A method and system for visual assessment of cable aging degree

Abstract

The application belongs to the technical field of image processing, and particularly relates to a cable aging degree visual evaluation method and system, which comprises the following steps: performing multi-scale analysis on a cable surface image to obtain Hessian matrices and eigenvalues of each pixel point under each scale; calculating a structure strength inhibition term according to two eigenvalues under the same scale; calculating an adaptive line response index according to a spot degree index and neighborhood direction consistency; calculating a final response value of each pixel point under each scale according to the adaptive line response index and the structure strength inhibition term; generating an enhanced image based on the maximum final response value of all pixel points; and performing morphological analysis on the enhanced image to evaluate the cable aging degree. The application can specifically identify and enhance the response of crack cross points, thereby ensuring the integrity of the detected crack network.

Description

Technical Field

[0001] This invention relates to the field of image processing technology. More specifically, this invention relates to a visual assessment method and system for the degree of cable aging. Background Technology

[0002] In the cable manufacturing industry, the aging resistance of a product is a core indicator for measuring cable quality and expected service life. To verify the reliability of newly developed insulation materials or monitor the quality consistency of batch products, cable manufacturers usually need to conduct rigorous accelerated aging tests in laboratory environments, such as thermal aging, ultraviolet aging, or ozone aging chamber tests. In these high-intensity simulated test environments, cable sheath materials will undergo accelerated degradation of physicochemical properties, and the most intuitive and critical failure symptom is the gradual hardening, embrittlement, and microcrack formation of the material surface.

[0003] As aging tests continue, the microcracks on the sheath surface rapidly expand and intertwine, forming a complex network crack structure. In experimental analysis, the ability to accurately quantify the density, average length, and network topology of these cracks (such as the number of crack intersections) is a key physical basis for researchers to evaluate the quality of material formulations and determine the weather resistance level of products.

[0004] To replace inefficient and subjective manual visual inspection, the industry is gradually adopting automated evaluation technology based on machine vision. Among them, the Hessian-Frangi algorithm, as a classic algorithm for enhancing blood vessels or linear structures, is often used for the automatic extraction of such cracks because it can effectively filter tubular or linear features using the eigenvalues ​​of the Hessian matrix.

[0005] However, the existing Hessian-Frangi algorithm has significant technical flaws when dealing with high-density crack networks generated by laboratory aging tests. The algorithm is designed to maximize the response of linear structures while suppressing the response of speckled structures. However, in the complex network formed by cable aging cracks, there are a large number of X-shaped or Y-shaped intersections. At these intersections, the local geometry of the image no longer presents an ideal unidirectional linearity, but exhibits isotropic characteristics similar to speckles. This causes the Hessian-Frangi algorithm to mistakenly identify these key physical connection points as noise or background speckles and suppress them, resulting in a sharp drop in response values ​​or even zero. This mismatch in the algorithm mechanism will incorrectly divide the crack network, which should be complete and connected, into a large number of unrelated short segments, resulting in a shorter calculated average crack length and an underestimated network complexity. This seriously affects the accuracy of manufacturers' assessment of cable aging and the reliability of experimental data. Summary of the Invention

[0006] To address the technical problem in existing technologies that misidentify crack intersections as spot-like structures and suppress them during linear structure detection, leading to crack network breakage and inaccurate assessment, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a visual assessment method for cable aging, comprising: acquiring cable surface images; performing multi-scale analysis on the cable surface images to obtain the Hessian matrix and eigenvalues ​​of each pixel at each scale; calculating the final response value of each pixel at each scale, wherein the final response value is positively correlated with the product of an adaptive linear response index and a structural strength suppression term; the adaptive linear response index includes a compensation term, wherein the compensation term is positively correlated with a speckle index and negatively correlated with neighborhood orientation consistency; the speckle index is the ratio of the absolute values ​​of two eigenvalues ​​at the same scale; the neighborhood orientation consistency is positively correlated with the mean cosine similarity of the second eigenvector of the central pixel and the second eigenvectors of each neighboring pixel within a neighborhood window centered on the corresponding pixel; the structural strength suppression term is positively correlated with effective line strength, wherein the effective line strength is the difference between the absolute values ​​of two eigenvalues ​​at the same scale; generating an enhanced image based on the maximum final response value of all pixels; and performing morphological analysis on the enhanced image to assess the cable aging degree.

[0008] This invention introduces an adaptive linear response exponent with a compensation term and a structural strength suppression term based on the difference of eigenvalues. This allows for a compensated response when the neighborhood direction consistency of pixels is low and the speckle density is high, i.e., pixels have typical characteristics of intersection points. This effectively maintains the continuity of complex crack networks at intersection points. At the same time, noise is suppressed through a new definition of effective line strength. The resulting enhanced image can more completely and realistically reflect the aging crack morphology of the cable surface, laying a solid foundation for subsequent accurate quantitative evaluation.

[0009] Preferably, the adaptive linear response exponent is calculated as follows: In the formula, For pixels at scale The adaptive linear response exponent under these conditions; For pixels at scale The neighborhood direction is consistent; For pixels at scale The spot density index below; It is a natural exponential function; The compensation coefficient; These are control parameters.

[0010] Unlike existing technologies that rely solely on Gaussian functions to suppress speckle responses, this invention adds a compensation term proportional to neighborhood direction consistency and speckle density on top of the original exponential function term. This makes the compensation term close to zero when processing linear structures, without affecting the original performance. However, when processing intersections, the compensation term provides a significant positive compensation value, thereby accurately enhancing the intersection signal and ensuring the topological integrity of the cracked network.

[0011] Preferably, the formula for calculating the effective linear strength is: In the formula, For pixels at scale The effective linear strength below, , Each pixel is at a different scale. The two eigenvalues ​​below, This indicates taking the absolute value. This indicates taking the maximum value.

[0012] Existing technologies typically use the sum of squares of eigenvalues ​​to measure structural strength. This method amplifies the strength of both linear and speckled structures equally, making it difficult to distinguish between them. This invention defines effective linear strength by calculating the difference between the absolute values ​​of two eigenvalues. For ideal linear structures, this difference is large, indicating high strength; for speckled or noise structures, this difference approaches zero, indicating low strength. This fundamentally enhances the preference for linear cracks while effectively suppressing the interference of speckled noise, thus improving the signal-to-noise ratio.

[0013] Preferably, the formula for calculating the structural strength suppression term is: In the formula, For pixels at scale The structural strength suppression term below; For pixels at scale Effective linear strength below; It is a natural exponential function; These are control parameters.

[0014] This invention utilizes effective line strength to construct a nonlinear suppression function, which can strongly suppress pixels with low effective line strength, i.e., pixels belonging to background noise, while retaining the response of pixels with high effective line strength, i.e., pixels belonging to cracks. Compared with simple linear thresholding, this method is smoother and more robust, effectively improving the algorithm's noise resistance and avoiding the erroneous amplification of non-crack noise points by the compensation term.

[0015] Preferably, morphological analysis of the enhanced image includes: binarizing the enhanced image using the Otsu algorithm to obtain a binary crack network image, wherein white pixels are crack pixels; and skeletonizing the white pixels in the binary crack network image to obtain the center line of a single crack pixel.

[0016] This invention transforms the enhanced grayscale image into a crack centerline with a width of one pixel, providing a standardized data foundation for subsequent accurate calculation of key aging indicators such as crack density, length, and intersection points.

[0017] Preferably, the indicators for assessing the degree of cable aging include crack density, average crack length, and intersection density.

[0018] Preferably, the crack density is equal to the ratio of the number of pixels contained in the center line of a single crack pixel to the total number of pixels in the cable surface image.

[0019] Preferably, the intersection density is equal to the ratio of the number of all branch points to the total number of pixels in the cable surface image, and the number of white pixels contained in the 8-neighborhood of each branch point is greater than or equal to 3.

[0020] Preferably, the method for obtaining the average crack length is as follows: count the lengths of all independent crack segments in the center line of a single pixel of the crack, and calculate the average value of all lengths as the average crack length; an independent crack segment is a line segment between two endpoints or branch points, and the length of an independent crack segment is the number of pixels contained in the line segment between two endpoints or branch points.

[0021] Secondly, the present invention provides a visual assessment system for cable aging, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned visual assessment method for cable aging is implemented.

[0022] By adopting the above technical solution, a computer program is generated from the above-mentioned visual assessment method for cable aging, and stored in a memory for loading and execution by a processor. Terminal devices are then manufactured based on the memory and processor for convenient use.

[0023] The beneficial effects of this invention are as follows:

[0024] This invention introduces an adaptive linear response exponent with a compensation term and a structural strength suppression term based on the difference of eigenvalues. This allows for a compensated response when the neighborhood direction consistency of pixels is low and the speckle density is high, i.e., pixels have typical characteristics of intersection points. This effectively maintains the continuity of complex crack networks at intersection points. At the same time, noise is suppressed through a new definition of effective line strength. The resulting enhanced image can more completely and realistically reflect the aging crack morphology of the cable surface, laying a solid foundation for subsequent accurate quantitative evaluation. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a visual assessment method for cable aging according to the present invention;

[0026] Figure 2 This is a schematic diagram showing an image of the cable surface;

[0027] Figure 3 This schematically illustrates the application of existing technology, namely the standard Hessian-Frangi algorithm, to... Figure 2 The enhanced image is obtained through processing;

[0028] Figure 4 This is an illustrative representation of the method of the present invention for... Figure 2 The enhanced image is obtained through processing;

[0029] Figure 5 It is shown schematically. Figure 3 Binary crack network image obtained after binarization processing;

[0030] Figure 6 It is shown schematically. Figure 4 Binary crack network image obtained after binarization processing. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0033] This invention discloses a visual assessment method for the degree of cable aging, referring to... Figure 1 This includes steps S1-S5:

[0034] S1: Acquire images of the cable surface.

[0035] Specifically, an industrial camera is used to vertically photograph the surface of the cable sheath to obtain the original image of the cable surface, which is a grayscale image.

[0036] S2: Perform multi-scale analysis on the cable surface image to obtain the Hessian matrix and eigenvalues ​​of each pixel at each scale.

[0037] It should be noted that since the crack thickness may be uneven at different locations, analysis at multiple scales can ensure the detection of cracks of different widths; the Hessian matrix is ​​a second-order partial derivative matrix used to describe the local geometry within the neighborhood of a pixel.

[0038] Specifically, across a range of scales Below, using Gaussian second-order partial derivatives , and Convolution is performed on each pixel in the cable surface image to obtain the scale of each pixel in the cable surface image. Hessian matrix under Among them, pixels In scale Hessian matrix under The formula for calculation is:

[0039]

[0040] In the formula, Gaussian second-order partial derivative For pixels The convolution result, Gaussian second-order partial derivative For pixels The convolution result, , All are Gaussian second-order partial derivatives For pixels The convolution result.

[0041] In this embodiment, scale The increment is from 0.5 to 5, with a step size of 0.5.

[0042] Thus, through multi-scale analysis, the algorithm is able to respond to cracks of different widths, whether they are fine early cracks or coarse main cracks.

[0043] Furthermore, the eigenvalues ​​of the Hessian matrix directly reflect the morphology of the local structure; therefore, for each pixel at the scale... Hessian matrix under Perform eigenvalue decomposition to obtain the pixel's scale. The two eigenvalues ​​below and Sort the eigenvalues ​​so that Simultaneously, calculate eigenvalues. The corresponding second feature vector Thus, the eigenvalues ​​are obtained. and It is the foundation for describing the local geometry of a pixel and a prerequisite for calculating all subsequent metrics.

[0044] In this embodiment, cracks are represented as dark lines with low grayscale values ​​in the image, corresponding to... It should be a positive value, therefore, if If the pixel is not considered a crack, it may be a bright background or a flat area, and its subsequent response is uniformly set to 0.

[0045] S3: Calculate the blob index and effective line intensity based on two feature values ​​at the same scale; calculate the neighborhood orientation consistency based on the cosine similarity between the second feature vector of the center pixel and the second feature vector of each neighboring pixel within the neighborhood window centered on the corresponding pixel.

[0046] First, calculate the spot size index. Spot density index It is one of the core metrics of the Hessian-Frangi algorithm, used to distinguish between linear and speckled structures; the eigenvalues ​​corresponding to linear structures. Eigenvalues ​​approaching 0, corresponding to the speckled structure and similar.

[0047] Specifically, based on eigenvalues and Calculate the spot density index The calculation formula is:

[0048]

[0049] In the formula, For pixels at scale The following spot density index, , Each pixel is at a different scale. The two eigenvalues ​​below, This indicates taking the absolute value.

[0050] Among them, the eigenvalues ​​corresponding to the linear structure The eigenvalues ​​approach 0; therefore, the speckle index corresponding to linear structures is close to 0, and the eigenvalues ​​corresponding to speckled structures are close to 0. and Therefore, the speckle intensity index corresponding to the speckled structure is close to 1; in summary, the speckle intensity index... A value close to 0 indicates that the pixel has very strong linear features, which is the speckle index. A value close to 1 indicates that the pixel has a very strong speckled feature.

[0051] Thus, by calculating the speckle index This achieves low linear regions and high values Preliminary differentiation of the intersection region of values.

[0052] It should be noted that the original structural strength indicators This will incorrectly amplify the intensity of speckled noise, therefore, through a bus-like response Compared with total speckled response The difference is used to construct a new effective line strength index. This indicator fundamentally reconstructs the definition of strength, favoring linear crack structures while suppressing the granular feel of speckled materials.

[0053] Specifically, based on eigenvalues and Calculate the effective linear strength The calculation formula is:

[0054]

[0055] In the formula, For pixels at scale The effective linear strength below, , Each pixel is at a different scale. The two eigenvalues ​​below, This indicates taking the absolute value. This indicates taking the maximum value.

[0056] For an ideal line segment, the bus-like response is... Much larger than the total speckled response Effective linear strength Approaching High intensity; bus-like response for ideal speckle or noise. and total speckled response Similar, effective linear strength As the value approaches zero, the intensity is suppressed.

[0057] It should be noted that this indicator fundamentally reconstructs the definition of "strength," making it naturally favor linear structures (cracks) while suppressing speckled structures (material graininess).

[0058] Finally, in the scenario of cable aging analysis, the true definition of a crack intersection point is not a spot, but a point where multiple lines converge in different directions. Therefore, the eigenvectors of the Hessian matrix are introduced for differentiation. The second eigenvector of the Hessian matrix... Physically, it points to the normal direction of the linear structure and has directionality; for a line segment structure: on a line segment, the second feature vector of all pixels... The resulting vector fields are highly consistent, pointing roughly in the same direction; for the intersection structure: within the neighborhood of an intersection, the second eigenvector The resulting vector field is chaotic and conflicting because multiple linear structures in different directions converge here.

[0059] Specifically, taking any pixel as the center pixel, we obtain its size as... The neighborhood window is used to calculate the relationship between the center pixel and each neighboring pixel in the window at different scales. The cosine similarity of the angle between the second feature vectors is calculated, and the average of the absolute values ​​of all cosine similarities is used as the pixel's scale. Neighborhood direction consistency If the neighborhood directions are consistent A value close to 1 indicates that the pixel belongs to a line segment with a consistent height. If the neighborhood direction is consistent... A value close to 0 indicates that the pixel is either a crossroads with conflicting directions or noise.

[0060] S4: Calculate the adaptive linear response index based on the speckle index and neighborhood orientation consistency; calculate the structural strength suppression term based on the effective linear strength, and multiply the adaptive linear response index by the structural strength suppression term to obtain the final crack response value.

[0061] Specifically, the adaptive linear response index is calculated based on the speckle count index and neighborhood orientation consistency. The calculation formula is:

[0062]

[0063] In the formula, For pixels at scale The adaptive linear response exponent under these conditions; For pixels at scale The neighborhood direction is consistent; For pixels at scale The spot density index below; It is a natural exponential function; This is the compensation coefficient, used to adjust the compensation intensity at the intersection. These are control parameters used to control the linear response term. Sensitivity, in this embodiment Take 0.5.

[0064] Among them, the compensation coefficient The value range is between [0.1, 1.0]. Too small, insufficient compensation; if Excessive size may amplify noise; in this embodiment, it is preferable to... Set it to 0.5.

[0065] It should be noted that traditional algorithms use For the intersection point, its speckle index As the value approaches 1, the term approaches 0, leading to a suppressed response; therefore, a blob index is introduced here. Proportional compensation term: For the linear part, the compensation term is almost 0 and has basically no effect. For the intersection part, the compensation term provides a significant compensation value, enabling the algorithm to identify the intersection and provide a compensation channel for it, ensuring that the point is not lost in the final response.

[0066] It should be further explained that by introducing a neighborhood orientation consistency index and its nonlinear multiplicative constraint relationship with the speckle count index, compensation is made only for the true intersections of conflicting orientation fields, rather than compensating for all specks: for a line segment, its neighborhood orientation consistency... Approaching 1, the spot size index Near to 0, adaptive linear response exponent Approaching 1 indicates a strong response; for intersections, the neighborhood directions are consistent. Approaching 0, the spot size index Near to 1, adaptive linear response exponent Approaching This is a compensation response; for speckle noise, its neighborhood direction consistency is... Approaching 0, the spot size index It is close to 1, but its structural strength index It will be very low, and therefore suppressed.

[0067] Furthermore, based on the effective linear strength, the structural strength suppression term is calculated. To suppress the background noise response, the calculation formula is:

[0068]

[0069] In the formula, For pixels at scale The structural strength suppression term below; For pixels at scale Effective linear strength below; It is a natural exponential function; These are control parameters used to control the sensitivity to effective line strength. The parameters, and control parameters Equal to all pixels at scale eigenvalues ​​under absolute value Half of the maximum value.

[0070] Thus, the structural strength suppression term ensures the algorithm's noise resistance and avoids the compensation term from erroneously amplifying non-crack noise points.

[0071] Finally, the pixels are scaled. The adaptive linear response exponent is multiplied by the structural strength suppression term to obtain the pixel's scale. Final crack response value Thus, the final crack response value obtained simultaneously takes into account linear response, intersection point compensation, and noise suppression, forming a complete crack detection logic under a single scale.

[0072] S5: Generate an enhanced image based on the maximum final response value of all pixels; perform morphological analysis on the enhanced image to assess the degree of cable aging.

[0073] Because cracks vary in thickness, a single pixel may respond at multiple scales. Therefore, only the scale with the strongest response is taken as the final result. Specifically, for any pixel, its final crack response value at all scales is calculated, and the maximum value among all final crack response values ​​is taken as the enhancement result for that pixel. Finally, the enhancement results of all pixels are combined to form an enhanced image.

[0074] In this way, multi-scale fusion ensures that cracks of different thicknesses can be clearly detected at their optimal scale and uniformly presented on the enhanced image.

[0075] For example, targeting Figure 2 Image of the cable surface shown:

[0076] (1) Using existing technology, namely the standard Hessian-Frangi algorithm, to... Figure 2 The enhanced image obtained after processing is as follows: Figure 3 As shown, in areas with dense crack networks, especially where X-shaped or Y-shaped intersections are formed, the image brightness is significantly darkened or even broken. This is because at the intersection points, the local geometry exhibits isotropic speckle characteristics. The standard Hessian-Frangi algorithm has a high speckle index, and the algorithm mistakenly identifies it as noise and suppresses it. This leads to the crack network being incorrectly cut into multiple short lines.

[0077] (2) The method of the present invention is used to... Figure 2 The enhanced image obtained after processing is as follows: Figure 4 As shown, the intersections of the cracks were significantly enhanced and preserved, the overall brightness was uniform, and the cracks exhibited a complete network topology. This is because the method introduces neighborhood direction consistency. At the intersections, although the speckle density is high, the neighborhood direction consistency is low, i.e., direction conflict. By using the compensation term in the calculation formula of the adaptive linear response index, positive response compensation is provided for these key connection points, thereby repairing the breakpoints.

[0078] Furthermore, the Otsu algorithm is used to binarize the enhanced image to obtain a binary crack network image, where white pixels represent crack pixels. The Otsu algorithm is a segmentation algorithm that automatically determines the threshold. Its core idea is to find the gray value that maximizes the inter-class variance between the foreground (crack) and background (non-crack) pixels as the threshold.

[0079] For example, targeting Figure 3 The enhanced image obtained by existing technology, as shown, is a binary crack network image obtained by binarization processing. Figure 5 As shown, although the linear part of the main crack was extracted, obvious breaks appeared at the key intersections of the crack network, such as the Y-shaped branch in the center of the image. This is because the standard algorithm misclassifies physically connected intersection nodes as isotropic speckle noise and suppresses them, resulting in the complete crack topology being incorrectly divided into unconnected independent line segments, which cannot truly reflect the severity of cable aging.

[0080] For example, targeting Figure 4 The enhanced image obtained by the method of the present invention, as shown, is a binary crack network image obtained by binarization processing. Figure 6 As shown; with Figure 5 In comparison, this image clearly reconstructs the complete crack network, especially at the crack intersections where effective connections and preservation are achieved, significantly improving the overall continuity of the lines. This is thanks to the fact that this solution introduces a neighborhood direction consistency index to accurately identify and compensate for the response signals at the intersections, successfully avoiding the generation of breakpoints, thereby accurately quantifying the topological complexity and physical extension length of the microcracks on the cable surface.

[0081] Morphological feature analysis of binary crack network images can yield quantitative indicators of cable aging. Specifically, skeletonization is performed on white pixels in the binary crack network image to obtain the center line of a single crack pixel. The skeletonization method includes, but is not limited to, the Zhang-Suen algorithm. Skeletonization is an operation that reduces a binary target to its topological center line, i.e., the width of a single pixel.

[0082] Finally, based on the statistical analysis of the center line of each crack pixel, the following indicators were obtained:

[0083] (1) Crack density It is equal to the ratio of the number of pixels contained in the center line of a single crack to the total number of pixels in the cable surface image.

[0084] (2) Average crack length The method for obtaining the length is as follows: count the length of all independent crack segments in the center line of a single crack pixel, and calculate the average of all lengths as the average crack length; an independent crack segment is a line segment between two endpoints or branch points, and the length of an independent crack segment is the number of pixels contained in the line segment between the two endpoints or branch points.

[0085] (3) Crosspoint density : Count the number of all branch points in the center line of a single pixel of a crack, where the number of white pixels within the 8-neighborhood of each branch point is greater than or equal to 3; calculate the ratio of the total number of branch points to the total number of pixels in the cable surface image, as the intersection point density. .

[0086] This invention also discloses a visual assessment system for cable aging, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a visual assessment method for cable aging according to the present invention.

[0087] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A method for visually assessing the degree of aging of a cable, characterized in that, The method comprises the following steps: collecting a cable surface image; performing multi-scale analysis on the cable surface image to obtain Hessian matrix and eigenvalues of each pixel point at each scale; calculating the final response value of each pixel point at each scale, wherein the final response value is positively correlated with the product of adaptive linear response index and structure strength inhibition term; the adaptive linear response index comprises a compensation term, which is positively correlated with a spot degree index and negatively correlated with neighborhood direction consistency, and the calculation formula of the adaptive linear response index is: ； wherein, is an adaptive line response index of the pixel point at scale ; is a neighborhood direction consistency of the pixel point at scale ; is a blob degree index of the pixel point at scale ; is a natural exponential function; is a compensation coefficient; is a control parameter; the calculation formula of the structure strength inhibition term is: ； wherein is a structure strength suppression term for the pixel point at scale ; is an effective line strength for the pixel point at scale ; is a natural exponential function; is a control parameter; the spot degree index is the ratio of the absolute values of two eigenvalues at the same scale; the neighborhood direction consistency is positively correlated with the average value of the cosine similarity between the second feature vector of the center pixel point and the second feature vector of each neighborhood pixel point in the neighborhood window with the corresponding pixel point as the center; the structure strength inhibition term is positively correlated with effective line intensity, and the effective line intensity is the difference between the absolute values of two eigenvalues at the same scale; generating an enhanced image based on the maximum final response value of all pixel points; performing morphological analysis on the enhanced image to evaluate the cable aging degree, comprising: performing binaryzation processing on the enhanced image using Otsu algorithm to obtain a binary crack network image, wherein the white pixel points are crack pixel points; performing skeletonization processing on the white pixel points in the binary crack network image to obtain a crack single-pixel center line.

2. The method of claim 1, wherein the method further comprises: the calculation formula of the effective line intensity is: ； In the formula, is the effective line intensity of the pixel point at scale , , are two eigenvalues of the pixel point at scale , denotes taking the absolute value, denotes taking the maximum value.

3. The method of claim 1, wherein the method further comprises: the evaluation index of the cable aging degree comprises crack density, crack average length and intersection point density.

4. The method of claim 3, wherein the step of determining the degree of aging of the cable is performed by a human operator. the crack density is equal to the ratio of the number of pixel points contained in the crack single-pixel center line to the total number of pixel points in the cable surface image.

5. The method of claim 3, wherein the step of determining the degree of aging of the cable is performed by a human operator. the intersection point density is equal to the ratio of the number of all branch points to the total number of pixel points in the cable surface image, wherein the number of white pixel points contained in the 8-neighborhood of the branch point is greater than or equal to 3.

6. The method of claim 3, wherein the step of visually evaluating the degree of cable aging is performed by a human operator. the crack average length is obtained by: counting the lengths of all independent crack segments in the crack single-pixel center line and calculating the average value of all lengths as the crack average length; the independent crack segment is a line segment between two end points or branch points, and the length of the independent crack segment is the number of pixel points contained in the line segment between the two end points or branch points.

7. A cable aging degree visual assessment system characterized by, The method comprises the following steps: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a cable aging degree visual evaluation method according to any one of claims 1-6 is realized.

Images