Forging overburning detection method based on microscopic image processing

Abstract

The present invention relates to the field of forging overburn detection, and in particular to a forging overburn detection method based on microscopic image processing, comprising: obtaining a microscopic image of a forging to be detected, obtaining an actual grain region according to the microscopic image, and using the edge line of the actual grain region as a grain boundary line; calculating the possibility that the actual grain region is an overburned grain, the possibility being negatively correlated with the radius shape index of the actual grain region; performing edge detection on the actual grain region to obtain multiple edge lines, determining whether adjacent edge lines are grain boundaries at the same time, and if so, fitting two grain boundaries to obtain a fitting line, and using the area enclosed by the grain boundary line and the fitting line as an ideal grain region. Calculating the degree of overburn of the forging to be detected, when the degree of overburn is greater than a threshold value, the forging to be detected is overburned. The present invention detects whether the forging is overburned by the degree of overburn, thereby improving the accuracy of the detection result and the detection efficiency.

Description

Technical Field

[0001] The invention relates to the field of forging overburn detection, and in particular to a forging overburn detection method based on microscopic image processing. Background Art

[0002] Overburning of forgings refers to excessive changes in the material microstructure and overheating of the forgings due to excessively high heating temperatures or long holding times during the forging process, which reduces the mechanical properties and service life of the forgings. Overburning detection of forgings is an important part of ensuring the quality of forgings, and it involves the evaluation of the internal structure and performance of the forgings.

[0003] Overburning of forgings can cause serious changes in the microstructure of the material, resulting in coarse grains and weakened grain boundaries, which will significantly reduce the strength and plasticity of the forgings. Overburned forgings are prone to folding and cracking during the forging process. In severe cases, they may even crack as soon as they are forged, rendering the forgings unusable. In addition, overburning has a significant effect on the fatigue properties of forgings, seriously affecting the safety and reliability of forgings. In applications with extremely high safety requirements, such as the aerospace and automotive industries, the use of overburned forgings may cause structural failures and even lead to serious safety accidents.

[0004] A Chinese patent with announcement number CN103592222B discloses a method for preparing a standard sample image for detecting overburned structure of a magnesium alloy casting. Four batches of mechanical property samples of a single-cast magnesium alloy are taken and placed in furnaces at different temperatures for solution treatment. A mechanical property test is carried out at room temperature, and the samples classified after the mechanical test are marked to prepare metallographic samples. After corrosion, the samples are magnified 200 to 400 times under an optical microscope for observation, and a metallographic image of the structure is taken. The metallographic image is selected as a standard sample image for detecting overburned structure within the solution temperature range according to the degree of metallographic overburning of the structure. The metallographic image can be used to compare and analyze the metallographic structure of the casting, thereby determining whether the structure of the casting is overburned.

[0005] However, the above detection method uses inspectors to observe microscopic images to determine whether the casting is overburned. This judgment method relies on the experience and skills of the inspectors. Different inspectors may come to different conclusions when inspecting the same casting, and the accuracy of the detection results is low. Summary of the invention

[0006] In order to solve the problem that the existing casting overburn detection method relies heavily on the experience and skills of the detection personnel, resulting in low accuracy of the detection results, the present invention provides a forging overburn detection method based on microscopic image processing.

[0007] The present invention provides a forging overburn detection method based on microscopic image processing, which adopts the following technical solutions:

[0008] Obtain a microscopic image of the forging to be inspected, obtain the actual grain area according to the microscopic image, and use the edge line of the actual grain area as the grain boundary line;

[0009] Calculate the probability that the actual grain area is an overburned grain, and the probability is negatively correlated with the radius shape index of the actual grain area;

[0010] Edge detection is performed on the actual grain region to obtain multiple edge lines, and it is determined whether adjacent edge lines are grain boundaries at the same time. If so, two grain boundaries are fitted to obtain a fitting line, and the area enclosed by the grain boundary line and the fitting line is used as the ideal grain region.

[0011] Calculate the overburning degree of the forging to be tested. When the overburning degree is greater than a threshold value, the forging to be tested is overburned.

[0012] The expression of overburning degree is:

[0013]

[0014] In the formula, Indicates the degree of overburning of forgings. represents the number of ideal grain regions, is the number of actual grain regions, Indicates The possibility that the actual grain area is an overburned grain, is the average area of ​​all actual grain regions, Represents the number of all fitted line pixels, Represents the number of all grain boundary pixels.

[0015] The overburning degree of the forging is calculated based on multiple factors such as the number of ideal grain areas and actual grain areas, the possibility of overburned grains, the average area of ​​the actual grain area, and the proportion of the number of pixels on the fitting line. The overburning degree is used to detect whether the forging is overburned, thereby improving the accuracy of the detection results and improving the detection efficiency.

[0016] Preferably, the method for calculating the possibility that the actual grain region is an overburned grain is:

[0017] The grain boundary lines of the actual grain area are detected to obtain the positions of the corner points, the distance from each corner point to the center point of the actual grain area is calculated, and the variance of multiple distances is calculated;

[0018] The expression for the probability is:

[0019]

[0020] in, Indicates the possibility that the actual grain area is an overburned grain. represents the area of ​​the actual grain region, represents the perimeter of the grain boundary line in the actual grain area, Indicates the number of corner points on the grain boundary line in the actual grain area, Represents the variance of the distance from the corner point to the center point on the grain boundary line of the actual grain area, The radius shape index representing the actual grain area, Represents the normalization function.

[0021] Preferably, the method for calculating the possibility that the actual grain region is an overburned grain is:

[0022] The grain boundary lines of the actual grain area are detected to obtain the positions of the corner points, the distance from each corner point to the center point of the actual grain area is calculated, and the variance of multiple distances is calculated;

[0023] The expression for the probability is: ;

[0024] in, Indicates the possibility that the actual grain area is an overburned grain. The radius shape index representing the actual grain area, Represents the normalization function.

[0025] The possibility that the actual grain region is an overburned grain is calculated through multiple factors, which improves the accuracy of the calculated possibility, thereby facilitating the judgment of whether the actual grain region is an overburned grain.

[0026] Preferably, the method for determining whether adjacent edge lines are simultaneously grain boundaries is:

[0027] Taking one of the endpoints on the edge line as the starting point, multiple pixel points are intercepted along the edge line direction;

[0028] Calculate the average gradient direction of multiple intercepted pixel points;

[0029] The normal direction of the gradient direction is taken as the extension direction, and the extension angle range is constructed based on the extension direction. The extension angle range is (α-β, α+β), where α is the angle between the extension direction and the horizontal direction, and β is the preset angle threshold;

[0030] In response to an endpoint of another edge line being within the extension angle range, the adjacent edge lines are simultaneously grain boundaries.

[0031] By judging whether two edge lines are grain boundaries at the same time, the influence of impurities on the fitting results is eliminated, thereby improving the accuracy of grain boundary fitting.

[0032] Preferably, the method of fitting two grain boundary lines to obtain a fitting line is:

[0033] Connect the pixel points between the endpoints of two grain boundaries to obtain a connecting line, construct multiple rays with the center of the actual grain area as the origin, intersect the rays with the connecting line to obtain an intersection point, calculate the distance between the intersection point and the origin, cluster the multiple distances obtained to obtain multiple clusters, take the intersection point corresponding to the distance in the cluster with the largest number of distances as the initial intersection point, and use the direct least squares method to fit the initial intersection point to obtain an ellipse;

[0034] Take any pixel point on the connecting line as the first reference point, and the pixel point on the ellipse closest to the first reference point as the second reference point; the pixel points on both sides of the first reference point on the connecting line correspond one-to-one to the pixel points on both sides of the second reference point on the ellipse, take the midpoint of the two corresponding pixel points as the fitting point, and connect the fitting points to obtain the fitting line.

[0035] The grain boundary line is obtained by screening, and the grain boundary line is fitted to obtain the fitting line, so that the overburning degree of the forging can be preliminarily obtained, which is convenient for understanding the quality of the forging.

[0036] Preferably, the method for obtaining the actual grain area according to the microscopic image is:

[0037] Obtain a microscopic image of the forging to be inspected, perform binarization on the microscopic image to obtain a binary image, and use any pixel point with a gray value of 255 in the binary image as a reference point;

[0038] In the microscopic image, the pixel point corresponding to the reference point is taken as the seed point, and the actual grain area is obtained by regional growth with the seed point as the center.

[0039] Preferably, the pixel point corresponding to the centroid of the area composed of pixels with a gray value of 255 in the binary image is used as the reference point.

[0040] Preferably, the Harris corner point detection algorithm is used to detect the grain boundary line of the actual grain region to obtain the position of the corner point.

[0041] By detecting the corner points of the grain boundaries in the actual grain area, the degree of overburning of the forging can be determined through the factor of the corner points, thereby improving the accuracy of the calculation of the degree of overburning.

[0042] The present invention has the following technical effects:

[0043] 1. The degree of overburning of forgings is calculated by multiple factors such as the number of ideal grain areas and actual grain areas, the possibility of overburned grains, the average area of ​​actual grain areas, and the proportion of the number of pixels of the fitting line. The degree of overburning is used to detect whether the forgings are overburned, thereby improving the accuracy of the detection results and the detection efficiency.

[0044] 2. In the process of fitting the grain boundary line, by judging whether the edge line is the grain boundary line, the interference caused by impurities in the forging is eliminated, the accuracy of the fitting result is improved, and the accuracy of the overburning degree calculation result is further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0046] Figure 1 It is a method flow chart of the forging overburning detection method based on microscopic image processing of the present invention.

[0047] Figure 2 It is a microscopic image of the overburned forging of the present invention.

[0048] Figure 3 It is a schematic diagram of the melting of the grain boundary in the microscopic image of the overburned forging of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0050] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0051] In the case of forgings without overburning, the white area in the binary image of the forging microscopic image is the actual grain area, and the black area is the edge of the actual grain area, which is also the grain boundary. When the forging is overburned, the grain boundary will break. Figure 3 As shown, at this time, there is no obvious grain boundary between the adjacent actual grain regions, so that the adjacent actual grains appear to be fused. Figure 2 As shown, this embodiment is described under this premise.

[0052] The embodiment of the present invention discloses a forging overburn detection method based on microscopic image processing, referring to Figure 1 , including the following steps, as follows:

[0053] S1: Obtain a microscopic image of the forging to be inspected, obtain the actual grain area according to the microscopic image, and use the edge line of the actual grain area as the grain boundary line.

[0054] The surface of the forging sample is ground and polished to obtain a smooth surface, and chemical or electrochemical corrosion treatment is performed to clearly display the microstructure of the casting, and the microscopic image of the forging is captured using the camera device of the microscope. The microscopic image is segmented by Otsu threshold to obtain a binary image, and the binary image is denoised. The binary image is divided into black pixels with a gray value of 0 and white pixels with a gray value of 255. The pixel corresponding to the centroid of the white area composed of pixels with a gray value of 255 in the binary image is used as a reference point, and the pixel corresponding to the reference point is used as a seed point. The actual grain area is obtained by regional growth with the seed point as the center.

[0055] S2: Calculate the probability that the actual grain area is an overburned grain. The probability is negatively correlated with the radius shape index of the actual grain area.

[0056] The radius shape index of the actual grain region is calculated. The radius shape index is also the Boyce-Clark radius shape index, which is used to measure the compactness of the actual grain region. Calculating the radius shape index of the actual grain region is a prior art, and the specific process is not repeated here. The Harris corner point detection algorithm is used to detect the grain boundary line of the actual grain region to obtain the position of the corner point, and the distance from the corner point to the center point of the actual grain region is calculated to obtain multiple distances, and then the variance of the multiple distances is calculated.

[0057] In this embodiment, the expression of possibility is:

[0058]

[0059] in, Indicates the possibility that the actual grain area is an overburned grain. represents the area of ​​the actual grain region, represents the perimeter of the grain boundary line in the actual grain area, Indicates the number of corner points on the grain boundary line in the actual grain area, Represents the variance of the distance from the corner point to the center point on the grain boundary line of the actual grain area, The radius shape index representing the actual grain area, Represents the normalization function.

[0060] The number of pixels included in the actual grain region is taken as the area of ​​the actual grain region, and the number of pixels included in the grain boundary line of the actual grain region is taken as its perimeter. It indicates the perimeter-to-area ratio of the actual grain region, and is used to measure the compactness of the shape of the actual grain region. The actual grain region of a normal forging is moderate in size and has clear grain boundaries, while the grain boundaries of the actual grain region of an overburned forging are blurred or broken, resulting in the fusion of adjacent actual grain regions and the abnormal growth of the shape of the actual grain region, which is less compact and regular than normal grains. The smaller it is, the more compact the shape of the actual grain area is, the greater the possibility that the actual grain area is a normal grain and the less likely it is an overburned grain. The larger it is, the less compact the shape of the actual grain area is, and the greater the possibility that the actual grain area is an overburned grain.

[0061] In the case of overburning, the grains grow abnormally, and grain boundaries often break between grains, which will cause multiple grain regions to grow into one region during the regional growth process. Then the number of corner points on the grain boundary line will increase, the irregularity will increase, and the distance from each corner point to the center point will also vary greatly, thereby increasing the variance of the distance from the corner point to the center point. Therefore, when and The larger the value, the more likely it is that the actual grain area is an overburned grain. The shape compactness of the actual grain area The smaller it is, the more irregular the shape of the actual grain area is, and the greater the possibility that the actual grain area is an overburned grain.

[0062] In other embodiments, the expression of the possibility is:

[0063]

[0064] in, Indicates the possibility that the actual grain area is an overburned grain. The radius shape index representing the actual grain area, Represents the normalization function.

[0065] S3: Perform edge detection on the actual grain region to obtain multiple edge lines, determine whether adjacent edge lines are grain boundaries at the same time, and if so, fit the two grain boundaries to obtain a fitting line, and use the area enclosed by the grain boundary line and the fitting line as the ideal grain region.

[0066] Canny edge detection is performed on the microscopic image to obtain multiple edge lines. Since the grain boundary lines may be broken in over-burned forgings, the detected grain boundary lines are open and discontinuous lines.

[0067] Since there will be some impurities in the actual grain area, almost all materials will contain a certain amount of impurities. These impurities can exist naturally during the processing of the material, or be introduced due to environmental factors during processing and use. Therefore, the detected edge lines include grain boundaries and lines formed by impurities. Tiny impurity particles or impurities with similar optical properties to the matrix material are difficult to distinguish directly. Therefore, in this case, the impurities in the actual grain area are similar to the characteristics of the melted grain boundaries and are difficult to distinguish. Impurities will interfere with the detection of the melted grain boundaries. Since the distribution of impurities is irregular and random. The distribution of the melted grain boundaries has certain rules. For example, the edge extension directions of adjacent grain boundaries are similar. Therefore, it is necessary to determine whether the adjacent edge lines are grain boundaries at the same time. If so, the two grain boundaries are fitted to obtain the fitting line.

[0068] S31: The method for determining whether adjacent edge lines are simultaneously grain boundaries is as follows:

[0069] Taking the endpoint on one of the edge lines as the starting point, multiple pixel points are intercepted along the edge line direction; the average gradient direction of the intercepted multiple pixel points is calculated; the normal direction of the gradient direction is the extension direction. It should be noted that the extension direction is toward the side of the other edge line endpoint. The extension angle range is constructed based on the extension direction. The extension angle range is (α-β, α+β), where α is the angle between the extension direction and the horizontal direction, and β is the preset angle threshold. The angle threshold is artificially set according to the actual situation. For example, the angle threshold is 20 degrees. In response to the endpoint of another edge line being within the extension angle range, the two adjacent edge lines are both grain boundaries. Otherwise, at least one of the two adjacent edge lines is not a grain boundary.

[0070] S32: The method for fitting two grain boundary lines to obtain a fitting line is:

[0071] Connect the pixel points between the endpoints of two grain boundaries to obtain a connecting line, construct multiple rays with the center of the actual grain area as the origin, intersect the rays with the connecting line to obtain an intersection, calculate the distance between the intersection and the origin, cluster the multiple distances obtained to obtain multiple clusters, take the intersection corresponding to the distance in the cluster with the largest number of distances as the initial intersection, and use the direct least squares ellipse fitting algorithm to fit the initial intersection to obtain an ellipse; take any pixel point on the connecting line as the first reference point, and take the pixel point on the ellipse closest to the first reference point as the second reference point; the pixel points on both sides of the first reference point on the connecting line correspond one-to-one to the pixel points on both sides of the second reference point on the ellipse, take the midpoint of the two corresponding pixel points as the fitting point, and connect the fitting points to obtain a fitting line.

[0072] Specifically, the first pixel point on the left side of the first reference point on the connecting line corresponds to the first pixel point on the left side of the second reference point on the ellipse, the second pixel point on the left side of the first reference point on the connecting line corresponds to the second pixel point on the left side of the second reference point on the ellipse, ...; the first pixel point on the right side of the first reference point on the connecting line corresponds to the first pixel point on the right side of the second reference point on the ellipse, the second pixel point on the right side of the first reference point on the connecting line corresponds to the second pixel point on the right side of the second reference point on the ellipse, .... According to the corresponding two pixel points, the corresponding midpoint is found, and multiple midpoints are connected to obtain the fitting line. The fitting line is the ideal grain boundary line when the simulated forging does not have overburning, and the area surrounded by the grain boundary line and the fitting line is the ideal grain area. It can also be understood that when the forging does not have overburning, the grain boundary line and the fitting line are ideal grain boundaries, and the area surrounded is the ideal grain area.

[0073] S4: Calculate the degree of overburning of the forging to be detected. When the degree of overburning is greater than a threshold value, the forging to be detected is overburned.

[0074] The expression of overburning degree is:

[0075]

[0076] In the formula, Indicates the degree of overburning of forgings. represents the number of ideal grain regions, is the number of actual grain regions, Indicates The possibility that the actual grain area is an overburned grain, is the average area of ​​all actual grain regions, Represents the number of all fitted line pixels, Represents the number of all grain boundary pixels.

[0077] When the number of ideal grain regions The actual number of grain areas The smaller the difference, the fewer grain boundaries are melted, and vice versa, the more grain boundaries are melted, causing more grain regions to fuse and the greater the degree of overburning. Reflects the melting ratio of the grain boundary line. The smaller it is, the fewer the melted grain boundaries are and the smaller the degree of overburning is. The larger it is, the more grain boundaries are melted and the greater the degree of overburning. is the average area of ​​all actual grain regions, when The larger it is, the more abnormally grown grains there are, that is, the greater the degree of overburning of the forging. Indicates the average overburning possibility of the actual grain area. The greater the overburning degree, the greater the degree of overburning of the forging. When the overburning degree is greater than the threshold, the forging to be detected is overburned. The threshold is set manually according to the actual situation, and the exemplary threshold is 10.

[0078] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0079] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting overburning of forgings based on microscopic image processing, characterized in that: Includes steps: Acquire a microscopic image of the forging to be inspected, obtain the actual grain region according to the microscopic image, and use the actual edge line of the actual grain region as the grain boundary line of the actual grain region; Calculate the probability that the actual grain area is an overburned grain, and the probability is negatively correlated with the radius shape index of the actual grain area; Perform edge detection on the actual grain region to obtain multiple edge lines, the detected edge lines include grain boundary lines and lines formed by impurities, determine whether adjacent edge lines among the detected edge lines are both grain boundary lines, and if so, fit the two grain boundary lines to obtain a fitting line; and take the area enclosed by the grain boundary lines and the fitting lines as the ideal grain region; The method for fitting two grain boundary lines to obtain a fitting line is as follows: connect the pixel points between the endpoints of the two grain boundary lines to obtain a connecting line, construct multiple rays with the center of the actual grain area as the origin, intersect the rays with the connecting line to obtain an intersection point, calculate the distance between the intersection point and the origin point, cluster the obtained multiple distances to obtain multiple cluster clusters, use the intersection point corresponding to the distance in the cluster with the largest number of distances as the initial intersection point, and use the direct least squares ellipse fitting algorithm to fit the initial intersection point to obtain an ellipse; use any pixel point on the connecting line as the first reference point, and use the pixel point on the ellipse closest to the first reference point as the second reference point; the pixel points on both sides of the first reference point on the connecting line correspond to the pixel points on both sides of the second reference point on the ellipse one by one, use the midpoint of the two corresponding pixel points as the fitting point, and connect the fitting points to obtain a fitting line; Calculate the degree of overburning of the forging to be tested. When the degree of overburning is greater than a threshold value, the forging to be tested is overburned. The expression of overburning degree is: In the formula, Indicates the degree of overburning of forgings. represents the number of ideal grain regions, is the number of actual grain regions, Indicates The possibility that the actual grain area is an overburned grain, is the average area of ​​all actual grain regions, Represents the number of all fitted line pixels, Represents the number of all grain boundary pixels.

2. The forging overburning detection method based on microscopic image processing according to claim 1 is characterized in that: The probability that the actual grain area is an overburned grain is calculated as: The grain boundary line of the actual grain area is detected to obtain the position of the corner point, the distance from each corner point to the center point of the actual grain area is calculated, and the variance of multiple distances is calculated; The expression for the probability is: in, Indicates the possibility that the actual grain area is an overburned grain. represents the area of ​​the actual grain region, represents the perimeter of the grain boundary line in the actual grain area, Indicates the number of corner points on the grain boundary line in the actual grain area, Represents the variance of the distance from the corner point to the center point on the grain boundary line of the actual grain area, The radius shape index representing the actual grain area, Represents the normalization function.

3. The forging overburning detection method based on microscopic image processing according to claim 1 is characterized in that: The probability that the actual grain area is an overburned grain is calculated as: The expression for the probability is: in, Indicates the possibility that the actual grain area is an overburned grain. The radius shape index representing the actual grain area, Represents the normalization function.

4. The forging overburning detection method based on microscopic image processing according to claim 1 is characterized in that: The method for determining whether adjacent edge lines among the detected edge lines are grain boundaries at the same time is as follows: Taking one of the endpoints on the edge line as the starting point, multiple pixel points are intercepted along the edge line direction; Calculate the average gradient direction of multiple intercepted pixel points; The normal direction of the gradient direction is taken as the extension direction, and the extension angle range is constructed based on the extension direction. The extension angle range is (α-β, α+β), where α is the angle between the extension direction and the horizontal direction, and β is the preset angle threshold; In response to an endpoint of another edge line being within the extension angle range, the adjacent edge lines are simultaneously grain boundaries.

5. The forging overburning detection method based on microscopic image processing according to claim 1 is characterized in that: The method to obtain the actual grain area based on the microscopic image is: Obtain a microscopic image of the forging to be inspected, perform binarization on the microscopic image to obtain a binary image, and use any pixel point with a gray value of 255 in the binary image as a reference point; In the microscopic image, the pixel point corresponding to the reference point is taken as the seed point, and the actual grain area is obtained by regional growth with the seed point as the center.

6. The forging overburning detection method based on microscopic image processing according to claim 5 is characterized in that: The pixel point corresponding to the centroid of the area composed of pixels with a gray value of 255 in the binary image is used as the reference point.

7. The forging overburning detection method based on microscopic image processing according to claim 2 is characterized in that: The Harris corner point detection algorithm is used to detect the grain boundary lines in the actual grain area to obtain the positions of the corner points.

Images