A phase morphology statistical method and system

Abstract

This invention relates to the field of phase morphology statistical technology, specifically to a phase morphology statistical method and system. The method includes: acquiring a first grayscale image of the alloy; performing a reconstruction operation on the first grayscale image to identify a first S-phase; performing an enhancement operation on the first grayscale image to identify a second S-phase; and outputting statistical results of the S-phase based on the first and second S-phases. This invention provides an automated analysis method for differentiated segmentation of different S-phase morphologies in STEM images of 2xxx series aluminum alloys, exhibiting good interactivity and parameter adjustability, and outputting complete size-area fraction distribution statistical results, thereby improving the efficiency and accuracy of S-phase characterization.

Description

Technical Field

[0001] This invention relates to the field of phase morphology statistical technology, specifically to a phase morphology statistical method and system. Background Technology

[0002] In existing technologies, the identification and analysis of S-phase precipitates in STEM images mainly relies on manual interpretation. This involves operators visually observing the image, identifying S-phase regions based on experience, and manually selecting and measuring each precipitate to obtain its morphology and size information. While this traditional method is intuitive, it suffers from significant efficiency and consistency issues. When processing large numbers of samples or performing statistical analysis on multiple fields of view of the same sample, manual methods are not only time-consuming and labor-intensive, but also prone to significant discrepancies between different operators, and even between the same operator at different times, making it difficult to guarantee data comparability and repeatability.

[0003] Alternatively, some existing methods primarily employ simple segmentation strategies based on grayscale thresholds. This method binarizes the image by setting one or more grayscale thresholds, thereby separating the foreground (precipitated phase) from the background (matrix). However, the grayscale distribution of the S-phase in STEM images often overlaps with the background, and the image brightness distribution is uneven, making it difficult for fixed thresholds to adapt to the segmentation needs of different regions. More importantly, simple threshold segmentation lacks the ability to differentiate between different morphologies of the S-phase: slender, needle-like S-phases are easily broken into multiple fragments during binarization due to local grayscale fluctuations; while ring-shaped S-phase structures formed by multiple small particles may break or adhere to adjacent structures due to improper threshold settings, losing their original morphological features.

[0004] In recent years, deep learning methods have made significant progress in image segmentation, and some researchers have attempted to apply convolutional neural networks to the analysis of material microstructures. However, deep learning methods typically require a large amount of labeled data for model training, while the annotation of STEM images is inherently dependent on specialized knowledge and time-consuming, making it difficult to quickly obtain sufficient training samples. Furthermore, models trained for specific alloy systems often have limited generalization ability, making it difficult to directly transfer to images with different imaging conditions. The black-box nature of these models also makes the segmentation results lack interpretability, hindering researchers from adjusting and optimizing them according to specific needs.

[0005] Therefore, there is an urgent need for a more accurate S-phase identification method. Summary of the Invention

[0006] The purpose of this invention is to provide a phase morphology statistical method and system that partially solves or alleviates the above-mentioned deficiencies in the prior art and can improve the accuracy of alloy S phase identification.

[0007] To solve the aforementioned technical problems, the present invention specifically adopts the following technical solution: A first aspect of the present invention is to provide a phase morphology statistical method, comprising the steps of: S101, Obtain the first grayscale image of the alloy; S102, performing a reconstruction operation on the first grayscale image to identify the first S-phase, wherein S102 includes: S1021, perform morphological reconstruction on the first grayscale image to obtain a second grayscale image; S1022, in the second grayscale image, geometric shapes that meet the filtering rules are identified as the first S-phase, wherein the filtering rules include: (1) Morphological filtering rules, wherein the morphological filtering rules include: , ; (2) Contrast screening rules: ; in, Minimum area threshold For the maximum area threshold, For the lower limit threshold of brightness, For the upper limit threshold of solidity, , , The area, average brightness, and solidity of the geometric figure; S103, perform an enhancement operation on the first grayscale image to identify the second S-phase, wherein S103 includes: S1031, the first S phase is deleted from the first grayscale image, forming a corresponding empty region, and the empty region is filled with reconstructed grayscale to obtain a third grayscale image; the reconstructed grayscale is determined based on the global grayscale value of the first grayscale image. S1032, Perform a morphological top-hat transformation operation on the third grayscale image to obtain a fourth grayscale image; S1033, perform threshold segmentation on the fourth grayscale image to obtain a binary image; S1034, A skeletonization algorithm is used to identify candidate S-phases from the binary image; S1035, identify the candidate S-phase whose length is greater than a preset length threshold as the second S-phase; S104, output the statistical results of the S phase based on the first S phase and the second S phase.

[0008] In some embodiments, S1021 includes: An edge detection algorithm is used to extract edge patterns from the first grayscale image, and at least one of the edge patterns has a broken edge; A morphological closing operation is performed on the edge graphics to connect at least some of the broken edges, thereby reconstructing a corresponding geometric shape; wherein the model of the morphological closing operation includes: ; in, The geometric figure, For the edge shape, B is a disk-shaped structural element, and ⊕ represents the dilation operation. This represents the erosion operation; The geometric shape is filled with holes to form the second grayscale image.

[0009] In some embodiments, the radius of B is 1-4 pixels.

[0010] In some embodiments, the operational model of the morphological cap transformation operation in S1032 includes: ; ; in, Represents the morphological opening operation. It is a disk-shaped structural element. The third grayscale image, This refers to the fourth grayscale image.

[0011] In some embodiments, The radius of is greater than the radius of B.

[0012] In some embodiments, The radius is 5-8 pixels.

[0013] In some embodiments, the reconstructed grayscale is the median of the global grayscale values ​​of the first grayscale image.

[0014] In some embodiments, prior to S101, the following step is also included: At least two photographs of the alloy were acquired at at least two imaging angles; The photo is divided into multiple image blocks; Calculate multiple standard deviations for multiple of the aforementioned patches; Calculate the variance of the multiple standard deviations; The photo with the smaller variance is selected as the recommended photo; The first grayscale image is generated based on the recommended photo.

[0015] A second aspect of the present invention is to provide a phase morphology statistical system, comprising: The image acquisition module is used to acquire the first grayscale image of the alloy; A reconstruction module is used to perform a reconstruction operation on the first grayscale image to identify the first S-phase, wherein the reconstruction module includes: A reconstruction unit is used to perform morphological reconstruction on the first grayscale image to obtain a second grayscale image; A filtering unit is configured to identify geometric shapes that meet filtering rules as a first S-phase in the second grayscale image, wherein the filtering rules include: (1) Morphological filtering rules, wherein the morphological filtering rules include: , ; (2) Contrast screening rules: ; in, Minimum area threshold For the maximum area threshold, For the lower limit threshold of brightness, For the upper limit threshold of solidity, , , The area, average brightness, and solidity of the geometric figure; An enhancement module is used to perform enhancement operations on the first grayscale image to identify the second S-phase, wherein the enhancement module includes: A mask filling unit is used to delete the first S-phase in the first grayscale image, thereby forming a corresponding empty region, and to fill the empty region with reconstructed grayscale to obtain a third grayscale image; the reconstructed grayscale is determined based on the global grayscale value of the first grayscale image. A transformation unit is used to perform a morphological top-hat transformation operation on the third grayscale image to obtain a fourth grayscale image; A threshold segmentation unit is used to perform threshold segmentation on the fourth grayscale image to obtain a binary image; A skeleton recognition unit is used to identify candidate S-phases from the binary image using a skeletonization algorithm; A phase recognition unit is used to identify the candidate S-phase with a length greater than a preset length threshold as the second S-phase; The statistics module is used to output the statistical results of the S phase based on the first S phase and the second S phase.

[0016] In some embodiments, the reconfiguration unit includes: An edge pattern extraction subunit is used to extract edge patterns from the first grayscale image using an edge detection algorithm, wherein at least one of the edge patterns has a broken edge; A morphological closing operation subunit is used to perform a morphological closing operation on the edge graphics to at least connect a portion of the broken edges, thereby reconstructing a corresponding geometric shape; wherein, the model of the morphological closing operation includes: ; in, The geometric figure, For the edge shape, B is a disk-shaped structural element, and ⊕ represents the dilation operation. This represents the erosion operation; The hole-filling subunit is used to fill holes in the geometric shape to form the second grayscale image.

[0017] Beneficial technical effects: For S-phases that are interleaved and have significant morphological differences, this application provides a step-by-step, hierarchical image processing method.

[0018] Specifically, this application first focuses on combining morphological reconstruction and contrast difference to identify the ring-shaped S phase relatively independently; then, it returns to the original image to remove the ring-shaped S phase, and then performs an enhanced operation on the remaining image (i.e., introducing multiple methods such as gap reconstruction, top-hat transformation, and threshold segmentation).

[0019] In other words, this application sets up two major image processing stages for S-phase recognition, and in each stage, it focuses on strengthening processing for a limited region and a limited number of objects, thereby reducing the difficulty of strengthening. Furthermore, through step-by-step and hierarchical collaborative strengthening processing, this application can reconstruct the image to obtain clearer phase information, while reducing the introduction of excessive noise into the original image during the strengthening process, which could introduce immeasurable errors.

[0020] This application introduces a multi-image angle photo selection mechanism in the image preprocessing stage. It evaluates image quality and imaging uniformity by using the standard deviation and variance of the image patches, and selects recommended photos with smaller variance and more stable imaging to generate grayscale images.

[0021] This proactive screening mechanism can reduce image distortion and grayscale deviation caused by imaging angle and uneven lighting as early as possible. It works in synergy with the subsequent step-by-step hierarchical recognition strategy to achieve a better balance between the difficulty and accuracy of S-phase recognition. It not only fully enhances the weak contrast of needle-like S-phase, but also effectively controls the introduction of noise and artifacts, ultimately making the overall statistical results more stable, reliable and less errory. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale. Obviously, the drawings described below are some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0023] Figure 1 A flowchart illustrating a method for differentially segmenting S-phase precipitates in STEM images based on morphological features, provided by this invention. Figure 2 A schematic diagram of the S-phase morphology distribution in a STEM image of a 2xxx series aluminum alloy provided by this invention; Figure 3 This is a schematic diagram of the spiral S-phase projection morphology in a STEM image of a 2xxx series aluminum alloy provided by the present invention. Figure 4 This is a schematic diagram of the needle-like S-phase projection morphology in a STEM image of a 2xxx series aluminum alloy provided by the present invention. Figure 5 This is a schematic diagram of the annular S-phase interactive interface provided by the present invention; Figure 6 This is a schematic diagram of the spiral S-phase detection and mask filling provided by the present invention; Figure 7 This is a schematic diagram of the interactive interface for needle-bar-shaped S-phase detection provided by the present invention; Figure 8 This is a schematic diagram of needle-bar S-phase detection and mask filling provided by the present invention; Figure 9 Statistical diagrams of the length and area distribution of spiral and needle-like S-phases provided by the present invention; Figure 10 A flowchart illustrating a phase morphology statistical method provided by the present invention; Figure 11 This is a schematic diagram of the structure of a phase morphology statistical system provided by the present invention; Figure 12 This is a schematic block diagram of the structure of a computer device provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0025] In this document, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0026] In this document, the terms "upper," "lower," "inner," "outer," "front," "rear," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the present invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0027] In this document, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0028] In this document, "and / or" includes any and all combinations of one or more of the listed related items.

[0029] In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0030] As used in this specification, the term "about" typically means + / -5% of the value, more typically + / -4% of the value, more typically + / -3% of the value, more typically + / -2% of the value, even more typically + / -1% of the value, and even more typically + / -0.5% of the value.

[0031] In this specification, certain embodiments may be disclosed in a range-bound format. It should be understood that this "range-bound" description is merely for convenience and brevity and should not be construed as a rigid limitation on the disclosed range. Therefore, the description of a range should be considered as having specifically disclosed all possible subranges and the individual numerical values ​​within those ranges. For example, a description of the range 1-6 should be considered as having specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., and the individual numbers within those ranges, such as 1, 2, 3, 4, 5, and 6. This rule applies regardless of the breadth of the range.

[0032] Definition of noun: Phase morphology refers to the crystal structure and organizational state of a material under different conditions.

[0033] S phase is a strengthening precipitate or supersaturated solid solution metastable phase in metallic alloys. Its morphology, distribution and stability have an important influence on the alloy properties.

[0034] Example 1: This invention designs differentiated segmentation strategies for two typical morphologies, annular and acicular S-phases, instead of using a uniform "one-size-fits-all" segmentation method. Secondly, it adopts a multi-step process combining edge detection, morphological processing, and feature selection to overcome the limitations of simple threshold segmentation. Thirdly, it introduces a skeletonization method to process acicular S-phases, fundamentally solving the problem of easy breakage of slender structures. Fourthly, it provides an interactive parameter adjustment interface, enabling the method to flexibly adapt to STEM images under different imaging conditions.

[0035] Please see Figure 10 This invention provides a phase morphology statistical method, comprising the following steps: S101, Obtain the first grayscale image of the alloy; S102, performing a reconstruction operation on the first grayscale image to identify the first S-phase, wherein S102 includes: S1021, perform morphological reconstruction on the first grayscale image to obtain a second grayscale image; S1022, in the second grayscale image, geometric shapes that meet the filtering rules are identified as the first S-phase, wherein the filtering rules include: (1) Morphological filtering rules, wherein the morphological filtering rules include: , ; (2) Contrast screening rules: ; in, Minimum area threshold For the maximum area threshold, For the lower limit threshold of brightness, For the upper limit threshold of solidity, , , The area, average brightness, and solidity of the geometric figure; S103, perform an enhancement operation on the first grayscale image to identify the second S-phase, wherein S103 includes: S1031, the first S phase is deleted from the first grayscale image, forming a corresponding empty region, and the empty region is filled with reconstructed grayscale to obtain a third grayscale image; the reconstructed grayscale is determined based on the global grayscale value of the first grayscale image. S1032, Perform a morphological top-hat transformation operation on the third grayscale image to obtain a fourth grayscale image; S1033, perform threshold segmentation on the fourth grayscale image to obtain a binary image; S1034, A skeletonization algorithm is used to identify candidate S-phases from the binary image; S1035, identify the candidate S-phase whose length is greater than a preset length threshold as the second S-phase; S104, output the statistical results of the S phase based on the first S phase and the second S phase.

[0036] The sulfur phase is an important strengthening phase in aluminum alloys. However, due to its diverse morphologies and complex, interwoven distribution, it is difficult to accurately segment and quantitatively analyze it.

[0037] For S-phases that are interleaved and have significant morphological differences, this application provides a step-by-step, hierarchical image processing method.

[0038] Specifically, this application first focuses on combining morphological reconstruction and contrast difference to identify the ring-shaped S phase relatively independently; then, it returns to the original image to remove the ring-shaped S phase, and then performs an enhanced operation on the remaining image (i.e., introducing multiple methods such as gap reconstruction, top-hat transformation, and threshold segmentation).

[0039] In other words, this application sets up two major image processing stages for S-phase recognition, and in each stage, it focuses on strengthening processing for a limited region and a limited number of objects, thereby reducing the difficulty of strengthening. Furthermore, through step-by-step and hierarchical collaborative strengthening processing, this application can reconstruct the image to obtain clearer phase information, while reducing the introduction of excessive noise into the original image during the strengthening process, which could introduce immeasurable errors.

[0040] Specifically, this application is designed with the following two stages: (1) First stage: The original image is initially processed by a low-destructive morphological reconstruction method, such as using a small-radius disk structure for image dilation and erosion, in order to carefully reconstruct the ring-shaped S phase that may be broken, and avoid introducing too much error; (2) Second stage: On the one hand, the enhancement area is restricted, that is, the first S-phase (i.e., the identified ring-shaped S-phase) is deleted from the original image. On the other hand, the enhancement object is guided, that is, a morphological top-hat transformation operation is used to focus on enhancing the longitudinal structure in order to achieve effective enhancement of the needle-shaped S-phase. Thus, by restricting the area and object, this embodiment is conducive to reducing the introduction of errors while achieving sufficient enhancement.

[0041] From another perspective, in this embodiment, a small disk is used in the first stage to bridge small gaps, while a large disk is used in the second stage to remove the background from the incomplete image (i.e., the image after removing the ring-shaped S-phase). This differentiated reconstruction method works synergistically to effectively improve the image reconstruction intensity while reducing the difficulty of reconstruction (such as reducing the introduction of excessive errors during the reconstruction process).

[0042] Specifically, the technical solution of the present invention brings the following beneficial effects.

[0043] 1) This invention employs a strategy combining Canny edge detection with morphological closure and hole filling for annular S-phases, effectively extracting the complete outline of the annular structure and avoiding structural fragmentation caused by simple threshold segmentation. Through solidity feature filtering, it can accurately distinguish between annular S-phases with hollow features and solid block regions. For needle-like S-phases, this invention innovatively introduces a skeletonization method. First, the candidate region is refined into skeleton lines of single-pixel width. The true needle-like structure is then selected based on the skeleton length, and its actual width is restored through morphological dilation. This "refine-then-restore" strategy fundamentally solves the problem of easy breakage of needle-like phases in traditional binarization, significantly improving the segmentation integrity of needle-like S-phases.

[0044] 2) This invention automates S-phase segmentation, significantly reducing analysis time. Traditional manual methods typically require tens of minutes or even longer to process a single STEM image, while the automated process of this invention can complete segmentation and statistics within seconds after parameter adjustment, improving processing efficiency by one to two orders of magnitude. This has significant practical value for research work that requires processing large numbers of samples or performing multi-field statistical analysis.

[0045] 3) This invention provides a graphical, interactive parameter adjustment interface, including edge detection thresholds, morphological processing parameters, and feature selection conditions, all of which can be adjusted in real time via sliders, and the segmentation effect can be previewed instantly. This design enables the method to flexibly adapt to STEM images with different imaging conditions, magnifications, and contrasts without retraining the model or modifying the program code, thus lowering the barrier to entry and improving the versatility of the method.

[0046] 4) The segmentation process of this invention is entirely based on deterministic image processing algorithms. The same parameter settings will inevitably produce the same segmentation results for the same image, completely eliminating the influence of subjective human factors. When it is necessary to compare different samples or different process conditions, using uniform parameter settings can ensure the consistency and comparability of the analysis results, providing a reliable data foundation for quantitative research.

[0047] 5) This invention not only outputs the segmented binary mask, but also automatically calculates the morphological parameters of various S-phases, including area and major axis length, and presents the statistical results intuitively in the form of a size-area fraction distribution histogram. This output format facilitates researchers' rapid understanding of the overall distribution characteristics of the S-phase, providing directly usable quantitative data for subsequent material property analysis and process optimization.

[0048] In some embodiments, the corresponding image can be output based on the statistical results of the first S-phase and the second S-phase, such as... Figure 8 As shown.

[0049] In some embodiments, the first grayscale image may refer to a preprocessed TEM / STEM image. Preprocessing may include converting a color image to a grayscale image, normalizing the grayscale image, and calculating the actual physical size (in nm / pixel) for each pixel.

[0050] In some embodiments, the second grayscale image is a binary mask image obtained by performing edge detection, morphological closing operation and hole filling on the first grayscale image.

[0051] In some embodiments, the reconstruction operation may refer to morphological reconstruction, including edge detection, closing operation, hole filling, and other operations.

[0052] In some embodiments, the first S-phase refers to the annular S-phase. The second S-phase is the pointer-shaped S-phase.

[0053] In some embodiments, the reconstructed grayscale can be the median grayscale of the image.

[0054] In some embodiments, the vacant area may also be referred to as the deduction area.

[0055] In some embodiments, the third grayscale image may be referred to as the residual image. The fourth grayscale image may be referred to as the enhanced image.

[0056] In some embodiments, the enhancement operation may refer to removing the annular S-phase region and filling it with median grayscale, performing a top-hat transformation to enhance the needle-like structure, performing threshold segmentation, skeletonization, length filtering, and dilation recovery, and finally identifying the needle-like S-phase.

[0057] In some embodiments, S1021 includes: An edge detection algorithm is used to extract edge patterns from the first grayscale image, and at least one of the edge patterns has a broken edge; A morphological closing operation is performed on the edge graphics to connect at least some of the broken edges, thereby reconstructing a corresponding geometric shape; wherein the model of the morphological closing operation includes: ; in, The geometric figure, For the edge shape, B is a disk-shaped structural element, and ⊕ represents the dilation operation. This represents the erosion operation; The geometric shape is filled with holes to form the second grayscale image.

[0058] Preferably, the edge detection algorithm can be the Canny edge detection algorithm, which can initially extract the edge contour of the annular S-phase to connect edge discontinuities (i.e. broken edges) that may occur in the STEM image due to contrast or noise.

[0059] In some embodiments, the radius of B is 1-4 pixels.

[0060] In some embodiments, the operational model of the morphological cap transformation operation in S1032 includes: ; ; in, Represents the morphological opening operation. It is a disk-shaped structural element. The third grayscale image, This refers to the fourth grayscale image.

[0061] In some embodiments, The radius of is greater than the radius of B.

[0062] In some embodiments, The radius is 5-8 pixels.

[0063] In some embodiments, the reconstructed grayscale is the median of the global grayscale values ​​of the first grayscale image.

[0064] In some embodiments, the method further includes the step of: before S101. At least two photographs of the alloy were acquired at at least two imaging angles; The photo is divided into multiple image blocks; Calculate multiple standard deviations for multiple of the aforementioned patches; Calculate the variance of the multiple standard deviations; The photo with the smaller variance is selected as the recommended photo; The first grayscale image is generated based on the recommended photo.

[0065] The applicant noted that when needle-like and ring-like structures are arranged in an interlaced pattern, they interfere with each other severely, significantly increasing the difficulty of recognition. The step-by-step, hierarchical processing method described in this application, however, can control the degree of error introduction while sufficiently enhancing the image. Specifically, in this embodiment, the invention can filter images from multiple scanning angles to select the most suitable images for multiple image processing. Furthermore, by filtering images with relatively uniform texture variations and brightness, the difficulty of image enhancement can be further reduced.

[0066] In other words, this application introduces a multi-image angle photo selection mechanism in the image preprocessing stage. It evaluates image quality and imaging uniformity by using the standard deviation and variance of the image patches, and selects recommended photos with smaller variance and more stable imaging to generate grayscale images.

[0067] This proactive screening mechanism can reduce image distortion and grayscale deviation caused by imaging angle and uneven lighting as early as possible. It works in synergy with the subsequent step-by-step hierarchical recognition strategy to achieve a better balance between the difficulty and accuracy of S-phase recognition. It not only fully enhances the weak contrast of needle-like S-phase, but also effectively controls the introduction of noise and artifacts, ultimately making the overall statistical results more stable, reliable and less errory.

[0068] Please see Figure 11 This invention provides a phase morphology statistical system, comprising: The image acquisition module is used to acquire the first grayscale image of the alloy; A reconstruction module is used to perform a reconstruction operation on the first grayscale image to identify the first S-phase, wherein the reconstruction module includes: A reconstruction unit is used to perform morphological reconstruction on the first grayscale image to obtain a second grayscale image; In some embodiments, candidate connected regions can be determined based on the second grayscale image, and the average brightness of each candidate connected region can be calculated based on the first grayscale image. A filtering unit is configured to identify geometric shapes that meet filtering rules as a first S-phase in the second grayscale image, wherein the filtering rules include: (1) Morphological filtering rules, wherein the morphological filtering rules include: , ; (2) Contrast screening rules: ; in, Minimum area threshold For the maximum area threshold, For the lower limit threshold of brightness, For the upper limit threshold of solidity, , , The area, average brightness, and solidity of the geometric figure; The average brightness is calculated from the pixel grayscale of the candidate connected component at the corresponding position in the first grayscale image.

[0069] An enhancement module is used to perform enhancement operations on the first grayscale image to identify the second S-phase, wherein the enhancement module includes: A mask filling unit is used to delete the first S-phase in the first grayscale image, thereby forming a corresponding void region, and to fill the void region with reconstructed grayscale to obtain a third grayscale image; the reconstructed grayscale is the median of the global grayscale values ​​of the first grayscale image. A transformation unit is used to perform a morphological top-hat transformation operation on the third grayscale image to obtain a fourth grayscale image; A threshold segmentation unit is used to perform threshold segmentation on the fourth grayscale image to obtain a binary image; A skeleton recognition unit is used to identify candidate S-phases from the binary image using a skeletonization algorithm; A phase recognition unit is used to identify the candidate S-phase with a length greater than a preset length threshold as the second S-phase; The statistics module is used to output the statistical results of the S phase based on the first S phase and the second S phase.

[0070] It should be understood that the phase morphology statistical system proposed in this invention can be used to implement the method steps described in any embodiment of this invention.

[0071] Aluminum alloys (such as the 2xxx series aluminum alloys) are Al-Cu-Mg series high-strength aluminum alloys with Cu and Mg as the main alloying elements. Due to their excellent specific strength, good damage tolerance, and fatigue resistance, they are widely used in the manufacture of load-bearing components such as aerospace structural parts, aircraft skins, and wings. The high strength of these alloys mainly comes from the S phase (Al2CuMg) precipitated during the aging process. As the most important strengthening phase in 2xxx series aluminum alloys, the precipitation behavior, morphological evolution, and spatial distribution characteristics of the S phase directly determine the final mechanical properties of the alloy. A deep understanding and accurate characterization of the microstructure of the S phase is of great value for optimizing alloy composition design, formulating reasonable heat treatment processes, and predicting the service performance of materials.

[0072] The precipitation morphology of the S phase in 2xxx series aluminum alloys is diverse and complex. Depending on the precipitation conditions and observation direction, the S phase can exhibit different morphological characteristics such as needle-like, plate-like, or ring-like forms. Among them, needle-like S phases grow along specific crystallographic directions and are distributed in slender rod-like shapes (e.g., ...). Figure 4 The image shows the different projected morphologies of the needle-like S phase. When the S phase precipitates along a specific crystal plane and projects from different directions, it appears as a ring or spiral structure in STEM (Scanning Transmission Electron Microscopy) images (e.g., ...). Figure 3 The spiral S-phase shown in the figure has different projected morphologies. This ring-shaped structure is actually formed by multiple fine S-phase particles arranged continuously along a specific orientation. Different morphologies of S-phase contribute differently to the strengthening effect of the alloy. Therefore, accurate segmentation and quantitative statistics of the two typical morphologies of S-phase, namely needle-like and ring-like, are of great significance for comprehensively evaluating the precipitation strengthening effect of the alloy.

[0073] Transmission electron microscopy (STEM) is the most direct and effective means of characterizing nanoscale S-phase precipitates, providing high-resolution images of the S-phase morphology, size, and distribution. However, traditional STEM image analysis mainly relies on manual identification and measurement, which is not only time-consuming and labor-intensive but also highly susceptible to the subjective factors of the operator, making it difficult to guarantee the consistency and repeatability of measurement results. The limitations of manual analysis become even more pronounced when statistical analysis of a large number of samples or comprehensive evaluation of multiple fields of view of the same sample is required.

[0074] In addition, please see Figure 2 The morphological features of the S phase in STEM images are quite complex. The needle-shaped S phase is slender and has weak contrast, making it easy to break during segmentation. The annular S phase is a spiral structure formed by multiple connected particles, which is easy to break or adhere to adjacent structures during morphological processing. These characteristics further increase the difficulty of accurate segmentation and identification.

[0075] Existing image analysis methods have several shortcomings when processing the S phase in STEM images of 2xxx series aluminum alloys. 1. Segmentation methods based on simple grayscale thresholds are sensitive to image brightness inhomogeneities, easily leading to S phase particle breakage or adhesion, especially for weakly contrasted needle-like S phases, which are difficult to extract completely. 2. While edge detection methods can capture S phase contours, they are prone to fracture in slender needle-like S phases and struggle to maintain the morphological integrity of ring structures formed by multiple connected particles. 3. Traditional morphological operations such as dilation and corrosion, while useful for connecting fractured areas or separating adhered particles, are global and cannot distinguish between fractures within the same S phase structure and the proximity of adjacent independent S phases, often introducing new errors. 4. Although deep learning methods have made significant progress in image segmentation, they typically require large amounts of labeled data for training, have limited adaptability to STEM images of specific alloy systems, and suffer from poor model interpretability.

[0076] In terms of quantitative statistics, traditional counting methods struggle to provide physically meaningful statistical results because the annular S-phase is a complex structure formed by the connection of multiple fine particles. Existing studies typically only report the average size or simple number density of the S-phase, lacking a systematic description of the area fraction distribution characteristics of S-phase clusters of different sizes.

[0077] However, according to precipitation strengthening theory, the strengthening effect of a material depends not only on the average size and total area fraction of the precipitates, but also on the size distribution morphology of the precipitates. Therefore, establishing a quantitative analysis method that can output the S-phase size-area fraction distribution curve is of great significance for accurately assessing the precipitation strengthening contribution of alloys and for a deeper understanding of the age-hardening mechanism.

[0078] Based on this, the present invention proposes an automated analysis method that can perform differentiated segmentation processing on S phases of different morphologies in STEM images of aluminum alloys (especially 2xxx series), has good interactivity and parameter adjustability, and can output complete size-area fraction distribution statistical results, so as to overcome the shortcomings of the prior art to a certain extent and improve the efficiency and accuracy of S phase characterization.

[0079] For example, this invention employs different image processing procedures for annular S-phases and acicular S-phases, rather than using a uniform method for all morphologies. For annular S-phases, the procedure is "edge detection → morphological closure → hole filling → solidity screening"; for acicular S-phases, the procedure is "cap enhancement → binarization → skeletonization → length screening → dilation recovery".

[0080] For example, this invention applies the skeletonization algorithm to the segmentation of needle-like precipitates. Through a three-step process of "extracting the skeleton first → filtering by length → then expanding and restoring", it solves the problem of needle-like phases being prone to breakage in traditional binarization.

[0081] For example, the present invention uses the solidity of the connected domain (region area / convex hull area) as a screening index to distinguish between annular S-phase with hollow features and solid block regions.

[0082] The following will illustrate the phase morphology statistical method proposed in this invention with specific examples: like Figure 1 As shown, this invention provides a method for differentially segmenting the S-phase precipitate in a STEM image based on morphological features, comprising the following steps: S201 Image Acquisition and Preprocessing STEM images of the 2xxx series aluminum alloys to be analyzed were acquired and preprocessed. Specifically, color images were converted to grayscale and normalized to the [0,1] range, and the image dimensions were obtained. Simultaneously input the image scale information to calculate the actual size corresponding to each pixel. .

[0083] First, read STEM image I. Determine if the image is a color image; if so, convert it to a grayscale image using a weighted average method. ;in, These are the pixel values ​​for the red, green, and blue channels, respectively. These are pixel coordinates. Get the image size. Where H is the image height (pixels), W is the image width (pixels), and I... gray (x,y) is the pixel value of the converted grayscale image at coordinates (x,y).

[0084] Simultaneously input the image scale information to calculate the actual size corresponding to each pixel: ; Where k is the scale factor, with units of nm / pixel; The actual length indicated on the scale, in nm; The scale bar is represented by pixels in the image. The grayscale image is normalized, mapping pixel values ​​to the [0,1] interval: ; in, and These are the minimum and maximum gray values ​​in the image, I. norm (x,y) represents the normalized image pixel values. The purpose of normalization is to eliminate the inconsistency in grayscale range caused by differences in imaging conditions between different STEM images, making it easier to use a unified threshold parameter for subsequent processing.

[0085] S202 Annular S-phase Edge Detection and Candidate Region Extraction The preprocessed image is segmented into annular S-phase regions using the first image segmentation module. The Canny edge detection algorithm is then used to extract image edges, with the high and low thresholds adjustable via an interactive interface. Morphological closure operations are performed on the edge images to connect broken edges, and hole filling is then applied to obtain candidate annular S-phase regions. Figure 5 As shown, users can adjust the high and low thresholds, solidity, eccentricity, etc. of the image through the interactive interface, thereby obtaining the image after edge processing and closure processing, as well as the calculated number and size of the annular S-phase.

[0086] The annular S-phase originates from the projection of the helical S-phase along the electron beam direction, and its morphological characteristic is a closed or nearly closed annular structure (see [link to relevant documentation]). Figure 6 This invention employs an edge detection-based segmentation strategy to identify ring-shaped S-phases. First, the normalized image is smoothed using Gaussian filtering to suppress noise interference. ; Among them, I smooth (x,y) represents the smoothed (filtered) output image, and * indicates a convolution operation. The standard deviation is Two-dimensional Gaussian kernel function: ; Where exp is the natural exponential function. In this embodiment, σ is set to a default value of 1. Then, the Canny edge detection algorithm is used to extract the image edges, resulting in a binary edge image E. ; in, These are the low and high thresholds for the Canny algorithm, with default values ​​of 0.1 and 0.3 respectively. Users can adjust these in real time via the interactive interface. Morphological closing operations are performed on the edge image E to connect broken edges caused by noise or insufficient contrast. ; in, The output image is the result of the closing operation; ⊕ represents the dilation operation. This represents the erosion operation, where B is a disk-shaped structuring element with a radius of . The default value is 3 pixels. The closing operation first dilates and then erodes, which can connect adjacent edge fragments while maintaining the overall shape of the edge.

[0087] The edge image after the closing operation is processed to fill holes, thereby obtaining a binary mask image of the candidate annular S-phase region. : ; Here, FillHoles is the hole-filling function. The principle of hole filling is: starting from the image boundary, the background area is flooded with fill; the unfilled internal areas are the holes surrounded by the edges. These holes are then filled as the foreground, resulting in a candidate region mask. See also... Figure 6 ,in, Figure 6 (a) shows the Canny edge detection results; Figure 6 (b) shows the result of morphological closure processing; Figure 6 (c) shows the morphological filling result; Figure 6 (d) shows the result of the annular mask stacking.

[0088] Screening of S203 annular S-phase morphological characteristics Morphological feature selection is performed on candidate annular S-phase regions, extracting the area, average brightness, and solidity features of each connected component. Based on area thresholds, lower brightness limits, and upper solidity limits, regions satisfying the morphological characteristics of annular S-phases are retained, resulting in an annular S-phase segmentation mask. Each selection parameter can be adjusted in real-time via an interactive interface. Binary images of the candidate regions are then processed. Connectivity component labeling is performed using the 8-neighborhood connectivity criterion, marking all connected foreground pixels as the same region. Let there be N connected components, denoted as... For each connected component Extract the following morphological features: (1) Area Connected components The total number of mid-foreground pixels is calculated using the following formula: ; (2) Average brightness: connected components The average gray value of all pixels in the normalized image The calculation formula is: ; (3) Solidity The ratio of the actual area of ​​a connected component to the area of ​​its convex hull is calculated using the following formula: ; in, For connected components The area of ​​the convex hull. The convex hull is the smallest convex polygon containing the connected region. The solidity value ranges from [0,1]. For solid convex regions (such as circles and rectangles), the solidity is close to 1; for regions with depressions or internal cavities (such as rings and C-shapes), the solidity is significantly less than 1. Due to its ring structure and the presence of a cavity in the center, the annular S-phase has a low solidity characteristic. Based on the above three characteristics, the screening criteria for the annular S-phase are set as follows: ; in, This is the minimum area threshold used to exclude noise and small-sized artifacts; the default value is 50 pixels. This is the maximum area threshold used to exclude large artifacts or overlapping regions; the default value is 10% of the total image area. This is the lower limit threshold for brightness, used to ensure that the filtered area is a high-contrast S-phase region rather than background noise; the default value is 0.3. This is the upper limit threshold for solidity, used to filter regions with ring-shaped or hollow features. The default value is 0.85.

[0089] Connected components that meet all screening criteria are determined to be cyclic S-phase, and a cyclic S-phase segmentation mask is generated. .

[0090] Extraction and Enhancement of S204 Needle-like S-phase Residual Images The annular S-phase segmentation mask is subtracted from the original image, and the subtracted region is filled with the median gray level of the image to obtain the residual image of the needle-like S-phase to be segmented. The residual image is then preprocessed using a second image segmentation module to enhance its longitudinal structure. A disk structuring element is used for morphological top-hat filtering to enhance the contrast between the needle-like S-phase and the background, resulting in an enhanced image. First, the original normalized image is calculated. Median gray level of all pixels : ; Where Median is the median operation function; Ω represents the entire pixel set of the image. The annular S-phase region is subtracted from the original image, and the subtracted region is filled with median gray to obtain the residual image. The purpose of using median grayscale filling is to avoid the annular S-phase region being misidentified as the needle-shaped S-phase in subsequent processing, while maintaining the continuity of the overall grayscale distribution of the image.

[0091] The needle-like S-phase is a slender, linear structure. Morphological top-hat transformation is used to enhance the bright, slender structure in the image. ; Among them, I tophat Output image for top-hat transformation; Represents the morphological opening operation. It is a disk-shaped structural element with a radius of . The default value is 5 pixels. The opening operation is defined as erosion followed by dilation: ; Top-hat transform can extract bright regions in an image that are smaller than the structuring element, effectively enhancing the contrast between the needle-like S-phase and the background. This results in an enhanced image. .

[0092] S205 needle-like S-phase skeleton extraction and segmentation The enhanced image is thresholded to obtain a binary image. A skeletonization algorithm is used to extract the morphological skeleton (i.e., skeleton lines, such as suspected candidate S-phases) from the binary image, and burr trimming is performed. Long skeletons belonging to the needle-like S-phase are selected based on a skeleton length threshold. The selected skeletons are then morphologically dilated to restore the actual width of the needle-like S-phase, resulting in a needle-like S-phase segmentation mask. First, the enhanced image... Thresholding segmentation is performed to obtain a binary image. : ; in, This is the brightness threshold; the default value is 0.4. Please refer to [link / reference]. Figure 7 Users can adjust parameters such as top cap, threshold, and skeleton length through the interactive interface.

[0093] The morphological skeleton of a binary image is extracted using a skeletonization algorithm. Skeletonization is a morphological thinning operation that shrinks a region to a single-pixel-wide centerline (skeleton) by iteratively eroding edge pixels, while preserving the region's topological connectivity. The skeleton is then trimmed to remove short branches at its endpoints. ; Where Prune is the pruning function; A pruning length threshold is set; branches shorter than this threshold will be removed. The default value is 10 pixels. Short branches are usually caused by noise or edge irregularities and do not represent true needle-like S-phase structures. Long skeletons belonging to the needle-like S-phase are selected based on the skeleton length threshold. The pruned skeleton... Perform connected component labeling. Suppose there are M skeleton segments in total. Calculate the length of the m-th skeleton segment: ; Retention length greater than minimum length threshold The skeleton segment, in this embodiment The default value is 15 pixels. Morphological dilation is performed on the filtered skeleton to restore the actual width of the needle-like S-phase, resulting in a needle-like S-phase segmentation mask. (Please see) Figure 8 , Figure 8 Figure (a) shows the result after removing the spiral S phase. Figure 8 (b) shows the result after the top cap reinforcement treatment. Figure 8 (c) is the result after binarization. Figure 8 (d) is the edge skeleton of the needle-like S phase. Figure 8 (e) shows the result after stacking needle-shaped masks. Figure 8 (f) is a superimposed image of spiral and needle-like features): ; Among them, S filtered This represents the filtered candidate region. It is a disk-shaped structural element with a radius of . The default value is 3 pixels. (Expansion radius) The setting should be based on the actual width of the needle-shaped S-phase, and users can adjust it through the interactive interface.

[0094] S206 Quantitative Statistics and Results Output Connected component labeling was performed on the segmentation masks for annular and acicular S-phases, and the area features of each phase cluster were extracted. The equivalent diameter of the annular S-phase and the major axis length of the acicular S-phase were calculated as size indicators. The area fraction distribution of S-phase clusters within different size ranges was statistically analyzed, and a size-area fraction distribution histogram was plotted. Quantitative analysis results were then output. (The annular S-phase mask was then analyzed.) Perform connected component labeling, assuming there are a total of There are several annular S-phase clusters. The area of ​​the j-th annular S-phase cluster is extracted as follows: ; Calculate the equivalent diameter The diameter of the region when its area is equivalent to a circle is used as a size index for the annular S-phase: ; Convert the equivalent diameter to the actual size. : ; Where k is the scale factor (nm / pixel) calculated in step S201.

[0095] Needle-shaped S-phase mask Corresponding skeleton Perform connected component labeling, assuming there are a total of There are 1 k-th needle-like S-phase clusters. The skeleton length is extracted from the k-th needle-like S-phase cluster. As a size indicator: ; Convert skeleton length to actual size : ; Calculate the area fraction distribution for each size interval (e.g.) Figure 9 (As shown). Let the total area of ​​the image be... (pixels).

[0096] For annular S-phase, the equivalent diameter (actual size) is divided into several intervals, and the area fraction of the m-th interval is: ; For acicular S-phase, the skeleton length (actual size) is divided into several intervals, and the area fraction of the k-th interval is... for: ; Calculate the total area fraction of the annular S-phase and the acicular S-phase: ; ; Draw size-area fraction distribution histograms: for the annular S-phase histogram, the horizontal axis is the equivalent diameter (nm) and the vertical axis is the area fraction (%); for the acicular S-phase histogram, the horizontal axis is the skeleton length (nm) and the vertical axis is the area fraction (%).

[0097] The segmentation mask and statistical data are exported as .mat and .xlsx files for easy subsequent analysis and processing. Simultaneously, a superimposed image of the segmentation results is generated, with annular S-phases marked in red and acicular S-phases marked in blue, superimposed on the original grayscale image for intuitive evaluation of the segmentation effect.

[0098] In some embodiments, this application also provides a schematic block diagram of the structure of a computer device, please see... Figure 12 Computer programs can be used in situations such as Figure 12 It runs on the computer device shown. Figure 12 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The memory may include non-volatile storage media and internal memory. The non-volatile storage media may store an operating system and computer programs. The computer programs include program instructions that, when executed, cause the processor to perform arbitrary methods. The processor provides computational and control capabilities to support the operation of the entire computer device. The internal memory provides an environment for the execution of the computer programs in the non-volatile storage media; when executed by the processor, these programs cause the processor to perform arbitrary methods. The network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 12The structures shown are merely block diagrams of a portion of the structure related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. It should be understood that the processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0099] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0100] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a computer terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0101] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A phase morphology statistical method, characterized in that, Including the following steps: S101, Obtain the first grayscale image of the alloy; S102, performing a reconstruction operation on the first grayscale image to identify the first S-phase, wherein S102 includes: S1021, perform morphological reconstruction on the first grayscale image to obtain a second grayscale image; S1022, in the second grayscale image, geometric shapes that meet the filtering rules are identified as the first S-phase, wherein the filtering rules include: (1) Morphological filtering rules, wherein the morphological filtering rules include: 、 ； (2) Contrast screening rules: ; in, Minimum area threshold, For the maximum area threshold, For the lower limit threshold of brightness, For the upper limit threshold of solidity, , , The area, average brightness, and solidity of the geometric figure; S103, perform an enhancement operation on the first grayscale image to identify the second S-phase, wherein S103 includes: S1031, the first S phase is deleted from the first grayscale image, forming a corresponding empty region, and the empty region is filled with reconstructed grayscale to obtain a third grayscale image; the reconstructed grayscale is determined based on the global grayscale value of the first grayscale image. S1032, Perform a morphological top-hat transformation operation on the third grayscale image to obtain a fourth grayscale image; S1033, perform threshold segmentation on the fourth grayscale image to obtain a binary image; S1034, A skeletonization algorithm is used to identify candidate S-phases from the binary image; S1035, identify the candidate S-phase whose length is greater than a preset length threshold as the second S-phase; S104, output the statistical results of the S phase based on the first S phase and the second S phase.

2. The method according to claim 1, characterized in that, S1021 includes: An edge detection algorithm is used to extract edge patterns from the first grayscale image, and at least one of the edge patterns has a broken edge; A morphological closing operation is performed on the edge graphics to connect at least some of the broken edges, thereby reconstructing a corresponding geometric shape; wherein the model of the morphological closing operation includes: ； in, The geometric figure, For the edge graphic, B is a disk-shaped structural element, and ⊕ represents the dilation operation. This represents the erosion operation; The geometric shape is filled with holes to form the second grayscale image.

3. The method according to claim 2, characterized in that, The radius of B is 1-4 pixels.

4. The method according to any one of claims 1-3, characterized in that, The operational model for the morphological top-hat transformation operation in S1032 includes: ； ； in, Represents the morphological opening operation. It is a disk-shaped structural element. The third grayscale image, This refers to the fourth grayscale image.

5. The method according to claim 4, characterized in that, The radius of is greater than the radius of B.

6. The method according to claim 5, characterized in that, The radius is 5-8 pixels.

7. The method according to claim 1, characterized in that, The reconstructed grayscale is the median of the global grayscale values ​​of the first grayscale image.

8. The method according to claim 1, characterized in that, Before S101, the following steps are also included: At least two photographs of the alloy were acquired at at least two imaging angles; The photo is divided into multiple image blocks; Calculate multiple standard deviations for multiple of the aforementioned patches; Calculate the variance of the multiple standard deviations; The photo with the smaller variance is selected as the recommended photo; The first grayscale image is generated based on the recommended photo.

9. A phase morphology statistical system, characterized in that, include: The image acquisition module is used to acquire the first grayscale image of the alloy; A reconstruction module is used to perform a reconstruction operation on the first grayscale image to identify the first S-phase, wherein the reconstruction module includes: A reconstruction unit is used to perform morphological reconstruction on the first grayscale image to obtain a second grayscale image; A filtering unit is configured to identify geometric shapes that meet filtering rules as a first S-phase in the second grayscale image, wherein the filtering rules include: (1) Morphological filtering rules, wherein the morphological filtering rules include: 、 ； (2) Contrast screening rules: ; in, Minimum area threshold, For the maximum area threshold, For the lower limit threshold of brightness, For the upper limit threshold of solidity, , , The area, average brightness, and solidity of the geometric figure; An enhancement module is used to perform enhancement operations on the first grayscale image to identify the second S-phase, wherein the enhancement module includes: A mask filling unit is used to delete the first S-phase in the first grayscale image, thereby forming a corresponding empty region, and to fill the empty region with reconstructed grayscale to obtain a third grayscale image; the reconstructed grayscale is determined based on the global grayscale value of the first grayscale image. A transformation unit is used to perform a morphological top-hat transformation operation on the third grayscale image to obtain a fourth grayscale image; A threshold segmentation unit is used to perform threshold segmentation on the fourth grayscale image to obtain a binary image; A skeleton recognition unit is used to identify candidate S-phases from the binary image using a skeletonization algorithm; A phase recognition unit is used to identify the candidate S-phase with a length greater than a preset length threshold as the second S-phase; The statistics module is used to output the statistical results of the S phase based on the first S phase and the second S phase.

10. The system according to claim 9, characterized in that, The reconfiguration unit includes: An edge pattern extraction subunit is used to extract edge patterns from the first grayscale image using an edge detection algorithm, wherein at least one of the edge patterns has a broken edge; A morphological closing operation subunit is used to perform a morphological closing operation on the edge graphics to at least connect a portion of the broken edges, thereby reconstructing a corresponding geometric shape; wherein, the model of the morphological closing operation includes: ； in, The geometric figure, For the edge graphic, B is a disk-shaped structural element, and ⊕ represents the dilation operation. This represents the erosion operation; The hole-filling subunit is used to fill holes in the geometric shape to form the second grayscale image.

Images