Wafer imaging image processing compensation method and system

Abstract

The application provides a wafer imaging image processing compensation method and system, and relates to the technical field of image processing. The method comprises the following steps: acquiring a wafer image collected by a wafer imaging system; performing image feature analysis on the wafer image to determine the difference in reflection characteristics of different regions in the wafer image; dividing the wafer image into at least two regions with different reflection characteristics according to the difference in reflection characteristics; determining corresponding image compensation parameters for different regions with different reflection characteristics; performing adaptive compensation processing on the wafer image based on the image compensation parameters to obtain a compensated wafer image; wherein the image compensation parameters are limited by a preset imaging consistency constraint condition during the determination process. The application can compensate for the differences between different regions with different reflection characteristics, avoid the problem of imbalance and inconsistency in traditional imaging, and improve the consistency of the imaging results.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a wafer imaging image processing compensation method and system. Background Technology

[0002] In semiconductor manufacturing, the wafer surface is typically composed of various materials, such as conductors, dielectrics, and multilayer thin-film structures. Different materials exhibit different reflective properties during optical imaging, resulting in significant differences in brightness, contrast, and edge response across different regions of the wafer in the image.

[0003] Existing wafer imaging processing methods mostly employ uniform image enhancement or fixed parameter compensation approaches, performing the same brightness adjustment, contrast enhancement, or edge processing on the entire wafer image. When faced with different material reflectivity characteristics, this processing method may result in local over-enhancement or under-enhancement, introducing false detections or missed detections, and it is difficult to achieve stable and consistent imaging results in complex wafer structures with mixed materials. Summary of the Invention

[0004] The technical problem to be solved by this application is to provide a wafer imaging image processing compensation method and system that can avoid inconsistencies in imaging results and inaccurate detection and measurement results caused by differences in the reflectivity of different regions and materials.

[0005] The technical solution adopted by this application to solve the above-mentioned technical problems is as follows:

[0006] This application provides a wafer imaging image processing compensation method, comprising: acquiring a wafer image collected by a wafer imaging system; performing image feature analysis on the wafer image to determine the differences in reflectance characteristics of different regions in the wafer image; dividing the wafer image into at least two reflectance characteristic regions according to the differences in reflectance characteristics; determining corresponding image compensation parameters for different reflectance characteristic regions; and performing adaptive compensation processing on the wafer image based on the image compensation parameters to obtain a compensated wafer image; wherein the image compensation parameters are subject to preset imaging consistency constraints during the determination process.

[0007] Optionally, the image features include at least one of grayscale distribution features, edge gradient features, and frequency domain features.

[0008] Optionally, the differences in reflectivity include differences in material reflectivity and differences in reflectivity of the graphic structure.

[0009] Optionally, dividing the wafer image into at least two reflective characteristic regions based on the differences in reflective properties specifically includes: extracting the normalized grayscale mean of a local region of the image. As a measure of material reflectance difference ,in This represents the average gray level of a local area. The maximum grayscale mean of the entire image; extracting the edge density of local regions of the image. As a measure of the difference in reflection of graphic structure ,in This represents the number of edge pixels in a local region extracted using the Canny edge detection operator. The total number of pixels in the local region; the material weighting coefficient is dynamically determined based on the known material type information and pattern spatial frequency of the local region. With graphic weight coefficient ,in According to the weighted formula Calculate the comprehensive score of reflection characteristics for each region. A comprehensive score based on the aforementioned reflection characteristics. The distribution of the wafer image is determined by using a clustering algorithm to divide the wafer image into at least two reflective regions.

[0010] Optionally, the dynamic determination of material weighting coefficients With graphic weight coefficient This includes: if the number of known material types corresponding to the local region is greater than 1, then the value is increased relative to the default value. and reduce If the pattern spatial frequency of the local region is higher than a preset threshold, then the value is reduced relative to the default value. and increase .

[0011] Optionally, the dynamic determination of material weighting coefficients With graphic weight coefficient It also includes: obtaining the texture complexity T by calculating the contrast of the gray-level co-occurrence matrix of the local region; the weighting formula is introduced into the texture complexity T to obtain: ;in, For texture weight coefficients, and .

[0012] Optionally, the texture weight coefficient The pattern regularity of the local region is adaptively determined; the lower the pattern regularity, the better. The larger the value, the better.

[0013] Optionally, the image compensation parameters include at least one of brightness compensation parameters, contrast compensation parameters, and edge enhancement parameters.

[0014] Optionally, the image compensation parameters are adaptively determined based on the degree of reflection difference in the corresponding reflective characteristic regions.

[0015] Optionally, the imaging consistency constraint is determined based on at least one of the differences in grayscale distribution, edge response, or contrast among regions with different reflective properties after compensation.

[0016] This application also provides a wafer imaging image processing compensation system for implementing the wafer imaging image processing compensation method described above, comprising: an image acquisition module for acquiring a wafer image acquired by a wafer imaging system; a difference analysis module for performing image feature analysis on the wafer image to determine the differences in reflectance characteristics of different regions; a region division module for dividing the wafer image into at least two reflectance characteristic regions based on the differences in reflectance characteristics; a parameter determination module for determining corresponding image compensation parameters for different reflectance characteristic regions, wherein the image compensation parameters are subject to preset imaging consistency constraints during the determination process; and an image compensation module for performing adaptive compensation processing on the wafer image based on the image compensation parameters.

[0017] This application provides a wafer imaging image processing compensation method and system. By dividing the region according to the difference in reflectance characteristics and determining the image compensation parameters for each region, differentiated compensation for regions with different reflectance characteristics is achieved, avoiding the problem of inconsistency in traditional imaging and improving the consistency of imaging results. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of the wafer imaging image processing compensation method of this application.

[0020] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0022] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this invention are only used to explain the relative positional relationship and movement of the components in a specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0023] In this application, unless otherwise expressly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; it can mean a mechanical connection or an electrical connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0024] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the meaning of "and / or" throughout the text includes three parallel solutions; for example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0025] Existing wafer imaging processing methods mostly employ uniform image enhancement or fixed parameter compensation approaches, performing the same brightness adjustment, contrast enhancement, or edge processing on the entire wafer image. This processing method, when faced with different material reflectivity characteristics, can lead to localized over-enhancement or under-enhancement, introducing false detections or missed detections, and it is difficult to achieve stable and consistent imaging results in complex wafer structures with mixed materials. To address the above problems, the embodiments of this application refer to... Figure 1 The following technical solutions are provided to overcome the above problems.

[0026] Please refer to Figure 1This application provides a wafer imaging image processing compensation method, comprising: acquiring a wafer image collected by a wafer imaging system; performing image feature analysis on the wafer image to determine the differences in reflectance characteristics of different regions in the wafer image; dividing the wafer image into at least two reflectance characteristic regions according to the differences in reflectance characteristics; determining corresponding image compensation parameters for different reflectance characteristic regions; and performing adaptive compensation processing on the wafer image based on the image compensation parameters to obtain a compensated wafer image; wherein the image compensation parameters are subject to preset imaging consistency constraints during the determination process.

[0027] Specifically, wafer images are acquired using a wafer imaging system. These images can be bright-field, dark-field, or other optical imaging images, and their source is not limited. After acquisition, image feature analysis is performed on the wafer images. The purpose of image feature analysis is to determine the differences in reflectance characteristics between different regions of the wafer image. Image feature analysis may include statistical analysis of grayscale distribution, local contrast, edge response intensity, or other features that characterize reflectance. Through this analysis, regions with significantly different reflectance intensities in the wafer image can be identified.

[0028] After determining the differences in reflectivity among different regions of the wafer image, the image is divided into at least two reflectivity regions based on these differences. A reflectivity region refers to a group of image regions that exhibit similar reflectivity in their imaging response. This region division avoids overcompensation or undercompensation issues that arise from applying a uniform compensation method to the entire image, resulting in more accurate compensation.

[0029] After region segmentation, corresponding image compensation parameters are determined for regions with different reflective characteristics. These parameters can be used to adjust regional brightness, contrast, or edge response intensity, and their specific form is not limited. The key is that different regions with different reflective characteristics require different image compensation parameters, thus achieving adaptive region compensation.

[0030] When determining image compensation parameters, preset imaging consistency constraints are introduced. These constraints limit the imaging performance of regions with different reflectivity characteristics after compensation, ensuring they meet predetermined consistency targets. For example, the grayscale distribution difference between regions with different reflectivity characteristics after compensation can be limited to a preset range, or the contrast difference between different regions can be limited to a target range. By introducing imaging consistency constraints, excessive enhancement of a single region can be prevented from disrupting the overall imaging balance, thereby improving the stability of the compensation results.

[0031] Subsequently, based on the image compensation parameters, adaptive compensation processing is performed on the wafer image to obtain the compensated wafer image. Adaptive compensation processing involves applying corresponding image compensation parameters to different reflective characteristic regions. After compensation, the imaging differences between different regions are effectively reduced, while retaining the effective structural information in the original image.

[0032] Understandably, in this embodiment, after acquiring the wafer image captured by the wafer imaging system, image feature analysis is performed on the wafer image, and the mean and variance of gray levels in each local region are statistically analyzed. The degree of difference in reflectivity is determined based on the statistical results. Regions with a mean gray level higher than the overall image average are classified as first reflectivity regions, and regions with a mean gray level lower than the overall image average are classified as second reflectivity regions. Subsequently, a first image compensation parameter is determined for the first reflectivity regions to appropriately reduce brightness; a second image compensation parameter is determined for the second reflectivity regions to appropriately increase brightness. When determining the first and second image compensation parameters, an imaging consistency constraint is set such that the average gray level difference between the two types of regions after compensation does not exceed a preset threshold. If the difference after compensation exceeds the preset threshold, the image compensation parameters are adjusted until the imaging consistency constraint is met. Finally, a compensated wafer image is obtained, in which the gray level distribution in different regions tends to be consistent, while maintaining edge details without being weakened.

[0033] This application provides a wafer imaging image processing compensation method. By dividing the region according to the difference in reflectance characteristics and determining the image compensation parameters for each region, differentiated compensation for regions with different reflectance characteristics is achieved, avoiding the problem of inconsistency in traditional imaging and improving the consistency of imaging results.

[0034] In one embodiment, the image features include at least one of grayscale distribution features, edge gradient features, and frequency domain features.

[0035] Specifically, grayscale distribution features are used to characterize the spatial brightness distribution of a wafer image. In practice, statistical quantities such as the grayscale mean, grayscale variance, and grayscale histogram distribution shape can be calculated for local areas of the wafer image to reflect the reflection intensity level of different areas. Since different materials or different structural regions have different reflectivity to incident light, their grayscale distribution features usually differ. Therefore, grayscale distribution features can effectively identify areas with significant differences in reflection intensity.

[0036] Edge gradient features are used to characterize the intensity of grayscale changes in a wafer image. In practice, gradient calculations can be performed on the wafer image to calculate the gradient magnitude of each pixel and to statistically analyze the average gradient intensity or gradient distribution within local regions. Regions with complex graphic structures typically have a higher frequency of grayscale changes, resulting in relatively strong edge gradient features; conversely, regions with relatively flat structures exhibit relatively weak edge gradient features. Therefore, edge gradient features can reflect changes in reflectivity caused by the graphic structure.

[0037] Frequency domain features are used to characterize the energy distribution of a wafer image in frequency space. In practice, frequency transformation can be performed on the wafer image to obtain the frequency component distribution and analyze the energy ratio in different frequency ranges. Regions with strong high-frequency components typically correspond to areas with intricate patterns or high-density structures, while regions dominated by low-frequency components typically correspond to flat or uniform areas. Since different structures have varying effects on the imaging response, frequency domain features can reflect regional differences from the perspective of the overall frequency structure.

[0038] By selecting at least one of grayscale distribution features, edge gradient features, and frequency domain features for image feature analysis, the imaging characteristics of wafer images can be characterized from multiple angles. Compared to methods that only use single grayscale information, this method can combine brightness information, structural change information, and frequency structure information to improve the accuracy of identifying differences in reflectivity. When only grayscale distribution features are used, it is suitable for scenarios with significant differences in material reflectivity; when edge gradient features are used, it is more suitable for scenarios with significant structural differences; when frequency domain features are used, it is suitable for scenarios that require analysis from the perspective of the overall structural frequency distribution. Selecting one or more image features according to actual application needs can enhance the adaptability of the method.

[0039] In one embodiment, the difference in reflectivity includes differences in material reflectivity and differences in reflectivity of graphic structures.

[0040] Specifically, in the wafer manufacturing process, different process layers may include metal layers, insulating layers, or semiconductor layers. These materials have different reflectivities of incident light, resulting in differences in grayscale levels in the wafer image. When the overall brightness of certain areas in a wafer image is significantly higher or lower than that of other areas, it can usually be attributed to differences in material reflection. Therefore, when performing image feature analysis, the existence of material reflection differences can be determined by statistically analyzing the grayscale distribution of different areas.

[0041] Pattern structure reflection differences refer to the differences in imaging response caused by variations in the geometry, linewidth density, or spatial distribution of pattern structures on a wafer surface. Even with the same material type, different pattern structures can lead to different optical scattering and interference effects, resulting in different grayscale variation patterns in wafer images. For example, densely patterned areas may produce more edge responses or high-frequency components, while sparsely patterned areas exhibit a smoother grayscale distribution. Pattern structure reflection differences can be identified by analyzing the degree of grayscale variation or structural complexity in wafer images.

[0042] In this embodiment, by clearly dividing the differences in reflection characteristics into material reflection differences and graphic structure reflection differences, the analysis of these differences can have a clear physical correspondence. Material reflection differences are mainly reflected in the overall reflection intensity difference, while graphic structure reflection differences are mainly reflected in the structural response difference. These two types of differences differ in their imaging manifestations, and distinguishing between them can avoid misjudgment during the analysis process. For example, if a certain area has a higher grayscale value, and this is accompanied by a change in structural complexity, it may belong to a graphic structure reflection difference; if the grayscale value is high but the structural features are basically consistent, it is more likely to belong to a material reflection difference.

[0043] In one embodiment, dividing the wafer image into at least two reflective characteristic regions based on the differences in reflective properties specifically includes: extracting the normalized grayscale mean of a local region of the image. As a measure of material reflectance difference ,in This represents the average gray level of a local area. The maximum grayscale mean of the entire image; extracting the edge density of local regions of the image. As a measure of the difference in reflection of graphic structure ,in This represents the number of edge pixels in a local region extracted using the Canny edge detection operator. The total number of pixels in the local region; the material weighting coefficient is dynamically determined based on the known material type information and pattern spatial frequency of the local region. With graphic weight coefficient ,in According to the weighted formula Calculate the comprehensive score of reflection characteristics for each region. A comprehensive score based on the aforementioned reflection characteristics. The distribution of the wafer image is determined by using a clustering algorithm to divide the wafer image into at least two reflective regions.

[0044] Specifically, the normalized gray-level mean M of a local region in the image is first extracted as a measure of material reflectance difference. , This represents the average gray level of a local area. The normalization process represents the maximum grayscale mean of the entire image. This normalization eliminates the influence of overall brightness level differences between different images, making images from different wafers comparable. Since different material regions typically have different reflectivities, their grayscale mean values ​​differ. Calculating the normalized grayscale mean M effectively characterizes the degree of material reflectance differences. Simultaneously, the edge density P of local image regions is extracted as a measure of reflectance differences in the graphic structure. , This represents the number of edge pixels in a local region extracted using the Canny edge detection operator. This represents the total number of pixels in the local area. Edge density reflects the complexity of the graphic structure and the frequency of spatial changes. Densely structured areas typically have a higher proportion of edge pixels, resulting in a larger P-value. Edge density P can be used to characterize the reflective differences of the graphic structure from a structural perspective.

[0045] After obtaining the material reflection difference metric M and the pattern structure reflection difference metric P, the material weighting coefficients are dynamically determined based on the known material type information and pattern spatial frequency of the local region. With graphic weight coefficient ,in By introducing material weighting coefficients With graphic weight coefficient The weights can be adjusted according to the importance of material and structural differences in the actual scene, so that the comprehensive score of reflection characteristics is more in line with the actual imaging characteristics.

[0046] Subsequently, according to the weighted formula Calculate the comprehensive reflection characteristic score S for each region. The comprehensive reflection characteristic score S comprehensively reflects the influence of material reflection differences and pattern structure reflection differences, and is a holistic quantitative representation of the imaging characteristics of a local area. A single index S allows for unified comparisons between different regions.

[0047] After obtaining the comprehensive reflectance score S for each region, the wafer image is divided into at least two reflectance characteristic regions based on the distribution of the comprehensive reflectance score S. The K-means clustering algorithm automatically groups regions according to the similarity of S values, grouping regions with similar comprehensive reflectance scores into the same reflectance characteristic region, thus achieving automated region division. This method avoids the incompatibility issues caused by manually setting fixed thresholds and improves the robustness of region division.

[0048] In this embodiment, the wafer image is divided into several local regions of fixed size. For each local region, its average grayscale value μ is calculated, and the maximum average grayscale value of the entire image is obtained. This allows for the calculation of the normalized grayscale mean M. Subsequently, the edge pixels of this local region are extracted using the Canny edge detection operator, and the number of edge pixels is counted. and the total number of pixels in the local area Calculate the edge density P. Based on current wafer process information, determine if material differences are significant. If material differences are significant, appropriately increase the material weighting coefficient. If the graphic density changes significantly, the graphic weight coefficient should be increased appropriately. In determining the material weighting coefficients With graphic weight coefficient Then, according to Calculate the comprehensive reflectance score S for each local region. Input the S values ​​of all local regions into a K-means clustering algorithm, and divide the region into at least two reflectance characteristic regions based on score similarity, such as a high reflectance characteristic region and a low reflectance characteristic region. Through the above steps, automatic division of reflectance characteristic regions based on multi-feature fusion is achieved.

[0049] Understandably, this embodiment constructs a material reflection difference metric M and a graphic structure reflection difference metric P, and uses a weighted approach to form a comprehensive reflection characteristic score S, thus achieving a shift from single-feature analysis to multi-feature fusion analysis and improving the accuracy of reflection characteristic difference identification. Simultaneously, through normalization processing and weight constraint relationships... This ensures the stability of the scoring system. Using clustering algorithms for region division allows for adaptive determination of the number and boundaries of reflective regions based on the actual image distribution, avoiding misclassification caused by simple threshold segmentation.

[0050] In one embodiment, the dynamic determination of material weighting coefficients With graphic weight coefficient This includes: if the number of known material types corresponding to the local region is greater than 1, then the value is increased relative to the default value. and reduce If the pattern spatial frequency of the local region is higher than a preset threshold, then the value is reduced relative to the default value. and increase .

[0051] Specifically, when the number of known material types corresponding to a local area is greater than 1, the material weighting coefficient will be increased. Reduce graphic weight coefficient In this case, the difference in reflectance characteristics is mainly caused by the type of material; therefore, the influence of material on image reflectance characteristics should be given more attention. When the pattern spatial frequency in a local region is higher than a preset threshold, the material weighting coefficient will be reduced. Increase the weight coefficient of the graph Because regions with higher spatial frequencies in a pattern usually mean that the pattern structure has a greater impact on reflection characteristics, the influence of the pattern structure needs to be considered more carefully during compensation.

[0052] In practical implementation, for example, in a wafer image, one area may be a metallic material and another a dielectric material, while the wafer image may also contain complex pattern structures. By implementing this technical solution, the following processing can be performed: In the metallic material region, the number of material types is greater than 1, and the pattern spatial frequency is low. Based on the differences in reflection characteristics in this region, the weighting coefficient is dynamically adjusted as follows: increase the material weighting coefficient ( =0.7), reduce the graphic weight coefficient ( =0.3). This ensures that the reflection characteristics of the metallic material area are compensated with high accuracy.

[0053] In the dielectric material region, the number of material types is 1, and the pattern spatial frequency is relatively high. The dynamic adjustment of the weighting coefficient is: decrease the material weighting coefficient ( =0.4), increase the graphic weight coefficient ( =0.6). This can better compensate for the influence of the pattern structure on the reflection characteristics and avoid insufficient compensation due to the complexity of the pattern structure.

[0054] In densely patterned areas, the number of material types may be one or more, and the pattern spatial frequency is high. Based on the pattern spatial frequency of this area, the system automatically adjusts the weighting coefficients, reducing the material weighting coefficients and increasing the graphic weighting coefficients. This dynamic adjustment ensures that the complexity of the pattern structure is adequately compensated.

[0055] In one embodiment, the dynamic determination of material weighting coefficients With graphic weight coefficient It also includes: obtaining the texture complexity T by calculating the contrast of the gray-level co-occurrence matrix of the local region; the weighting formula is introduced into the texture complexity T to obtain: ;in, For texture weight coefficients, and .

[0056] Specifically, the gray-level co-occurrence matrix of the region is first calculated, and then contrast information is extracted from the matrix as a measure of texture complexity. The texture complexity T is incorporated into the calculation of a comprehensive score for reflectivity, forming a new weighted formula: ,in Here, represents the texture weighting coefficient. This weighted formula calculates a comprehensive reflection characteristic score S for each region by comprehensively considering differences in material reflection, differences in reflection due to graphic structure, and texture complexity. According to the weighted formula, the comprehensive reflection characteristic score for each region is affected not only by differences in material reflection and differences in reflection due to graphic structure, but also by texture complexity T. By incorporating texture complexity into the calculation, the actual reflection characteristics of the image can be more accurately reflected, especially in regions with more complex textures, thereby optimizing the compensation effect.

[0057] Understandably, in this embodiment, by introducing texture complexity as a new metric, the compensation method can more comprehensively consider the structural features of the image. Especially in areas with complex textures, it can more accurately adjust compensation parameters, thereby improving the overall quality of image compensation. Texture complexity, as a measure reflecting the texture features of an image, helps to identify and process regions with different textures during image processing, avoiding the inaccuracies caused by ignoring texture differences in traditional methods. In areas with complex textures, traditional compensation methods based on material and graphic structure may not be able to fully handle details. However, by introducing texture complexity T, more refined compensation can be provided in these areas, resulting in a more balanced and natural compensation effect.

[0058] In one embodiment, the texture weight coefficient The pattern regularity of the local region is adaptively determined; the lower the pattern regularity, the better. The larger the value, the better.

[0059] Specifically, the lower the pattern regularity, the more irregular the pattern structure in that region. Therefore, the influence of texture complexity needs to be enhanced during the compensation process, thereby increasing the texture weight coefficient. The larger the value, the higher the texture weight coefficient; conversely, a higher pattern regularity indicates a more regular pattern structure in the region, which reduces the impact of texture complexity. The value is relatively small. Through this adaptive adjustment, the texture weight coefficient can be dynamically adjusted according to the structural features of the actual image. This allows for more precise compensation.

[0060] It should be noted that by adaptively determining the texture weight coefficients based on pattern regularity, the complexity of the pattern structure can be fully considered in image processing. Particularly in regions with low pattern regularity, the weight of texture complexity is appropriately increased, resulting in more detailed and accurate compensation in these areas. Conversely, in regions with high pattern regularity, the weight of texture complexity is relatively small, thus avoiding overcompensation. In this way, the texture weight coefficients can be dynamically adjusted during the compensation process, improving the accuracy and adaptability of image compensation.

[0061] In practical implementation, for example, a wafer image may contain two regions: one with low pattern regularity and the other with high pattern regularity. In the low-regularity region, due to its more complex and irregular pattern structure, a texture weight coefficient for that region is calculated. The texture complexity T has a greater impact on the overall reflection characteristic score S, resulting in more detailed and accurate compensation in this area. In areas with higher regularity, the pattern structure is relatively simple, and the texture weight coefficient... The smaller size means that the compensation process is less dependent on texture complexity, thus avoiding unnecessary overcompensation.

[0062] In one embodiment, the image compensation parameters include at least one of brightness compensation parameters, contrast compensation parameters, and edge enhancement parameters.

[0063] Specifically, brightness compensation parameters are mainly used to adjust the overall brightness of the image, solving the problem of insufficient or excessive brightness caused by differences in imaging systems or materials; contrast compensation parameters are used to enhance the contrast of the image, helping to highlight details and improve the visual effect of the image; edge enhancement parameters are mainly used to enhance the edge details of the image, making the image clearer in detail, especially for imaging complex patterns or small structures.

[0064] Understandably, by selecting appropriate image compensation parameters based on the differences in reflectivity of different regions, more precise adaptive image compensation can be achieved. In this way, during wafer image processing, different regions can select the most suitable compensation method based on their specific reflectivity characteristics, thereby obtaining a more balanced image quality. Especially when dealing with regions of different materials or pattern structures, it avoids over-compensation or under-compensation problems caused by a uniform compensation method. Furthermore, by introducing brightness compensation parameters, contrast compensation parameters, and edge enhancement parameters, the compensation method can optimize the image from multiple aspects, thereby effectively improving imaging quality. Compared to the single compensation method in traditional approaches, this technical solution introduces multiple compensation parameters, making the compensation effect more comprehensive. Particularly in regions of different materials or complex patterns, it can more finely adjust image characteristics, ensuring the consistency and accuracy of image quality.

[0065] In one embodiment, the image compensation parameters are adaptively determined based on the degree of reflection difference in the corresponding reflective characteristic region.

[0066] Specifically, the degree of reflection difference may vary for different regions with different reflective characteristics. In regions with significant reflection differences, higher compensation is required, necessitating increased image compensation parameters (such as brightness compensation, contrast compensation, and edge enhancement parameters). Conversely, in regions with smaller reflection differences, lower compensation is needed, allowing for a reduction in compensation parameters to avoid overcompensation. Adaptive methods for determining compensation parameters can flexibly adjust the intensity of image compensation based on the degree of reflection difference in different regions of the actual image, thereby improving the accuracy of the compensation effect. For example, in regions with significant reflection differences, stronger compensation may be needed to balance the imaging differences between different regions; while in regions with smaller reflection differences, the intensity of compensation can be reduced to avoid unnecessary enhancement or distortion of the image.

[0067] Understandably, by adaptively determining image compensation parameters based on the degree of reflection difference in reflective regions, flexible compensation adjustments can be made according to the actual situation of the image, ensuring that the compensation effect for each region accurately matches its reflective characteristics. This approach avoids the uneven compensation phenomenon caused by using fixed compensation parameters, thereby improving the overall quality of the wafer image.

[0068] In one embodiment, the imaging consistency constraint is determined based on at least one of the differences in grayscale distribution, edge response, or contrast among regions with different reflective properties after compensation.

[0069] Specifically, grayscale distribution differences: After image compensation, the grayscale distribution differences in regions with different reflective properties can be calculated. If the grayscale distribution in some regions differs too much from other regions, it may cause visual inconsistency. Therefore, by constraining the compensated grayscale distribution, obvious brightness imbalances can be avoided. Edge response differences: Edge response differences reflect the sharpness of the image's edges. If the edge response in some regions is too strong or too weak, it may cause the edges in the image to be unclear or overly prominent, affecting image quality. Therefore, imaging consistency constraints can optimize the edge details of the image by controlling the differences in edge response after compensation, keeping them consistent. Contrast differences: Contrast differences in different regions may lead to inconsistent visual effects in the image. By constraining contrast differences, excessively high or low contrast in certain areas of the image can be avoided, thus maintaining the visual balance of the entire image.

[0070] In practice, a wafer image contains two regions, one with significantly different reflectance characteristics and the other with relatively similar reflectance characteristics. First, compensation processing is performed on the images of both regions. During the compensation process, the differences in grayscale distribution, edge response, and contrast between the two regions are calculated. If a significant difference in grayscale distribution exists between the two regions after compensation, the imaging consistency constraint automatically adjusts the compensation intensity to ensure that the grayscale distributions of the two regions are closer. Similarly, if the edge response of a certain region is too strong, the compensation intensity is appropriately reduced through the constraint of edge response difference to ensure a smooth transition of the edges. The constraint of contrast difference prevents any region from having excessively high or low contrast, thus making the entire image more visually consistent and harmonious.

[0071] This application also provides a wafer imaging image processing compensation system for implementing the wafer imaging image processing compensation method described above, comprising: an image acquisition module for acquiring a wafer image acquired by a wafer imaging system; a difference analysis module for performing image feature analysis on the wafer image to determine the differences in reflectance characteristics of different regions; a region division module for dividing the wafer image into at least two reflectance characteristic regions based on the differences in reflectance characteristics; a parameter determination module for determining corresponding image compensation parameters for different reflectance characteristic regions, wherein the image compensation parameters are subject to preset imaging consistency constraints during the determination process; and an image compensation module for performing adaptive compensation processing on the wafer image based on the image compensation parameters.

[0072] The above description is merely a specific implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in the embodiments of this application should be covered within the protection scope of the embodiments of this application. Therefore, the protection scope of the embodiments of this application should be determined by the protection scope of the claims.

Claims

1. A wafer imaging image processing compensation method, characterized by, include: Acquire wafer images from a wafer imaging system; Image feature analysis is performed on the wafer image to determine the differences in reflectance characteristics in different regions of the wafer image; Based on the differences in reflectance characteristics, the wafer image is divided into at least two reflectance characteristic regions, specifically including: extracting the normalized grayscale mean of local regions of the image. As a measure of material reflectance difference ,in This represents the average gray level of a local area. The maximum grayscale mean of the entire image; extracting the edge density of local regions of the image. As a measure of the difference in reflection of graphic structure ,in This represents the number of edge pixels in a local region extracted using the Canny edge detection operator. The total number of pixels in the local region; the material weighting coefficient is dynamically determined based on the known material type information and pattern spatial frequency of the local region. With graphic weight coefficient ,in According to the weighted formula Calculate the comprehensive score of reflection characteristics for each region. A comprehensive score based on the aforementioned reflection characteristics. The distribution of the wafer image is divided into at least two reflective regions using a clustering algorithm. For regions with different reflectivity, determine the corresponding image compensation parameters; Based on the image compensation parameters, adaptive compensation processing is performed on the wafer image to obtain the compensated wafer image; The image compensation parameters are subject to preset imaging consistency constraints during the determination process. These constraints are determined based on at least one of the differences in grayscale distribution, edge response, or contrast among regions with different reflective characteristics after compensation.

2. The wafer imaging image processing compensation method according to claim 1, characterized in that, The image features include at least one of grayscale distribution features, edge gradient features, and frequency domain features.

3. The wafer imaging image processing compensation method according to claim 1, characterized in that, The differences in reflectivity include differences in material reflectivity and differences in reflectivity of graphic structures.

4. The wafer imaging image processing compensation method according to claim 1, characterized in that, The dynamic determination of material weight coefficients With graphic weight coefficient ,include: If the number of known material types corresponding to the local region is greater than 1, then the value is increased relative to the default value. and reduce ; If the pattern spatial frequency of the local region is higher than a preset threshold, then the value is reduced relative to the default value. and increase .

5. The wafer imaging image processing compensation method according to claim 1, characterized in that, The dynamic determination of material weight coefficients With graphic weight coefficient It also includes: The texture complexity T is obtained by calculating the contrast of the gray-level co-occurrence matrix of the local region; The weighted formula, incorporating the texture complexity T, yields: ; in, For texture weight coefficients, and .

6. The wafer imaging image processing compensation method according to claim 5, characterized in that, The texture weight coefficient The pattern regularity of the local region is adaptively determined; the lower the pattern regularity, the better. The larger the value, the better.

7. The wafer imaging image processing compensation method according to claim 1, characterized in that, The image compensation parameters include at least one of brightness compensation parameters, contrast compensation parameters, and edge enhancement parameters.

8. The wafer imaging image processing compensation method according to claim 1, characterized in that, The image compensation parameters are adaptively determined based on the degree of reflection difference in the corresponding reflective characteristic regions.

9. A wafer imaging image processing compensation system, characterized in that, The wafer imaging image processing compensation method for implementing any one of claims 1 to 8 includes: The image acquisition module is used to acquire wafer images captured by the wafer imaging system; The difference analysis module is used to perform image feature analysis on the wafer image to determine the differences in reflectance characteristics in different regions; The region division module is used to divide the wafer image into at least two reflective characteristic regions based on the differences in reflective characteristics. The parameter determination module is used to determine the corresponding image compensation parameters for different reflective characteristic regions, and the image compensation parameters are subject to preset imaging consistency constraints during the determination process. An image compensation module is used to perform adaptive compensation processing on the wafer image based on the image compensation parameters.

Images