Method, system, and program for calculating silica dispersion evaluation value

Abstract

To provide a calculation method of a dispersion evaluation value of silica, a system and a program which can improve accuracy of an evaluation value by reducing a harmful effect given by a substance other than silica.SOLUTION: A calculation method of a dispersion evaluation value of silica specifies a first diameter of a circle that the best matches a plurality of pixel blocks having relatively higher strength than that of the periphery in an SEM image being a grayscale image obtained by imaging a rubber surface, removes the pixel block smaller than a circle of the first diameter and the pixel block equal to or larger than a circle of the second diameter larger than the first diameter in the plurality of pixel blocks having the relatively higher strength than that of the periphery in the SEM image, generates a first image obtained by converting the pixel block included in the circle being equal to or greater than the first diameter and less than the second diameter into a round shape, binarizes the first image, and calculates a dispersion evaluation value of silica on the basis of the binarized first image.SELECTED DRAWING: Figure 13

Description

[Technical Field]

[0001] The present disclosure relates to a method, system, and program for calculating a dispersion evaluation value of silica. [Background technology]

[0002] To evaluate the dispersion of silica in rubber, human eyes are used to evaluate the dispersion based on scanning electron microscope (SEM) images, taking advantage of the fact that silica appears strong (white) in the images. However, manual evaluation is subject to variations in experience and individual differences, so a quantitative evaluation method is desired.

[0003] Paragraph 0056 of Patent Document 1 describes that the average aggregate area of ​​vulcanized rubber is determined. Patent Document 1 also describes that the SEM image is binarized using a threshold determined by the Otsu method, and the aggregate area of ​​the silica portion is calculated based on the binarized image. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 5894583 Summary of the Invention [Problem to be solved by the invention]

[0005] However, SEM images of rubber surfaces show different intensity variations depending on the imaging conditions, even when the object being imaged is the same rubber, and it is thought that the accuracy of the evaluation value will be impaired unless appropriate binarization is performed. The Otsu method described in Patent Document 1 can determine a relatively appropriate threshold value for binarization if the image has two peaks in its intensity histogram, but SEM images of rubber surfaces do not have clear intensity peaks, so it cannot be said that the threshold value determined simply using the Otsu method is appropriate.

[0006] Furthermore, since SEM images contain substances other than silica that appear white, there is a risk that substances other than silica will be converted into white pixels through binarization and evaluated as silica, so it is preferable to address the adverse effects of substances other than silica.

[0007] The present disclosure provides a method, system, and program for calculating a dispersion evaluation value of silica that can reduce the adverse effects of substances other than silica and improve the accuracy of the evaluation value. [Means for solving the problem]

[0008] The method for calculating a silica dispersion evaluation value of the present disclosure is a method executed by one or more processors, which identifies a first diameter of a circle that best matches a plurality of pixel clusters that have a relatively higher intensity than their surroundings in an SEM image, which is a grayscale image of a rubber surface; removes pixel clusters that are smaller than a circle of the first diameter and pixel clusters that are equal to or larger than a circle of a second diameter larger than the first diameter from among the plurality of pixel clusters that have a relatively higher intensity than their surroundings in the SEM image; generates a first image in which pixel clusters included in circles that are equal to or larger than the first diameter and smaller than the second diameter are converted into a circular shape; binarizes the first image; and calculates a silica dispersion evaluation value based on the binarized first image. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a block diagram showing a system according to an embodiment of the present invention. [Figure 2] 1 is a flowchart showing a process executed by the system. [Figure 3] Figure showing an SEM image of a rubber surface. [Figure 4] FIG. 4 is a graph showing an intensity histogram of the image main body portion shown in FIG. 3. [Figure 5] FIG. 10 shows an intensity histogram of an SEM image after contrast processing. [Figure 6] FIG. [Figure 7] FIG. [Figure 8]FIG. 10 is an explanatory diagram regarding opening processing in morphological processing. [Figure 9] FIG. 10 is a diagram showing the results of an opening process using a disk element of radius (R) and the results of an opening process using a disk element of radius (R+1). [Figure 10] This is a diagram showing an image obtained by performing opening processing on the SEM image shown in Figure 3 using a disk element with a radius of R = 7. [Figure 11] Graph showing the intensity integration value of each image. [Figure 12] 10 is a graph showing differences in intensity integrated values. [Figure 13] FIG. 10 is an explanatory diagram relating to a process of generating a first image from a second image and a third image. [Figure 14] FIG. 4 is a diagram showing a binarized first image. [Figure 15] FIG. 10 is a diagram showing second regions set so as not to overlap each other. [Figure 16] FIG. 10 is a diagram showing the process of setting second regions that are allowed to partially overlap with each other. [Figure 17] A figure showing a histogram of the number of white pixels in the second regions in an example where the second regions shown in Figure 15 do not overlap each other, using a solid line, and a histogram of the number of white pixels in the second regions in an example where the second regions shown in Figure 16 overlap each other, using a dotted line. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, an embodiment of the present disclosure will be described with reference to the drawings.

[0011] [system] The system 1 of this embodiment calculates the silica dispersion evaluation value from an SEM image of a rubber surface. The SEM image is a grayscale image captured with an electron microscope. In a grayscale image, each pixel has an intensity, and the intensity represents shades of black. The intensity gradation is, for example, 256 for 8 bits and 65536 for 16 bits. Low intensity is black, and high intensity is white. Single silica particles or particle aggregates appear as relatively white pixels in the SEM image.

[0012] As shown in FIG. 1, the system 1 includes an SEM image acquisition unit 10, a first diameter determination unit 11, a second diameter determination unit 12, a first image generation unit 13, a binarization execution unit 14, a second region setting unit 15, an evaluation value calculation unit 16, an enlargement processing unit 17, and a contrast processing unit 18. These units 10-18 are implemented by software and hardware working together as the processor 1a executes a processing routine shown in FIG. 2, which is pre-stored in a computer equipped with a processor 1a, a memory 1b, various interfaces, etc. In this embodiment, each unit is implemented by the processor 1a in a single device, but this is not limited to this. For example, the processing of each unit may be distributed over a network, with multiple processors executing the processing. That is, one or more processors execute the processing. The memory 1b stores the SEM image, the first image, the second image, the third image, a binarized image of the first image, position information of the second region (coordinates, identifier of the second region), the number of white pixels in the second region, and a silica dispersion evaluation value.

[0013] The SEM image acquisition unit 10 acquires a grayscale SEM image of the rubber surface and stores it in memory 1b. Figure 3 shows an SEM image of the rubber surface, a grayscale image in which intensity is expressed in 16-bit gradations. The imaging conditions are listed at the bottom of Figure 3, but the determination of the first diameter, generation of the first image, binarization, and calculation of the silica dispersion evaluation value are performed on the main image excluding the lower part. Hereinafter, SEM images other than those in Figure 3 or images obtained by binarizing SEM images refer to images corresponding to the main image excluding the lower part of Figure 3. Figure 4 shows an intensity histogram of the main image portion shown in Figure 3. The horizontal axis of Figure 4 represents intensity, and the vertical axis represents the number of pixels. The left end of the horizontal axis of Figure 4 represents 0 (black), and the right end of the horizontal axis represents 65535 (white). As shown in Figure 4, the intensity distribution of the SEM image of the rubber surface is uniform and has no clear peak, making it difficult to properly identify silica pixels using a single appropriate threshold value that applies to the entire image. Therefore, image processing before binarization involves removing pixels that appear to be other than silica.

[0014] The enlargement processing unit 17 executes a process to enlarge the size of the SEM image. In this embodiment, the enlargement processing unit 17 enlarges an SEM image of 1025 × 1536 pixels to 3072 × 4608 pixels, which is approximately three times the size. In the process described below, pixel blocks consisting of multiple pixels are assumed to represent silica, and processing is performed on the pixel blocks. Therefore, silica that was previously represented by a single pixel is now represented by a pixel block through the enlargement process, and the amount of silica to be processed increases, making it easier to capture silica. The enlargement processing unit 17 can be omitted. The enlargement ratio is preferably 2x or more and 4x or less.

[0015] The contrast processing unit 18 performs contrast processing on the SEM image enlarged by the enlargement processing unit 17. The contrast processing is processing to flatten the intensity histogram (histogram flattening). FIG. 5 shows the intensity histogram of the SEM image after the contrast processing. As shown in FIG. 5, the intensity distribution is flat (uniform). If the intensity distribution (histogram) is flat, the intensity change in the edge portion becomes uniform throughout the entire image, improving the stability of the morphology processing (identification of the first diameter and replacement with a circle) described later. In this embodiment, the contrast processing is performed on the SEM image enlarged by the enlargement processing unit 17, but this is not limiting. The enlargement processing unit 17 may also perform enlargement processing on the SEM image on which the contrast processing unit 18 has performed contrast processing.

[0016] The first diameter determination unit 11 determines the first diameter of the circle that best matches a plurality of pixel clusters in the SEM image that have a relatively higher intensity than the surrounding areas. Silica is dispersed in the SEM image, and the silica particles are circular. Therefore, the circle that best matches the white pixel cluster in the SEM image is likely to be silica. Here, the first diameter of the circle is determined to be the diameter of the silica particle. In this specification, "diameter" is used for image processing, so it is preferable to express it in terms of the number of pixels. In this specification, "diameter" means radius unless otherwise specified.

[0017] Specifically, to identify the first radius of the circle, an opening process, a morphological process in which a circle is used as the structuring element, is used. The opening process involves performing an erosion process in which a circle of radius R is used as the structuring element, followed by an expansion process in which a circle of the same radius R is used as the structuring element. In this specification, a circular structuring element may be referred to as a "disk element."

[0018] The erosion process will be briefly explained using FIG. 6. FIG. 6 is an explanatory diagram of the erosion process. As shown in FIG. 6, assume that the original image has a rectangular cluster of white pixels on a black background. As shown in the figure, the following operation is performed while moving a disk element (a circular structural element) one pixel at a time relative to the original image. Each time the disk element is moved, the operation of outputting the pixel located at the center of the disk element as the lowest intensity value within the disk element to the corresponding pixel (at the same position) in the new image is repeated. If the original image has N pixels, the operation of outputting the pixel located at the center of the disk element as the lowest intensity value within the disk element to the corresponding pixel (at the same position) in the new image is executed N times. Because only the original image and not the new image are referenced within the disk element, the pixels of the new image that have been output do not affect the erosion process. As a result, pixels with relatively high intensity in the original image are eroded into a disk shape. By performing morphological contraction processing using a circle of radius R as the structuring element, pixel clusters and individual pixels smaller than radius R can be removed, making it possible to use this for noise reduction. In the example in Figure 6, a rectangular cluster of white pixels is changed to a cluster of white pixels that has been contracted to a size according to the set structuring element.

[0019] The expansion process will be briefly explained using FIG. 7. FIG. 7 is an explanatory diagram of the expansion process. As shown in FIG. 7, assume that the original image has a rectangular block of white pixels on a black background. As shown in the figure, the following operation is performed while moving a disk element (a circular structural element) one pixel at a time relative to the original image. Each time the disk element is moved, the pixel located at the center of the disk element is output as the highest intensity value within the disk element to the corresponding pixel (at the same position) in the new image, and this operation is repeated. If the original image has N pixels, this operation of outputting the pixel located at the center of the disk element as the highest intensity value within the disk element to the corresponding pixel (at the same position) in the new image is executed N times. Because only the original image and not the new image are referenced within the disk element, the pixels of the new image that have been output do not affect the expansion process. As a result, pixels with relatively high intensity in the original image are expanded into a circular shape. By performing the expansion process, even if a pixel block has an irregular shape that is not circular, it can be replaced with a pixel block that is close to a circle of radius R, as long as the pixel block is approximately the same size as a circle of radius R.

[0020] The opening process will be described using the example of FIG. 8. FIG. 8 is an explanatory diagram of the opening process of morphological processing. As shown in FIG. 8, the original image is assumed to contain white pixel clusters of various shapes and sizes on a black pixel base. As shown in FIG. 8, when a contraction process is performed on the original image using a disk element of a predetermined radius R, white pixel clusters smaller than the disk element of the predetermined radius R are removed. At the same time, white pixel clusters larger than the disk element of the predetermined radius R are reduced in size. Furthermore, as shown in FIG. 8, when a dilation process is performed on the reduced image using a disk element of a predetermined radius R, the white pixel clusters that have been replaced with circular shapes or shapes with rounded corners are expanded and returned to their original size.

[0021] FIG. 9 shows the results of opening processing using a disk element of diameter (R) and the results of opening processing using a disk element of diameter (R+1). As shown in FIG. 9, the original image contains a black base and white pixel clusters of various shapes and sizes. The original image contains white pixels corresponding to silica in a circle of diameter R, white pixels (noise) smaller than the circle of diameter R, and white pixels larger than diameter R (substances other than silica; noise). In reality, the diameter R of the white pixels of a circle or a shape similar to a circle corresponding to silica is unknown, and the first diameter determination unit 11 determines a representative value of diameter R using the following algorithm.

[0022] When an opening process is performed on the original image using a disk element of radius R, as shown in the image in the middle of Figure 9, pixel blocks smaller than a circle of radius R are removed from the processed image, and circular pixel blocks of radius R (shown as circles in the figure) and pixel blocks larger than radius R (shown as polygons in the figure) remain. When opening processing is performed on the original image using a disk element of diameter (R+1), pixel clusters smaller than a circle of diameter (R+1) (both pixel clusters with a diameter less than R and pixel clusters with a circular diameter of R) are removed from the processed image, and non-silica pixel clusters larger than diameter R (pixel clusters with a diameter of R+1 or more) remain, as shown in the bottom part of Figure 9. R+1 means that 1 pixel has been added to R (pixels).

[0023] Here, the representative value of the silica diameter is R T Assume that the radius R=R T The image (middle image in Figure 9) after opening processing with the disk element of R contains pixel blocks (circles in the figure) corresponding to silica and noise such as large substances other than silica (shown as polygons in the figure). T The image opened with a +1 disk element (the bottom image in Figure 9) contains noise (shown as polygons in the figure) from large substances other than silica. The difference in the intensity integrated values ​​between the two images corresponds to the pixel block that corresponds to silica, and the difference in the intensity integrated values ​​is the smallest. On the other hand, the diameter R = 3 (3 <R T) in the image that has been opened with the disk element, as in the original image in Figure 9, there are pixel clusters that correspond to silica (circular in the figure), noise such as large substances other than silica (shown as polygons in the figure), and pixels larger than a circle with a radius of R = 3 but smaller than R T The image obtained by opening with a disk element of radius R=4 contains pixel clusters that correspond to noise smaller than a circle. The same applies to the image obtained by opening with a disk element of radius R=4. Therefore, the difference in intensity integrated values ​​between the two images (R=3, R=3+1) corresponds to pixel clusters that correspond to silica and pixel clusters of smaller size. In this case, the difference in intensity integrated values ​​is not the smallest. Therefore, the diameter R with the smallest difference in intensity integration value between an image subjected to opening processing with a disk element of diameter (R) and an image subjected to opening processing with a disk element of diameter (R+1) can be regarded as the representative value.

[0024] The first diameter specifying unit 11 uses the above algorithm to specify a first diameter, which is a representative value. Specifically, as shown in FIG. 2, in step ST2, the first diameter specifying unit 11 sets radii R=R1 to Rn with an increment of 1, and performs opening processing on the SEM image using disk elements of each radius R to obtain images 1 to n. n can be set arbitrarily. In this embodiment, n=22. The SEM image is an image obtained by the SEM image obtaining unit 10 in step ST1. FIG. 10 shows an image obtained by performing opening processing on the SEM image shown in FIG. 3 using disk elements of radius R=7.

[0025] In the next step ST3, the first diameter determination unit 11 calculates the intensity integrated values ​​L1 to Ln for each of the images 1 to n after the opening process. The intensity integrated value is a value obtained by integrating the pixel intensities of all pixels in each image. The intensities are integrated for the number of pixels that make up one image. In this embodiment, n=22, but this is not limited to this. Figure 11 is a graph showing the intensity integrated values ​​for each image. The horizontal axis of the figure represents the radius R [number of pixels] of the disk element, and the vertical axis of the figure represents the intensity integrated value. It can be seen that the larger the radius R of the disk element, the more white pixels are removed by the opening process, and therefore the smaller (i.e., darker) the intensity integrated value becomes.

[0026] In the next step ST4, the first diameter determination unit 11 calculates the difference [L k+1 -L k ] is calculated, where k = 1 to (n-1). In other words, the difference is calculated as [L2-L1], [L3-L2], ..., [L 21 -L 20 ], [L 22 -L 21 ] is calculated as follows. FIG. 12 is a graph showing the difference in the intensity integrated value. The horizontal axis of the graph shows the radius R [number of pixels] of the disk element, and the vertical axis of the graph shows the difference in the intensity integrated value. According to FIG. 12, it can be seen that the difference in the intensity integrated value is smallest when the radius R of the disk element is 7.

[0027] In the next step ST5, the first diameter determining unit 11 determines the diameter of the silica particle (first diameter) as the diameter that minimizes the difference between the intensity integrated values. In the example of Fig. 12, the first diameter determining unit 11 determines that the first diameter is a radius R = 7.

[0028] The second diameter determination unit 12 determines a second diameter that is larger than the first diameter. In this embodiment, the second diameter is calculated by adding one pixel to the first diameter. Of course, this is not limited to this. If the second diameter is larger than the first diameter, the difference between the first and second diameters is not limited to one pixel, but can be set to a predetermined width such as two or three pixels. However, sufficient accuracy can be ensured if the difference between the first and second diameters is a minimum of one pixel. It is preferable that the difference between the first and second diameters be a maximum of five pixels or less.

[0029] The first image generation unit 13 generates a first image. The first image is an image obtained by removing pixel clusters smaller than a circle with a first diameter (R=7) and pixel clusters larger than a circle with a second diameter (R=8) that is larger than the first diameter from among multiple pixel clusters that have a relatively higher intensity than their surroundings in the SEM image, and converting pixel clusters included in circles with a diameter equal to or larger than the first diameter (R=7) and smaller than the second diameter (R=7) into circular shapes. The first image generation unit 13 includes a second image generation unit 13a, a third image generation unit 13b, and a difference processing unit 13c to generate the first image. Converting pixel clusters into circular shapes means converting pixel clusters into pixel clusters that are circular or close to circular.

[0030] In step ST6 (FIG. 2), the second image generation unit 13a performs opening processing of morphology processing on the SEM image, with the structural elements (disk elements) being circles with a first diameter (R=7), to generate a second image. The second image corresponds to the image shown in the middle of FIG. 9 when R=7. The white pixels included in the second image are pixel clusters corresponding to silica particles and pixel clusters larger than the silica particles.

[0031] In step ST7 (FIG. 2), the third image generation unit 13b performs opening processing of the morphology process on the SEM image, with the structural elements (disk elements) being circles with a second diameter (R=8), to generate a third image. The third image corresponds to the image shown in the lower part of FIG. 9 when R=8. The white pixels included in the third image are pixel clusters larger than the silica particles.

[0032] In step ST8 (FIG. 2), the subtraction processing unit 13c generates a first image based on the difference between the second image and the third image. FIG. 13 is an explanatory diagram related to the process of generating the first image from the second image and the third image. Specifically, as shown in FIG. 13, the subtraction processing unit 13c can obtain the first image by subtracting the intensity of each pixel of the third image from the intensity of each pixel of the second image. This subtraction process removes pixel clusters in the second image that are larger than silica particles.

[0033] In step ST9 (FIG. 2), the binarization execution unit 14 binarizes the first image. The binarization method may be the Otsu method, or may simply use the half value of the intensity range, the average intensity, or the mode intensity. FIG. 14 shows the binarized first image. Comparing the SEM image shown in FIG. 3 with the binarized first image shown in FIG. 14, it can be seen that the large whitish pixel cluster (a different substance larger than silica) in the center of the first image shown in FIG. 14 has turned black and has been properly removed.

[0034] In step ST10 (FIG. 2), the second region setting unit 15 divides the binarized image to set multiple second regions. As shown in FIG. 15, multiple second regions Ar2 may be set for the binarized image so that the second regions Ar2 do not overlap each other, or as shown in FIG. 16, the second regions Ar2 may be set so that some overlap is permitted. In the example of FIG. 15, one binarized image is divided into 64 regions, and 64 second regions Ar2 are set. The second regions Ar2 are smaller than the binarized image and contain multiple pixels. Since the number of second regions Ar2 after setting will be too large to illustrate, FIG. 16 shows the process of setting the second regions Ar2. When overlapping is permitted for the second regions shown in FIG. 16, each second region is shifted by sliding it vertically and horizontally by five pixels. The amount of shift (overlap amount) of the second regions Ar2 is not limited to five pixels and can be changed as long as it is uniform within the image. The number of second regions Ar2 when overlapping of the second regions Ar2 shown in Fig. 16 is permitted is greater than the number of second regions Ar2 when overlapping of the second regions Ar2 is not permitted as shown in Fig. 15. The size of the second region Ar2 may be u times the size of the binarized image (u is a number greater than 0 and less than 1).

[0035] In step ST11 (FIG. 2), the evaluation value calculation unit 16 counts the white pixels for each second region Ar2 set by the second region setting unit 15, and calculates the silica dispersion evaluation value based on the number of white pixels in each second region Ar2. As a specific example, if the number of second regions is N, the white pixels in each of the N second regions are counted. For k=1 to N, the number of white pixels in the k-th second region is denoted by X. k The number of white pixels in each second region is X1, X2, ..., X N-1 ,X N It is expressed as: In this embodiment, the evaluation value calculation unit 16 calculates the coefficient of variation CV as a dispersion evaluation value. The coefficient of variation CV (evaluation value) is calculated by multiplying the number of white pixels (X k The coefficient of variation CV can be calculated based on the following formula: S is the standard deviation, and X is the av indicates the average number of white pixels in all second regions.

number

[0036] Figure 17 shows a histogram of the number of white pixels in the second regions for an example in which the second regions shown in Figure 15 do not overlap, as shown in Figure 15, and a histogram of the number of white pixels in the second regions for an example in which the second regions overlap, as shown in Figure 16, as shown in Figure 16. The horizontal axis represents the number of white pixels in one second region, and the vertical axis represents the number of second regions. Comparing the dot-dash line and the solid line, the distorted distribution shape of the solid line is smoothed out and approaches a symmetrical shape in the dot-dash line. Increasing the number of second regions when performing variance evaluation is thought to increase the statistical amount and improve the accuracy of the variance evaluation value. Comparing the dot-dash line and the solid line, the degree of variance of the white pixels is similar, indicating that accurate evaluation is possible using either method of setting the second regions.

[0037] [Calculation method for silica dispersion evaluation value] The method of calculating the dispersion evaluation value of silica, which is executed by the above system 1, will be described with reference to FIG.

[0038] First, in step ST1, the SEM image acquisition unit 10 acquires an SEM image, which is a grayscale image of a rubber surface. The enlargement processing unit 17 may perform enlargement processing to enlarge the SEM image acquired by the SEM image acquisition unit 10. The contrast processing unit 18 may perform contrast processing to flatten the intensity histogram. The first diameter specifying unit 11 executes steps ST2 to ST5 to specify the first diameter of the circle that best matches a plurality of pixel clusters that have relatively higher intensity than the surrounding areas in the SEM image. The first image generating unit 13 generates the first image by executing steps ST6 to ST8. In step ST9, the binarization execution unit 14 binarizes the first image to obtain a binarized image. By executing steps ST10 to ST11, the dispersion evaluation value of silica is calculated based on the binarized first image.

[0039] As described above, the method for calculating the silica dispersion evaluation value of this embodiment is a method executed by one or more processors, and may include identifying a first diameter of a circle that best matches a plurality of pixel clusters that are relatively more intense than their surroundings in an SEM image, which is a grayscale image of a rubber surface; removing pixel clusters that are smaller than a circle of the first diameter and pixel clusters that are equal to or larger than a circle of a second diameter that is larger than the first diameter from among the plurality of pixel clusters that are relatively more intense than their surroundings in the SEM image; generating a first image in which pixel clusters included in circles that are equal to or larger than the first diameter and smaller than the second diameter are converted into a circular shape; binarizing the first image; and calculating the silica dispersion evaluation value based on the binarized first image. In this way, the first diameter of the circle that best matches the multiple pixel clusters that are thought to be silica contained in large quantities in the rubber is determined, pixel clusters smaller than a circle of the first diameter that are thought to be smaller than silica are removed, pixel clusters that are circular or larger than a second diameter that are thought to be larger than silica are removed, and pixel clusters that are larger than the first diameter but smaller than the second diameter that are thought to be silica are replaced with circles, thereby improving the ability to identify silica in subsequent processes, increasing the possibility of removing substances other than silica, and making it possible to calculate an appropriate silica dispersion evaluation value.

[0040] Although not particularly limited, as in the present embodiment, a preferred example may be adopted in which the second image is generated by performing an opening process of a morphology process with the structuring element being a circle of a first diameter, the third image is generated by performing an opening process of a morphology process on the SEM image with the structuring element being a circle of a second diameter, and the first image is generated based on the difference between the second image and the third image.

[0041] Although not particularly limited, as in this embodiment, an enlargement process for enlarging the image may be performed on the SEM image, and the first diameter may be determined based on the SEM image after the enlargement process. The enlargement process increases the amount of silica represented by multiple pixels, making it easier to capture silica and allowing an appropriate silica dispersion evaluation value to be calculated.

[0042] Although not particularly limited, as in this embodiment, contrast processing for flattening the intensity histogram may be performed on the SEM image, and the first diameter may be identified based on the SEM image after the contrast processing. If the intensity distribution becomes flat, the intensity change at the edge portion becomes uniform (flat), which makes it possible to improve the stability of the first diameter determination process and subsequent processes.

[0043] Although not particularly limited, as in this embodiment, the binarized image may be divided into multiple second regions Ar2, the white pixels may be counted for each second region Ar2, and the silica dispersion evaluation value may be calculated based on the number of white pixels in each second region Ar2. By dividing the binarized image into multiple second regions Ar2 and then evaluating the dispersion, the degree of bias in the white regions within the entire image can be evaluated with high accuracy, and the accuracy of the silica dispersion evaluation value can be improved.

[0044] Although not particularly limited, the plurality of second regions Ar2 may partially overlap each other, as in this embodiment. In this way, the number of second regions Ar2 increases, and the number of white pixels in each local second region Ar2 increases, so that the accuracy of the dispersion evaluation value of silica can be improved.

[0045] The program according to this embodiment is a program that causes one or more computers to execute the above method. By executing these programs, it is possible to obtain the effects of the above-mentioned methods.

[0046] Although the embodiments of the present disclosure have been described above with reference to the drawings, the specific configurations should not be considered to be limited to these embodiments. The scope of the present disclosure is defined not only by the description of the above embodiments but also by the claims, and further includes all modifications within the meaning and scope of the claims.

[0047] (1) When the binarization method of the above embodiment is implemented, the method for calculating the silica dispersion evaluation value of the binarized image is not limited to the method described in this specification, and various methods can be adopted.

[0048] (2) In the above embodiment, the enlargement processing of the SEM image by the enlargement processing unit 17 and the contrast processing by the contrast processing unit 18 are performed, but these may be omitted. Only the enlargement processing or only the contrast processing may be performed.

[0049] The structures employed in the above-described embodiments can be employed in any other embodiment. The specific configurations of the components are not limited to the above-described embodiments, and various modifications are possible within the scope of the present disclosure.

[0050] For example, the order of execution of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings, can be implemented in any order, as long as the output of a previous process is not used in a subsequent process. Even if the flow in the claims, specifications, and drawings is explained using terms such as "first" and "next" for convenience, this does not mean that the processes must be executed in this order.

[0051] 1 are realized by executing a predetermined program on one or more processors, but each unit may also be configured with a dedicated memory or dedicated circuit. In the system 1 of the above embodiment, each unit is implemented in the processor 1a of a single computer, but each unit may be distributed and implemented on multiple computers or in the cloud. In other words, the above method may be executed on one or more processors.

[0052] System 1 includes a processor 1 a. For example, processor 1 a can be a central processing unit (CPU), a microprocessor, or other processing unit capable of executing computer-executable instructions. System 1 also includes memory 1 b for storing data for system 1. In one example, memory 1 b includes computer storage media, such as RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, DVD or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device, or any other medium that can be used to store desired data and that can be accessed by system 1. [Explanation of symbols]

[0053] 11...first diameter identification unit, 12...second diameter determination unit, 13...first image generation unit, 13a...second image generation unit, 13b...third image generation unit, 13b...third image generation unit, 14...binarization execution unit, 15...second area setting unit, 16...evaluation value calculation unit, 17...enlargement processing unit, 18...contrast processing unit.

Claims

1. 1. A method executed by one or more processors, comprising: In an SEM image, which is a grayscale image of a rubber surface, the diameter of a white pixel having a circular or similar shape, which corresponds to silica, is defined as R. An opening process is performed using the diameter R and a diameter (R+1) to generate each image, and the diameter R that minimizes the difference between the intensity integrated values ​​of the images is identified as a first diameter, which is the diameter of the silica particle; generating a second image by performing an opening process using the circle of the first diameter as a structuring element; generating a third image by performing an opening process using a circle having a second diameter that is one pixel larger than the first diameter as a structuring element; generating a first image by subtracting the intensity of each pixel of the third image from the intensity of each pixel of the second image; binarizing the first image; A method for calculating a dispersion evaluation value of silica, the method comprising: calculating a dispersion evaluation value of silica based on the binarized first image.

2. The method of claim 1 , further comprising: performing a magnification process on the SEM image to magnify the image; and identifying the first diameter based on the SEM image after the magnification process.

3. The method of claim 1 or 2, further comprising: performing a contrast process on the SEM image to flatten an intensity histogram; and identifying the first diameter based on the SEM image after the contrast process.

4. 4. The method according to claim 1, further comprising: dividing the binarized image into a plurality of second regions; counting white pixels in each of the second regions; and calculating a dispersion evaluation value of silica based on the number of white pixels in each of the second regions.

5. The method of claim 4 , wherein the second regions partially overlap each other.

6. A system comprising one or more processors for carrying out the method according to any one of claims 1 to 5.

7. A program that causes one or more processors to execute the method according to any one of claims 1 to 5.

Images