Magnetic domain image statistical method for oriented silicon steel powder grain method

Abstract

The invention relates to the technical field of image processing, and discloses an oriented silicon steel powder grain method magnetic domain image statistical method, which comprises the following steps: obtaining an original image, and carrying out binarization processing on the original image to generate a black and white image; fitting scattered black pixel points in the black-and-white image into a linear shape to obtain a linear image; converting the linearity in the linear image into a straight line to obtain a linear image; the linear image is scanned to calculate a domain number and an average magnetic domain width. In order to solve the problem that data in a large number of images are difficult to count or the statistical efficiency is low when a magnetic domain observation image using a powder pattern method is adopted, an image recognition and enhancement method is adopted to process the image to be in a distortionless and convenient-to-count form, and main observation parameters in the image are recognized, recorded and output quickly.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more particularly to a statistical method for magnetic domain images using the powder texture method for oriented silicon steel. Background Technology

[0002] The performance changes of grain-oriented silicon steel caused by external influences are generally manifested through its material characteristic parameters, such as unit loss and magnetic permeability. However, it is difficult to investigate the changes in grain-oriented silicon steel under external influences solely from the perspective of macroscopic material characteristic parameter changes, as it is impossible to obtain patterns that match the macroscopic parameter measurement results. Therefore, in order to explain the changes in macroscopic parameters from a mechanistic perspective and to accurately calculate or predict changes in material characteristic parameters, it is necessary to explore the changes in the magnetic domains within the grain-oriented silicon steel material from a mesoscopic perspective.

[0003] Magnetic domain observation methods currently include powder microscopy, magnetic fluid microscopy, transmission electron microscopy, and X-ray diffraction. Among these, powder microscopy offers advantages such as convenience and speed, a short cycle from sample preparation to observation completion, low sample quality requirements, low operational difficulty, and low cost.

[0004] However, due to the characteristics of magnetic powder, its imaging effect is worse than that of microscopy, and the magnetic domains themselves are relatively dense, which undoubtedly brings great trouble to the subsequent data statistics work of the experimenters, especially when batch processing of observation images will consume a lot of time. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method for statistical analysis of magnetic domain images using powder texture method for oriented silicon steel, so as to solve one or more problems in the prior art.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows: A statistical method for magnetic domain images using powder texture analysis of oriented silicon steel includes the following steps: Obtain the original image and perform binarization on the original image to generate a black and white image; A linear image is obtained by fitting scattered black pixels in a black and white image into a line. Transform the linear elements in a linear image into straight lines to obtain a linear image; The linear image is scanned to calculate the domain number and average domain width.

[0007] Furthermore, the binarization process of the original image to generate a black and white image includes the following steps: The pixels are scanned one by one according to the preset scanning direction to obtain the grayscale value of each pixel. The threshold is determined based on the grayscale values ​​of all pixels. Determine whether the grayscale value of each pixel is greater than the threshold. If the value is greater than the specified value, the corresponding pixel will be identified as white. If the value is less than or equal to the value, the corresponding pixel will be identified as black.

[0008] Furthermore, determining the threshold based on the grayscale values ​​of all pixels includes the following steps: The percentage of pixels at each grayscale value in the entire original image was calculated. Calculate the inter-class variance of the average gray level of the "black" region, the average gray level of the "white" region, and the overall average gray level of the image when each gray level is used as the threshold, based on the respective proportions. Select the gray level corresponding to the maximum value of the inter-class variance as the threshold.

[0009] Furthermore, the process of transforming linearity in a linear image into a straight line to obtain a linear image includes the following steps: Slide a window of a preset shape along the scanning direction with a preset step size until the entire linear image has been traversed. Each time the window is swiped, it checks whether the proportion of black pixels within the window exceeds a preset ratio. If so, adjust the corresponding area of ​​the entire window to black; If not, adjust the corresponding area of ​​the entire window to white.

[0010] Furthermore, after traversing the entire linear image, the process further includes the following steps: Determine if the preset step size is equal to the length of the linear image. If not, extract a length value from the preset data stack, update the preset step size to the length value, and slide the window of the preset shape along the scanning direction with the preset step size again; the data stack in the initial state stores multiple length values ​​in descending order of data, with the largest length value being the length of the linear image.

[0011] Furthermore, the step of fitting scattered black pixels in a black and white image into a line to obtain a linear image includes the following steps: The average length of vertically continuous black pixels and the average length of horizontally continuous black pixels are determined based on the black and white image, and then structural elements are generated. A dilation operation is performed based on the structuring element and the black-and-white image to generate a linear image.

[0012] Compared with the prior art, the beneficial technical effects of the present invention are as follows: In view of the problem that it is difficult to statistically analyze a large amount of data in magnetic domain observation images using the powder method or that the statistical efficiency is slow, the image recognition and enhancement method is used to process the image to a form that is not distorted and is easy to analyze, and the main observation parameters are identified, recorded and output relatively quickly. Attached Figure Description

[0013] Figure 1 The flowchart shown is a statistical method for magnetic domain images of oriented silicon steel using powder texture method according to an embodiment of the present invention. Figure 2 A schematic diagram of a magnetic domain observation image using the powder method provided in an embodiment of the present invention is shown. Figure 3 A schematic diagram of a black and white image after binarization processing provided by an embodiment of the present invention is shown. Figure 4 A schematic diagram of the optimized linear image provided by an embodiment of the present invention is shown. Figure 5 A schematic diagram of a linear black and white striped image provided in an embodiment of the present invention is shown. Detailed Implementation

[0014] A statistical method for magnetic domain images using powder texture analysis of oriented silicon steel, see [link to relevant documentation]. Figure 1 This includes the following steps: S100. Obtain the original image and perform binarization on the original image to generate a black and white image.

[0015] S200. Fit the scattered black pixels in the black and white image into a line to obtain a linear image.

[0016] S300: Transform the linear elements in the linear image into straight lines to obtain a linear image.

[0017] S400: Scan the linear image to calculate the domain number and average domain width.

[0018] The original image is a magnetic domain observation image obtained using the powder method. The original image can be referenced as follows: Figure 2 .

[0019] First, binarize the original image to convert the grayscale values ​​of each pixel into only two colors: black (grayscale value 0) and white (grayscale value 255). A black and white image can be referenced. Figure 3 To reduce the amount of information needed for subsequent image recognition and highlight scattered points, interference from pixels with similar grayscale values ​​at the edges of magnetic domain walls that are difficult to identify is eliminated in advance.

[0020] Then, the scattered images of the magnetic domain walls adsorbed by magnetic powder are interconnected according to their orientation trends, transforming the "point-shaped" image into a "line-shaped" image. The resulting linear image can be referenced. Figure 4 .

[0021] Due to the characteristics of powder imaging and the uncontrollable distribution of magnetic particles, the image still contains interference such as discontinuities, distortions, and blemishes. Therefore, it is necessary to completely convert the image into a linear black and white striped image. Specific images can be found in [reference needed]. Figure 5 The optimized image can then be used to quickly calculate the domain number and average domain width based on the distribution of black and white stripes.

[0022] In one embodiment, binarizing the original image to generate a black and white image includes the following steps: S110. Scan each pixel one by one according to the preset scanning direction to obtain the grayscale value of each pixel.

[0023] S120. Determine the threshold based on the grayscale values ​​of all pixels.

[0024] S130. Determine whether the gray value of each pixel is greater than the threshold.

[0025] S140. If the value is greater than the specified value, the corresponding pixel will be identified as white.

[0026] S150. If it is less than or equal to, the corresponding pixel will be identified as black.

[0027] The scanning direction is generally perpendicular to the magnetic domain direction, and is preset by the staff based on the magnetic domain direction in the original image.

[0028] By setting a threshold, all pixels are classified into two categories: one category has a gray value less than or equal to the threshold; the other category has a gray value greater than the threshold. The gray values ​​of the former are uniformly adjusted to 0, and the gray values ​​of the latter are uniformly adjusted to 255, so that the image contains only black and white pixels.

[0029] The threshold can be set manually, but in order to improve accuracy, the threshold is determined based on the original image in this embodiment.

[0030] Specifically, the threshold is determined based on the grayscale values ​​of all pixels, including the following steps: S121. Calculate the percentage of pixels at each gray level in the entire original image.

[0031] S122. Calculate the inter-class variance of the average gray level of the "black" region, the average gray level of the "white" region, and the total average gray level of the image when each gray level is used as the threshold, based on the respective proportions.

[0032] S123. Select the gray value corresponding to the maximum value of the inter-class variance as the threshold.

[0033] By scanning in step S110, the gray value of each pixel in the original image can be obtained, and then the number of gray levels L (generally L=256) of the entire image can be counted, as well as the number of pixels ni under each gray level, where i is the corresponding gray level, and the total number of pixels is N.

[0034] Calculate the percentage of pixels at each gray level in the entire original image: Pi = ni / N. Then, for each possible threshold t (ranging from 0 to L-1), divide the pixels into two categories: C0 (gray level less than or equal to t) and C1 (gray level greater than t).

[0035] The Otsu algorithm is used to calculate the inter-class variance of the average gray level of the "black" region, the average gray level of the "white" region, and the overall average gray level of the image.

[0036] In one embodiment, fitting scattered black pixels in a black and white image into a line to obtain a linear image includes the following steps: S210. Determine the average length of vertically continuous black pixels and the average length of horizontally continuous black pixels based on the black and white image, and then generate structural elements.

[0037] S220. Perform a dilation operation based on the structuring element and the black-and-white image to generate a linear image.

[0038] The two-dimensional dimensions of the structuring element are the average length L1 of consecutive black pixels horizontally and the average length L2 of consecutive black pixels vertically, with its center position denoted as (L1 / 2, L2 / 2). L1 is obtained by scanning the black and white image along the width direction to count the lengths of each consecutive black pixel and then calculating the average. L2 is obtained by scanning the black and white image along the length direction to count the lengths of each consecutive black pixel and then calculating the average. To simplify the calculation, this step only needs to scan a portion of the black and white image, such as scanning the area within 1 / 8 of the image length at the top of the image in the width direction. Considering the uncertainty of the domain wall position in the length direction, the scanning range can be appropriately increased, such as scanning the area within 1 / 6 of the image width at the leftmost end of the image.

[0039] Furthermore, the expansion operation can be represented by the following mathematical formula: Assuming image A is a binary image, its pixel value a(x,y) takes the value 0 (representing background) or 1 (representing target), and the pixel value b(x',y') of structuring element B also takes the value 0 or 1. In the resulting image C of the dilation operation A⊕B, the pixel value c(x,y) is calculated as follows: c(x,y)=max{a(x-x',y-y')∧(x',y'):(x',y')∈B} In this context, ⊕ represents the dilation operation, max represents the maximum value operation, and ∧ represents the logical AND operation. In actual calculation, for each pixel (x, y) in image A, the center of the structuring element B is aligned with that pixel. All pixels (x', y') within structuring element B are traversed, and a(x-x', y-y')∧(x', y') is calculated. If any pixel within the area covered by the structuring element makes a(x-x', y-y')∧(x', y') = 1, then the dilated pixel value c(x, y) is 1; otherwise, it is 0. In this way, the target region in the binary image expands along the shape and direction of the structuring element.

[0040] In one embodiment, transforming linearity in a linear image into a straight line to obtain a linear image includes the following steps: S310. Slide a window of a preset shape along the scanning direction with a preset step size until the entire linear image is traversed.

[0041] S320: Each time the window is swiped, determine whether the proportion of black pixels in the window exceeds a preset ratio.

[0042] S330. If so, adjust the corresponding area of ​​the entire window to black.

[0043] S340. If not, adjust the corresponding area of ​​the entire window to white.

[0044] The window is a rectangle with length 'a' and width 'b', and its width direction is consistent with the scanning direction. In this implementation, length 'a' is 1 / 6 of the image length, and to avoid repeated scanning, the preset step size is equal to the width 'b'.

[0045] The preset ratio is set manually. When the proportion of pixels with a grayscale value of 0 is greater than the preset ratio, all pixels in the window are judged as black; when the proportion of pixels with a grayscale value of 255 is less than or equal to the preset ratio, all pixels in the window are judged as white.

[0046] This process, continuing until the entire linear image has been traversed, includes the following steps: S311. Determine whether the preset step size is the length of the linear image.

[0047] S312. If not, extract a length value from the preset data stack, update the preset step size to the length value, and slide the window of the preset shape along the scanning direction with the preset step size again.

[0048] In the initial state, the data stack stores multiple length values ​​in descending order of data size, with the largest length value being the length of the linear image. In this implementation, the initial data stack stores image lengths of 1 / 5, 1 / 4, 1 / 3, 1 / 2, and 1 times the original image length.

[0049] After performing the first step S310, the window length is changed to 1 / 5 of the image length and the above operation is repeated. Then, the window length is changed to 1 / 4, 1 / 3, 1 / 2, and 1 times the image length for iterative scanning. The purpose is to completely convert the image into a straight black and white striped image, and the step-by-step iteration can improve the conversion accuracy.

[0050] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0051] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A statistical method for magnetic domain images using powder etching in oriented silicon steel, characterized in that, Includes the following steps: Obtain the original image and perform binarization on the original image to generate a black and white image; A linear image is obtained by fitting scattered black pixels in a black and white image into a line. Transform the linear elements in a linear image into straight lines to obtain a linear image; The linear image is scanned to calculate the domain number and average domain width.

2. The method for statistical analysis of magnetic domain images using powder texture analysis of oriented silicon steel as described in claim 1, characterized in that: The process of binarizing the original image to generate a black and white image includes the following steps: The pixels are scanned one by one according to the preset scanning direction to obtain the grayscale value of each pixel. The threshold is determined based on the grayscale values ​​of all pixels. Determine whether the grayscale value of each pixel is greater than the threshold. If the value is greater than the specified value, the corresponding pixel will be identified as white. If the value is less than or equal to the value, the corresponding pixel will be identified as black.

3. The method for statistical analysis of magnetic domain images using powder texturing of oriented silicon steel as described in claim 2, characterized in that, Determining the threshold based on the grayscale values ​​of all pixels includes the following steps: The percentage of pixels at each grayscale value in the entire original image was calculated. Calculate the inter-class variance of the average gray level of the "black" region, the average gray level of the "white" region, and the overall average gray level of the image when each gray level is used as the threshold, based on the respective proportions. Select the gray level corresponding to the maximum value of the inter-class variance as the threshold.

4. The method for statistical analysis of magnetic domain images using powder etching in oriented silicon steel as described in claim 1, characterized in that, The process of transforming linear elements in a linear image into straight lines to obtain a linear image includes the following steps: Slide a window of a preset shape along the scanning direction with a preset step size until the entire linear image has been traversed. Each time the window is swiped, it checks whether the proportion of black pixels within the window exceeds a preset ratio. If so, adjust the corresponding area of ​​the entire window to black; If not, adjust the corresponding area of ​​the entire window to white.

5. The method for statistical analysis of magnetic domain images using powder etching in oriented silicon steel as described in claim 4, characterized in that, After traversing the entire linear image, the process further includes the following steps: Determine if the preset step size is equal to the length of the linear image. If not, extract a length value from the preset data stack, update the preset step size to the length value, and slide the window of the preset shape along the scanning direction with the preset step size again; the data stack in the initial state stores multiple length values ​​in descending order of data, with the largest length value being the length of the linear image.

6. The method for statistical analysis of magnetic domain images using powder texturing of oriented silicon steel as described in claim 1, characterized in that... The process of fitting scattered black pixels in a black and white image into a line to obtain a linear image includes the following steps: The average length of vertically continuous black pixels and the average length of horizontally continuous black pixels are determined based on the black and white image, and then structural elements are generated. A dilation operation is performed based on the structuring element and the black-and-white image to generate a linear image.

Images