Automatic Identification Method of Precipitated Phases in Single-Crystal Superalloys Based on Scanning Electron Microscopy Images

Abstract

This invention relates to an automatic identification method for precipitates in single-crystal superalloys based on scanning electron microscopy images, belonging to the field of image processing technology. It solves the problem of misjudgment or missed judgment in the automatic identification and segmentation of precipitates in single-crystal superalloys in existing technologies. The specific steps include: performing grayscale processing on the acquired scanning electron microscopy image of the single-crystal superalloy to obtain grayscale images of the precipitate grains; performing binarization processing on the grayscale images using a first grayscale threshold and a second grayscale threshold respectively to obtain corresponding first and second binary images; matching all grain regions on the two denoised binary images, and obtaining a single-layer binary image of the corresponding upper grain based on the grains with determined correspondences; classifying the grains in the single-layer binary image to identify individual grains, partially adhered grains, and / or completely adhered grains, thus achieving efficient, stable, and objective automated identification of precipitates.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and particularly relates to a single crystal high-temperature alloy precipitated phase automatic identification method based on scanning electron micrographs. BACKGROUND

[0002] Nickel-based single crystal high-temperature alloy is widely used as a basic material for key equipment such as turbine blades of an aero-engine and hot end components of a gas turbine due to its excellent high-temperature mechanical properties and oxidation corrosion resistance. The high-temperature mechanical properties (such as creep strength and fatigue life) of the nickel-based high-temperature alloy are directly affected by the internal microstructure characteristics. The microstructure of the nickel-based high-temperature alloy is mainly composed of two phases of a γ-Ni matrix phase and a γ'-Ni3Al precipitated phase (also referred to as a γ' precipitated phase or a γ' phase). The γ' phase is coherently precipitated on the matrix and is an important strengthening phase. During the service process, due to the synergistic effect of temperature and stress, the γ' phase is directionally coarsened along a certain direction to form a raft-like structure, which significantly reduces the high-temperature mechanical properties of the alloy. The actual γ' phase precipitation and rafting microstructure are relatively complex. Temperature field, plastic flow, force field, ordering and crystal defects may all affect the morphology, distribution and volume change of the γ' phase, and further affect the high-temperature mechanical properties of the alloy. Therefore, quantitative identification and statistical analysis of the γ' phase are of great significance in material performance evaluation, process optimization and product quality control.

[0003] In the traditional analysis method, the observation and measurement of the metal microstructure mainly depend on the metallographic preparation and image acquisition technology. After corrosion, grinding and polishing, the images observed and collected by a metallographic microscope or a scanning electron microscope (SEM) have become the main means for studying the microstructure of materials such as grain, precipitated phase and defect distribution. However, the current analysis of these images mostly adopts artificial quantitative metallographic methods such as grid intercept method and micrometer eyepiece method. These methods are tedious, time-consuming and extremely dependent on the experience of operators, resulting in large errors in statistical results. In addition, due to the limited field of view of single observation, it is difficult to meet the engineering application requirements of large sample and rapid detection.

[0004] In recent years, with the development of digital image processing technology and computational geometry analysis method, automatic analysis has gradually become a hot spot for quantitative research of metallographic images. The existing automatic recognition method depends on the traditional algorithm, but in practical application, due to the influence of uneven illumination and noise interference of microscopic image during imaging of metallographic sample, the boundary of gamma' precipitated phase (showing grain-like, the precipitated phase is called grain hereinafter) in the image is often blurred and discontinuous; in addition, the surface matrix needs to be removed by corrosion during the preparation process of the metallographic sample, and the corrosion time is too long, which leads to excessive removal of the matrix, and then the gamma' phase is accumulated at different depths, which presents the phenomenon of overlapping of front and rear layers (i.e. upper layer and lower layer), so that the traditional algorithm is prone to misjudgment or omission during automatic recognition and segmentation, which affects the accuracy of the measurement result and the reliability of the statistical data. SUMMARY

[0005] In view of the above analysis, the embodiments of the present application aim to provide a single crystal superalloy precipitated phase automatic recognition method based on scanning electron microscopic images, to solve the problem of misjudgment or omission during automatic recognition and segmentation of single crystal superalloy precipitated phase in the prior art.

[0006] The purpose of the present application is mainly realized by the following technical solutions:

[0007] The embodiments of the present application provide a single crystal superalloy precipitated phase automatic recognition method based on scanning electron microscopic images, comprising the following steps:

[0008] The collected scanning electron microscopic images of single crystal superalloy are subjected to gray scale processing to obtain a gray scale image of precipitated phase grains;

[0009] The gray scale image is subjected to binaryzation processing using a first gray scale threshold value and a second gray scale threshold value to obtain a corresponding first binary image and a second binary image; wherein the second gray scale threshold value is greater than the first gray scale threshold value;

[0010] The first binary image and the second binary image are subjected to denoising processing respectively;

[0011] All grain regions on the two denoised binary images are corresponded, and a single layer binary image corresponding to the upper layer grains is obtained based on the grains with determined correspondence relationship;

[0012] The grains in the single layer binary image are classified to identify separate grains, partially adhered grains and / or completely adhered grains.

[0013] Further, the single layer binary image corresponding to the upper layer grains comprises:

[0014] The area of the grains with correspondence relationship in the denoised first and second binary images is used to identify completely upper layer grains;

[0015] respectively, to obtain corresponding output images I closed and I opened ;

[0016] based on the output images I closed and I opened , identify the upper layer grains and the lower layer grains in the upper and lower overlapping layers;

[0017] remove the lower layer grains on the denoised second binary image, and retain the completely upper layer grains and the upper layer grains in the upper and lower overlapping layers, to obtain the single layer binary image.

[0018] Further, the identifying the upper layer grains and the lower layer grains in the upper and lower overlapping layers comprises:

[0019] based on the denoised second binary image with labeled grain numbers, obtain output images I closed and I opened ;

[0020] sequentially judge each grain with the same number in the output images I closed and I opened : if a ratio of an area of the grain in the output image I opened to an area of the grain in the output image I closed is greater than a preset lower layer threshold, the grain is an upper layer grain in the upper and lower overlapping layers; otherwise, the grain is a lower layer grain.

[0021] Further, the identifying the completely upper layer grains comprises:

[0022] traverse each grain in the denoised second binary image, and if the following conditions are met, the grain belongs to a completely upper layer grain:

[0023]

[0024] wherein, Area(a) is an area of a grain in the denoised second binary image; Area(b) is an area of a corresponding grain in the denoised first binary image; E1 is a preset first level threshold.

[0025] Further, the identifying the single grain comprises:

[0026] sequentially calculate an aspect ratio of each grain in the single layer binary image, if the aspect ratio of the grain meets a preset first shape threshold range, identify the grain as a single grain, otherwise, the grain is a connected grain.

[0027] Further, the identifying the partially adhered grains and the completely adhered grains comprises:

[0028] Based on the single-layer binary image, the identified adhered grains are extracted to obtain an adhered binary image;

[0029] The edge of each adhered grain in the adhered binary image is calculated to obtain a minimum convex hull corresponding to each adhered grain;

[0030] If any adhered grain and the minimum convex hull corresponding to the adhered grain satisfy a shape condition, the adhered grain is identified as a partially adhered grain; otherwise, the adhered grain is identified as a completely adhered grain.

[0031] Further, the shape condition comprises:

[0032] The distance value of each pixel point in the adhered grain to the boundary pixel point of the adhered grain is calculated to obtain a maximum distance value h of the adhered grain region;

[0033] The distance value of each pixel point in the minimum convex hull corresponding to the adhered grain to the boundary pixel point of the adhered grain is calculated to obtain a maximum distance value h' of the minimum convex hull;

[0034] If is less than a preset second shape threshold, the shape condition is satisfied.

[0035] Further, the first binary image and the second binary image are obtained by:

[0036] The gray histogram of the gray image is obtained;

[0037] Based on the gray histogram, the first binary image is obtained by segmenting the gray image using a first gray threshold determined by the maximum inter-class variance;

[0038] The second binary image is obtained by segmenting the gray image using a second gray threshold which is the peak value of the gray value in the gray histogram.

[0039] Further, the first binary image and the second binary image are respectively subjected to denoising processing, comprising:

[0040] For any one of the binary images, the area of each connected region of grains in the binary image is calculated;

[0041] The area of each connected region in the binary image is traversed, and if the area of the connected region is less than a preset connected threshold, the connected region is removed;

[0042] The pixel points at the edges of the binary image are traversed, and if the pixel at the edge is an incomplete boundary, the connected region is removed.

[0043] Further, corresponding to all grain regions on the two denoised binary images, including:

[0044] Respectively determine the region coordinates of all grains on the two binary images;

[0045] Traverse each grain region in the denoised second binary image, and determine the corresponding grain of each grain in the denoised first binary image based on the following manner:

[0046] According to the coordinates of any pixel point in the grain region in the denoised second binary image, find the corresponding pixel point coordinates in the denoised first binary image, and the grain where the corresponding pixel point is located has a corresponding relationship with the grain in the denoised second binary image.

[0047] Compared with the prior art, the present application can at least achieve one of the following beneficial effects:

[0048] 1. The present application proposes a method for automatically identifying and extracting micro-morphology parameters of precipitated key from scanning electron micrography by statistical analysis of a large amount of image data, and through comparative analysis of two binary images obtained by strategy segmentation, the binary image of the upper layer precipitated phase is extracted, and the precipitated phase in the image is identified and analyzed, so that automatic identification and statistical analysis of the precipitated phase of single crystal superalloy are finally realized, the detection efficiency is greatly improved, the morphology characteristics of the precipitated phase are accurately identified, and the volume fraction, size, distribution and shape parameters of the precipitated phase are quantitatively analyzed.

[0049] 2. A reasonable segmentation threshold is selected to obtain two binary images, the area of the grain (i.e. the precipitated phase) in the two binary images is compared in combination with the morphological characteristics of the grain, and the complete upper layer grain is screened; the obtained single layer binary image is respectively subjected to morphological closing operation and opening operation, the area of the grain in the obtained corresponding image is used to identify the overlapping region and the lower layer grain, advanced image processing and geometric technology is introduced to fully exploit the brightness and geometric characteristics presented in the SEM image due to sample preparation and imaging conditions, the overlapping interference of the lower layer precipitated phase to the upper layer precipitated phase is overcome, subjective errors existing in the manual measurement process are avoided, and the accuracy of the data is ensured.

[0050] 3. The shape type of the precipitated phase (i.e. the identification of the grain adhesion condition) in the single layer binary image from which the upper layer precipitated phase is extracted is combined with the set characteristics of the grain shape, different judgment indexes are used to calculate the adhesion degree, the division of separate grains, edge adhesion grains and completely adhesion grains is realized, and then the grain size and shape are quantitatively counted, which not only shortens the analysis period, but also establishes a quantitative correlation between the microstructure and the mechanical properties, and provides data support for the quantitative research of material organization evolution.

[0051] 4. The noise points on the binary image are removed by using the determination of the connected region, so that a clearer image of the grain boundary is obtained, and the incomplete grains at the image boundary are removed, the interference of the blurred edges and the noise of the precipitated phase on the image is reduced, and the accuracy of the recognition of the precipitated phase is improved.

[0052] The above technical solutions can be combined with each other in the present application to realize more preferred combination solutions. Other features and advantages of the present application will be described in the subsequent description, and some advantages will become apparent from the description, or will be understood by implementing the present application. The purposes and other advantages of the present application can be realized and obtained from the contents specifically indicated in the description and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0053] The accompanying drawings are included to provide a further understanding of the present application, and are incorporated herein and constitute a part of the detailed description. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application. In the drawings:

[0054] Figure 1 A flowchart of the single crystal superalloy precipitated phase automatic recognition method based on a scanning electron micrograph image according to an embodiment of the present application;

[0055] Figure 2 A gray scale diagram of a single crystal superalloy electron micrograph according to an embodiment of the present application;

[0056] Figure 3 An example diagram of grains in a binary image obtained by using different threshold values according to an embodiment of the present application;

[0057] Figure 4 A schematic diagram of a preprocessed binary image according to an embodiment of the present application;

[0058] Figure 5a A schematic diagram of a single grain classified by a single crystal superalloy precipitated phase according to an embodiment of the present application;

[0059] Figure 5b A schematic diagram of a partially adhered grain classified by a single crystal superalloy precipitated phase according to an embodiment of the present application;

[0060] Figure 5c A schematic diagram of a completely adhered grain classified by a single crystal superalloy precipitated phase according to an embodiment of the present application. DETAILED DESCRIPTION

[0061] The preferred embodiments of the present application will be specifically described below with reference to the accompanying drawings, wherein the drawings constitute a part of the present application, and are used to explain the principles of the embodiments of the present application, and are not used to limit the scope of the present application.

[0062] Embodiment 1

[0063] One specific embodiment of the present application discloses a single crystal superalloy precipitated phase automatic identification method based on scanning electron micrograph, as shown in the formula (I), comprising the following steps: Figure 1

[0064] Step S1, the collected single crystal superalloy scanning electron micrograph is grayed to obtain the gray image of the precipitated phase grain.

[0065] Step S2, the gray image is binarized by using the first gray threshold and the second gray threshold respectively to obtain the corresponding first binary image and the second binary image; wherein the second gray threshold is greater than the first gray threshold.

[0066] Step S3, the first binary image and the second binary image are denoised respectively.

[0067] Step S4, the corresponding relationship of all grain regions on the two denoised binary images is determined, and the single layer binary image corresponding to the upper layer grain is obtained based on the determined grain.

[0068] Step S5, the grains in the single layer binary image are classified to identify the single grain, the partially adhered grain and / or the completely adhered grain.

[0069] Through the above method, two binary images are divided by using different thresholds, the upper layer precipitated phase is identified by comparing the grain regions (i.e. precipitated phase) in the two denoised binary images, and the shape classification of the precipitated phase in the single layer binary image is carried out to automatically and accurately identify the precipitated phase and its state in the scanning electron micrograph.

[0070] In the research of metal alloy in material science, the single crystal superalloy has the phenomenon of uneven distribution of precipitated phase composition in space under high temperature service, and the corresponding scanning electron micrograph has a dark background part of the matrix phase and a bright white precipitated phase which may appear upper and lower layers, even upper and lower layers overlap; at the same time, the brightness of upper and lower layers is different due to the depth difference of precipitated phase; among them, the upper layer precipitated phase is closer to the surface, has higher corrosion degree, stronger reflected electron signal and higher brightness, while the lower layer is located in the matrix gap, the signal attenuation is significant, and the brightness contrast is reduced (i.e. the upper layer is brighter, and the lower layer is darker).

[0071] For example, the single crystal alloy is a nickel-based single crystal. Specifically, in step S1, the collected scanning electron micrograph is grayed to convert each pixel value of the nickel-based single crystal material precipitated phase organization image into a single brightness value, remove the color information of the image, retain only the brightness information, and obtain a gray image, as shown in the formula (II). Figure 2

[0072] For example, the gray processing is realized by weighted average, as shown in the following formula:​​

[0073] Gray = 0.2989R + 0.5870G + 0.1440B,

[0074] where R, G and B represent the red, green and blue channel values of each pixel in the original image, respectively.

[0075] Specifically, in step S2, the gray-scale image is divided by selecting a reasonable gray-scale threshold to obtain two binary images, which are used for subsequent removal of the influence of the lower layer grains. Since the second gray-scale threshold is greater than the first gray-scale threshold, the grain regions in the second binary image obtained by the second gray-scale threshold are all included in the corresponding grain regions in the first binary image. Specifically, the following steps are included:

[0076] S21, obtaining a gray-scale histogram of the gray-scale image, i.e., counting the number of pixels of each gray-scale value in the image;

[0077] It should be noted that when binarizing, the pixel distribution of the gray-scale image is usually analyzed using a gray-scale histogram. Since there is a difference in the upper and lower layer gray-scale values in the collected scanning electron micrograph, the gray-scale histogram will present a form with two significant peaks. The left side of the bimodal histogram will have a peak of low gray-scale values, corresponding to the background part of the image; the right side will have a peak of high gray-scale values, corresponding to the foreground part of the image. Based on the pixel distribution of the gray-scale image, a suitable threshold is selected for segmentation to obtain a binary image, wherein the gray-scale threshold is between 0 and 255. If the gray-scale value of a pixel is greater than or equal to the gray-scale threshold, the pixel is classified as foreground (usually assigned a value of 255, i.e., white); if the gray-scale value of a pixel is less than the threshold, the pixel is classified as background (usually assigned a value of 0, i.e., black).

[0078] S22, based on the gray-scale histogram, determining the best threshold using the maximum inter-class variance to obtain a first gray-scale threshold T1, which is used to segment the gray-scale image to obtain a first binary image;

[0079] For example, each gray-scale value in the range of 0-255 is sequentially used as a threshold, and the inter-class variance between the foreground and background of the image segmented by the threshold is calculated. The threshold determined by maximizing the inter-class variance between the foreground and background is used as the first gray-scale threshold for the segmentation of the gray-scale image.

[0080] S23, using the peak value of the maximum gray-scale value in the gray-scale histogram as a second gray-scale threshold T2 to segment the gray-scale image to obtain a second binary image.

[0081] It should be noted that the gray value in the histogram is larger and larger from left to right, and the brightness is higher. The core of the threshold value determined by the maximum inter-class variance is to maximize the difference between the foreground and the background of the image. For the histogram with double peaks, the threshold value determined in this way will appear between the two peaks. Since the second gray threshold selects the right peak with the largest gray value in the double-peak histogram, the second gray threshold will be greater than the first gray threshold. Further, the grain area in the second binary image (corresponding to the foreground part) is relatively smaller than the corresponding grain area in the first binary image.

[0082] In the above manner, the gray-scale image is segmented by using the preferred peak threshold combined with the maximum inter-class variance threshold to form a contrast analysis of the binary image, as shown in Figure 3 The first row (a) shows a contrast example of the upper layer grain at the corresponding position in the initial gray-scale image, the first binary image and the second binary image, the second row (b) shows a contrast example of the overlapping grain at the corresponding position in the above three images, and the third row (c) shows a contrast example of the lower layer grain at the corresponding position in the above three images. It can be seen that this method can present obvious differences between different layer grains, can effectively distinguish the overlapping area, and solves the problem that the conventional algorithm is difficult to distinguish the overlapping area, thereby laying a foundation for subsequent grain recognition.

[0083] Specifically, in step S3, the two binary images are denoised in the same way; wherein the specific steps of denoising any binary image include:

[0084] S31, the area of all connected regions of the binary image is calculated respectively;

[0085] Specifically, the connected region H i in the binary image is marked, where i=1, 2,..., n represents the index of each connected region, and the area A(H i ) of the connected region H i is calculated.

[0086] S32, the area of each connected region in the binary image is traversed, and if the area of the connected region is less than a preset connected threshold, the connected region is removed;

[0087] Exemplarily, the preset connected threshold A min is 10 pixels; for the area of any connected region, if A(H i ) is less than A min , the connected region H i is regarded as a noise point and is removed; the area of each connected region in the binary image is traversed to obtain an image with clearer grain boundary details. Through analysis of the resolution size of a large number of scanning electron micrographs, the magnification size of the microscope when sampling the image, and the structure of the grain itself, it is determined that the threshold value is 10 pixel values.

[0088] S33. Traverse each pixel at the edge of the binary image. If the pixel at the edge is an incomplete boundary, remove the connected region to remove incomplete grains at the image boundary.

[0089] For example, in a binary image where pixels are only 0 or 1, a pixel value of 1 represents a grain. By traversing the pixels at the image edges, when a pixel value is 1, its connected region is deleted. This ultimately removes grains from the image edges.

[0090] The above denoising process removes interference noise and incomplete grains at the boundaries from the acquired image, ensuring the quantization accuracy of subsequent algorithm research.

[0091] Specifically, in step S4, the steps for identifying the grain region (i.e., the upper grain) suitable for subsequent quantization and classification based on the two denoised binary images, and obtaining the corresponding single-layer binary image of the upper grain, include:

[0092] S41. Perform positional mapping on all grain regions in the two denoised binary images to ensure comparison of grains at the same location in both images. This includes:

[0093] S411. Determine the region coordinates of all grains in the two denoised binary images respectively;

[0094] For example, since the spacing between many grains in an image is narrow, the 4-neighborhood method is used to determine the connected regions as the target grain regions, thereby obtaining the coordinates of the grain regions. The 4-neighborhood method means that for any pixel (x, y) in the image, its four directly adjacent pixels (top, bottom, left, and right) are defined as its 4-neighborhood. The coordinates of these four neighboring pixels can be represented as: (x-1, y) (top); (x+1, y) (bottom); (x, y-1) (left); (x, y+1) (right). If two pixels are within the 4-neighborhood and have the same grayscale attribute, they are considered connected, and the coordinates of the target grain region are determined.

[0095] S412. Match the coordinates of any grain in the denoised second binary image with the coordinates of each grain in the denoised first binary image.

[0096] For example, because each grain region in the second binary image is smaller than the corresponding grain region in the first binary image—in other words, the position of the grains in the same coordinate system remains unchanged in both images—only the grain region in the second binary image is contained within the corresponding grain region in the first binary image, based on this characteristic, the following judgment is made for any grain in the denoised second binary image to determine the corresponding grain in the denoised first binary image:

[0097] For any grain in the denoised second binary image, take any pixel point coordinate in the grain region, find the corresponding pixel point coordinate in the denoised first binary image, and the grain where the corresponding pixel point is located has a corresponding relationship with the grain in the denoised second binary image.

[0098] S413, traverse each grain region in the denoised second binary image, and determine the corresponding grain of each grain in the denoised first binary image.

[0099] S42, according to the analysis of the grains in the collected nickel-based single crystal material precipitated phase organization image, the grains are divided into three categories, which are single grain, upper and lower layer overlapping grain and lower layer grain. Based on the grains determined by the corresponding relationship, the upper and lower layer situation is identified, including:

[0100] S421, using the area of the corresponding relationship grain in the denoised first and second binary images, to identify the completely upper layer grain.

[0101] For example, traverse each grain in the denoised second binary image, if the following conditions are met, the grain belongs to the completely upper layer grain:

[0102]

[0103] Wherein, Area(a) is the area of the grain in the denoised second binary image; Area(b) is the area of the corresponding grain in the denoised first binary image; E1 is a preset upper layer threshold.

[0104] It should be noted that, as Figure 3 shown, for the completely upper layer grain with higher brightness, the area difference of the grain in the two binary images obtained by using different threshold segmentation is small, and combined with the geometric characteristics, the grain data in a large number of images is statistically analyzed to determine that the upper layer threshold is 90%, which can be adjusted according to actual situation.

[0105] S422, respectively, the denoised second binary image is subjected to morphological closing operation and opening operation to obtain the corresponding output images I closed and I opened ;

[0106] Specifically, the closing operation is completed by two-step sequence combination of dilation and erosion, which is represented as:

[0107]

[0108] Wherein, I closed is the output image after morphological closing operation; I is the denoised second binary image; B is the structure element; Represents morphological expansion; Represents morphological corrosion.

[0109] Using the above method, image I is first dilated to expand bright areas and fill in holes and broken areas. Then, an erosion operation is performed on the dilated image to remove noise caused by the dilation, preserving the integrity of the original object's outline, resulting in the output image I. closed .

[0110] The opening operation is completed by a sequential combination of two steps: erosion and expansion, represented as:

[0111]

[0112] First, an erosion operation is performed on image I to remove small noise, reduce the size of objects, and break up small connected regions. Then, a dilation operation is performed on the eroded image to restore the main target area left after erosion, without recovering the original noise, resulting in the output image I. opened .

[0113] S423, Based on the output image I closed and output image I opened The algorithm identifies the upper and lower grains in the overlapping layers, removes the lower grains from the denoised second binary image, and retains the complete upper grains and the upper grains in the overlapping layers to obtain a single-layer binary image of the corresponding upper grain, such as... Figure 4 As shown.

[0114] For example, before performing opening and closing operations on the denoised second binary image, the positions of the grains in the image are pre-numbered, and the output image I is then processed sequentially. closed and output image I opened The same numbered grains are identified in the output image I: if the grain is in the same output image I... opened The area of ​​the region in the output image I closed If the area of ​​a region in a given layer satisfies the following condition, then the grain is the upper grain in the overlapping layers; otherwise, the grain is the lower grain:

[0115]

[0116] Among them, Area(a closed ) is the output image I closed Area of ​​the middle grain a; Area(a) opened ) is the output image I opened The area of ​​the region corresponding to grain a; E2 is the preset second-level threshold, which is determined to be 95% based on statistical analysis, and can be adjusted according to the actual situation.

[0117] Further, after the extraction of the upper layer of the phase region, the area thereof can be calculated to provide data support for the quantitative study of the material organization evolution.

[0118] Specifically, in step S5, the grain types in the scanning electron micrograph of the single crystal high-temperature alloy material include three types: single grains, grains with edge adhesion, and completely adhered grains with smooth edges. The steps for identifying the grain types in the single-layer binary image include:

[0119] S51, identify single grains and adhered grains;

[0120] For example, the aspect ratio is used for identification: the aspect ratio of each grain in the single-layer binary image is calculated in sequence. If the aspect ratio of the grain meets the preset first shape threshold range, the grain is identified as a single grain, otherwise, the grain is an adhered grain. The identified single grains are as shown in FIG. 5B. Figure 5a

[0121] The aspect ratio is usually the ratio of the long axis to the short axis of the grain, and is expressed as: '

[0122] In material science, the γ precipitated phase single grain is in a square state, and the adhered grain is usually composed of two or more single grains, so the first shape threshold range is set to [1-1.5], that is, when 1≤Aspect≤1.5, the grain is determined to be a single grain.

[0123] S52, calculate and classify the adhesion degree of the adhered grain, identify the grain with edge adhesion and / or complete adhesion, specifically including:

[0124] S521, based on the single-layer binary image, extract the identified adhered grain to obtain an adhered binary image;

[0125] S522, calculate the edge of each adhered grain in the adhered binary image to obtain the minimum convex hull corresponding to each adhered grain;

[0126] For example, the edge pixels of the adhered grain are solved by the divide and conquer method of two-dimensional point set convex hull to obtain the convex hull. This algorithm uses a recursive method to find the "external" points that constitute the convex hull, and by constantly "eliminating" internal points, the minimum convex polygon containing all points is finally obtained, which is the convex graph of the adhered grain.

[0127] S523, if any adhered grain and its corresponding minimum convex hull meet the shape condition, the adhered grain is identified as a partially adhered grain, as shown in FIG. 5D; otherwise, it is a completely adhered grain, as shown in FIG. 5E. Figure 5b Figure 5c ​​​

[0128] Specifically, the distance transform is performed on the adhesion binary image to obtain an image containing distance information. The region of the originally adhesion grain pixel points (i.e. the grain interior) is obtained after the distance transform, which can be distinguished from the different distances of the background region.

[0129] For example, the Euclidean distance is used to calculate the distance value of each pixel point in any adhesion grain to the pixel point of the adhesion grain boundary (i.e. the background region), and the maximum distance value is selected and recorded as h (because it is the distance value from the grain boundary, the distance value at the geometric center of the grain is the maximum).

[0130] Similarly, the distance value of each pixel point of the minimum convex hull corresponding to the adhesion grain to the pixel point of the adhesion grain boundary is calculated to obtain the maximum distance value h' of the minimum convex hull (which is the depth of the adhesion of the grain).

[0131] h'

[0132] If h is less than a preset second morphology threshold, the morphology condition is satisfied, and the grain is determined to be partially adhesion. The preset second morphology threshold is determined to be 90% according to statistical analysis, which can be adjusted according to actual conditions.

[0133] Further, after the grain shape type is determined, the size, shape and other specific properties of the precipitated phase can be further counted to ensure the accuracy of the data.

[0134] Compared with the prior art, the single crystal superalloy precipitated phase automatic recognition method based on a scanning electron micrograph image provided by the embodiment overcomes the overlapping interference of the lower precipitated phase relative to the upper precipitated phase in the scanning electron micrograph image, extracts the two-phase diagram of the upper precipitated phase, and can automatically and rapidly identify the state of the γ' phase in the image, and finally realizes efficient and accurate extraction of the morphology characteristics of the precipitated phase. On the one hand, the automatic analysis method of the method helps to reduce human intervention and improve the accuracy and consistency of the data; on the other hand, through the identification of the adhesion degree of the grain, the quantitative statistical analysis of the precipitated phase is realized, which provides a scientific basis for optimizing the heat treatment process of the single crystal superalloy and improving its service performance.

[0135] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the program can be stored in a computer readable storage medium. The computer readable storage medium includes a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.

[0136] The above merely describes preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for automatic identification of precipitates in single crystal superalloys based on scanning electron micrographs, characterized in that, The method comprises the following steps: The collected single crystal high-temperature alloy scanning electron micrograph is grayed to obtain a gray image of precipitated phase grains; The gray image is binarized using a first gray threshold and a second gray threshold to obtain a corresponding first binary image and a second binary image; the second gray threshold is greater than the first gray threshold; The first binary image and the second binary image are denoised respectively; Corresponding is performed on all the grain regions of the two denoised binary images, and a single-layer binary image corresponding to the upper-layer grains is obtained based on the grains with the determined correspondence; wherein the single-layer binary image contains complete upper-layer grains and upper-layer grains in the upper-and-lower-overlapping layer; based on the denoised second binary image with the labeled grain numbers, an output image is obtained and the output image ; each grain with the same number in the output image and the output image is judged in sequence: if the ratio of the area of the grain in the output image to the area of the grain in the output image is greater than a preset lower-layer threshold, the grain is an upper-layer grain in the upper-and-lower-overlapping layer; otherwise, the grain is a lower-layer grain; The grains in the single-layer binary image are classified to identify separate grains, partially adhered grains and / or completely adhered grains.

2. The method according to claim 1, wherein, The single-layer binary image corresponding to the upper layer grains is obtained, comprising: The area of the corresponding relationship grain in the denoised first binary image and second binary image is used to identify completely upper layer grains; Respectively, the second binary image after denoising is carried out morphological close operation and open operation, obtain the corresponding output image and ; based on the output image and output image identifying upper and lower grains in the upper and lower overlapping layers The lower layer grains in the denoised second binary image are removed, and the completely upper layer grains and the upper layer grains in the upper and lower overlapping layers are retained to obtain the single-layer binary image.

3. The method of claim 2, wherein the method is characterized by: The completely upper layer grains are identified, comprising: Each grain in the denoised second binary image is traversed, and if the following conditions are met, the grain belongs to the completely upper layer grain: ， wherein, is a region area of the crystal grain in the denoised second binary image; is a region area of the corresponding crystal grain in the denoised first binary image; is a preset first level threshold.

4. The method of claim 1, wherein the method is characterized by: The separate grains are identified, comprising: The aspect ratio of each grain in the single-layer binary image is calculated in sequence, and if the aspect ratio of the grain meets the preset first shape threshold range, the grain is identified as a separate grain, otherwise, the grain is an adhered grain.

5. The method of claim 4, wherein the method is characterized by: The partially adhered grains and the completely adhered grains are identified, comprising: Based on the single-layer binary image, the identified adhered grains are extracted to obtain an adhered binary image; The edges of each adhered grain in the adhered binary image are calculated to obtain the minimum convex hull corresponding to each adhered grain; If any adhered grain and its corresponding minimum convex hull meet the shape condition, the adhered grain is identified as a partially adhered grain; otherwise, it is a completely adhered grain.

6. The method of claim 5, wherein the method is characterized by: The shape condition is met, comprising: Calculate the distance value from each pixel point in the any one of the coherent grains to the pixel point on the boundary of the coherent grain, to obtain the maximum distance value of the coherent grain ; Calculate the distance value from each pixel point of the minimum convex hull corresponding to the conglutination crystal grain to the boundary pixel point of the conglutination crystal grain, and obtain the maximum distance value corresponding to the minimum convex hull ; If The shape condition is satisfied if the shape is smaller than a preset second shape threshold.

7. A method for automatic identification of precipitates in single crystal superalloys based on scanning electron micrographs according to any one of claims 1 to 6, characterized in that, The first binary image and the second binary image are obtained, comprising: The gray histogram of the gray image is obtained; The first binary image is obtained by segmenting the gray image using the first gray threshold determined by the maximum inter-class variance based on the gray histogram; The second binary image is obtained by segmenting the gray image using the second gray threshold which is the peak value of the maximum gray value in the gray histogram.

8. The method of claim 1, wherein the method is characterized by: The first binary image and the second binary image are denoised respectively, comprising: For any binary image, the area of all connected regions of grains in the binary image is calculated respectively; Each connected region area in the binary image is traversed, and if the area of the connected region is less than a predetermined connected threshold, the connected region is removed; Each edge pixel point in the binary image is traversed, and if the edge pixel is an incomplete boundary, the connected region is removed.

9. The method of claim 1, wherein the method is characterized by: The areas of all grain regions on the two denoised binary images are corresponded, comprising: The region coordinates of all grains on the two binary images are determined respectively; Each grain region in the denoised second binary image is traversed, and the corresponding grain of each grain in the denoised first binary image is determined based on the following method: According to any pixel point coordinate in the grain region in the denoised second binary image, a corresponding pixel point coordinate in the denoised first binary image is found, and the grain where the corresponding pixel point is located is in a corresponding relationship with the grain in the denoised second binary image.

Images