Image processing method and device

Abstract

The invention provides an image processing method and device, and belongs to the technical field of image processing, and the method comprises the steps: obtaining a plurality of target images, which are preliminarily detected to be free of appearance defects, of a target product, and determining a preliminary reference image of the target images; determining an image area to be subjected to gray scale compensation from the preliminary reference image as a first compensation area; determining a second compensation area corresponding to the first compensation area on the target image; performing gray scale compensation on the second compensation area of each target image based on the gray scale value of the pixel in the first compensation area; performing registration based on pixel positions on all the target images after gray scale compensation, determining a pixel with the highest gray scale value and a pixel with the lowest gray scale value on each pixel position, and generating a corresponding high-order reference image and a corresponding low-order reference image; and based on the high-order reference image and the low-order reference image, performing defect identification on the to-be-detected image. According to the invention, the influence caused by uneven illumination can be eliminated, and the defect detection precision is improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing technologies, and in particular, to an image processing method and apparatus. Background Art

[0002] Currently, the more commonly used methods for Wafer inspection are the subtraction comparison methods of Golden Die and Die To Die. The former generates a standard reference Die image by selecting the OK Die images on the Wafer, and the gray-scale difference between each Die to be detected and this reference image is compared. If the gray-scale difference exceeds the set threshold, the difference at that location is considered a defect position. The latter compares the Die to be detected with other Dies within the field of view, and determines whether it is a defect by the occurrence times of inconsistent positions. There are various ways to select other Dies for comparison in the Die To Die method, such as Dies in the 4-field, 8-field, or all Dies within the field of view. The more Dies are selected, the longer the calculation time. Since there is no standard, defects cannot be determined 100%. Therefore, Golden Die is more commonly used in the industry.

[0003] Implementing Golden Die is relatively simple, but making the standard reference image is crucial, and the selection of the gray-scale difference threshold is also important. Because the surface of the object to be measured is not uniform and there are height differences, uneven illumination will also occur with the same lighting method. When comparing with the standard reference image, the local gray-scale difference of the OK Die may also exceed the set threshold, resulting in false detection. Summary of the Invention

[0004] The present invention provides an image processing method and apparatus to solve the defect of easy false detection due to uneven illumination in the prior art, and can effectively avoid the influence of normal gray-scale fluctuations of products.

[0005] In a first aspect, the present invention provides an image processing method, including: obtaining a plurality of target images of a target product that are preliminarily detected as having no appearance defects, and determining a preliminary reference image of the target images; the target images are gray-scale images including a preset target; determining an image area to be subjected to gray-scale compensation from the preliminary reference image as a first compensation area; and determining a second compensation area corresponding to the first compensation area on the target images by means of pixel position registration; performing gray-scale compensation on the second compensation area of each target image based on the gray-scale values of the pixels in the first compensation area; performing registration based on pixel positions on all the target images after gray-scale compensation, determining the pixels with the highest gray-scale value and the pixels with the lowest gray-scale value at each pixel position to respectively generate corresponding high-order reference images and low-order reference images; and performing defect recognition on the image to be detected based on the high-order reference images and the low-order reference images.

[0006] An image processing method provided by the present invention performs gray-scale compensation on the second compensation region of each target image based on the gray-scale values of the pixels in the first compensation region, including: determining the average value of the gray-scale values of the pixels in the first compensation region as the first gray-scale average value; and determining the average value of the gray-scale values of the pixels in the second compensation region as the second gray-scale average value; taking the ratio of the first gray-scale value to the second gray-scale value as the gray-scale compensation parameter; multiplying the gray-scale values of the pixels in the second compensation region of each target image by the compensation parameter to achieve gray-scale compensation for each target image.

[0007] An image processing method provided by the present invention for determining a preliminary reference image of the target image includes: registering multiple target images based on pixel positions; obtaining the gray-scale average value of the pixel points at each pixel position; and constructing a preliminary reference image based on the gray-scale average values at each pixel position.

[0008] An image processing method provided by the present invention for determining an image region to be gray-scale compensated from the preliminary reference image includes: determining a gray-scale value range; determining the pixel points with gray-scale values within the gray-scale value range from the preliminary reference image through the gray-scale value range; and determining the image region to be gray-scale compensated based on the pixel points with gray-scale values within the gray-scale value range.

[0009] An image processing method provided by the present invention for defect recognition of a test image based on the high-order reference image and the low-order reference image includes: registering the test image, the high-order reference image, and the low-order reference image based on pixel positions; determining that any pixel has a bright defect when the gray-scale value of any pixel on the test image is greater than the gray-scale value of the pixel at the same pixel position on the high-order reference image; and determining that any pixel has a dark defect when the gray-scale value of any pixel on the test image is less than the gray-scale value of the pixel at the same pixel position on the low-order reference image.

[0010] An image processing method provided by the present invention for obtaining multiple target images of a target product that are preliminarily detected as having no appearance defects includes: obtaining a grayscale image of the target product; extracting multiple sub-images containing a preset target and having the same size from the grayscale image; and performing a preliminary detection on the sub-images, and taking the sub-images without appearance defects as target images.

[0011] In an image processing method provided by the present invention, the target product is a wafer, and the preset target is a semiconductor chip.

[0012] In a second aspect, the present invention further provides an image processing apparatus, including:

[0013] An image acquisition module, configured to acquire multiple target images of a target product that are preliminarily detected as having no appearance defects, and determine a preliminary reference image of the target images; the target images are grayscale images including a preset target.

[0014] A compensation area determination module, configured to determine an image area to be subjected to grayscale compensation from the preliminary reference image as a first compensation area; and, determine a second compensation area corresponding to the first compensation area on the target image by means of pixel position registration.

[0015] A grayscale compensation module, configured to perform grayscale compensation on the second compensation area of each target image based on the grayscale values of the pixels in the first compensation area.

[0016] A high-order and low-order reference image generation module, configured to perform registration based on pixel positions on all the target images after grayscale compensation, determine the pixels with the highest grayscale value and the pixels with the lowest grayscale value at each pixel position, so as to generate corresponding high-order reference images and low-order reference images respectively.

[0017] A defect recognition module, configured to perform defect recognition on a test image based on the high-order reference image and the low-order reference image.

[0018] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the steps of the image processing method as described in any one of the above are implemented.

[0019] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the image processing method as described in any one of the above are implemented.

[0020] The image processing method and device provided by the present invention have the following beneficial effects compared with the prior art:

[0021] (1) The present invention proposes a method for performing grayscale compensation on target images that are preliminarily detected as having no appearance defects, eliminates the influence caused by partial uneven illumination, and reduces false detection caused by normal grayscale fluctuations on the product surface; at the same time, high-order and low-order reference images are generated for defect recognition, which can more accurately determine whether there are actual defects, further reduce the probability of false detection, and improve the defect detection accuracy.

[0022] (2) The present invention proposes an industrial defect detection method applicable to repeating units, especially products such as Wafer wafers with multiple Dies in the shooting field of view; however, it is not limited to such products, and some steps in this solution can also be used for non-periodic substrate line defect detection; it has the characteristics of simple implementation, fewer parameters, and wide applicability. Brief Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0024] Figure 1 is a schematic flowchart of the image processing method provided by the present invention;

[0025] Figure 2 is a schematic diagram of the grayscale image of the target product provided by the present invention;

[0026] Figure 3 is a schematic diagram of the first compensation area provided by the present invention;

[0027] Figure 4 is a schematic structural diagram of an image processing apparatus provided by the present invention;

[0028] Figure 5 is a schematic structural diagram of an electronic device provided by the present invention. Detailed Description of the Embodiments

[0029] In order to make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0030] It should be noted that in the description of the embodiments of the present invention, the terms "include", "comprise" or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device including the element. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0031] The terms "first", "second", etc. in this application are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of this application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are usually of the same category, and do not limit the number of objects. For example, the first object can be one or more.

[0032] The following will describe the image processing method and device provided by the embodiments of the present invention in conjunction with Figures 1-5 the following.

[0033] Figure 1 is a schematic flowchart of the image processing method provided by the present invention, as Figure 1 shown, including but not limited to the following steps:

[0034] Step 101: Obtain a plurality of target images of the target product that are initially detected as having no appearance defects, and determine a preliminary reference image of the target images; the target images are grayscale images containing a preset target.

[0035] Among them, the target images initially detected as having no appearance defects refer to a batch of images that appear to have no obvious defects on the surface determined through rapid screening means before more detailed analysis or final quality inspection. This process is usually to reduce the amount of data for subsequent processing and ensure that the images used to establish the reference model are relatively "clean", that is, they do not contain obvious manufacturing defects.

[0036] Optionally, the target product in the present invention can be a wafer, and the preset target can be a semiconductor chip. Figure 2 is a schematic diagram of the grayscale image of the target product provided by the present invention, as Figure 2 shown, the images within the red frame are the images detected with obvious defects during the preliminary detection process, and the images within the green frame are the target images without appearance defects required by the present invention.

[0037] As an optional embodiment, the image processing method provided by the present invention for obtaining a plurality of target images of the target product that are initially detected as having no appearance defects includes but not limited to the following steps:

[0038] (1) Obtain the grayscale image of the target product:

[0039] The present invention can use a high-resolution camera or a scanning device to image the target product (such as a semiconductor chip on a wafer).

[0040] (2) Extract a plurality of sub-images from the grayscale image that contain the preset target and have the same size:

[0041] The present invention preprocesses the obtained grayscale images, including operations such as cropping and scaling, to ensure that all images have the same size. Using template matching, edge detection, or other computer vision techniques, a preset target (such as a semiconductor chip) in the image is identified and located to extract a sub-image containing the preset target.

[0042] (3) Perform a preliminary detection on the sub-image, and use the sub-image without appearance defects as the target image:

[0043] The present invention can apply automated defect detection algorithms (such as techniques based on threshold segmentation, morphological operations, texture analysis, etc.) to perform a preliminary detection on each image containing the preset target extracted. Analyze the gray-scale changes, edge continuity, texture features, etc. in the image to identify obvious appearance defects, such as cracks, scratches, stains, particle contamination, etc. According to the detection results, mark the images with obvious defects and exclude them, and retain those images determined to have no appearance defects as the target images.

[0044] The preliminary reference image of the target image can be a target image without appearance defects preset artificially, or an image generated based on the target image without appearance defects. As an optional embodiment, the method for determining the preliminary reference image of the target image provided by the present invention includes, but is not limited to, the following steps:

[0045] (1) Perform registration on multiple target images based on pixel positions:

[0046] The present invention can use computer vision algorithms to detect feature points in each target image and match these feature points with the feature points in the reference image to achieve registration.

[0047] (2) Obtain the gray-scale mean of the pixel points at each pixel position, and the specific calculation method is as follows:

[0048]

[0049] where Avg(x,y) is the gray-scale mean of the pixel point at the position (x,y), N is the number of target images, and I i (x,y) represents the gray-scale value of the i-th target image at the position (x,y).

[0050] (3) Construct a preliminary reference image based on the gray-scale mean at each pixel position:

[0051] Create a new image with the same size as the registered target image. For each pixel position (x,y) in the new image, set its gray-scale value to the previously calculated mean Avg(x,y).

[0052] Step 102: Determine the image area to be grayscale compensated from the preliminary reference image as the first compensation area; and, determine the second compensation area corresponding to the first compensation area on the target image by means of pixel position registration.

[0053] In this step, the area for grayscale compensation can be determined by threshold selection or by manually determining the area.

[0054] Optionally, the method for determining the image area to be grayscale compensated from the preliminary reference image includes but is not limited to the following steps:

[0055] (1) Determine the grayscale value range:

[0056] The grayscale value range can be determined based on experience. For example, based on experience or through statistical analysis (such as calculating the average grayscale value and standard deviation), a reasonable grayscale value range is set. This range should be able to cover the grayscale values of most normal pixels and exclude those significantly abnormal (too high or too low) grayscale values.

[0057] (2) Determine the pixel points in the preliminary reference image whose grayscale values are within the grayscale value range:

[0058] Compare each pixel in the preliminary reference image with the set grayscale value range according to its grayscale value.

[0059] (3) Determine the image area to be grayscale compensated according to the pixel points whose grayscale values are within the grayscale value range.

[0060] It can be understood that the image area to be grayscale compensated in the present invention is the set of pixel points whose grayscale values are within the grayscale value range.

[0061] Step 103: Perform grayscale compensation on the second compensation area of each target image based on the grayscale values of the pixels in the first compensation area.

[0062] It should be noted that there are various ways of grayscale compensation in step 103, such as the method of linear stretching, the method of histogram matching (so that the gray level distribution of the second compensation area of the target image is as close as possible to the gray level distribution of the first compensation area of the reference image), etc. The present invention can select a suitable method to use the grayscale values of the pixels in the first compensation area to perform grayscale compensation on the second compensation area according to needs.

[0063] As an optional embodiment, performing grayscale compensation on the second compensation area of each target image based on the grayscale values of the pixels in the first compensation area includes but is not limited to the following steps:

[0064] (1) Determine the average grayscale value of the pixels within the first compensation region as the first grayscale average value; and determine the average grayscale value of the pixels within the second compensation region as the second grayscale average value.

[0065] By obtaining the grayscale value of each pixel within the first compensation region, the average value of the grayscale values of all pixels can be calculated as the first grayscale average value G1; similarly, by obtaining the grayscale value of each pixel within the second compensation region, the average value of the grayscale values of all pixels can be calculated as the second grayscale average value G2.

[0066] (2) Use the ratio of the first grayscale value to the second grayscale value as the grayscale compensation parameter.

[0067] Let the grayscale compensation parameter be K, then the grayscale compensation parameter K = G1 / G2.

[0068] (3) Multiply the grayscale value of the pixels within the second compensation region of each target image by the compensation parameter to achieve grayscale compensation for each target image.

[0069] Step 104: Perform registration based on pixel positions for all the target images after grayscale compensation, and determine the pixel with the highest grayscale value and the pixel with the lowest grayscale value at each pixel position to respectively generate the corresponding high-order reference image and low-order reference image.

[0070] The present invention can use an image registration algorithm to accurately register all images, ensure that the preset target in each image is located at the same position, and ensure that the pixel positions between all images are completely corresponding for subsequent pixel-by-pixel comparison.

[0071] For each pixel position, traverse all the registered target images, record the grayscale value at this position, and for each pixel position, find the pixels with the highest and lowest grayscale values among all the images.

[0072] Create a new image, the grayscale value of each pixel position of which is taken from the pixel with the highest grayscale value at this position among all the target images, which is the high-order reference image; similarly, create a new image, but this time the grayscale value of each pixel position is taken from the pixel with the lowest grayscale value at this position among all the target images, which is the low-order reference image.

[0073] Step 105: Based on the high-order reference image and the low-order reference image, perform defect recognition on the image to be measured.

[0074] Compare the image to be tested with the high-order reference image and the low-order reference image pixel by pixel. For each pixel position, perform the following operations: Check whether the gray scale value of the image to be tested exceeds the range defined by the high-order reference image and the low-order reference image. If the gray scale value in the image to be tested is higher than the high-order reference image or lower than the gray scale value of the low-order reference image, mark the pixel as a potential defect.

[0075] Optionally, based on the high-order reference image and the low-order reference image, perform defect identification on the image to be tested, including but not limited to the following steps:

[0076] (1) Perform registration based on pixel positions on the image to be tested, the high-order reference image, and the low-order reference image;

[0077] (2) When the gray scale value of any pixel in the image to be tested is greater than the gray scale value of the pixel at the same pixel position on the high-order reference image, determine that any pixel has a bright defect;

[0078] (3) When the gray scale value of any pixel in the image to be tested is less than the gray scale value of the pixel at the same pixel position on the low-order reference image, determine that any pixel has a dark defect.

[0079] To more clearly illustrate the technical solution of the present invention, the following takes the target product as a wafer and the preset target as a semiconductor chip as an example to introduce a specific implementation manner:

[0080] Step 1: Cut out individual Die images (i.e., target images) from the captured image of the wafer according to a fixed size, and select the OK defect-free Dies for standby;

[0081] The Die image is a single repeating unit in the captured image of the wafer. In the present invention, a Die is an unencapsulated semiconductor chip cut from the wafer, such as Figure 2 As shown, there are multiple repeating units in one field of view, and each repeating unit is a Die.

[0082] Step 2: Through registration and alignment of the Dies selected in Step 1, calculate the average value of the gray levels of a certain number of Dies according to each corresponding pixel to generate a preliminary standard reference Die image.

[0083] Step 3: Determine the area for gray scale compensation by means of threshold selection or manual determination of the area.

[0084] Among them, the area for gray scale compensation can determine the area that needs to be compensated for the preliminary standard reference Die image through the range of gray scales Gray1 to Gray2 and the manually selected range, which is called the first compensation area. Figure 3 This is a schematic diagram of the first compensation area provided by the present invention, such as Figure 3As shown, the area within the red frame is the first compensation area. According to the registration alignment in Step 2, the area corresponding to the gray-scale compensation of the OK Die image selected in Step 1 can also be obtained, which is called the second compensation area.

[0085] Step 4: Calculate the gray-scale mean value G1 of the first compensation area for the preliminary standard reference Die image in Step 2, and calculate the gray-scale mean value G2 of the second compensation area for the Die image selected in Step 1; perform gray-scale compensation on the gray scale of the Die image selected in Step 1 according to the ratio of G1 / G2.

[0086] Step 5: Perform the same registration alignment on the Die image after gray-scale compensation in Step 4 as in Step 2. Sort and count the pixels at each position of the aligned Die image from low to high, and record the lowest gray scale and the highest gray scale of each pixel point. Generate a low-order reference Die image (low-order reference image) from the lowest gray scale of all pixel positions according to the original position, and generate a high-order reference Die image (high-order reference image) from the highest gray scale of all pixel positions according to the original position.

[0087] Step 6: Compare the Die image to be tested with the low-order reference Die image in Step 5, and the defect lower than the corresponding position is a dark defect; compare the Die image to be tested with the high-order reference Die image in S105, and the defect higher than the corresponding position is a bright defect.

[0088] On the other hand, Figure 4 is a schematic structural diagram of an image processing device provided by the present invention. As Figure 4 shown, the device includes:

[0089] An image acquisition module 410, configured to acquire a plurality of target images of a target product that are preliminarily detected as having no appearance defects, and determine a preliminary reference image of the target image; the target image is a grayscale image including a preset target;

[0090] A compensation area determination module 420, configured to determine an image area to be subjected to gray-scale compensation from the preliminary reference image as the first compensation area; and determine a second compensation area corresponding to the first compensation area on the target image by means of pixel position registration;

[0091] A gray-scale compensation module 430, configured to perform gray-scale compensation on the second compensation area of each target image based on the gray-scale values of the pixels within the first compensation area;

[0092] A high / low-order reference image generation module 440, configured to perform registration based on pixel positions on all the target images after gray-scale compensation, determine the pixel with the highest gray-scale value and the pixel with the lowest gray-scale value at each pixel position, so as to generate corresponding high-order reference images and low-order reference images respectively;

[0093] A defect recognition module 450 is configured to perform defect recognition on a to-be-tested image based on the high-order reference image and the low-order reference image.

[0094] It should be noted that the image processing device provided in the embodiment of the present invention can execute the image processing method described in any of the above embodiments during specific operation, and details thereof are not described in this embodiment.

[0095] The image processing method and device provided by the present invention have the following beneficial effects compared with the prior art:

[0096] (1) The present invention proposes a method for gray-scale compensation for a target image initially detected as having no appearance defects, eliminating the influence caused by partial uneven illumination and reducing false detection caused by normal gray-scale fluctuations on the product surface. At the same time, high-order and low-order reference images are generated for defect recognition, which can more accurately determine whether there are actual defects, further reducing the probability of false detection and improving the detection accuracy.

[0097] (2) The present invention proposes an industrial defect detection method applicable to repetitive units, especially products such as Wafer wafers with multiple Dies in the shooting field of view. However, it is not limited to such products, and some steps of this solution can also be used for the detection of non-periodic substrate line defects. It has the characteristics of simple implementation, fewer parameters, and wide applicability.

[0098] Figure 5 is a schematic structural diagram of the electronic device provided by the present invention, as Figure 5As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communication bus 540. Among them, the processor 510, the communications interface 520, and the memory 530 complete mutual communication through the communication bus 540. The processor 510 may call logic instructions in the memory 530 to execute an image processing method, which includes: obtaining a plurality of target images of a target product that are initially detected as having no appearance defects, and determining a preliminary reference image of the target images; the target images are grayscale images containing a preset target; determining an image area to be subjected to grayscale compensation from the preliminary reference image as a first compensation area; and determining a second compensation area corresponding to the first compensation area on the target images by means of pixel position registration; performing grayscale compensation on the second compensation area of each target image based on the grayscale values of the pixels in the first compensation area; performing registration based on pixel positions on all the target images after grayscale compensation, determining the pixels with the highest grayscale value and the pixels with the lowest grayscale value at each pixel position to respectively generate corresponding high-order reference images and low-order reference images; and performing defect recognition on a test image based on the high-order reference images and the low-order reference images.

[0099] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the image processing methods provided in the above embodiments. The method includes: obtaining a plurality of target images of a target product that are initially detected as having no appearance defects, and determining a preliminary reference image of the target images; the target images are grayscale images containing a preset target; determining an image area to be subjected to grayscale compensation from the preliminary reference image as a first compensation area; and determining a second compensation area corresponding to the first compensation area on the target images by means of pixel position registration; performing grayscale compensation on the second compensation area of each target image based on the grayscale values of the pixels in the first compensation area; performing registration based on pixel positions on all the target images after grayscale compensation, determining the pixels with the highest grayscale value and the pixels with the lowest grayscale value at each pixel position to respectively generate corresponding high-order reference images and low-order reference images; and performing defect recognition on a test image based on the high-order reference images and the low-order reference images.

[0100] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute the image processing method provided in each of the above embodiments. The method includes: obtaining a plurality of target images of a target product that are initially detected as having no appearance defects, and determining a preliminary reference image of the target images; the target images are grayscale images containing a preset target; determining an image region to be subjected to grayscale compensation from the preliminary reference image as a first compensation region; and determining a second compensation region corresponding to the first compensation region on the target images by means of pixel position registration; performing grayscale compensation on the second compensation region of each target image based on the grayscale values of the pixels in the first compensation region; performing registration based on pixel positions on all the target images after grayscale compensation, and determining the pixel with the highest grayscale value and the pixel with the lowest grayscale value at each pixel position to respectively generate corresponding high-order reference images and low-order reference images; and performing defect recognition on a test image based on the high-order reference images and the low-order reference images.

[0101] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. An image processing method, characterized in that: include: Acquire a plurality of target images of a target product that are initially detected as having no appearance defects, and determine a preliminary reference image of the target image; The target image is a grayscale image containing a preset target; Determining an image area to be subjected to grayscale compensation from the preliminary reference image as a first compensation area; And, determining a second compensation area corresponding to the first compensation area on the target image by pixel position registration; Based on the grayscale values ​​of the pixels in the first compensation area, performing grayscale compensation on the second compensation area of ​​each target image; Perform pixel position-based registration on all target images after grayscale compensation, determine the pixel with the highest grayscale value and the pixel with the lowest grayscale value at each pixel position, and generate corresponding high-order reference images and low-order reference images respectively; Defect recognition is performed on the image to be tested based on the high-order reference image and the low-order reference image.

2. The image processing method according to claim 1, characterized in that: Based on the grayscale values ​​of the pixels in the first compensation area, grayscale compensation is performed on the second compensation area of ​​each target image, including: Determine an average value of the grayscale values ​​of the pixels in the first compensation area as a first grayscale average value; and determine an average value of the grayscale values ​​of the pixels in the second compensation area as a second grayscale average value; Using the ratio of the first grayscale value to the second grayscale value as a grayscale compensation parameter; The grayscale value of the pixel in the second compensation area of ​​each target image is multiplied by the compensation parameter to achieve grayscale compensation for each target image.

3. The image processing method according to claim 1, characterized in that: Determining a preliminary reference image of the target image includes: Perform pixel-based registration on multiple target images; Get the grayscale mean of the pixel at each pixel position; A preliminary reference image is constructed based on the grayscale mean at each pixel position.

4. The image processing method according to claim 1, characterized in that: Determining an image area to be subjected to grayscale compensation from the preliminary reference image comprises: Determine the grayscale value range; Determining pixel points whose grayscale values ​​are within the grayscale value range from the preliminary reference image through the grayscale value range; The image area to be subjected to grayscale compensation is determined according to the pixel points whose grayscale values ​​are within the grayscale value range.

5. The image processing method according to claim 1, characterized in that: Based on the high-order reference image and the low-order reference image, defect recognition is performed on the image to be tested, including: Perform pixel position-based registration on the image to be tested, the high-order reference image, and the low-order reference image; In the case where the grayscale value of any pixel on the image to be tested is greater than the grayscale value of the pixel at the same pixel position on the high-order reference image, determining that any pixel has a bright defect; and When the grayscale value of any pixel on the image to be tested is less than the grayscale value of the pixel at the same pixel position on the low-level reference image, it is determined that any pixel has a dark defect.

6. The image processing method according to claim 1, characterized in that: Acquire multiple target images of the target product that are initially detected as having no appearance defects, including: Obtain a grayscale image of the target product; Extracting multiple sub-images containing preset targets and having the same size from the grayscale image; Perform a preliminary inspection on the sub-images and take the sub-images without appearance defects as the target images.

7. The image processing method according to claim 1, characterized in that: The target product is a wafer, and the preset target is a semiconductor chip.

8. An image processing device, characterized in that: include: An image acquisition module, used to acquire a plurality of target images of a target product that are initially detected as having no appearance defects, and determine a preliminary reference image of the target image; The target image is a grayscale image containing a preset target; A compensation area determination module, used to determine an image area to be subjected to grayscale compensation from the preliminary reference image as a first compensation area; And, determining a second compensation area corresponding to the first compensation area on the target image by pixel position registration; A grayscale compensation module, configured to perform grayscale compensation on a second compensation area of ​​each target image based on grayscale values ​​of pixels in the first compensation area; A high- and low-order reference image generation module is used to perform pixel position registration on all target images after grayscale compensation, determine the pixel with the highest grayscale value and the pixel with the lowest grayscale value at each pixel position, and generate corresponding high-order reference images and low-order reference images respectively; The defect recognition module is used to perform defect recognition on the image to be tested based on the high-order reference image and the low-order reference image.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the image processing method according to any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the image processing method according to any one of claims 1 to 7 are implemented.

Images