Quantitative method for representing dispersibility of CNTs in resin-based composite material system

Abstract

The invention provides a quantitative method for representing dispersibility of CNTs in a resin-based composite material system, and relates to the technical field of image processing, the method comprises the following steps: obtaining a two-dimensional distribution image of CNTs in a resin-based composite material, applying gamma gray scale transformation to a gray scale image, and generating a gradient amplitude image; binarization is realized by using a function im2bw; identifying a target area formed by a pixel set which has the same pixel gray value and is communicated with each other; dividing the target area into rectangular image infinitesimal elements according to the common divisor of the transverse and longitudinal resolutions; calculating the area proportion of the CNTs in each image infinitesimal, flattening the area proportion matrix of all the infinitesimal into a vector Q, and calculating the variance and the standard deviation; according to the method, the problems that in the prior art, due to the fact that a unified, objective and quantitative evaluation means is lacked, the dispersity of the CNTs in the resin-based composite material is often analyzed depending on artificial experience or subjective judgment, the subjectivity of a characterization result is high, and the real distribution condition of the CNTs in a matrix is difficult to accurately reflect are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and particularly relates to a quantitative method for representing dispersion of CNTs in a resin-based composite material system. BACKGROUND

[0002] At present, the evaluation method for dispersion of CNTs in a resin-based composite material system usually relies on electronic or optical imaging technology (such as ultraviolet-visible spectroscopy, optical microscopy, scanning electron microscopy SEM and transmission electron microscopy TEM), first acquires the distribution characteristics of CNTs in a resin matrix, and then combines indexes such as aggregation degree and distribution uniformity to perform qualitative analysis.

[0003] For example, Yanhui Wang and other scholars qualitatively evaluated the dispersion effect of MWCNTs in an epoxy resin when evaluating the dispersion of MWCNTs functionalized by o-benzene triol in the epoxy resin by using ultraviolet-visible spectroscopy and optical microscopic observation method; Xiaohui Bao and other scholars qualitatively observed the dispersion state of functionalized CNTs in an epoxy resin when exploring the thermal degradation behavior of a modified carbon nanotube reinforced epoxy resin composite by using scanning electron microscopy (SEM); Amit K. Chakraborty and other scholars systematically represented the dispersion state of CNTs in a bisphenol F-based epoxy resin from mixing, curing to the formation of a final composite material by using an optical microscope.

[0004] In summary, the technical problem in the prior art is that the representation method based on imaging technology combined with artificial observation relies on subjective experience of researchers, it is difficult to objectively and uniformly define the uniformity of the distribution of CNTs in a resin matrix, and it is difficult to accurately reflect the improvement effect of different modification or dispersion processes on the dispersion state of CNTs by using a quantitative index. SUMMARY

[0005] The purpose of the present application is to provide a quantitative method for representing dispersion of CNTs in a resin-based composite material system, so as to solve the technical problem in the prior art that the representation method based on imaging technology combined with artificial observation relies on subjective experience of researchers, it is difficult to objectively and uniformly define the uniformity of the distribution of CNTs in a resin matrix, and it is difficult to accurately reflect the improvement effect of different modification or dispersion processes on the dispersion state of CNTs by using a quantitative index.

[0006] In view of the above problems, the application provides a quantitative method for characterizing the dispersion of CNTs in a resin-based composite system, wherein the image collected from a high-resolution imaging device is used to obtain a two-dimensional distribution image of CNTs in the resin-based composite, the two-dimensional distribution image is a TIF image containing an Alpha channel, and the Alpha channel is deleted to convert the image into a grayscale image; a gamma grayscale transformation is applied to the grayscale image to achieve nonlinear enhancement of the image contrast; the enhanced image is input into an image derivation method based on a Sobel operator to calculate a convolution factor Gx along the horizontal direction and a convolution factor Gy along the vertical direction, and a gradient amplitude image is generated to obtain a CNTs edge feature image; the edge feature image is converted into a binary image using a function im2bw, and an area threshold constraint condition is applied to the binary image to screen out small-area noise regions; the binary image that meets the area threshold condition is subjected to connected domain recognition and segmentation to identify a target region formed by a pixel set with the same pixel grayscale value and mutual connectivity; resolution analysis is performed on the target region, the target region is set to have a horizontal resolution W and a vertical resolution H, and the target region is divided into m x n rectangular image units according to the common divisor of the horizontal and vertical resolutions; the area proportion of CNTs in each image unit is calculated, and all the area proportion matrices of the units are flattened into a vector Q, and the variance and standard deviation are calculated to quantify the dispersion uniformity of CNTs in the two-dimensional resin plane.

[0007] The one or more technical solutions provided in the application have at least the following beneficial effects: By automatically extracting the features of the two-dimensional microscopic image of CNTs and combining standard deviation analysis, the problem of relying on subjective judgment of researchers in the traditional method is avoided, and the uniformity of the distribution of CNTs in the resin-based composite system is uniformly and objectively defined. A quantitative index based on standard deviation is established, which can directly reflect the improvement effect of different modification methods or dispersion processes on the dispersion state of CNTs, overcoming the lack of quantitative index in the prior art. By combining image processing and data analysis, the repeatability of the characterization process and the stability of the analysis results are ensured, thereby improving the scientificity and reliability of the evaluation of CNTs dispersion. The method of the application is not only suitable for the dispersion evaluation of CNTs in an epoxy resin matrix, but also can be extended to other types of resin-based composite systems, and has strong universality and application value. By providing accurate and intuitive quantitative results, the method of the application can provide a scientific basis for the optimization of CNTs dispersion process, and is helpful to promote the development and application of CNTs modification, dispersion and composite material preparation and other related technologies.

[0008] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0010] Figure 1 This is an image (original image) showing the distribution characteristics of CNTs inside the CNTs / PPESK composite film obtained in Embodiment 1 of the present invention. Figure 2 for Figure 1 Image after gamma transformation; Figure 3 for Figure 2 The image calculated using the Sobel operator; Figure 4 for Figure 3 Image after thresholding and binarization; Figure 5 for Figure 4 Image after noise reduction; Figure 6 for Figure 5 The image after image region segmentation and standard deviation calculation; Figure 7 This is the image before processing using the wet ball milling method in Embodiment 2 of the present invention; Figure 7 Image (a) shows the distribution of CNTs in the thin film without ball milling treatment; Figure 7 Image (b) shows the distribution of CNTs in the film after ball milling for 2 hours; Figure 7 (c) in the image represents the distribution of CNTs in the film after ball milling for 4 hours; Figure 7 (d) in the image represents the distribution of CNTs in the film after ball milling for 6 hours; Figure 7 (e) in the image represents the distribution of CNTs in the film after ball milling for 8 hours; Figure 7(f) of FIG. 1 shows the distribution image of CNTs in the thin film after ball milling for 10 hours; Figure 7 (g) of FIG. 1 shows the distribution image of CNTs in the thin film after ball milling for 12 hours; Figure 7 (h) of FIG. 1 shows the distribution image of CNTs in the thin film after ball milling for 14 hours; Figure 7 (i) of FIG. 1 shows the distribution image of CNTs in the thin film after ball milling for 16 hours; Figure 7 (j) of FIG. 1 shows the distribution image of CNTs in the thin film after ball milling for 18 hours; Figure 7 (k) of FIG. 1 shows the distribution image of CNTs in the thin film after ball milling for 20 hours; Figure 7 FIG. 2 shows the image of the ball milling method using wet ball milling for Example 2 of the present application; Figure 8 (a) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film without ball milling; Figure 8 (b) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 2 hours; Figure 8 (c) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 4 hours; Figure 8 (d) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 6 hours; Figure 8 (e) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 8 hours; Figure 8 (f) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 10 hours; Figure 8 (g) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 12 hours; Figure 8 (h) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 14 hours; Figure 8 (i) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 16 hours; Figure 8 (j) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 18 hours; Figure 8 (k) of FIG. 2 shows the dispersion characterization result image of CNTs in the thin film after ball milling for 20 hours; Figure 9 FIG. 3 shows the flow chart of Example 1 of the present application. DETAILED DESCRIPTION

[0011] The application solves the technical problems in the prior art that it is difficult to objectively determine the uniformity of CNTs distribution due to the dependence on manual observation and the lack of quantitative indicators, resulting in the lack of scientificity and universality of dispersion evaluation by providing a quantitative method for characterizing the dispersion of CNTs in a resin-based composite system. The area distribution characteristics of CNTs in each local region are determined by image element division and black pixel ratio calculation of the two-dimensional microscopic image of CNTs, and a standard deviation quantitative characterization model is constructed, thereby improving the accuracy, objectivity and repeatability of dispersion evaluation.

[0012] The technical solutions in the application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the application, rather than all the embodiments of the application. It should be understood that the application is not limited by the example embodiments described herein. Based on the embodiments of the application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the application. In addition, it should be noted that, for the convenience of description, only parts related to the application are shown in the drawings, not all.

[0013] Embodiment one The application provides a quantitative method for characterizing the dispersion of CNTs in a resin-based composite system, as shown in the accompanying Figure 9 The quantitative method for characterizing the dispersion of CNTs in a resin-based composite system specifically includes the following steps: An image from a high-resolution imaging device is collected to obtain a two-dimensional distribution image of CNTs in a resin-based composite material, the two-dimensional distribution image is a TIF image containing an Alpha channel, and the Alpha channel is deleted to convert the image into a gray-scale image.

[0014] Specifically, a high-resolution imaging device (such as a scanning electron microscope or a high-resolution optical microscope) is used to image a resin-based composite sample to obtain a two-dimensional distribution image containing CNTs distribution information. The two-dimensional distribution image is saved in TIF format and contains an Alpha channel. First, the fourth channel (Alpha channel) value is deleted, and the four-channel image (RGB+Alpha) is converted into a gray-scale image. For example, a 3D digital optical microscope produced by Haoshi Corporation with a model number of RH-2000 is used to obtain a distribution characteristic image of CNTs inside a CNTs / PPESK composite film with a magnification of 1000 times, a mass fraction of 5wt%, and a thickness of 1mm, as shown in the accompanying Figure 1 Through this processing step, the influence of color dimension on image analysis can be effectively reduced, and the performance of CNTs dispersion characteristics in the image is enhanced.

[0015] A gamma gray scale transformation is applied to the gray scale image to achieve nonlinear enhancement of image contrast.

[0016] According to the plurality of heterogeneous storage units, data synchronization conflict identification is performed on the plurality of source heterogeneous data sets to determine the data units to be synchronized.

[0017] Further, the application further includes the following steps: obtaining a gray scale image processed by an Alpha channel; setting a gamma coefficient γ and a compensation coefficient eps according to a target image contrast requirement; applying a gamma transformation formula to each pixel gray scale value g as f1(g)=f1(g)= ; wherein g is the gray scale value of the initial image pixel, f1(g) is the gray scale value of the image after gray scale transformation; and outputting the gamma-enhanced gray scale image for subsequent edge extraction or dispersity analysis.

[0018] Specifically, in order to enhance the contrast between CNTs and the resin background, a gamma gray scale transformation method is adopted; the transformation formula is to perform a multiplication operation on each pixel value on the original image, and in the calculation process, the influence of the offset is usually considered, so the expression is usually written as formula (1). Wherein g is the gray scale value of the initial image pixel, f1(g) is the gray scale value of the image after gray scale transformation; eps represents the compensation coefficient, which is automatically set according to the expected gray scale range; γ represents the Gamma coefficient, which is used to adjust the gray scale mapping relationship; f1(g)= .

[0019] When γ<1, the dark details can be enhanced, and when γ>1, the bright region can be suppressed. Through gamma enhancement processing, the identifiable of CNTs distribution characteristics in the image can be effectively improved, and the accuracy and robustness of dispersity analysis can be improved. For example, input the gray scale image obtained in step 100 into an image processing software (such as MATLAB or Python OpenCV library), apply a gamma transformation formula to each pixel gray scale value g, f1(g)= , set γ as 5, achieve nonlinear enhancement of image contrast. Output the enhanced gray scale image, which can obviously highlight the boundary between CNTs and background resin.

[0020] Input the enhanced image into an image derivation method based on a Sobel operator, calculate the convolution factor along the horizontal direction and the convolution factor along the vertical direction, and generate a gradient amplitude image to obtain a CNTs edge feature image, as shown in the accompanying Figure 2 .

[0021] Further, the application further comprises the following steps: setting a position pointer of a top-left pixel of the gray image as (0, 0), setting (i, j) as a pixel position of an i-th row and a j-th column in the image; For the pixel point (i, j), a 3*3 neighborhood matrix A(i, j) is selected with the pixel point (i, j) as the center, and a differential gray value along a horizontal direction at the pixel point (i, j) is calculated (i, j) and a differential gray value along a vertical direction at the pixel point (i, j) (i, j) is calculated, and the formula is as follows: = ×A(i, j); wherein A(i, j) is a 3*3 gray matrix with the pixel point (i, j) as the center; = ×A(i, j); Based on and , a gradient amplitude value is calculated to form a gradient amplitude value image, and the formula is as follows: ; wherein f2(i, j) is a gray value of a pixel at the pixel point (i, j) in the derived image, and gray values of all pixel points constitute the gradient amplitude value image; The gradient amplitude value image is output as a CNT edge feature image, which is used for subsequent binarization and dispersity analysis.

[0022] Specifically, after the image contrast is nonlinearly enhanced through a gamma coefficient, the image gray range is adjusted; then, an image derivation method based on a Sobel operator is applied to the image after gray transformation, as shown in FIG. 2. The method can be expressed as formula (2): Figure 3 ; Setting (i, j) as a pixel position of an i-th row and a j-th column in the image, setting a position pointer of a top-left pixel of the gray image as (0, 0), f2(i, j) is a gray value of a pixel at the pixel point (i, j) in the derived image, for the pixel point (i, j), a 3*3 neighborhood matrix A(i, j) is selected with the pixel point (i, j) as the center, and a differential gray value along a horizontal direction at the pixel point (i, j) is calculated (i, j) and a differential gray value along a vertical direction at the pixel point (i, j) (i, j), = ×A(i, j); wherein A(i, j) is a 3*3 gray matrix with the pixel point (i, j) as the center; ​ = ×A(i, j); wherein, and are the Sobel operator convolution factors along the horizontal and vertical directions, respectively, and the mathematical expression is: ; The mathematical expression of A(i, j) is: ; After pixel-by-pixel calculation, a complete gradient amplitude image is obtained. For example, the obtained gamma-enhanced gray image is input into an edge detection algorithm based on the Sobel operator, to calculate the convolution factor along the horizontal direction and the convolution factor along the vertical direction , to generate a gradient amplitude image. The output result is a CNTs edge feature image, in which the CNTs contour is clear and identifiable, and can be used for subsequent binarization processing. Compared with the original gray image, the CNTs edge contour of the gradient amplitude image is clearer, and can accurately reflect the distribution boundary of CNTs in the resin matrix composite material, providing a reliable data basis for subsequent quantitative evaluation of dispersity based on edge distribution density, edge connectivity and other indicators.

[0023] The edge feature image is converted into a binarized image by using the function im2bw, and an area threshold constraint condition is applied to the binarized image, for screening out small-area noise regions.

[0024] Further, the present application further comprises the following steps: normalizing the gradient amplitude image to a preset pixel value range to generate a normalized image ; determining a binarization threshold T from the normalized image, the binarization threshold T being a fixed pixel value; calling the function im2bw to convert the normalized gradient image into a binarized image f3(i, j) according to the binarization threshold T, and the formula is: ; The binarized image f3(i, j) is subjected to connected component labeling to obtain a connected component set, each connected component being a set of pixels that are connected to each other in the image; a pixel area threshold of the connected component is set, and each connected component set is screened, if the pixel area of the current connected component is less than the threshold, the current connected component is removed from the binarized image, and if the pixel area of the current connected component is greater than or equal to the threshold, the current connected component is retained as a candidate target region.

[0025] Specifically, the obtained binarized image is further subjected to connected component analysis and area thresholding to remove noise regions and extract effective edge regions related to the CNT distribution. First, the image calculated by Sobel is transformed to grayscale to facilitate the extraction of CNT pixels. Then, the grayscale image is converted to a binarized image using the formula for binarization, as shown in the appendix. Figure 4 , Figure 5 As shown, the function "im2bw" is used to convert the grayscale image to a binary image and fill the holes in the binary image.

[0026] The gradient magnitude image is normalized to a preset pixel value range to generate a normalized image. ; A binarization threshold T is determined based on the normalized image, wherein the binarization threshold T is a fixed pixel value; The function `im2bw` is called to convert the normalized gradient image into a binary image `f3(i, j)` based on the binarization threshold `T`. The formula is as follows: ; Connected component labeling is performed on the binarized image f3(i,j) to obtain a set of connected components, where each connected component is a set of pixels in the image that are connected to each other. Set the pixel area threshold t for connected components gray For each set of connected components, a filter is applied. If the pixel area of ​​the current connected component is less than a threshold, it is removed from the binarized image; if the pixel area of ​​the current connected component is greater than or equal to the threshold, it is retained as a candidate target region. A connected component is a set of pixels that are spatially adjacent to each other in the image and have the same pixel value. By labeling connected components, the binarized image can be divided into several independent connected regions, each representing a candidate structure in the image. For example, the edge image is input into the function im2bw, with a threshold... =0.2, converting the image into a binary image. An area threshold constraint is applied to the binary image, setting the minimum area to 50 pixels, to remove isolated small noise regions, resulting in a denoised binary image. The denoised binary image is then subjected to connected component identification and segmentation, marking interconnected regions with the same pixel value to form independent target regions. Each connected component represents a cluster or distribution region of CNTs, providing an accurate regional basis for micro-element analysis. Significant impurity noise signals are observed in the image, exhibiting high density and small-scale (sub-millimeter) spatial distribution. To suppress this type of noise interference, this study applies an area threshold constraint to the binary image based on connected component analysis theory. Connected component identification and extraction is a crucial step in this image processing, used to identify and segment connected regions in the image.

[0027] The target region is analyzed for resolution. Let the horizontal resolution of the target region be L and the vertical resolution be W. The target region is divided into m×n rectangular image elements according to the common divisor of the horizontal and vertical resolutions.

[0028] Furthermore, this application also includes the following steps: performing resolution analysis on the target region, assuming the horizontal pixel resolution of the image is L and the vertical pixel resolution is W; dividing the image into several rectangular micro-regions according to the resolution, wherein the number of horizontal pixels in each micro-region is denoted as . The vertical pixel count is denoted as h. h is preferably chosen as the common divisor of L and W; the image is divided into m×n infinitesimal regions, where m= n= For each rectangular micro-region, set the coordinates of its top-left vertex in the image as follows: Set the column index of the micro-element region to , For column index, The row index is used; the x and y coordinates of the top left vertex of the image element are... ,as follows: Based on the length and width of the set micro-element Cropping is performed using the horizontal and vertical coordinates, resulting in a number of micro-elements per image. The total number of columns and rows of the segmented micro-elements in the total field of view of the image are respectively set as follows: and The formula for calculating is: = × Each micro-element region is trimmed according to the set length and width, resulting in a total number of micro-elements of m×n.

[0029] Specifically, this invention proposes a second part of quantitative characterization—standard deviation analysis. This analysis still includes two parts—"region segmentation" and "numerical calculation". The first step, region segmentation, involves segmenting the image obtained from the "CNTs feature extraction" part into regions. The number of pixels in the horizontal and vertical directions of a unit region, i.e., the length and width of the image micro-element, is usually taken as the common divisor of the number of pixels in the horizontal and vertical directions of the total resolution.

[0030] First, set the length and width of the micro-regions according to the image resolution (let the image length be L and the width be W). Then, set the coordinates of the top-left vertex of each rectangular image micro-region in the image to , and set the column index and row index of the image micro-region to and , respectively. Perform resolution analysis on the target region, assuming the horizontal pixel resolution of the image is L and the vertical pixel resolution is W. Based on the resolution, divide the image into several rectangular micro-regions, and denot the horizontal pixel count of each micro-region as . The vertical pixel count is denoted as h. h is preferably chosen as the common divisor of L and W; the image is divided into m×n infinitesimal regions, where m= n= For each rectangular micro-region, set the coordinates of its top-left vertex in the image as follows: Set the column index of the micro-element region to , For column index, The row index is used; the x and y coordinates of the top left vertex of the image element are... ,as follows: Based on the length and width of the set micro-element The image is cropped using its x and y coordinates, resulting in a number of micro-elements per image. The total number of columns and the total number of rows of the segmented micro-elements in the total field of view of the image are set respectively. and The formula for calculating is: = × Each micro-element region is cropped according to the set length and width, resulting in a total of m×n micro-elements. Each rectangular image micro-element consists of a group of adjacent pixel units, which maintain a consistent resolution division accuracy in both the horizontal and vertical directions, as shown in the attached figure. Figure 6 As shown. For example, resolution analysis is performed on each target region. Assuming a horizontal resolution of 1000 pixels and a vertical resolution of 1000 pixels, the target region is divided into 10 × 10 = 100 rectangular image elements according to the common divisor of the horizontal and vertical resolutions. Each element is 100 × 100 pixels in size. This division method ensures that the elements have regular shapes and uniform sizes, thus facilitating subsequent region feature extraction and image analysis processing.

[0031] Calculate the area ratio of CNTs within each image micro-element, and flatten all micro-element area ratio matrices into a vector Q. Calculate the variance and standard deviation to quantify the uniformity of CNT dispersion in the two-dimensional resin plane.

[0032] Furthermore, this application also includes the following steps: for each image micro-element (p, q), calculate the area ratio of the black pixel block; ×100%; of which, The percentage of the area of ​​black pixels, i.e., the first... Liede The proportion of black pixels in a row image micro-element to the total number of pixels in the micro-element; For the first Liede The total number of black pixel blocks in a row image micro-element h is the width of the image micro-element, i.e., the number of horizontal pixels in each micro-element region, and h is the height of the image micro-element, i.e., the number of vertical pixels in each micro-element region. Liede The proportion of black pixels in a row image micro-element to the total number of pixels in the micro-element; For the first Liede The total number of black pixel blocks in a row image micro-element is the width of the image micro-element, i.e., the number of pixels horizontally in each micro-element region; h is the height of the image micro-element, i.e., the number of pixels vertically in each micro-element region; the area proportions of all micro-elements are flattened into a vector Q in row or column order for subsequent statistical analysis; the average area proportion is calculated. : Where μ is the average percentage of black pixels in the image micro-element. The total number of columns of the image elements. The total number of rows of image elements; calculate the dispersibility index of CNTs in the resin-based composite system. : ;in, The dispersion index of CNTs in resin-based composite systems; based on The size of CNTs is used to assess their dispersibility in the matrix. The closer the value is to 0, the more uniform the distribution of CNTs in the matrix.

[0033] Specifically, numerical calculation: quantifying the dispersion of CNTs in the resin matrix is ​​an innovative aspect of this study and a new research method in the process of dispersion analysis and research of composites of reinforcing phase / matrix type; using equation (9) to calculate the area ratio of CNTs in each divided micro-element region, and calculating the variance of the obtained data. In the process of standard deviation calculation, since the quantification target is the uniformity of CNTs dispersion in the "two-dimensional resin plane", it is necessary to first flatten the obtained CNTs pixel block ratio matrix into a vector and then calculate the variance and standard deviation. For each image micro-element (p, q), calculate the area ratio of black pixel blocks. ×100%; of which, The percentage of the area of ​​black pixels, i.e., the first... Liede The proportion of black pixels in a row image micro-element to the total number of pixels in the micro-element; For the first Liede The total number of black pixel blocks in a row image micro-element h is the width of the image micro-element, that is, the number of horizontal pixels in each micro-element region, and h is the height of the image micro-element, that is, the number of vertical pixels in each micro-element region.

[0034] The dispersion of CNTs in the resin matrix composite system is characterized by the standard deviation formula, and the area proportion of all micro elements is expanded into a vector Q in row or column order for subsequent statistical analysis; the average value of the area proportion is calculated : ; Wherein, μ is the average value of the proportion of black pixel blocks in the image micro element, is the total number of columns of the image micro element, is the total number of rows of the image micro element; the dispersion index of CNTs in the resin matrix composite system is calculated : ; Wherein, is the dispersion index of CNTs in the resin matrix composite system; According to the size of , the dispersion of CNTs in the matrix is evaluated, and when is closer to 0, it indicates that the distribution of CNTs in the matrix is more uniform. For example, the area ratio of CNTs in each micro element to the total pixels of the micro element is calculated, for example, the CNTs pixels of a certain micro element account for 1100 / 10000, and the total proportion is 0.11. The area proportion of 100 micro elements is expanded into a vector, and the variance and standard deviation of the vector are calculated. For example, the variance is 0.0004 and the standard deviation is 0.02, which is used to quantify the uniformity of the dispersion of CNTs in the two-dimensional plane of the composite film. The smaller the variance / standard deviation, the more uniform the distribution of CNTs.

[0035] In summary, the quantitative method for characterizing the dispersion of CNTs in the resin matrix composite system provided by the present application has the following beneficial effects: By collecting the TIF image containing the Alpha channel, removing the Alpha channel, and converting it to a grayscale image, the integrity and processability of the original image information are ensured, providing a high-quality data foundation for subsequent processing. Nonlinear enhancement of the grayscale image using gamma grayscale transformation effectively improves the contrast between CNTs and the background, making the edge and micro-distribution features clearer. Using gradient amplitude calculation based on the Sobel operator to extract CNT edge features can accurately identify the outline of CNTs in the two-dimensional plane, improving the accuracy of edge detection. Binaryzation of the edge image and application of area threshold constraints effectively eliminate small noise areas, improve the signal-to-noise ratio, and ensure that subsequent analysis is only for valid CNT areas. By using connected component recognition and segmentation methods, the regions in the binary image that are connected to each other and have the same pixel value are clearly segmented, forming independent target areas and providing accurate area information for quantitative analysis. Based on the resolution of the target area, the elements are divided, and the rectangular image elements are generated according to the common divisor rule of horizontal and vertical resolution, making the CNT area ratio statistics more uniform and repeatable. By calculating the CNT area ratio of each element and flattening it into a vector, and then calculating the variance and standard deviation, the quantitative evaluation of the uniformity of CNT dispersion is realized, overcoming the subjectivity and low accuracy of traditional manual observation methods. The method of the present application can objectively and quantitatively evaluate the dispersion of CNTs in the resin-based composite material system while ensuring the accuracy of image processing, improving the accuracy and repeatability of the analysis, and providing reliable data support for composite material performance evaluation and process optimization.

[0036] Example Two The present application provides a quantitative method for characterizing the dispersion of CNTs in a resin-based composite material system. The mechanical grinding action of a planetary ball mill is used to reduce the length of carboxylated CNTs, thereby improving the dispersion of CNTs in PPESK. The method includes the following steps: The planetary ball mill used in this experiment is an oil-sealed silent vertical planetary ball mill produced by Guangzhou Guirui Company, model YXQM-2L, with a maximum rotational speed of 800 rpm / min. The grinding tank is a corundum grinding tank with a volume of 250 ml from the same manufacturer as the ball mill. The grinding balls are zirconia balls with sizes of 3 mm, 5 mm, and 8 mm, with a weight ratio of 5:3:2. The carbon nanotubes used in this experiment are carboxylated CNTs produced by Zhongke Nanometer Nanometer, model TNMC3, with lengths of 10-30 microns. The NMP dispersant is a white powder dispersant produced by Zhongke Times Nanometer, model TNNDIS, which is suitable for dispersing carbon nanotubes in NMP (N-methyl pyrrolidone) solvent.

[0037] Process flow: 1. The ball milling method of wet ball milling is adopted; firstly, the CNTs / NMP dispersant / NMP is configured as 2 g:4 g:80 ml of CNTs / NMP dispersion liquid: firstly, 2 g of CNTs, 4 g of NMP dispersant and 80 ml of NMP solvent are weighed and added into a beaker in turn, and then the dispersion system is subjected to 5 min of ultrasonic pre-dispersion treatment, so that the CNTs are uniformly dispersed in the NMP. And it is added into a 250 ml ball mill tank to serve as a ball milling medium (the dispersion system after filling the CNTs / NMP solution accounts for about one third of the total volume of a single mill tank). The ball milling process program is set as a ball milling cycle sequence of forward rotation 1 h+intermittent 30 min+reverse rotation 1 h+intermittent 30 min, the public rotation speed is 500 rpm / min, and 11 groups of ball milling experiments are carried out for 2 h, 4 h, 6 h, 8 h, 10 h, 12 h, 14 h, 16 h, 18 h and 20 h of ball milling time respectively; and the ball milling parameters (such as ball milling speed, ball size and proportion) of each group of parallel experiments are consistent except for the ball milling time. Finally, the CNTs with typical length are obtained through the above ball milling process. And the CNTs (5wt%) with different lengths and distributions obtained by the above horizontal experiment are laid with PPESK to form a composite film with a thickness of 1 mm, and a high-magnification 3D optical microscope is used to observe the distribution state of the CNTs in the film.

[0038] 2. The quantitative characterization of the dispersion of CNTs is carried out by using the characterization means provided in the application, and the effect is shown in the attached Figure 7 According to the attached Figure 7 , those skilled in the art can see that the distribution characteristics of CNTs obtained by optical imaging technology can only be roughly judged by qualitative analysis method to improve the dispersion of CNTs (in PPESK resin) in example two and the influence of ball milling time on the dispersion of CNTs. Therefore, it can be seen that only by using simple electron or optical imaging to analyze the dispersion of CNTs in the resin-based composite material system or to guide the update iteration of the "dispersion improvement process" is lack of scientificity and objectivity. Figure 7 (a)-(k) are the distribution images of CNTs in the film after ball milling (the numerical value in the lower left corner of each figure is the ball milling time), Figure 7 (a) represents the distribution image of CNTs in the film without ball milling treatment, Figure 7 (b) represents the distribution image of CNTs in the film after ball milling treatment for 2 hours, Figure 7 (c) represents the distribution image of CNTs in the film after ball milling treatment for 4 hours, Figure 7 (d) represents the distribution image of CNTs in the film after ball milling treatment for 6 hours, Figure 7 (e) represents the distribution image of CNTs in the film after ball milling treatment for 8 hours, Figure 7(f) represents the distribution image of CNTs in the thin film after ball milling for 10 hours, Figure 7 (g) represents the distribution image of CNTs in the thin film after ball milling for 12 hours, Figure 7 (h) represents the distribution image of CNTs in the thin film after ball milling for 14 hours, Figure 7 (i) represents the distribution image of CNTs in the thin film after ball milling for 16 hours, Figure 7 (j) represents the distribution image of CNTs in the thin film after ball milling for 18 hours, Figure 7 (k) represents the distribution image of CNTs in the thin film after ball milling for 20 hours. The above images are obtained by RH-2000 3D digital optical microscope produced by Haoshi Corporation: initial resolution is 1920x1200, magnification is 1000.

[0039] 3. The attached Figure 8 After the image is subjected to CNTs two-dimensional feature extraction and standard deviation analysis, the CNTs in the field of view are more obvious than the resin background, and the effect is shown in the attached Figure 8 figure, and the non-CNTs "impurities" interfering with "feature extraction and standard deviation analysis" are removed, and the standard deviation data results given in the lower left corner of the image can be used to accurately quantitatively analyze the uniformity of CNTs distribution in each field of view and the span difference of the uniformity of CNTs distribution in the same ball milling time span, which is beneficial to the optimization of ball milling process parameters and the selection of parameters for other dispersion improvement methods similar to "ball milling". Figure 8 (a)-(k) are the dispersion characterization results of CNTs in the thin film (the numerical value in the lower left corner of each figure is the distribution standard deviation obtained by the image processing program), Figure 7 (a) is the dispersion characterization result of the distribution image shown in (a), Figure 8 (a) is the dispersion characterization result of the distribution image shown in (a), Figure 7 (b) is the dispersion characterization result of the distribution image shown in (b), Figure 8 (b) is the dispersion characterization result of the distribution image shown in (b), Figure 7 (c) is the dispersion characterization result of the distribution image shown in (c), Figure 8 (c) is the dispersion characterization result of the distribution image shown in (c), Figure 7 (d) is the dispersion characterization result of the distribution image shown in (d), Figure 8 (d) is the dispersion characterization result of the distribution image shown in (d), Figure 7 (e) is the dispersion characterization result of the distribution image shown in (e), Figure 8 (e) is the dispersion characterization result of the distribution image shown in (e), Figure 7 (f) is the dispersion characterization result of the distribution image shown in (f), Figure 8 (f) is the dispersion characterization result of the distribution image shown in (f), Figure 7 (g) is the dispersion characterization result of the distribution image shown in (g), Figure 8 (g) is the dispersion characterization result of the distribution image shown in (g), Figure 7 (h) is the dispersion characterization result of the distribution image shown in (h), Figure 8(h) the dispersion of the distribution image shown in (i) is characterized by Figure 7 (i) the dispersion of the distribution image shown in (i) is characterized by Figure 8 (i) the dispersion of the distribution image shown in (i) is characterized by Figure 7 (j) the dispersion of the distribution image shown in (j) is characterized by Figure 8 (j) the dispersion of the distribution image shown in (j) is characterized by Figure 7 (k) the dispersion of the distribution image shown in (k) is characterized by Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure 8 Figure 7 Figure (k) the dispersion of the distribution image shown in (k) is characterized by

[0040] 4、In summary, the present application can realize the accurate quantitative characterization of the dispersion of different contents and different types of CNTs in resin matrix composite system through "feature extraction of CNTs in two-dimensional plane" + "standard deviation analysis".

[0041] Obviously, for those skilled in the art, without departing from the principles of the present application, some improvements and modifications can be made to the present application, and these improvements and modifications also fall within the protection scope of the present application.

Claims

1. A quantitative method for characterizing the dispersion of CNTs in a resin-based composite system, characterized in that, The method comprises the following steps: Collecting images from a high-resolution imaging device to obtain a two-dimensional distribution image of CNTs in a resin-based composite material, the two-dimensional distribution image being a TIF image containing an Alpha channel, and deleting the Alpha channel to convert the image into a grayscale image; Applying a gamma grayscale transformation to the grayscale image to achieve nonlinear enhancement of the image contrast; The enhanced image is input into an image derivation method based on Sobel operator to calculate a convolution factor along a horizontal direction and a convolution factor along a vertical direction and generate a gradient amplitude image to obtain a CNTs edge feature image; Converting the edge feature image into a binary image using the function im2bw, and applying an area threshold constraint to the binary image to filter out small-area noise regions; Performing connected component recognition and segmentation on the binary image that meets the area threshold condition to identify target regions formed by a collection of pixels with the same grayscale value and mutual connectivity; Performing resolution analysis on the target regions, setting the horizontal resolution of the target region as L and the vertical resolution as W, and dividing the target region into m×n rectangular image elements according to the greatest common divisor of the horizontal and vertical resolutions; Calculating the area proportion of CNTs in each image element, and flattening the area proportion matrix of all elements into a vector Q to calculate the variance and standard deviation for quantifying the dispersion uniformity of CNTs in the two-dimensional resin plane.

2. A quantitative method of characterizing the dispersion of CNTs in a resin-based composite system according to claim 1, wherein, Applying a gamma grayscale transformation to the grayscale image to achieve nonlinear enhancement of the image contrast, comprising: Obtaining a grayscale image after Alpha channel processing; According to the contrast requirement of the target image, set the gamma coefficient and the compensation coefficient eps Applying a gamma transformation formula to each pixel grayscale value g as: f1(g) = g ; Wherein g is the grayscale value of the initial image pixel, f1(g) is the grayscale value of the image after grayscale transformation; Outputting the gamma-enhanced grayscale image for subsequent edge extraction or dispersion analysis.

3. A quantitative method of characterizing the dispersion of CNTs in a resin-based composite system as claimed in claim 1, wherein, The method for deriving the image based on the Sobel operator, calculating the convolution factor Gx along the horizontal direction and the convolution factor Gy along the vertical direction, and generating a gradient amplitude image to obtain a CNTs edge feature image, comprising: Setting the position pointer of the top-left corner pixel of the grayscale image as (0, 0), and setting (i, j) as the pixel position of the i-th row and j-th column in the image; For the pixel point (i, j), a 3x3 neighborhood matrix A(i, j) is selected with (i, j) as the center, and the differential gray value along the horizontal direction at (i, j) is calculated (i, j) and the differential gray value along the vertical direction at (i, j) is calculated (i, j), and the formula is: = xA(i, j); Wherein A(i, j) is a 3×3 grayscale matrix centered at pixel point (i, j); = ×A(i, j); based on with the gradient magnitude is calculated, forming a gradient magnitude image, with the formula: ； Wherein f2(i, j) is the grayscale value of the pixel at (i, j) in the derived image, and the grayscale values of all pixel points constitute a gradient amplitude image; Outputting the gradient amplitude image as a CNTs edge feature image for subsequent binarization and dispersion analysis.

4. The method of quantifying the dispersion of CNTs in a resin-based composite system according to claim 1, wherein, Converting the edge feature image into a binary image using the function im2bw, and applying an area threshold constraint to the binary image to filter out small-area noise regions, comprising: normalizing the gradient magnitude image into a preset pixel value range to generate a normalized image ; Determining a binary threshold T from the normalized image, the binary threshold T being a fixed pixel value; Calling the function im2bw to convert the normalized gradient image into a binary image f3(i, j) according to the binary threshold T, the formula being: ； Performing connected component labeling on the binary image f3(i, j) to obtain a connected component set, each connected component being a collection of mutually connected pixels in the image; Set the pixel area threshold of the connected domain, filter each connected domain set, if the pixel area of the current connected domain is less than the threshold, remove it from the binary image, if the pixel area of the current connected domain is greater than or equal to the threshold, keep it as a candidate target region.

5. The method of quantifying the dispersion of CNTs in a resin-based composite system according to claim 1, wherein, Resolution analysis is performed on the target region, the horizontal resolution of the target region is L, the vertical resolution is W, and the target region is divided into m×n rectangular image units according to the greatest common divisor of the horizontal and vertical resolutions, including: Resolution analysis is performed on the target region, the horizontal pixel resolution of the image is L, and the vertical pixel resolution is W; According to the resolution, the image is divided into several rectangular microelement regions, the number of horizontal pixels of each microelement region is denoted as , and the number of vertical pixels is denoted as h, the and h are preferably the common divisor of L and W; The image is divided into m x n microelement regions, where m = 2k, n = 2k, and k is an integer. , ; For each rectangular microelement region, set the coordinates of its upper left vertex in the image as Set the column number index of the microelement region as, For the column number index, For the row number index; The horizontal and vertical coordinates of the top-left vertex of the image pixel are As follows: ； According to the length and width of the microelement set The number of microelements in each image is obtained by clipping the horizontal and vertical coordinates of The total number of columns and rows of the segmented microelements in the total field of view of the image is set to and The calculation formula is: = × ； Crop each microelement region with the set length and width to obtain a total microelement number of m×n.

6. A quantitative method of characterizing the dispersion of CNTs in a resin-based composite system according to claim 1, wherein, Calculate the area proportion of CNTs in each image microelement, and flatten all microelement area proportion matrices into a vector Q to calculate the variance and standard deviation, which are used to quantify the dispersion uniformity of CNTs in the two-dimensional resin plane, including: For each image microelement (p, q), calculate the black pixel block area proportion; ×100％； in, The percentage of the area of ​​black pixels, i.e., the first... Liede The proportion of black pixels in a row image micro-element to the total number of pixels in the micro-element; For the first Liede The total number of black pixel blocks in a row image micro-element h is the width of the image micro-element, that is, the number of horizontal pixels in each micro-element region; h is the height of the image micro-element, that is, the number of vertical pixels in each micro-element region. Flatten the area proportions of all microelements into a vector Q in row or column order for subsequent statistical analysis; Average of the area ratio : ； wherein μ is an average value of the proportion of black pixel blocks in the image elements, is the total number of columns of the image elements, is the total number of rows of the image elements; Calculating the dispersibility index of CNTs in resin-based composite systems : ； wherein, is the dispersibility index of the CNTs in the resin-based composite system; According to the size of the CNTs in the matrix, when the value is closer to 0, it indicates that the distribution of the CNTs in the matrix is more uniform.

Images