image.

3. The method according to claim 1, characterized in that The step of performing image matching between the second window image and the first window image, and determining, in the test image, a region that matches the first window image as a target region specifically includes: Using the first window image as a window, traverse the second window image to determine each candidate region; For each candidate region, determine the matching degree between the candidate region and the first window image; According to the matching degrees of the candidate regions and the first window image, select a target region from the candidate regions.

4. The method according to claim 3, characterized in that The step of, for each candidate region, determining the matching degree between the candidate region and the first window image specifically includes: According to the gradient values of the pixels in the candidate region and the gradient values of the pixels in the first window image, determine the gradient matching value between the candidate region and the first window image through a gradient template matching method; According to the gradient matching value, determine the matching degree between the candidate region and the first window image.

5. The method according to claim 3, wherein The step of, for each candidate region, determining the matching degree between the candidate region and the first window image specifically includes: For each candidate region, according to the gray values of the pixels in the candidate region and the gray values of the pixels in the first window image, determine the gray matching value between the candidate region and the first window image by the method of squared difference; Determine the matching degree between the candidate region and the first window image according to the gray matching value.

6. The method according to claim 1, wherein The step of obtaining the surface image of the wafer to be detected specifically includes: Obtain the original images of the wafers to be detected scanned by each machine tool; Call the camera calibration parameters of each machine tool, and determine the image correction matrix according to the camera calibration parameters, where the camera calibration parameters include the camera orthogonal angle and the flat field coefficient; Call the camera parameters of each machine tool, where the camera parameters include the camera resolution of each machine tool, the lens magnification of each machine tool, the distortion coefficient of each machine tool, and the camera size of each machine tool; Correct the original image according to the camera parameters and the image correction matrix so that the size of each pixel in the corrected original image is the same, and obtain the surface image; Obtain at least the surface image of one wafer to be detected among the wafers to be detected.

7. The method according to claim 1, characterized in that, The step of comparing each die image with the template image one by one to implement defect detection specifically includes: For each die image, determine the gray difference image between the die image and the template image; According to the gray difference image, determine the detection results of the pixels in the gray difference image through a preset gray detection range; Determine the defect detection result according to the detection results of the pixels in the gray difference image.

8. The method according to claim 7, characterized in that, The step of determining the detection results of the pixels in the gray difference image through a preset gray detection range according to the gray difference image specifically includes: For each pixel in the gray difference image, determine whether the gray difference of the pixel falls within the preset gray detection range; If so, determine that the pixel has no defect as the detection result of the pixel; If not, determine the defect type of the pixel according to the size relationship between the gray difference of the pixel in the gray difference image and the gray detection range, and determine the detection result of the pixel according to the defect type of the pixel.

9. The method according to claim 8, wherein Before determining whether the gray difference of each pixel in the gray difference image falls within the preset gray detection range, the method further includes: Determine the gradient image of the template image; For each pixel in the gray difference image, determine the gradient value corresponding to the pixel in the gradient image according to the position of the pixel in the gray difference image, and determine the gray detection range of the pixel according to the gradient value and the preset gray detection range.

10. The method according to claim 9, wherein The step of determining the defect type of the pixel according to the size relationship between the gray value of the pixel in the gray difference image and the gray detection range specifically includes: When the gray difference of the pixel in the gray difference image is less than the gray detection range, the defect type of the pixel is a dark defect; When the gray difference of the pixel in the gray difference image is greater than the gray detection range, the defect type of the pixel is a bright defect.

11. The method according to claim 8, characterized in that, The step of determining the defect detection result according to the detection results of each pixel in the grayscale difference image specifically includes: Mark the defect regions in the grain image according to the detection results of each pixel in the grayscale difference image, and determine the defect types of the defect regions; Determine the image features of the defect regions; According to the image features of the defect regions and the defect types of the defect regions, determine the process defect categories to which the defect regions belong as the defect detection results.

12. A wafer detection method, characterized in that, Including: Obtain the surface image of the wafer to be detected, wherein the surface image includes the grain images of a plurality of grains; Select one grain image from each grain image as the reference image, and use the other grain images as the test images; Determine the image matching positions aligned with the reference image in each test image, and align and fuse the test images to obtain a template image; Compare each grain image one by one according to the template image to achieve defect detection; Among them, determining the image matching position includes: First matching: Determine a first marking point in the reference image, and determine the positioning point aligned with the first marking point in the test image through image matching as the image matching position of the test image; If the first matching fails, perform the second matching: Continuously reduce the reference image and the test image in the same proportion, perform image matching between the reduced images, and determine the image matching position in the test image that is aligned with the reference image; Determining a first marking point in the reference image and determining the positioning point aligned with the first marking point in the test image through image matching includes: According to the position of the first marking point in the reference image and the preset position relationship between the marking point and the grain, calibrate a first window image in the reference image; For each test image, based on the surface image of the wafer to be detected, determine the distance deviation between the test image and the reference image, and according to the distance deviation and the pixel size of the test image, determine the pixel deviation between the test image and the reference image. According to the pixel deviation and the position of the first marking point in the reference image, determine the position of the second marking point in the test image; According to the preset size of the first window image and the deviation of the machine movement and camera calibration, determine the size of the second window image; Perform image matching between the second window image and the first window image, determine the area in the test image that matches the first window image as the target area, and in the target area, determine the positioning point in the test image that is aligned with the first marking point, wherein the second window image is larger than the first window image, and the position relationship of the first marking point in the first window image is the same as the preset position relationship between the marking point and the grain, so as to ensure that more grain parts are included in the first window image, and the marking points in the grains are specific marks or test points provided for reference and positioning on a single grain, rather than feature points.

13. A wafer inspection device, characterized in that, Including: A first acquisition module, configured to acquire a surface image of a wafer to be detected, wherein the surface image includes grain images of a plurality of grains; A first selection module, configured to select one grain image from the grain images as a reference image, and use the other grain images as test images; A positioning point module, configured to calibrate a first window image in the reference image according to the position of a first marking point in the reference image and a preset positional relationship between the marking point and the grain. For each test image, based on the surface image of the wafer to be detected, determine the distance deviation between the test image and the reference image, and determine the pixel deviation between the test image and the reference image according to the distance deviation and the pixel size of the test image. According to the pixel deviation and the position of the first marking point in the reference image, determine the position of a second marking point in the test image; determine the size of a second window image according to the preset size of the first window image and the deviations of the machine stage movement and camera calibration; perform image matching between the second window image and the first window image, and determine, in the test image, a region that matches the first window image as a target region, and determine, in the target region, a positioning point in the test image that aligns with the first marking point, wherein the second window image is larger than the first window image, and the positional relationship of the first marking point in the first window image is the same as the preset positional relationship between the marking point and the grain, so as to ensure that more grain parts are included in the first window image, and the marking points in the grains are specific marks or test points provided on a single grain for reference and positioning, rather than feature points; A template image module, configured to align the test images according to the positioning points in the test images, and fuse them to obtain a template image; A first detection module, configured to compare each grain image one by one according to the template image to implement defect detection.

14. A wafer inspection device, characterized in that, Including: A second acquisition module, configured to acquire a surface image of a wafer to be detected, wherein the surface image includes grain images of a plurality of grains; A second selection module, configured to select one grain image from the grain images as a reference image, and use the other grain images as test images; An image matching module, configured to determine an image matching position aligned with the reference image in each test image, and align and fuse the test images to obtain a template image; wherein, determining the image matching position includes: the first matching: determining a first marker point in the reference image, and determining a positioning point aligned with the first marker point in the test image through image matching as the image matching position of the test image; if the first matching fails, perform the second matching: continuously reduce the reference image and the test image in the same proportion, perform image matching between the reduced images, and determine the image matching position in the test image that is aligned with the reference image. The first matching includes: according to the position of the first marker point in the reference image and the preset position relationship between the marker point and the crystal grain, calibrating a first window image in the reference image; for each test image, based on the surface image of the wafer to be detected, determining the distance deviation between the test image and the reference image, and according to the distance deviation and the pixel size of the test image, determining the pixel deviation between the test image and the reference image, and according to the pixel deviation and the position of the first marker point in the reference image, determining the position of the second marker point in the test image; according to the preset size of the first window image and the deviation of the machine movement and camera calibration, determining the size of the second window image; performing image matching between the second window image and the first window image, determining the area that matches the first window image in the test image as the target area, and in the target area, determining the positioning point in the test image that is aligned with the first marker point. Wherein, the second window image is larger than the first window image, and the position relationship of the first marker point in the first window image is the same as the preset position relationship between the marker point and the crystal grain, so as to ensure that more crystal grain parts are included in the first window image. The marker point in the crystal grain is a specific marker or test point that provides reference and positioning on a single crystal grain, rather than a feature point; A second defect detection module, configured to compare each crystal grain image one by one according to the template image to achieve defect detection.

15. A computer-readable storage medium, characterized in that, The storage medium has a computer program, and when the computer program is executed by a processor, the method described in any one of the above claims 1 to 12 is implemented.

16. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that when the processor executes the program, the method described in any one of the above claims 1 to 12 is implemented.

Images