A wafer detection method, device, storage medium and equipment

Abstract

In a wafer inspection method, device, storage medium and equipment provided in this specification, each grain image of the wafer to be inspected is obtained, and a reference image is selected from the images, and other grain images are used as test images. By continuously reducing the reference image and the test image in the same proportion, an image pyramid corresponding to the reference image and the test image is obtained. By image matching between the image pyramids, the image matching position of the test image and the reference image is determined. According to the image matching position of each test image, each test image is aligned and fused to obtain a template image, and defect detection is performed on each grain image based on the template image. By image matching of the image pyramid, the image matching position of the test image is determined, and then each test image is aligned and a template image is obtained. The cost is low and the efficiency is high, and the template image is a defect-free grain image fused based on each grain image collected in actual production, which improves the accuracy of defect detection.

Description

Technical Field

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

[0002] Defect detection on wafers can promptly discover and solve problems in the semiconductor chip production process, and is an important part of improving the yield rate of semiconductor chips.

[0003] At present, wafers are generally inspected through an automated optical inspection system (AOI), and the grain images with fewer defects are manually screened out from each grain image as a template image, and then the grain (Die) image in the wafer to be inspected is matched with the template image to determine the inspection result of the grain image. However, with the update of semiconductor chip production technology, manual repeated screening of template images is required, which is costly and inefficient, and the template image is not an ideal defect-free grain image, resulting in poor accuracy in determining the inspection result.

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

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

[0006] This manual adopts the following technical solutions:

[0007] This specification provides a wafer detection method, the method comprising:

[0008] Acquire a surface image of a wafer to be inspected, wherein the surface image includes grain images of a plurality of grains;

[0009] Select one grain image from each grain image as a reference image and use the other grain images as test images;

[0010] By continuously reducing the reference image and the test image at the same ratio, an image pyramid of the reference image and an image pyramid of each test image are obtained;

[0011] For each test image, determining an image matching position in the test image that is aligned with the reference image based on an image pyramid of the test image and an image match between the image pyramid of the reference image;

[0012] According to the image matching positions of the test images, the test images are aligned and fused to obtain a template image;

[0013] The grain images are compared one by one according to the template image to achieve defect detection.

[0014] Optionally, the step of determining, for each test image, an image matching position in the test image aligned with the reference image based on an image pyramid of the test image and an image matching between the image pyramid of the reference image, specifically includes:

[0015] For each test image, according to the image pyramid of the test image and the image pyramid of the reference image, a residual matrix between the same layers of the two image pyramids is determined, and according to the residual matrix, a first matching position of each layer in the image pyramid of the test image is obtained;

[0016] An image matching position in the test image that is aligned with the reference image is determined according to the first matching positions of each layer.

[0017] Optionally, the step of determining, for each test image, an image matching position in the test image aligned with the reference image based on an image pyramid of the test image and an image matching between the image pyramid of the reference image, specifically includes:

[0018] For each test image, determine a residual matrix according to a top image of an image pyramid of the test image and a top image of an image pyramid of the reference image;

[0019] According to the residual matrix, a second matching position of the test image at the top layer is obtained;

[0020] For each layer of the image pyramid of the test image, determine the image matching area of ​​the layer of the test image according to the second matching position of the test image in the layer above the layer, and the deviation of the machine motion and the camera calibration;

[0021] Determine a residual matrix of the image matching area according to the image of the layer in the image pyramid of the test image and the image corresponding to the layer in the image pyramid of the reference image, and obtain a second matching position of the test image in the layer according to the residual matrix;

[0022] The second matching position of the test image at the bottom layer in the image pyramid is used as the image matching position in the test image that is aligned with the reference image.

[0023] Optionally, the step of acquiring the surface image of the wafer to be inspected specifically includes:

[0024] Obtaining the original images of each wafer to be inspected obtained by scanning by each machine;

[0025] Calling the camera calibration parameters of each machine, wherein the calibration parameters include the camera orthogonal angle and the flat field coefficient;

[0026] Determining an image correction matrix according to the calibration parameters;

[0027] Calling the camera parameters of each machine, wherein the camera parameters include the camera resolution of each machine, the magnification of each machine, the distortion coefficient of each machine, and the camera size of each machine;

[0028] Correcting 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 consistent, thereby obtaining a surface image;

[0029] A surface image of at least one wafer to be inspected among the wafers to be inspected is obtained.

[0030] Optionally, the step of comparing each grain image one by one according to the template image to realize defect detection specifically includes:

[0031] For each grain image, determining a grayscale difference image between the grain image and the template image;

[0032] According to the grayscale difference image, determining the detection result of each pixel in the grayscale difference image through a preset grayscale detection range;

[0033] The defect detection result of the grain is determined according to the detection result of each pixel in the grayscale difference image.

[0034] Optionally, the step of determining the detection result of each pixel in the grayscale difference image according to a preset grayscale detection range specifically includes:

[0035] Determining a gradient image of the template image;

[0036] For each pixel in the grayscale difference image, determine the gradient value corresponding to 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 a preset grayscale detection range;

[0037] Determine whether the grayscale difference of the pixel falls within the grayscale detection range of the pixel;

[0038] If yes, the pixel is considered defect-free as the detection result of the pixel;

[0039] If not, the defect type of the pixel is determined based on the grayscale difference of the pixel in the grayscale difference image and the size relationship of the grayscale detection range, and the detection result of the pixel is determined based on the defect type of the pixel.

[0040] Optionally, the step of determining the defect detection result of the grain according to the detection result of each pixel in the grayscale difference image specifically includes:

[0041] Marking a defective area in the grain image according to a detection result of each pixel in the grayscale difference image, and determining a defect type of the defective area;

[0042] determining image features of the defective area;

[0043] According to the image features of the defect area and the defect type of the defect area, the process defect category to which the defect area belongs is determined as the defect detection result of the die.

[0044] This specification provides a wafer inspection device, including:

[0045] An acquisition module, used for acquiring a surface image of a wafer to be inspected, wherein the surface image includes grain images of a plurality of grains;

[0046] A selection module is used to select one grain image from each grain image as a reference image and use other grain images as test images;

[0047] A pyramid module, used for obtaining an image pyramid of the reference image and an image pyramid of each test image by continuously reducing the reference image and the test image at the same ratio;

[0048] An image matching module, for each test image, determining an image matching position in the test image that is aligned with the reference image based on an image pyramid of the test image and an image match between the image pyramids of the reference image;

[0049] A template image module, used to align the test images according to the image matching positions of the test images, and fuse them to obtain a template image;

[0050] The defect detection module is used to compare the grain images one by one according to the template image to achieve defect detection.

[0051] This specification provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, a wafer detection method is implemented.

[0052] The present specification provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a wafer detection method when executing the program.

[0053] At least one of the above-mentioned technical solutions adopted in this specification can achieve the following beneficial effects: in a wafer inspection method provided in this specification, a surface image of a wafer to be inspected is obtained, and from each grain image of the surface image, a grain image is selected as a reference image, and other grain images are used as test images, and the image pyramid of the reference image and the image pyramid of the test image are obtained by continuously reducing the reference image and the test image at the same ratio, and then for each test image, according to the image matching between the image pyramid of the test image and the image pyramid of the reference image, the image matching position where the test image is aligned with the reference image is determined, and according to the image matching position of each test image, each test image is aligned and fused to obtain a template image, and then defect detection is performed on each grain image based on the template image.

[0054] It can be seen from the above method that the image matching position of the test image is determined by image matching between the image pyramid of the test image and the image pyramid of the reference image, and then the test images are aligned and fused to obtain a template image. This method has low cost and high efficiency, and the template image is a defect-free grain image fused based on the images of each grain collected in actual production, which improves the accuracy of defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The illustrative embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification. In the drawings:

[0056] Figure 1 A schematic diagram of a wafer inspection method provided in this specification;

[0057] Figure 2 A schematic diagram of determining an image matching position provided in this specification;

[0058] Figure 3 A schematic diagram of determining a residual matrix based on a reference point provided in this specification;

[0059] Figure 4 A schematic diagram of a wafer inspection device provided in this specification;

[0060] Figure 5 A schematic diagram of the electronic device structure corresponding to a wafer detection method provided in this specification. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0062] The process of executing a wafer inspection method in this specification involves processing image data. Therefore, in the embodiments of this specification, the wafer inspection process can be executed by a server. Of course, this specification does not limit which device executes the wafer inspection process. For example, a personal computer, a mobile terminal and other devices perform wafer inspection, or in actual production, the wafer inspection process can also be executed by a machine. For the convenience of description, the following description is based on the server as the execution subject.

[0063] At present, AOI is generally used to detect defects on wafers. AOI is applied to defect detection in advanced packaging. Its working principle is to capture the grain image through a camera system, such as a charge coupled device (CCD) camera system, and after a series of algorithm processing such as an image processing card and a computer software system, it is compared with the template image to find defects and generate a report. However, with the update of semiconductor chip production technology, manual repeated screening of template images is required, which is costly and inefficient, and the template image is not an ideal defect-free grain image, resulting in poor accuracy in determining the detection results.

[0064] Based on this, this specification provides a wafer inspection method. In actual production, the process of the wafer inspection method can be executed by a machine. The machine includes a high-resolution CCD camera, an optical lens and a multi-angle color light source, that is, the machine has different lighting modes such as bright field and dark field, as well as adjustable wavelength and brightness. At the same time, a high-precision, high-speed motion platform cooperates with the camera to synchronously scan and acquire wafer images. While meeting various wafer surface defect detection needs, the template image can also be determined through image matching based on the collected images of each grain to reduce labor costs and improve efficiency.

[0065] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0066] Figure 1 A schematic diagram of a wafer inspection method provided in this specification includes the following steps:

[0067] S100: Acquire a surface image of a wafer to be inspected, wherein the surface image includes grain images of a plurality of grains.

[0068] In one or more embodiments of this specification, it is not limited to which specific device is used to implement a wafer detection method, such as a personal computer, a mobile terminal, and a server. However, since the subsequent steps involve operations such as data processing, and such operations with high requirements for computing resources are generally performed by a server, this specification will also be described as an example of a server implementing a wafer detection method. Among them, the server can be a single device, or composed of multiple devices, such as a distributed server, and this specification does not limit this.

[0069] In one or more embodiments of this specification, a wafer is a basic material in semiconductor manufacturing, usually a round sheet made of silicon, and a die refers to a single integrated circuit unit formed on a wafer through photolithography and other process steps during wafer processing. In order to determine the reference image and the test image in subsequent steps, in this step, the server needs to obtain the surface image of the wafer to be tested.

[0070] Specifically, the server can obtain a surface image of the wafer to be inspected, which includes grain images of several grains. This specification does not limit the method of obtaining the surface image, such as a scanning electron microscope, an optical microscope, etc. The surface image can also be captured by a camera system, and then the grain image can be processed by a series of algorithms such as an image processing card and a computer software system.

[0071] S102: selecting one grain image from each grain image as a reference image, and using the other grain images as test images.

[0072] In one or more embodiments of the present specification, in order to determine the image pyramid of the reference image and the image pyramid of the test image in subsequent steps, and then determine the image matching positions where each test image is aligned with the reference image, in this step, a reference image needs to be selected from each grain image as the target of image matching.

[0073] Specifically, the server may select one grain image from each grain image obtained in step S100 as a reference image, and use other grain images as test images. Among them, the technology of segmenting each grain image from the surface image may be performed by edge detection or other mature technologies in the art. The specific method of selecting the reference image is not limited in this specification, such as random selection, selecting the first segmented grain image, etc.

[0074] S104: Obtaining an image pyramid of the reference image and an image pyramid of each test image by continuously reducing the reference image and the test image at the same ratio.

[0075] In one or more embodiments of the present specification, in order to determine the image matching position where each test image is aligned with the reference image in the subsequent steps, in this step, the server needs to continuously reduce the reference image and each test image by the same proportion to obtain the image pyramid of the reference image and the image pyramid of each test image.

[0076] Specifically, the server can obtain the image pyramid of the reference image and the image pyramid of each test image by continuously reducing the reference image and each test image at the same ratio. The specific method of "continuously reducing at the same ratio" is not limited in this specification. For example, the images of the bottom layer and the top layer of the image pyramid can be reduced recursively in a way of accumulating powers of fixed multiples, as follows:

[0077] K=M k

[0078] Where K is the reduction ratio, M is a fixed multiple, and k is a non-negative integer. The reduction ratio of the bottom layer of the image pyramid is M 0 , then the reduction ratio of the second-to-last layer is M 1 , so as to reduce the scale when determining the image pyramid upwards.

[0079] S106: For each test image, determine an image matching position in the test image that is aligned with the reference image according to the image pyramid of the test image and the image matching between the image pyramid of the reference image.

[0080] In one or more embodiments of the present specification, due to deviations in machine motion and camera calibration, and in order to align the test images in subsequent steps and fuse them to obtain a template image, in this step, the server needs to determine, for each test image, an image matching position where the test image is aligned with the reference image based on the image pyramid of the test image and the image matching between the image pyramid of the reference image.

[0081] Specifically, the server may determine the residual matrix between the same layers of the two image pyramids for each test image according to the image pyramid of the test image and the image pyramid of the reference image. For each layer in the image pyramid of the test image, the image matching position of the layer is obtained according to the residual matrix of the layer. The image matching position of the test image aligned with the reference image is determined according to the first matching position of each layer in the image pyramid of the test image. The specific method of "determining the image matching position according to the first matching position of each layer" is not limited in this specification, such as taking the average value, median, etc. of each first matching position. At the same time, after determining the image matching position where the test image is aligned with the reference image, it is also necessary to reversely enlarge the first matching position of each layer of the image pyramid according to the reduction ratio of the layer, so as to obtain the image matching position of the original size where the test image is aligned with the reference image based on the enlarged first matching positions.

[0082] Determine the residual matrix between the same layers of the two image pyramids as follows:

[0083] First, for each layer of the image pyramid, the server slides the reference image of the layer on the test image of the layer to determine each sliding position.

[0084] Secondly, the server may determine the sum of the residuals of each pixel in the overlapped portion between the test image of the layer and the reference image of the layer at each sliding position, and use the ratio of the sum of the residuals to the number of pixels in the overlapped portion as the normalized residual, wherein the residual is the sum of the absolute values ​​of the differences between the pixel values ​​of each pixel in the overlapped portion.

[0085] Finally, the server can obtain the residual matrix of the test image of the layer according to the normalized residuals of each sliding position.

[0086] It should be noted that the similarity between the test image and the reference image at each sliding position is measured by determining the normalized residual. Since the normalized residual characterizes the similarity between the test image and the reference image, and the similarity is negatively correlated with the normalized residual. Therefore, the sliding position corresponding to the minimum normalized residual (maximum similarity) in the residual matrix can be determined as the first matching position of the test image. Of course, since the image resolution of each layer in the image pyramid is different and the accuracy is limited, the specific method of "determining the first matching position according to the residual matrix" is not limited in this specification. For example, each normalized residual below a preset threshold can also be determined, and the first matching position can be randomly selected from the sliding positions corresponding to each normalized residual.

[0087] S108: Aligning the test images according to the image matching positions of the test images, and fusing them to obtain a template image.

[0088] In one or more embodiments of the present specification, in order to compare each grain image one by one in the subsequent steps, in this step, the server needs to align each test image according to the image matching position of each test image, and fuse them to obtain a template image.

[0089] Specifically, the server may align each test image with the reference image according to the image matching position of the test image, and fuse them to obtain a template image.

[0090] Optionally, the server may determine the image matching position of each test image aligned with the reference image as the image matching position of each test image aligned with the reference image, thereby aligning the test images and fusing them to obtain the template image. The specific method of aligning the test images is not limited, for example, the test images are sorted, and for each test image, the image matching position of the subsequent test image aligned with the test image is determined based on the image matching position of the test image aligned with the reference image and the image matching position of the subsequent test image aligned with the reference image.

[0091] It should be noted that generally, a large number of grain images are involved in the fusion, and for each pixel, the defective grayscale value is small. Therefore, even if defective grain images are involved in the fusion, the obtained template image is close to defect-free and can be used for defect detection in subsequent steps.

[0092] At the same time, the specific method of fusing the aligned test images to obtain the template image is not limited in this specification. For example, the defective grayscale value of each pixel in each test image can be eliminated for each pixel, thereby further improving the accuracy of the template image. In one or more embodiments of this specification, for each test image, the test image is aligned with the reference image according to the image matching position of the test image, and for each pixel, the normal value detection range is determined according to the grayscale value corresponding to the pixel in each test image, and the average grayscale value of the pixel is determined according to the grayscale value corresponding to the pixel in the normal value detection range, and the template image is determined according to the average grayscale value of each pixel. Among them, the mean and standard deviation of the grayscale value of the pixel in each test image can be determined, and the normal value detection range is determined according to the mean and standard deviation. If the mean is Mean and the standard deviation is std, the normal value detection range can be (Mean-n*std)~(Mean+n*std), n is a positive integer, which can be set according to actual needs. Of course, the specific method of determining the normal value detection range is not limited in this specification. In one or more embodiments of the present specification, for each test image, the test image is aligned with the reference image according to the image matching position of the test image, and for each pixel, the probability distribution of the grayscale value of the pixel in each test image is determined, and the grayscale values ​​with a probability lower than a preset threshold are removed, and the average grayscale value of the pixel is determined according to each grayscale value in the probability distribution with a probability greater than the preset threshold, and the template image is determined according to the average grayscale value of each pixel. The preset threshold can be set according to actual needs. Of course, when the template image is fused, the reference image can also participate in the alignment of each test image.

[0093] S110: comparing the grain images one by one according to the template image to achieve defect detection.

[0094] In one or more embodiments of the present specification, in this step, defect detection may be performed on the grain image of each grain in the wafer to be inspected according to the template image determined in step S106.

[0095] Specifically, first, the server may determine, for each grain image of the grain, a grayscale difference image between the grain image and the template image.

[0096] Secondly, for each pixel in the grayscale difference image, determine whether the grayscale difference of the pixel falls within the preset grayscale detection range. If the grayscale difference of the pixel falls within the preset grayscale detection range, the pixel is considered to be defect-free as the detection result of the pixel. If the grayscale difference of the pixel does not fall within the preset grayscale detection range, the defect type of the pixel is determined, and then the detection result of the pixel is determined based on the defect type.

[0097] Finally, based on the detection results of each pixel in the grayscale difference image, mark the defect area in the grain image, and determine the defect type and image features of the defect area. According to the defect type and image features of the defect area, determine the process defect category to which the defect area belongs, and determine the defect detection result of the grain according to the process defect category of each defect area. Among them, the image features include the shape, length, width, area, etc. of the defect area. The process defect category can be divided into large defects, scratches, etc. For example, when the length of the defect area exceeds the preset threshold, the defect area is a large defect. According to the size relationship between the grayscale difference of the pixel and the grayscale detection range, determine the defect type of the pixel, as follows:

[0098] When the grayscale difference of the pixel in the grayscale difference image is less than the preset 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 preset grayscale detection range, the defect type of the pixel is a bright defect.

[0099] It should be noted that the grayscale value range is 0 to 255. When determining the grayscale difference image, in order to facilitate the observation of the defects of each pixel, the grain image and the template image can be subtracted, and then the grayscale value of each pixel obtained after the subtraction is added with a fixed grayscale value (such as 128, which is only used for example and is not a limitation on the fixed grayscale value). If the grayscale difference after adding the grayscale value of each pixel to the fixed grayscale value exceeds the grayscale value range, the endpoint grayscale value adjacent to the grayscale value of the pixel can be selected in the grayscale value range as the grayscale difference of the pixel, and then the grayscale difference image of the grain can be obtained based on the grayscale difference of each pixel.

[0100] In the above method, for each test image, the image matching position is determined according to the image matching between the image pyramid of the test image and the image pyramid of the reference image, and then based on the image matching position of each test image and the reference image, each test image is aligned and fused to obtain a template image, which has low cost and high efficiency. Moreover, the template image is a defect-free grain image fused based on each grain image collected in actual production. Therefore, when defect detection is performed on each grain image according to the template image, the accuracy is high.

[0101] Specific reference Figure 2 As shown, the image pyramid of the test image and the image pyramid of the reference image are determined by continuously scaling down the image pyramid at the same ratio as provided in this specification, and the first matching position of each layer is obtained by image matching between the same layers of the image pyramid, and then the image matching position for aligning the test image with the reference image is determined according to each first matching position.

[0102] In addition, for the process of determining the residual matrix in step S106, this specification provides determining a reference area in the reference image, and based on the reference area, sliding the reference image on the test image to determine the residual matrix, as follows:

[0103] In one or more embodiments of the present specification, the server may determine a reference area in the reference image, and by sliding the reference area on the test image, determine the "overlapping portion of the reference image and the test image" corresponding to each sliding position, and the normalized residual of the overlapping portion at each sliding position, thereby obtaining a residual matrix. The specific scope of the reference area is not limited in the present specification. In a preferred embodiment, the reference area may be determined based on the "deviation of the machine motion and camera calibration", and the reference area is the central area of ​​the reference image, wherein the machine motion and camera calibration cause a deviation between the test image and the reference image, but the deviation does not cause the "reference area" to deviate from the "overlapping portion when the test image and the reference image are aligned".

[0104] In the above, the reference area of ​​the reference image is determined, and based on the reference area, the reference image is slid on the test image to obtain a residual matrix, which reduces the amount of calculation of the normalized residuals of each sliding position of "the reference area is outside the test image" in step S106, and further improves the efficiency of subsequent determination of the image matching position.

[0105] Specific reference Figure 3 As shown, taking the reference area at the center of the reference image as an example, based on the reference area, the reference image is slid on the test image to determine the normalized residuals of the "overlapping part of the reference image and the test image" corresponding to each sliding position, thereby obtaining the "number of pixels of the test image" normalized residuals, thereby determining the residual matrix. Among them, this figure is only a schematic diagram, and this specification does not limit the method of sliding the reference image on the test image based on the reference area to only the sliding path in the figure, which can be set according to actual needs, such as starting to slide from the center of the test image.

[0106] In addition, for the process of determining the image matching position in step S106, this specification provides for each layer of the image pyramid of the test image, according to the second matching position of the previous layer of the layer, and the deviation of the machine motion and camera calibration, to determine the image matching area of ​​the test image in the layer, and then to match the image matching areas in the layer, specifically as follows:

[0107] In one or more embodiments of the present specification, first, the server may determine the residual matrix for each test image according to the top image of the image pyramid of the test image and the top image of the image pyramid of the reference image, and obtain the second matching position of the test image at the top layer according to the residual matrix. Since the "top layer" has no upper layer, the second matching position of the top layer is consistent with the first matching position of the top layer in S106. At the same time, the specific process of determining the residual matrix refers to step S106 and will not be repeated here.

[0108] Secondly, the server can determine the image matching area of ​​the test image in each layer of the image pyramid of the test image in turn according to the second matching position of the test image in the previous layer of the layer and the preset area range. The specific value of the area range is not limited in this specification. In a preferred embodiment, the area range can be "deviation of machine motion and camera calibration". At the same time, the area range of each layer can also be different and can be set according to needs.

[0109] Finally, the server can obtain the second matching position of the test image in each layer of the image pyramid according to the test image of the layer and the reference image of the layer, through the residual matrix of the reference image in the image matching area of ​​the "test image of the layer", and according to the residual matrix. The second matching position of the test image in the bottom layer of the image pyramid is used as the image matching position in the test image that is aligned with the reference image. The determination of the residual matrix is ​​not repeated here, and since the second matching position of the bottom layer has been combined with the second matching positions of each layer, the second matching position determined in the bottom layer can be used as the image matching position of the test image.

[0110] It should be noted that in S106, the determination of the residual matrix of each layer requires the sliding of the reference image on the "complete" test image. Therefore, in the above, for each layer of the image pyramid, the image matching area of ​​the test image of this layer is determined based on the "second matching position" determined by the previous layer of this layer. While combining the "second matching position" of the previous layer, it is only necessary to slide the reference image within the image matching area of ​​the test image, which improves efficiency.

[0111] In addition, when the surface image of the wafer to be inspected is obtained in step S100, the distortion of the image and the differences in camera resolution and lens magnification between the machines can be corrected based on the camera calibration parameters of each machine, so that the actual size represented by the images collected by each machine is consistent. In this way, if the wafers to be inspected are of the same product, that is, the wafers to be inspected have the same pattern, a template image obtained based on the surface image of a wafer to be inspected can be used as the defect detection of the grain image in each wafer to be inspected, thereby reducing the workload of repeatedly determining the template image.

[0112] In one or more embodiments of the present specification, the server can obtain the original image of each wafer to be inspected obtained by scanning each machine, and call the camera calibration parameters of each machine, and determine the image correction matrix according to the camera calibration parameters, wherein the camera calibration parameters include the camera orthogonal angle and the flat field coefficient. The camera parameters of each machine are called, and the camera parameters include the camera resolution of each machine, the lens magnification of each machine, the distortion coefficient of each machine, and the camera size of each machine. Then, based on the camera parameters and the image correction matrix, the original image of each wafer to be inspected is corrected so that the size of each pixel in the corrected original image is consistent, and a surface image is obtained. At least the surface image of one wafer to be inspected among the wafers to be inspected is obtained for determining the template image in the subsequent steps.

[0113] It should be noted that in actual production, since the machine includes a high-resolution CCD camera, an optical lens, and a multi-angle color light source, that is, the machine has different lighting modes such as bright field and dark field, and the wavelength and brightness are adjustable. Therefore, in a preferred embodiment, when acquiring the surface image of each wafer, it is necessary to collect it under the same lighting mode, the same wavelength, and the same brightness.

[0114] In addition, in step S110, when performing defect detection on each pixel in the grayscale difference image based on the preset grayscale detection range, since there is a gradient change relationship between each pixel in the template image, the server can determine the grayscale detection range of each pixel based on the gradient value of each pixel in the gradient image of the template image, as follows:

[0115] In one or more embodiments of the present specification, the server may determine the gradient image of the template image in step S108, and for each test image, determine the grayscale difference image between the test image and the template image, and then for each pixel in the grayscale difference image, determine the gradient value corresponding to the pixel in the gradient image of the template image according to the position of the pixel in the grayscale difference image, and determine the grayscale detection range of the pixel based on the gradient value and a preset grayscale detection range.

[0116] 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:

[0117] The preset grayscale detection range is: G dark ~G bright

[0118] Determine the gradient image G of the template image grad , then determine the gradient value G corresponding to the pixel in the gradient image grad(x, y), based on the gradient value and the preset grayscale detection range, the grayscale detection range of the pixel is determined to be G dark -k*G grad (x,y)~G bright +k*G grad (x,y).

[0119] Among them, G dark is the preset dark defect threshold, G bright is the preset bright defect threshold, and k is the gradient coefficient, which can be set according to actual needs.

[0120] 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, such as Figure 4 shown.

[0121] An acquisition module 500 is configured to acquire a surface image of a wafer to be inspected, wherein the surface image includes grain images of a plurality of grains;

[0122] A selection module 501 selects one grain image from each grain image as a reference image and uses other grain images as test images;

[0123] A pyramid module 502, which obtains an image pyramid of the reference image and an image pyramid of each test image by continuously reducing the reference image and the test image by the same proportion;

[0124] An image matching module 503 determines, for each test image, an image matching position in the test image that is aligned with the reference image based on an image pyramid of the test image and an image matching between the image pyramids of the reference image;

[0125] The template image module 504 aligns the test images according to the image matching positions of the test images, and fuses them to obtain a template image;

[0126] The defect detection module 505 compares the grain images one by one according to the template image to achieve defect detection.

[0127] Optionally, the image matching module 503 is specifically used to determine, for each test image, a residual matrix between the same layers of the two image pyramids according to the image pyramid of the test image and the image pyramid of the reference image, and obtain, according to the residual matrix, a first matching position of each layer in the image pyramid of the test image; and determine, according to the first matching position of each layer, an image matching position in the test image that is aligned with the reference image.

[0128] Optionally, the image matching module 503 is further used to determine, for each test image, a residual matrix according to a top image of an image pyramid of the test image and a top image of an image pyramid of the reference image; obtain a second matching position of the test image at the top layer according to the residual matrix; determine, in sequence, for each layer of the image pyramid of the test image, an image matching region of the layer of the test image according to the second matching position of the test image in a layer above the layer and deviations of machine motion and camera calibration; determine a residual matrix of the image matching region according to the image of the layer in the image pyramid of the test image and an image corresponding to the layer in the image pyramid of the reference image, and obtain a second matching position of the test image in the layer according to the residual matrix; and use the second matching position of the test image at the bottom layer in the image pyramid as the image matching position in the test image aligned with the reference image.

[0129] Optionally, the acquisition module 500 is specifically used to acquire the original image of each wafer to be inspected obtained by scanning each machine; call the camera calibration parameters of each machine, wherein the calibration parameters include the camera orthogonal angle and the flat field coefficient; determine the image correction matrix according to the calibration parameters; call the camera parameters of each machine, wherein the camera parameters include the camera resolution of each machine, the magnification of each machine, the distortion coefficient of each machine and the camera size of each machine; 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 consistent, and obtain a surface image; and obtain the surface image of at least one wafer to be inspected among the wafers to be inspected.

[0130] Optionally, the defect detection module 505 is specifically used to determine, for each grain image, a grayscale difference image between the grain image and the template image; based on the grayscale difference image, determine the detection result of each pixel in the grayscale difference image through a preset grayscale detection range; based on the detection result of each pixel in the grayscale difference image, determine the defect detection result of the grain.

[0131] Optionally, the defect detection module 505 is specifically used to determine the gradient image of the template image; for each pixel in the grayscale difference image, determine the gradient value corresponding to 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 a preset grayscale detection range; judge whether the grayscale difference of the pixel falls within the grayscale detection range of the pixel; if so, the pixel is considered to be defect-free as the detection result of the pixel; if not, determine the defect type of the pixel according to the grayscale difference of the pixel in the grayscale difference image and the size relationship of the grayscale detection range, and determine the detection result of the pixel according to the defect type of the pixel.

[0132] Optionally, the defect detection module 505 is specifically used to mark the defect area in the grain image according to the detection results of each pixel 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 of the grain.

[0133] This specification also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 A wafer inspection method is provided.

[0134] This manual also provides Figure 5 The schematic structure diagram of the electronic device shown in FIG. Figure 5 As mentioned above, at the hardware level, the device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 Of course, in addition to the software implementation, this specification does not exclude other implementations, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0135] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using 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 logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, 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 are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0136] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers 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 a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0137] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may 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 a combination of any of these devices.

[0138] For the convenience of description, the above device is described in 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.

[0139] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0140] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0141] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0142] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0143] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0144] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0145] 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.

[0146] 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.

[0147] 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.

[0148] 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.

[0149] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. A wafer detection method, characterized in that: include: Acquire a surface image of a wafer to be inspected, wherein the surface image includes grain images of a plurality of grains; Select one grain image from each grain image as a reference image and use the other grain images as test images; By continuously reducing the reference image and the test image at the same ratio, an image pyramid of the reference image and an image pyramid of each test image are obtained; For each test image, determining an image matching position in the test image that is aligned with the reference image based on an image pyramid of the test image and an image match between the image pyramid of the reference image; According to the image matching positions of the test images, the test images are aligned and fused to obtain a template image; The grain images are compared one by one according to the template image to achieve defect detection.

2. The method according to claim 1, characterized in that: The step of determining, for each test image, an image matching position in the test image that is aligned with the reference image based on the image matching between the image pyramid of the test image and the image pyramid of the reference image, specifically comprises: For each test image, according to the image pyramid of the test image and the image pyramid of the reference image, a residual matrix between the same layers of the two image pyramids is determined, and according to the residual matrix, a first matching position of each layer in the image pyramid of the test image is obtained; An image matching position in the test image that is aligned with the reference image is determined according to the first matching positions of each layer.

3. The method according to claim 1, characterized in that The step of determining, for each test image, an image matching position in the test image that is aligned with the reference image based on the image matching between the image pyramid of the test image and the image pyramid of the reference image, specifically comprises: For each test image, determine a residual matrix according to a top image of an image pyramid of the test image and a top image of an image pyramid of the reference image; According to the residual matrix, a second matching position of the test image at the top layer is obtained; For each layer of the image pyramid of the test image, determine the image matching area of ​​the layer of the test image according to the second matching position of the test image in the layer above the layer, and the deviation of the machine motion and the camera calibration; Determine a residual matrix of the image matching area according to the image of the layer in the image pyramid of the test image and the image corresponding to the layer in the image pyramid of the reference image, and obtain a second matching position of the test image in the layer according to the residual matrix; The second matching position of the test image at the bottom layer in the image pyramid is used as the image matching position in the test image that is aligned with the reference image.

4. The method according to claim 1, characterized in that The step of obtaining the surface image of the wafer to be inspected specifically includes: Obtaining the original images of each wafer to be inspected obtained by scanning by each machine; Calling the camera calibration parameters of each machine, wherein the calibration parameters include the camera orthogonal angle and the flat field coefficient; Determining an image correction matrix according to the calibration parameters; Calling the camera parameters of each machine, wherein the camera parameters include the camera resolution of each machine, the magnification of each machine, the distortion coefficient of each machine, and the camera size of each machine; Correcting 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 consistent, thereby obtaining a surface image; A surface image of at least one wafer to be inspected among the wafers to be inspected is obtained.

5. The method according to claim 1, characterized in that: The step of comparing the grain images one by one according to the template image to realize defect detection specifically includes: For each grain image, determining a grayscale difference image between the grain image and the template image; According to the grayscale difference image, determining the detection result of each pixel in the grayscale difference image through a preset grayscale detection range; The defect detection result of the grain is determined according to the detection result of each pixel in the grayscale difference image.

6. The method according to claim 5, characterized in that The step of determining the detection result of each pixel in the grayscale difference image according to the preset grayscale detection range specifically includes: Determining a gradient image of the template image; For each pixel in the grayscale difference image, determine the gradient value corresponding to 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 a preset grayscale detection range; Determine whether the grayscale difference of the pixel falls within the grayscale detection range of the pixel; If yes, the pixel is considered defect-free as the detection result of the pixel; If not, the defect type of the pixel is determined based on the grayscale difference of the pixel in the grayscale difference image and the size relationship of the grayscale detection range, and the detection result of the pixel is determined based on the defect type of the pixel.

7. The method according to claim 5, characterized in that: The step of determining the defect detection result of the grain according to the detection result of each pixel in the grayscale difference image specifically includes: Marking a defective area in the grain image according to a detection result of each pixel in the grayscale difference image, and determining a defect type of the defective area; determining image features of the defective area; According to the image features of the defect area and the defect type of the defect area, the process defect category to which the defect area belongs is determined as the defect detection result of the die.

8. A wafer inspection device, characterized in that: include: An acquisition module, used for acquiring a surface image of a wafer to be inspected, wherein the surface image includes grain images of a plurality of grains; A selection module is used to select one grain image from each grain image as a reference image and use other grain images as test images; A pyramid module, used for obtaining an image pyramid of the reference image and an image pyramid of each test image by continuously reducing the reference image and the test image at the same ratio; An image matching module, for determining, for each test image, an image matching position in the test image that is aligned with the reference image based on an image pyramid of the test image and an image matching between the image pyramids of the reference image; A template image module, used to align the test images according to the image matching positions of the test images, and fuse them to obtain a template image; The defect detection module is used to compare the grain images one by one according to the template image to achieve defect detection.

9. A computer-readable storage medium, characterized in that: The storage medium contains a computer program, and when the computer program is executed by a processor, the method described in any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the method according to any one of claims 1 to 7 is implemented when the processor executes the program.

Images