Image processing method, device, medium and product

Abstract

The application discloses an image processing method and device, medium and product, and relates to the technical field of semiconductor defect detection. The method comprises the following steps: dividing a plurality of region images corresponding to a plurality of feature regions from a scanning image of any die in a target wafer, all patterns in one feature region having the same shape and periodic arrangement characteristics; performing Fourier transform on the plurality of region images and longitudinal integration to obtain a plurality of frequency spectrum profile curves; determining configuration parameters by identifying target peaks in the plurality of frequency spectrum profile curves, the configuration parameters comprising the position and width of the target peaks; and configuring a hardware filter according to the configuration parameters, and performing global filtering processing on the scanning image of the target wafer by using the hardware filter. According to the scheme, Fourier filtering based on software is used to guide Fourier filtering based on hardware, noise in the periodic pattern region on the wafer surface is adaptively filtered out, and the adaptability and processing efficiency of filtering are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of semiconductor defect detection, and particularly relates to an image processing method, device, medium and product. BACKGROUND

[0002] In the scene of detecting wafer defects in dark field, due to the interference effect caused by the scattering of defects on incident light, speckle noise appears in the detection image. This speckle noise is particularly obvious in the high-density repeated pattern area of the wafer surface, which can significantly reduce the signal-to-noise ratio and affect the accurate identification of defects.

[0003] At present, the prior art solution adopts hardware-based Fourier filtering to filter specific frequency components in the frequency domain to suppress speckle noise and reduce background interference. However, the hardware-based Fourier filtering is a global filtering, which can only apply the same filtering strategy to the entire wafer, and has certain limitations in adaptability and flexibility, which may cause insufficient filtering (such as residual background interference) or excessive filtering (such as loss of valuable defect information), thereby affecting the detection effect of defects.

[0004] Therefore, how to improve the filtering adaptability and accurately suppress the speckle noise has become a technical problem to be solved at present. SUMMARY

[0005] The embodiments of the present application provide an image processing method, device, medium and product, which can improve the filtering adaptability and accurately suppress the speckle noise.

[0006] In a first aspect of the embodiments of the present application, an image processing method is provided, which includes: dividing a plurality of region images corresponding to a plurality of feature regions from a scanning image of any die in a target wafer, all patterns in a feature region having the same shape and periodic arrangement characteristics; performing Fourier transform and longitudinal integration on the plurality of region images to obtain a plurality of frequency spectrum profile curves; determining configuration parameters by identifying target peaks in the plurality of frequency spectrum profile curves, the configuration parameters including the position and width of the target peaks; configuring a hardware filter according to the configuration parameters, and performing global filtering processing on the scanning image of the target wafer by using the hardware filter.

[0007] In a second aspect, the embodiment of the present application provides an image processing device, which comprises: a division module configured to divide a plurality of region images corresponding to a plurality of feature regions from a scanning image of any die in a target wafer, all patterns in a region image having the same shape and periodic arrangement characteristics; a virtual Fourier transform module configured to perform Fourier transform and vertical integration on the plurality of region images to obtain a plurality of spectral profile curves; a configuration parameter determination module configured to determine a configuration parameter by identifying a target peak in the plurality of spectral profile curves; and a hardware Fourier filter module configured to configure a hardware filter according to the configuration parameter and perform global filter processing on the scanning image of the target wafer by using the hardware filter.

[0008] In a third aspect, the embodiment of the present application provides an electronic device, which comprises a memory and a program or instruction stored in the memory and executable on a processor, and the program or instruction is executed by the processor to implement the image processing method provided in any one of the above aspects of the embodiment of the present application.

[0009] In a fourth aspect, the embodiment of the present application provides a readable storage medium, and the readable storage medium stores a program or instruction, and the program or instruction is executed by a processor to implement the image processing method provided in any one of the above aspects of the embodiment of the present application.

[0010] In a fifth aspect, the embodiment of the present application provides a computer program product, and the instructions in the computer program product are executed by a processor of an electronic device to enable the electronic device to perform the image processing method provided in any one of the above aspects of the embodiment of the present application.

[0011] In the image processing method provided by the embodiment of the present application, a plurality of feature regions having periodic arrangement patterns in the scanning image of any die in a target wafer are identified, Fourier transform calculation is performed on the region image corresponding to each feature region, a target peak representing the feature region is found in the obtained spectral profile curve, the position and width of the target peak are taken as configuration parameters, a hardware filter is configured according to the configuration parameters, and global filter processing is performed on the target wafer. The method uses software Fourier filter to guide hardware Fourier filter, realizes adaptive filtering of noise in the periodic pattern region on the wafer surface, significantly improves the adaptability and processing efficiency of the filter, effectively enhances the noise filtering effect, and thus improves the accuracy and reliability of defect detection. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiments of the present application will be briefly introduced. Those skilled in the art can obtain other drawings according to these drawings without creating any creative labor.

[0013] Figure 1is a flowchart of an image processing method provided by an embodiment of the present application;

[0014] Figure 2 is a flowchart of an image processing method provided by an embodiment of the present application;

[0015] Figure 3 is a flowchart of an image processing method provided by an embodiment of the present application;

[0016] Figure 4 is a flowchart of an image processing method provided by an embodiment of the present application;

[0017] Figure 5 (a) and (b) show structural diagrams of local positions of crystal grains provided by embodiments of the present application;

[0018] Figure 6 shows a schematic diagram of a spectrum profile curve generation process provided by an embodiment of the present application;

[0019] Figure 7 shows a schematic diagram of superimposed spectrum profile curves provided by an embodiment of the present application;

[0020] Figure 8 shows a schematic diagram of determining a target peak provided by an embodiment of the present application;

[0021] Figure 9 is a structural diagram of an image processing device provided by an embodiment of the present application;

[0022] Figure 10 is a schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0023] The features and exemplary embodiments of various aspects of the present application will be described in detail below, in order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, but not to limit the present application. For those skilled in the art, the present application can be implemented without some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by showing examples of the present application.

[0024] It is to be noted that, in the present document, relational terms such as first and second, and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a", "comprises...", or "comprising...", does not, without more constraints, preclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0025] It is to be noted that, in the present document, relational terms such as first and second, and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a", "comprises...", or "comprising...", does not, without more constraints, preclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0026] First, the technical terms related to one or more embodiments of the present application are explained.

[0027] Wafer, a silicon wafer processed by a specific process, is the basic material for semiconductor manufacturing.

[0028] Die, a single chip unit separated from a wafer by cutting, contains a complete integrated circuit or functional module.

[0029] Dark field detection, an optical detection technique, is mainly used for detecting small defects on the wafer surface. Its principle is to use oblique incident light to illuminate the sample and collect the scattered light for detection.

[0030] Speckle noise, a common noise phenomenon in optical imaging, is caused by coherent light (such as laser) illuminating rough surfaces or passing through scattering media. It appears as randomly distributed bright and dark spots in the image, which reduces the resolution and contrast of the image and causes interference in detection and analysis.

[0031] Hardware-based Fourier filtering, a technique that directly performs Fourier transform and filtering on the physical layer using optical systems. Its basic principle is to perform Fourier transform on the incident light through a lens, place a spatial filter (such as a blocking sheet, a diffraction grating, or a spatial light modulator) in the Fourier plane (i.e. the frequency domain), and then recover the image by inverse Fourier transform.

[0032] In Fourier filtering, a mask is used to select or suppress specific frequency components in the frequency domain.

[0033] Software-based Fourier filtering is a technique that implements Fourier transform and filtering in the digital domain through algorithms. Its basic principle is to use a computer to perform a fast Fourier transform on a signal or image, filter out specific frequency components in the frequency domain, and then recover the signal or image through an inverse transform.

[0034] When using dark-field inspection technology to detect defects in wafers, lasers or highly coherent light sources are typically used. When light shines on the wafer surface, defects on the surface cause the incident light to scatter, and defect detection is achieved by collecting the scattered light. However, in areas with high-density repeating patterns on the wafer surface, due to the regularity and strong periodicity of the surface structure, stable interference modes are more easily formed between the scattered light fields in different areas. This makes speckle noise more pronounced, leading to drastic fluctuations in the intensity of the background signal, reducing the signal-to-noise ratio, and affecting the accuracy of subsequent defect identification.

[0035] Currently, there are methods to reduce noise interference using hardware-based Fourier filtering techniques. For example, a hardware Fourier filter can be placed in the optical path system to shield noise components at specific frequencies. Because this method is based on physical optical filtering and does not require computer computation, it can achieve real-time processing and has a high processing speed.

[0036] However, once a hardware Fourier filter is designed, its filtering strategy is fixed and can only perform global filtering on the entire wafer image. It cannot be adaptively adjusted according to the pattern characteristics of different regions, which has certain limitations in terms of adaptability and flexibility. This may lead to insufficient filtering (such as residual background interference) or excessive filtering (such as loss of valuable defect information), thereby affecting the defect detection effect.

[0037] To address the aforementioned technical problems, this application provides an image processing method, apparatus, medium, and product. In the image processing method provided in this application, software Fourier filtering guides hardware Fourier filtering, achieving adaptive noise removal in periodically arranged patterned regions on the wafer surface. This not only dynamically sets the filtering strategy based on pattern characteristics, improving adaptability, but also maintains high processing efficiency, significantly enhancing the noise removal effect, thereby improving the accuracy and reliability of defect detection.

[0038] For example, the image processing method provided by the embodiment of the present application can be applied to the production line of a semiconductor manufacturing enterprise, and used for defect detection of semiconductor devices generated in the production process. In actual application, first, a wafer defect image is acquired, and based on the image of any typical die area in the wafer, the image frequency spectrum characteristics are analyzed by using software Fourier transform, and the related parameters corresponding to the filtering strategy are determined. Subsequently, hardware-based Fourier filtering is performed according to the parameters, and the wafer defect image is filtered to remove noise, and a high-quality image after noise suppression is obtained. Finally, based on the high-quality image, further processing such as defect recognition, classification and quantization is performed, so as to improve the accuracy and reliability of defect detection, and provide support for production quality control.

[0039] It should be noted that the application scenarios described in the above embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. It can be known by those skilled in the art that, with the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems. The image processing method provided by the embodiments of the present application can be applied to various application scenarios in which wafer images need to be removed from noise.

[0040] The image processing method provided by the embodiments of the present application will be introduced below. In actual application, the execution subject of the image processing method of the embodiments of the present application can be an electronic device.

[0041] Figure 1 A flowchart of the image processing method provided by one embodiment of the present application is shown. As shown in Figure 1 The method includes steps S110 to S140.

[0042] S110, from the scanning image of any die in the target wafer, a plurality of region images corresponding to feature regions are divided, and all patterns in a region image have the same shape and periodic arrangement characteristics.

[0043] The target wafer is a wafer to be subjected to defect detection.

[0044] It should be understood that a wafer includes a plurality of repeatedly appearing dies, and the dies have consistency in structure and pattern.

[0045] The scanning image of the die is an original image of the die acquired by a scanning device.

[0046] In one embodiment, before step S110 is performed, a pre-scanning operation can be performed first, that is, a scanning device is used to completely scan any die in the target wafer to acquire an overall image thereof.

[0047] Exemplarily, the scanning device can be an imaging device such as an optical microscope or an electron microscope.

[0048] In an implementation, multiple high-magnification local images of the die can be taken by the scanning device, and the multiple local images can be combined into a complete scan image of the die by image stitching technology, so as to ensure the integrity and clarity of image details and improve the accuracy of overall imaging.

[0049] It can be understood that, since the dies on the wafer have high repeatability, a filtering strategy only needs to be designed for one die, and the strategy can be applied to all dies, thereby achieving efficient processing of the entire wafer.

[0050] For the sake of understanding and description, the die used for filtering strategy design will be referred to as a target die hereinafter, but it should be understood that the naming is only for the convenience of description and does not limit the selection of a specific die.

[0051] In an implementation, multiple feature regions in the scan image of the target die can be identified according to the periodicity of the pattern shape and arrangement, and then multiple region images corresponding to the multiple feature regions can be divided from the scan image of the die.

[0052] A feature region is a region in which the shapes of the patterns are the same and the arrangement has periodic characteristics. That is, the shapes of all the patterns in a region image are the same and the arrangement has periodic characteristics.

[0053] “Same shape” means that all the patterns in the feature region are basically consistent in geometric appearance, including boundary contour, size ratio, etc.

[0054] “Arrangement has periodic characteristics” means that the patterns of the same shape are repeatedly arranged in space according to certain rules, such as equidistant arrangement, matrix arrangement, or periodic distribution in a specific direction (horizontal or vertical). Such an arrangement makes the spatial distribution pattern of the patterns in the region predictable and repeated at different positions.

[0055] It can be understood that the high-density repetitive pattern region mentioned in the background section can be regarded as an example of the feature region in the embodiments of the present application.

[0056] For example, Figure 5 FIG. 1 shows a structure schematic diagram of a local position of a die provided by an embodiment of the present application.

[0057] Figure 5 The middle part includes two local positions (a) and (b) of the die.

[0058] From Figure 5 As can be seen from (a), the dividing line X is divided into two regions:

[0059] Region 1 contains multiple identical long rectangular line patterns, which are arranged in the horizontal direction, and the overall arrangement is 9x1, i.e. 9 in the horizontal direction and 1 in the vertical direction.

[0060] Region 2 contains multiple identical short rectangular line patterns, which are arranged in a 4x4 manner, i.e. 4 in both the horizontal and vertical directions.

[0061] As can be seen from Figure 5 (b), it is divided into two regions by the division line Y:

[0062] Region 3 contains multiple identical short rectangular line patterns, which are arranged in a 4x12 manner, i.e. 4 in the horizontal direction and 12 in the vertical direction.

[0063] Region 4 contains multiple identical long rectangular line patterns, which are arranged in a 9x1 manner, i.e. 9 in the horizontal direction and 1 in the vertical direction.

[0064] As can be seen from the figure, the patterns in regions 1, 2, 3 and 4 have the same shape and periodic arrangement characteristics, and the specific characteristics such as pattern shape, size, number and spacing between them are not exactly the same, so regions 1, 2, 3 and 4 can be regarded as characteristic regions.

[0065] Region image, a scanning image corresponding to a specific characteristic region, divided from a grain scanning image.

[0066] In one embodiment, the method of obtaining a region image can include the following steps:

[0067] S111, in the scanning image of the target grain, multiple characteristic regions are identified according to the pattern shape and arrangement periodicity characteristics, and their boundaries are determined;

[0068] S112, the identified characteristic regions are encoded and identified;

[0069] S113, based on the encoded identification, the scanning image of the target grain is divided according to the characteristic regions, and multiple corresponding region images are generated, each of which contains only the scanning information of a specific characteristic region.

[0070] In one implementation, step S111 can use a pre-trained visual recognition algorithm, i.e. use a visual recognition algorithm to automatically identify the shape and arrangement of all patterns in the target grain scanning image, and identify multiple characteristic regions accordingly. This method is suitable for large-scale automatic processing scenarios.

[0071] In an implementation, step S111 can be implemented based on a projection analysis method, i.e., projecting the scan image of the target grain in a fixed projection direction (e.g., horizontal or vertical), and identifying the feature region with periodic arrangement characteristics by analyzing the shape features of the projection result. This method is suitable for scenarios where the pattern arrangement is obvious.

[0072] In an implementation, step S111 can be implemented by manual annotation or by using a semi-automatic auxiliary tool, i.e., by manually annotating or using a semi-automatic auxiliary tool, the target grain scan image is interactively operated to identify and divide multiple feature regions. This method is suitable for high-precision processing requirements of specific grain structures or as a supplementary verification means for automatic division.

[0073] S120, Fourier transform and longitudinal integration are performed on the multiple region images to obtain multiple frequency spectrum profile curves.

[0074] In an implementation, step S120 can include steps S121-S122.

[0075] S121, Fourier transform is performed on the multiple region images to obtain multiple frequency spectrum images.

[0076] It can be understood that the region image and the frequency spectrum image are one-to-one corresponding. That is, for each region image in the multiple region images, Fourier transform is performed on the region image to obtain the Fourier frequency spectrum image corresponding to the region image. That is, for the multiple region images divided in the grain image, Fourier transform can be performed one by one to extract the frequency domain information of each region image and obtain the corresponding Fourier frequency spectrum image.

[0077] The frequency spectrum image is a two-dimensional matrix, and each element in the matrix represents the amplitude of the corresponding frequency component.

[0078] Fourier transform is used to convert a spatial domain image into a frequency domain representation, where different frequency components correspond to different scale features of the pattern.

[0079] Since the pattern in the region image has a periodic arrangement, a specific frequency component in the corresponding frequency spectrum image will exhibit a clear periodic peak.

[0080] In an implementation, Fast Fourier Transform (FFT) algorithm can be used to improve computational efficiency.

[0081] Exemplarily, before performing Fourier transform, the region image can be normalized to reduce noise interference.

[0082] S122, longitudinal integration is performed on the multiple frequency spectrum images to obtain multiple frequency spectrum profile curves.

[0083] It can be understood that the spectral image and the spectral profile curve are one-to-one corresponding. That is, for each of the plurality of spectral images, the spectral image is vertically integrated to obtain the spectral profile curve corresponding to the spectral image.

[0084] The vertical integration is to accumulate the pixel gray value of the spectral image in the vertical direction (Y direction) to generate a curve reflecting the spectral energy distribution.

[0085] The curve obtained by vertical integration is the vertical integration gray profile curve of the spectral image (referred to as the spectral profile curve). The horizontal axis of the curve represents the frequency (or pixel position), and the vertical axis represents the accumulation result of the gray value. The peak position and width reflect the distribution characteristics of the main frequency components in the spectral image.

[0086] It can be understood that the spectral image is obtained by transforming the region image, and the main frequency components thereof reflect the arrangement characteristics of the pattern in the region image. Since the pattern of the feature region has a periodic arrangement, its structure exhibits a repeated pattern in the spatial domain, and corresponds to a specific frequency component in the spectral image. Therefore, the main frequency component in the spectral image is actually a form of the periodic structure of the pattern in the feature region in the frequency domain.

[0087] S130, determining the configuration parameter by identifying the target peak in the plurality of spectral profile curves.

[0088] The configuration parameter includes the position and width of the target peak.

[0089] The target peak is a frequency point with a significant peak value in the plurality of spectral profile curves.

[0090] The position of the target peak is the position coordinate corresponding to the peak value of the target peak, which corresponds to the horizontal coordinate (i.e. frequency or pixel position) of the peak value in the spectral profile curve.

[0091] The width of the target peak is the horizontal span of the target peak on the spectral profile curve, that is, the frequency range when the peak value on both sides falls to a certain threshold.

[0092] In one implementation, all spectral profile curves are analyzed, all peaks in each curve are identified, and all peaks in the plurality of spectral profile curves are determined as target peaks. For each target peak, its position and width are obtained as configuration parameters. This method is suitable for scenarios that require comprehensive analysis of all frequency components in the spectral image, ensuring that each significant frequency feature is captured.

[0093] In an implementation, peaks in each frequency profile curve are identified, and peaks satisfying preset conditions are determined as target peaks. The preset conditions may, for example, include the amplitude size, width, position range, etc. of the peaks, for screening out frequency components most representative and effective for a specific task. For example, only peaks with an amplitude exceeding a certain threshold and a moderate width are identified as target peaks, and other peaks not meeting the conditions are ignored. This method can reduce the interference of noise or unimportant frequency components on the final configuration parameters, and is suitable for scenarios with strict requirements on specific frequency characteristics.

[0094] S140, configuring a hardware filter according to the configuration parameters, and performing global filtering processing on the scan image of the target wafer by using the hardware filter.

[0095] It should be noted that the embodiments of the present application set up a special hardware device in the optical path to implement Fourier filtering, so as to efficiently process and convert frequency components in the optical signal.

[0096] For example, the special hardware device may include an optical Fourier transform module, an optical processing unit (OPU), an application-specific integrated circuit (ASIC), or the like.

[0097] It can be understood that the special hardware device described above includes a hardware filter, and the hardware filter includes a filter mask (i.e. mask). For example, the mask may be a specific pattern generated by an optical filter, a spatial light modulator, or a digital micro-mirror device (DMD), for selectively filtering or adjusting optical signals of different frequency components.

[0098] In this step, the configuration parameters provide the required settings and control information for the hardware-based Fourier filtering process, so that the hardware can perform global filtering processing on the scan image of the target wafer.

[0099] In order to apply the configuration parameters to the hardware-based Fourier filtering, the configuration parameters can be transmitted to the hardware device through a configuration menu of a user interface or a hardware control system.

[0100] For example, the configuration parameters can be manually input by a user or automatically input by an automatic system, and saved to the device memory, to ensure that the hardware can perform subsequent operations according to the configuration parameters.

[0101] The hardware device reads the configuration parameters in the configuration menu, and performs a Fourier filtering operation based on the configuration parameters. Specifically, the scanning image is converted from the spatial domain to the frequency domain, and a filtering operation is applied to suppress certain frequency components. The filtering process adjusts the parameters of the filter according to the configuration parameters, such as setting the frequency bandwidth, filtering strength, etc.

[0102] After completing the Fourier transform, the filtering process acts on the entire scanning image, removing unnecessary noise, highlighting certain frequency components, or enhancing periodic features in the image. This improves the quality of the scanning image of the target wafer, facilitating subsequent defect detection.

[0103] The image processing method proposed in the embodiments of the present application identifies a plurality of feature regions with periodic arrangement patterns in the scanning image of any die in the target wafer, performs Fourier transform calculation on the region image corresponding to each feature region, finds the target peak representing the feature region in the obtained frequency spectrum profile curve, and uses the position and width of the target peak as configuration parameters to configure the hardware filter and perform global filtering processing on the target wafer. This method uses software Fourier filtering to guide hardware Fourier filtering, realizes adaptive noise filtering in periodic pattern regions on the wafer surface, significantly improves the adaptability and processing efficiency of the filter, and effectively enhances the noise filtering effect, thereby improving the accuracy and reliability of defect detection.

[0104] It should be noted that the method of the present application fully combines the advantages of hardware Fourier filtering and software Fourier filtering, and solves the respective limitations of the two: hardware Fourier filtering has a faster processing speed, but is limited by a fixed filtering strategy, has poor adaptability and flexibility, and cannot be dynamically adjusted according to different image features; while software Fourier filtering has strong flexibility and can adjust the filtering strategy according to different needs, but its calculation is slow and difficult to meet the requirements of real-time analysis.

[0105] To make up for this deficiency, the method uses configuration parameters representing filtering frequencies calculated by software Fourier filtering to guide the position adjustment of the mask in the hardware Fourier filter, so that it is directly applied in the hardware filtering process. In this way, software calculation provides the necessary flexibility and adaptability, while the hardware accelerator accelerates the execution of frequency domain filtering through logical circuits. It is particularly emphasized that, compared with the scheme of using a virtual filter in software Fourier filtering, the present scheme directly completes frequency domain filtering in the optical path using a hardware filter, avoiding dependence on external CPUs / GPUs, and having the advantages of high real-time performance, fast calculation efficiency, and strong noise resistance. In this way, the entire filtering process not only dynamically adjusts the filtering strategy in combination with the changes in the feature regions, but also efficiently applies the adjusted strategy to the filtering of the entire wafer image, realizing the combination of high efficiency and flexibility.

[0106] Figure 2 Fig. 1 shows a flowchart of an example of a method for image processing according to an embodiment of the present application.

[0107] It should be understood that Figure 2 The illustrated embodiment can be regarded as an example of step S140.

[0108] As Figure 2 The method can comprise the following steps.

[0109] S210, configuring the position and coverage range of the filter mask in the hardware filter according to the configuration parameters.

[0110] The filter mask in the hardware filter is configured according to the position and width of the target peak in the configuration parameters.

[0111] Specifically, the position of the target peak is used to determine the position of the filter mask in the spectral image, and the width of the target peak is used to set the coverage range of the filter mask.

[0112] Through the configuration, the filter mask can accurately cover the area corresponding to the target peak in the spectral image, thereby ensuring that the filter can suppress the interference frequency components in the correct position and appropriate range.

[0113] S220, performing Fourier transform to convert the scan image of the target wafer to the frequency domain to obtain a spectral image of the target wafer.

[0114] S230, in the frequency domain, applying the configured filter mask to suppress the interference frequency components in the spectral image to obtain an optimized spectral image.

[0115] Specifically, the mask will apply filtering to the part of the spectral image that coincides with the interference frequency components to suppress these unwanted frequency components.

[0116] It can be understood that the interference frequency components of this step are actually frequency components of the periodic pattern.

[0117] S240, performing Fourier transform to restore the optimized spectral image to the spatial domain to obtain a filtered target image.

[0118] As can be seen from the above, step S210 is a configuration process, and steps S220-S240 are an application process, which can be summarized as: using the configured filter mask to perform global filtering processing on the scan image of the target wafer to obtain a filtered target image.

[0119] For example, Figure 6 Fig. 2 shows a schematic diagram of an example of a spectral profile curve generation process according to an embodiment of the present application.

[0120] In this example, the feature region a and the feature region b correspond to different region images a and b respectively.

[0121] The region image a corresponding to the feature region a is subjected to Fourier transform to obtain a spectrum image a.

[0122] Through Fourier transform, the image can be converted from the spatial domain to the frequency domain, and the information of each frequency component in the image is extracted.

[0123] The spectrum image a is subjected to longitudinal integration to obtain a spectrum profile curve a.

[0124] The longitudinal integration process extracts the frequency distribution characteristics of the image in a specific direction by summing each column in the spectrum image, forming a spectrum profile curve.

[0125] Similarly, the region image b corresponding to the feature region b is subjected to Fourier transform to obtain a spectrum image b, and the spectrum image b is subjected to longitudinal integration to obtain a spectrum profile curve b.

[0126] The horizontal axis of the spectrum profile curves a and b represents the frequency, and the vertical axis represents the cumulative result of the gray value.

[0127] Through the above steps, the region image can be converted into a spectrum profile curve, which further provides basic data for subsequent filtering processing.

[0128] In the embodiment, the filter mask in the hardware filter is configured according to the configuration parameters, the position and coverage of the mask are determined, and the global filtering processing is applied to the scan image of the target wafer by using the configured filter mask, and finally the filtered target image is obtained. By using the configuration parameters obtained in the foregoing, the scan image of the entire target wafer is subjected to global filtering processing at the hardware level, the noise in the periodic pattern region is accurately filtered out, and the adaptability and processing efficiency of the filtering are significantly improved.

[0129] Specifically, after the spectrum image is obtained by Fourier transform on the scan image of the target wafer, the configured filter mask is applied to suppress the interference frequency components, and the spectrum image is optimized, and then the optimized spectrum image is restored to the spatial domain through inverse Fourier transform, so as to obtain the target image. Through this process, the interference frequency components corresponding to multiple feature regions can be accurately filtered out, effective noise suppression can be realized, important features in the image can be retained, and the image quality and the accuracy of defect detection can be significantly improved.

[0130] Figure 3 A flowchart of an image processing method provided by an embodiment of the application is shown.

[0131] It should be understood that Figure 3 The embodiments shown can be regarded as Figure 1One example of step S130 in the embodiment.

[0132] As shown in the figure, the method can include the following steps. Figure 3

[0133] S310, superimpose a plurality of spectral profile curves to obtain a comprehensive spectral profile curve.

[0134] Add the gray value (intensity value) of the same frequency position in each spectral profile curve to obtain a new spectral profile curve.

[0135] For example, in the process of intensity superposition, weighted average (if different spectral curves have different importance) or simple addition (each curve is equally important) can be used to synthesize the final comprehensive spectral profile curve.

[0136] For example, before intensity superposition, each curve can be normalized to make the intensity range of different curves consistent, so as to avoid the influence of some curves on the result being too large.

[0137] Figure 7 A schematic diagram of superimposing spectral profile curves provided by the embodiment of the application is shown.

[0138] As in the previous example, the gray value (intensity value) of the same frequency position in the spectral profile curve a and the spectral profile curve b is added to obtain a new comprehensive spectral profile curve c.

[0139] S320, identify all local peaks in the comprehensive spectral profile curve, and determine the local peaks whose peak values meet the preset requirements as target peaks.

[0140] Local peaks represent significant frequency components in the spectral image, and their shapes are higher than those of adjacent points and form a local peak value.

[0141] In one implementation, the process of identifying local peaks includes:

[0142] On the comprehensive spectral profile curve, find all local maxima higher than adjacent points, and these local extreme values are the peaks. For example, a peak detection algorithm (such as difference method or sliding window method) can be used to determine the position of each local peak.

[0143] After identifying all local peaks, the target peaks are screened out through a preset condition. For example, a minimum peak amplitude, a maximum width, or a frequency range condition is set, and only those local peaks that meet these requirements are retained as target peaks. Such screening helps to remove noise or irrelevant frequency components and only retains frequency features valuable for subsequent processing.

[0144] Figure 8 ​A schematic diagram for determining the target peak is shown.

[0145] In the foregoing manner, the peaks 1-6 are preliminarily identified in the comprehensive frequency spectrum profile curve c. In this example, the preset requirement for determining the target peak is that the peak value amplitude is not less than 20000. On this basis, the peaks 1-6 are determined to be target peaks respectively.

[0146] Exemplarily, before identifying all local peaks in the comprehensive frequency spectrum profile curve, the comprehensive frequency spectrum profile curve can be smoothed based on a preset window size.

[0147] The smoothing can be implemented by using a sliding window.

[0148] The preset window size determines the neighborhood range considered in each smoothing. It should be understood that the preset window size can be set according to the frequency distribution and noise characteristics, which is not limited in the present application.

[0149] Exemplarily, the spectral data in the sliding window can be weighted and averaged by using a smoothing algorithm (such as Gaussian smoothing) to generate the smoothed data points.

[0150] The sliding window slides on the comprehensive frequency spectrum profile curve, and gradually smoothes each part of the curve until the smoothing is completed.

[0151] Before identifying the local peak, the comprehensive frequency spectrum profile curve is smoothed first, which effectively reduces the interference of noise and makes the spectral characteristics more clear and prominent.

[0152] S330, obtaining the position and width of the target peak, and determining as a configuration parameter.

[0153] This step has been introduced in step S130, and will not be repeated here.

[0154] In the method of the embodiment, the comprehensive frequency spectrum profile curve obtained by superimposing the intensities of the plurality of frequency spectrum profile curves can present the overall characteristics of the frequency components in all the frequency spectrum profile curves, thereby enhancing the identification ability of the spectral characteristics. In addition, the peaks meeting the preset requirement are selected as the target peak by setting the screening condition, thereby further improving the accuracy and effectiveness of the spectral characteristics. Through the above steps, the identification of the local peak is more accurate, the misidentification and noise interference are avoided, and the position and width of the final target peak are more accurate, thereby improving the reliability and precision of the configuration parameter extraction.

[0155] In one embodiment, after the global filtering processing of the scanning image of the target wafer, an image quality evaluation mechanism can be introduced to ensure the effectiveness of the filtering operation. Specifically, the method of the present application can further comprise:

[0156] Every preset sampling period, the filtered local scanning image of the target wafer is extracted and evaluated to obtain an image quality evaluation value.

[0157] Exemplarily, the image quality evaluation value can be obtained by analyzing multiple visual features of the local scanning image, including brightness, contrast, saturation, and orthogonality of the pattern, etc.

[0158] According to the comparison result of the image quality evaluation value and the preset image quality index, it is determined whether the filtering is completed.

[0159] Correspondingly, the preset image quality index can be a comprehensive index preset in combination with the brightness, contrast, saturation, orthogonality of the pattern in the image, etc.

[0160] Exemplarily, the preset image quality index can be transmitted to the hardware device as a control basis through the configuration parameter.

[0161] In the case where the image quality evaluation value is less than the preset image quality index, it indicates that the image quality has not yet reached the expected standard, and the step of obtaining the configuration parameter will be re-executed, and the filtering operation will be performed again according to the new configuration parameter.

[0162] It should be understood that this process includes re-measuring the selection or width of the target peak in the spectral profile curve, analyzing the position from the left end to the right end of the target peak, and adjusting the position of the filtering mask according to the change, further guiding the hardware to perform the filtering operation.

[0163] In the case where the image quality evaluation value is greater than or equal to the preset image quality index, it indicates that the image quality has met the requirements, and the configuration parameter remains unchanged, and the filtering operation continues to be performed.

[0164] In this embodiment, by extracting and evaluating the local scanning image every preset sampling period after performing the hardware-based Fourier filtering, obtaining the image quality evaluation value, and comparing it with the preset image quality index, the change of the image quality can be dynamically monitored, and it is automatically determined whether the filtering is completed, thereby improving the adaptability and accuracy of the processing. When the image quality requirement is not met, the peak characteristics in the spectral profile curve are re-analyzed, the configuration parameter is adjusted, and the hardware filter is guided to perform the filtering operation again, thereby ensuring the continuous optimization of the image processing quality to reach the expected image quality standard, and ensuring that the final target image meets the quality requirements.

[0165] Figure 4 A flowchart of an image processing method provided by an embodiment of the present application is shown.

[0166] It should be understood that Figure 4The embodiment can be regarded as an example of introducing the complete process of the method of the application.

[0167] As Figure 4 shown, the method can include the following steps.

[0168] S410, start.

[0169] S420, pre-acquire a complete image of a die in the target wafer.

[0170] S430, divide all feature regions in the complete image of the die, and code each feature region.

[0171] S440, perform Fourier transform on the region image corresponding to each coded feature region to obtain a spectrum image.

[0172] S450, perform longitudinal integration on the spectrum image to construct a spectrum profile curve.

[0173] S460, synthesize all spectrum profile curves to identify a plurality of target peaks, and take the positions and widths of the plurality of target peaks and a preset image quality index Q as configuration parameters.

[0174] S470, write the above configuration parameters into a configuration menu of a hardware filter.

[0175] S480, start full-wafer scanning filter processing.

[0176] S490, sample and check the filtered image, and calculate an image quality evaluation value Q1.

[0177] S500, judge whether Q1≥Q is satisfied.

[0178] If yes, execute S600, continue the image detection process; if no, return to execute S440 until Q1≥Q.

[0179] Based on the image processing method. Accordingly, the application also provides specific embodiments of an image processing device.

[0180] As Figure 9 shown, the image processing device 1000 provided by the embodiments of the application includes the following modules.

[0181] The division module 1001 is configured to divide a plurality of region images corresponding to feature regions from a scanning image of any die in a target wafer.

[0182] All patterns in a region image have the same shape and periodic arrangement characteristics.

[0183] The virtual Fourier transform module 1002 is configured to perform Fourier transform on the plurality of area images and longitudinally integrate to obtain a plurality of spectral profile curves.

[0184] The configuration parameter determination module 1003 is configured to determine the configuration parameter by identifying a target peak in the plurality of spectral profile curves.

[0185] The hardware Fourier filter module 1004 is configured to configure a hardware filter according to the configuration parameter, and perform global filtering processing on the scanning image of the target wafer by using the hardware filter.

[0186] The image processing method is based on the hardware Fourier filter.

[0187] Figure 10 A hardware structure schematic diagram of an electronic device provided by an embodiment of the present application is shown.

[0188] The electronic device can include a processor 7001 and a memory 7002 storing computer program instructions.

[0189] Specifically, the processor 7001 can include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or can be configured to implement one or more integrated circuits of the embodiments of the present application.

[0190] The memory 7002 can include a mass storage for data or instructions. By way of example and not limitation, the memory 7002 can include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. Where appropriate, the memory 7002 can include removable or non-removable (or fixed) media. Where appropriate, the memory 7002 can be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, the memory 7002 is a non-volatile solid-state memory.

[0191] The processor 7001 reads and executes the computer program instructions stored in the memory 7002 to implement any of the image processing methods in the above embodiments.

[0192] In one example, the electronic device can further include a communication interface 7003 and a bus 7004. As shown, the processor 7001, the memory 7002, and the communication interface 7003 are connected by the bus 7004 and complete communication among each other. Figure 5 In one example, the electronic device can further include a communication interface 7003 and a bus 7004. As shown, the processor 7001, the memory 7002, and the communication interface 7003 are connected by the bus 7004 and complete communication among each other.

[0193] The communication interface 7003 is mainly configured to implement communication between the modules, devices, units and / or equipment in the embodiments of the present application.

[0194] The bus 7004 includes hardware, software, or both, that couples components of the electronic device to each other in a communicative manner. By way of example and not limitation, the bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or another suitable bus or combination of two or more of these. Where appropriate, the bus 7004 can include one or more buses. Although the present embodiments describe and show a particular bus, the present application contemplates any suitable bus or interconnect.

[0195] In addition, in combination with the image processing method in the above embodiments, the present embodiments can provide a computer storage medium for implementation. The computer storage medium has computer program instructions stored thereon; the computer program instructions are executed by a processor to implement any of the image processing methods in the above embodiments.

[0196] In addition, in combination with the image processing method in the above embodiments, the present embodiments can provide a computer program product for implementation, and the instructions in the computer program product are executed by the processor of the electronic device to enable the electronic device to execute the image processing method provided by any of the above embodiments of the present application.

[0197] It should be noted that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of well-known methods are omitted herein. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order of the steps, after understanding the spirit of the present application.

[0198] The functions noted in the description of the structural block diagrams above can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, and the like. When implemented in software, the elements of the present application are program or code segments that are used to perform the required tasks. The program or code segments can be stored in a machine-readable medium, or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. A "machine-readable medium" includes any medium that can store or transport information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and the like. The code segments can be downloaded via computer networks such as the Internet, intranets, and the like.

[0199] It is also important to note that the examples mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the steps mentioned above, that is, the steps can be performed in the order mentioned in the examples, or in an order different from the examples, or several steps can be performed simultaneously.

[0200] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other processing device to operate in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function / act specified in the flowchart and / or block diagram block or blocks. The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other processing device to operate in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function / act specified in the flowchart and / or block diagram block or blocks.

[0201] The above merely describes a specific implementation of the present application. Those skilled in the art can clearly understand the specific working processes of the system, modules and units described above for the convenience and brevity of description, and can refer to the corresponding processes in the foregoing method embodiments, which will not be described herein again. It should be understood that the protection scope of the present application is not limited to this, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. An image processing method, characterized by, The method comprises the following steps: Divide a plurality of feature region corresponding region images from the scanning image of any die in the target wafer, all patterns in a feature region have the same shape and periodic arrangement characteristics; Perform Fourier transform on the plurality of region images to obtain a plurality of frequency spectrum images, and perform longitudinal integration on the plurality of frequency spectrum images to obtain a plurality of frequency spectrum profile curves; Determine the configuration parameters by identifying the target peak in the plurality of frequency spectrum profile curves, the configuration parameters including the position and width of the target peak; Configure the hardware filter according to the configuration parameters, and perform global filtering processing on the scanning image of the target wafer by using the hardware filter; The method of determining the configuration parameters by identifying the target peak in the plurality of frequency spectrum profile curves comprises: Superimpose the plurality of frequency spectrum profile curves to obtain a comprehensive frequency spectrum profile curve; Identify all local peaks in the comprehensive frequency spectrum profile curve, and determine the local peak with a peak value meeting the preset requirement as the target peak; wherein the local peak represents a significant frequency component in the frequency spectrum image, the shape of the local peak is higher than that of the adjacent points, and a local peak value is formed; Obtain the position and width of the target peak to determine the configuration parameters; The configuration parameters further include a preset image quality index; After configuring the hardware filter according to the configuration parameters and performing global filtering processing on the scanning image of the target wafer by using the hardware filter, the method further comprises: Extract and evaluate the local scanning image of the target wafer after filtering every preset sampling period to obtain an image quality evaluation value; Judge whether the filtering is completed according to the comparison result of the image quality evaluation value and the preset image quality index; In the case that the image quality evaluation value is less than the preset image quality index, re-execute the steps of obtaining the configuration parameters, configuring the hardware filter according to the configuration parameters, and performing global filtering processing on the scanning image of the target wafer by using the hardware filter; In the case that the image quality evaluation value is greater than or equal to the preset image quality index, continue to perform the filtering operation.

2. The method of claim 1, wherein, The method of configuring the hardware filter according to the configuration parameters and performing global filtering processing on the scanning image of the target wafer by using the hardware filter comprises: Configure the position and coverage range of the filter mask in the hardware filter according to the configuration parameters; Perform global filtering processing on the scanning image of the target wafer by using the configured filter mask to obtain a filtered target image.

3. The method of claim 2, wherein, The method of performing global filtering processing on the scanning image of the target wafer by using the configured filter mask to obtain a filtered target image comprises: Perform Fourier transform to convert the scanning image of the target wafer to the frequency domain to obtain a frequency spectrum image of the target wafer; In the frequency domain, apply the configured filter mask to suppress the interference frequency components in the frequency spectrum image to obtain an optimized frequency spectrum image; Perform inverse Fourier transform to restore the optimized frequency spectrum image to the spatial domain to obtain the target image.

4. The method of claim 1, wherein, Before identifying all local peaks in the comprehensive frequency spectrum profile curve, the method further comprises: The comprehensive frequency spectrum profile curve is smoothed based on a preset window size.

5. An electronic device, comprising: The device comprises a processor and a memory storing computer program instructions; The processor executes the computer program instructions to implement the image processing method of any one of claims 1-4.

6. A computer-readable storage medium, characterized in that, The computer program instructions stored on the computer readable storage medium are executed by the processor to implement the image processing method of any one of claims 1-4.

7. A computer program product, characterised in that, The instructions in the computer program product are executed by the processor of the electronic device to enable the electronic device to perform the image processing method of any one of claims 1-4.

Images