Image-based mechanical terminal crack detection method, medium and equipment

Abstract

The invention relates to an image-based mechanical terminal crack detection method, which comprises the following steps of S100, shooting a first image of a mechanical terminal through visual acquisition equipment, determining an incident component and a reflection component according to the positions of the visual acquisition equipment and the mechanical terminal, and processing noisy points through Fourier transform to obtain a second image; s200, performing wavelet decomposition, morphological analysis and wavelet transformation on the second image to obtain an image displaying cracks, and segmenting the image according to the number of the cracks to obtain a third image; and S300, judging the damage degree of the crack for the third image through the fractal dimension. According to the mechanical terminal crack detection method based on the image, the accuracy and efficiency of crack detection are effectively improved by combining multiple image processing technologies.

Description

Technical Field

[0001] The present invention relates to the technical field of image detection, and in particular to an image-based mechanical terminal crack detection method, medium and equipment. Background Art

[0002] Mechanical terminals are key components in cable and other connection situations. Their safety and reliability directly affect the normal operation of the entire electrical system. Cracks are one of the common forms of damage to mechanical terminals. If they are not discovered and handled in time, they may lead to electrical failures or even safety accidents. Traditional crack detection methods usually rely on manual inspection, which is not only inefficient, but also easily affected by human factors, making it difficult to ensure the accuracy and consistency of detection.

[0003] With the continuous development of image processing technology, image-based crack detection methods have gradually become an effective alternative. This method collects images of mechanical terminals and uses image processing algorithms to analyze and process the images to automatically identify and detect cracks. However, existing image-based crack detection methods still face certain challenges when dealing with complex backgrounds and noise interference, making it difficult to improve detection efficiency while ensuring detection accuracy. Summary of the invention

[0004] The purpose of the present invention is to provide an image-based mechanical terminal crack detection method, medium, and device to solve the above-mentioned technical problems.

[0005] To achieve the above object, the present invention provides the following technical solutions: A method for detecting cracks in mechanical terminals based on an image, characterized in that it comprises the following steps: S100, taking a first image of the mechanical terminal by a visual acquisition device, determining an incident component and a reflected component according to the positions of the visual acquisition device and the mechanical terminal, and processing noise by Fourier transform to obtain a second image; S200, performing wavelet decomposition, morphological analysis and wavelet transform on the second image to obtain an image showing cracks, and segmenting the image according to the number of cracks to obtain a third image; S300: judging the damage degree of the crack by using the fractal dimension of the third image.

[0006] The present invention further arranges that the steps of obtaining the second image by processing the noise through the illumination reflection model and Fourier transform in step S100 are as follows: S1001, performing logarithmic transformation on the first image; S1002, transforming the logarithmically transformed image from the spatial domain to the frequency domain through Fourier transformation; S1003, constructing a filter K (p, b) to filter and obtain an enhanced visual image; S1004, using Fourier transform again to convert the visual image in step S1003 from the frequency domain to the spatial domain; S1005, performing exponential transformation on the image in S1004 to obtain a second image; The present invention further arranges that, S1001, when performing logarithmic transformation on the first image, the formula is as follows: lng(x,y)=lnj(x,y)+ lne(x,y); Where: g(x,y) is the light intensity function, j(x,y) is the incident component, e(x,y) is the reflected component, and 0<j(x,y)<∞, 0<e(x,y)<1; When the logarithmically transformed image is converted from the spatial domain to the frequency domain by Fourier transform in S1002, the following formula is used: G[ln g(x,y)]=G[ln j(x,y)]+G[ln e(x,y)]; Convert the above formula to: G(p,b) = J(p,b) + E(p,b); Where: J(p,b), E(p,b) and G(p,b) are the Fourier transforms of ln j(x,y), ln e(x,y) and ln g(x,y); E(p,b) represents the spectrum function, which is concentrated in the high frequency band; J(p,b) represents the spectrum function, which is concentrated in the low frequency band; In S1003, the high-frequency spectrum function E(p, b) is enhanced by the filter K(p, b) to obtain an enhanced visual image A(p, b): A(p,b) =K(p,b)G(p,b) =K(p,b)J(p,b) + K(p,b)E(p,b); When Fourier transform is used in S1004 to transform the image from the frequency domain to the spatial domain, k ( x , y ) is set to A ( p , b ) is the inverse Fourier transform of: k ( x , y ) = G -1 [ A ( p , b )]= kj ( x , y ) + ke ( x ,y ); Where kj(x,y) and ke(x,y) represent the transformation of the filter in the incident component and the reflected component; G -1 stands for Inverse Fourier Transform; In S1005, the above formula is exponentially modified to obtain the processed second image: h(x,y) = exp| k(x,y) | .

[0007] The present invention is further configured to obtain high-frequency detail components when performing wavelet decomposition on the second image, and the formula is as follows: I0 (x,y) = 2^(1 / 2) · {I H [h(x,y)] ×I N [h(x,y)]+ I D [h(x,y)]}; I1 (x,y) = I(x,y) + I H [h(x,y)]+ I N [h(x,y)]+ I D [h(x,y)]; In the formula, I H represents the high-frequency horizontal component; I D represents the high-frequency diagonal component; I N represents the high-frequency vertical component; H , D and N Represent the horizontal, diagonal and vertical directions respectively; I 0 ( x , y ) represents the low-frequency approximate component; I 1 ( x , y ) represents the high-frequency detail component.

[0008] The present invention is further configured to perform morphological analysis on the high-frequency component, corrode the high-frequency detail component through the structural element to obtain the crack boundary, and fill the internal area of ​​the obtained crack boundary in the form of expansion after obtaining the crack boundary; when obtaining the crack boundary, the high-frequency detail component is subtracted from the corrosion to obtain the boundary, and the formula is as follows: I(x,y)= I1 (x,y)- [I1 (x,y)ΔS]; In the above formula, S represents a structural element; △ represents a difference operation; The formula for expansion is: I i O(x,y)= (I i O-1 (x,y)⊙S)∩I e’ ′ (x,y); where I i (x,y)∈I(x,y); In the above formula, I e’ ′ (x,y) represents the complement of the high-frequency component; O represents the number of iterations; ⊙ represents the XOR operation.

[0009] The present invention further arranges that after dilation, an operation of obtaining connected components is performed on the image, and the formula is as follows: i O (x,y)=(I i O-1 (x,y)⊙S)∩I e’ (x,y),I e’ (x,y) is the high frequency component, when I i O (x,y) = I i O (x,y)-1, get the connected component: Y=OI i O (x, y), and then refine the image through the structural element, the formula is: Y'=Y-((I e’ ′ (x,y) △S1)∩(I e’ (x,y) △S2)), S1, S2∈S; where S1 is a set of elements associated with the object; S2 is a set of elements related to the corresponding background; repeat the above process until the refined image of Y' no longer changes.

[0010] The present invention is further configured to extract morphological skeleton features of cracks from the refined image.

[0011] The present invention is further configured to judge the degree of damage of the third image by fractal dimension, and the method is as follows: step 1: establish an a×b matrix A; step 2: fill the matrix A with edge pixel values ​​showing a crack image with a size of (a+l)×(b+l), select a pixel point containing a×b from the image with a size of (a+l)×(b+l) and select a rectangular area as an area in the image; step 3: divide each rectangular area into s×s areas again, s selects 2 1 , 2 2 and 2 3, and calculate r at different scales, r = s / a; Step 4: At each scale r, calculate n(i,j) = q-v+1 for each grid, where q and v are the box numbers of the maximum and minimum grayscale values ​​of the grid, respectively, and then move the grid; Step 5: After calculating n(i,j) for each grid, find b= ; Step 6: Use the least squares method to fit -log(r) and log(b) to calculate the slope, which is used as the fractal dimension D of the crack image in this area; Step 7: Traverse the pixel points in A row by row, repeat steps 2 to 6, calculate the fractal dimension D of each rectangular area, calculate the average value of the fractal dimension D of each rectangular area as the fractal dimension of the crack image, and judge the degree of damage according to the size of the fractal dimension.

[0012] The present invention also provides a computer device, which includes a memory and a processor. The memory is used to store computer instructions, and the processor executes the image-based mechanical terminal crack detection method by executing the computer instructions.

[0013] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the image-based mechanical terminal crack detection method.

[0014] Beneficial effects of the invention: The image-based mechanical terminal crack detection method of the invention effectively improves the accuracy and efficiency of crack detection by combining multiple image processing technologies. Specifically, the method has the following advantages: using the illumination reflection model and Fourier transform to process noise points, effectively reducing the impact of background noise on crack detection; through wavelet decomposition and morphological analysis, it can clearly display the crack image and accurately segment the crack area; using fractal dimension to judge the damage degree of the crack, it provides an effective means for quantitative evaluation of the crack. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Picture 1 Schematic diagram of a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will describe the implementation methods of the present application in detail with the help of accompanying drawings and examples, so that the implementation process of how the present application uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly.

[0017] Example 1

[0018] like Picture 1 As shown, a method for detecting cracks in mechanical terminals based on an image comprises the following steps: S100, taking a first image of the mechanical terminal by a visual acquisition device, and processing the noise of the first image by an illumination reflection model and Fourier transform to obtain a second image, wherein the illumination reflection model has an incident component and a reflection component, and the incident component and the reflection component are determined according to the positions of the visual acquisition device and the mechanical terminal; S200, performing wavelet decomposition, morphological analysis and wavelet transform on the second image to obtain an image showing cracks, and segmenting the image according to the number of cracks to obtain a third image; S300: judging the damage degree of the crack by using the fractal dimension of the third image.

[0019] The steps of obtaining the second image by processing the noise through the illumination reflection model and Fourier transform in step S100 are as follows: S1001, performing logarithmic transformation on the first image; S1002, transforming the logarithmically transformed image from the spatial domain to the frequency domain through Fourier transformation; S1003, constructing a filter K (p, b) to filter and obtain an enhanced visual image; S1004, using Fourier transform again to convert the visual image in step S1003 from the frequency domain to the spatial domain; S1005, performing exponential transformation on the image in S1004 to obtain a second image; The specific steps of obtaining the second image are as follows: S1001. When logarithmic transformation is performed on the first image according to the illumination reflection model g(x,y)= j(x,y) ×e(x,y), the formula is as follows: lng(x,y)=lnj(x,y)+ lne(x,y); Where: g(x,y) is the light intensity function, j(x,y) is the incident component, e(x,y) is the reflected component, and 0<j (x,y)<∞, 0<e(x,y)<1; When the logarithmically transformed image is converted from the spatial domain to the frequency domain by Fourier transform in S1002, the following formula is used: G[ln g(x,y)]=G[ln j(x,y)]+G[ln e(x,y)]; Convert the above formula to: G(p,b) = J(p,b) + E(p,b); Where: J(p,b), E(p,b) and G(p,b) are the Fourier transforms of ln j(x,y), ln e(x,y) and ln g(x,y); E(p,b) represents the spectrum function, which is concentrated in the high frequency band; J(p,b) represents the spectrum function, which is concentrated in the low frequency band; In S1003, the high-frequency spectrum function E(p, b) is enhanced by the filter K(p, b) to obtain an enhanced visual image A(p, b): A(p,b) =K(p,b)G(p,b) =K(p,b)J(p,b) + K(p,b)E(p,b); S1004, when Fourier transform is used to transform the image from the frequency domain to the spatial domain, k ( x , y ) is set to A ( p , b ) is the inverse Fourier transform of: k ( x , y ) = G -1 [ A ( p , b )]= kj ( x , y ) + ke ( x , y ); Where kj(x,y) and ke(x,y) represent the transformation of the filter in the incident component and the reflected component; G -1 stands for Inverse Fourier Transform; In S1005, the above formula is exponentially modified to obtain the processed second image: h(x,y) = exp| k(x,y) | .

[0020] When the second image is subsequently decomposed by wavelet, the high-frequency detail component is obtained, and the formula is as follows: I0 (x,y) = 2^(1 / 2) · {I H [h(x,y)] ×I N [h(x,y)]+ I D [h(x,y)]}; I1 (x,y) = I(x,y) + I H [h(x,y)]+ I N [h(x,y)]+ I D[h(x,y)]; In the formula, I H represents the high-frequency horizontal component; I D represents the high-frequency diagonal component; I N represents the high-frequency vertical component; H , D and N Represent the horizontal, diagonal and vertical directions respectively; I 0 ( x , y ) represents the low-frequency approximate component; I 1 ( x , y ) represents the high-frequency detail component.

[0021] In the present invention, the low-frequency component is omitted, and only the high-frequency component is subjected to morphological analysis. The high-frequency detail component is corroded by the structural element to obtain the crack boundary. After the crack boundary is obtained, the internal area of ​​the obtained crack boundary is filled in the form of expansion; when obtaining the crack boundary, the high-frequency detail component is subtracted from the corrosion to obtain the boundary, and the formula is as follows: I(x,y)= I1 (x,y)- [I1 (x,y)ΔS]; In the above formula, S represents a structural element; △ represents a difference operation; The formula for expansion is: I i O (x,y)= (I i O-1 (x,y)⊙S)∩I e’ ′ (x,y); where I i (x,y)∈I(x,y); In the above formula, I e’ ′ (x,y) represents the complement of the high-frequency component; O represents the number of iterations; ⊙ represents the XOR operation.

[0022] The present invention further arranges that after dilation, an operation of obtaining connected components is performed on the image, and the formula is as follows: i O (x,y)=(I i O-1 (x,y)⊙S)∩I e’ (x,y),I e’ (x,y) is the high frequency component, when I i O (x,y) = I i O(x,y)-1, get the connected component: Y=OI i O (x, y), and then refine the image through the structural element, the formula is: Y'=Y-((I e’ ′ (x,y) △S1)∩(I e’ (x,y) △S2)), S1, S2∈S; where S1 is a set of elements associated with the object; S2 is a set of elements related to the corresponding background; repeat the above process until the refined image of Y' no longer changes.

[0023] The morphological skeleton features of the cracks are extracted from the refined image. The specific method is as follows: the refined image is eroded again multiple times, Y'' = (… (Y'△S o' ) △S o' + 1 ) △…) △S O , S O Represents the Oth corrosion, and the morphological skeleton feature element after corrosion is: Y o '' = Y'' - Y''◎So, ◎ represents the Hadamard product, and the morphological skeleton features are:

[0024] Where O is max{o|I e' (x, y) △oS ≠∅}; The damage degree of the third image is judged by the fractal dimension, and the method is as follows: Step 1: Establish an a×b matrix A; Step 2: Fill the edge pixel values ​​of the matrix A with a crack image of size (a+l) ×(b+l), where l=2, select a pixel point containing a×b from the image of size (a+l) ×(b+l) and select a rectangular area as an area in the image; Step 3: Divide each rectangular area into s×s areas again, s selects 2 1 , 2 2 and 2 3 , and calculate r at different scales, r = s / a; Step 4: At each scale r, calculate n(i,j) = q-v+1 for each grid, where q and v are the box numbers of the maximum and minimum grayscale values ​​of the grid, respectively, and then move the grid; Step 5: After calculating n(i,j) for each grid, find b= ; Step 6: Use the least squares method to fit -log(r) and log(b) to calculate the slope, which is used as the fractal dimension D of the crack image in this area; Step 7: Traverse the pixel points in A row by row, repeat steps 2-6, calculate the fractal dimension D of each rectangular area, and calculate the average value `D of the fractal dimension D of each rectangular area as the fractal dimension of the crack image. The larger the `D value, the more serious the damage of the mechanical terminal. Therefore, the degree of damage is judged according to the size of the fractal dimension.

[0025] Example 2

[0026] The present invention also provides a computer device, which includes a memory and a processor. The memory is used to store computer instructions, and the processor executes the image-based mechanical terminal crack detection method by executing the computer instructions.

[0027] Example 3

[0028] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the image-based mechanical terminal crack detection method.

[0029] The image-based mechanical terminal crack detection method of the present invention effectively improves the accuracy and efficiency of crack detection by combining multiple image processing technologies. Specifically, the method has the following advantages: the illumination reflection model and Fourier transform are used to process noise points, effectively reducing the influence of background noise on crack detection; through wavelet decomposition and morphological analysis, the crack image can be clearly displayed and the crack area can be accurately segmented; the fractal dimension is used to judge the damage degree of the crack, providing an effective means for quantitative evaluation of the crack.

[0030] For example, certain words are used in the specification and claims to refer to specific components. Those skilled in the art should understand that hardware manufacturers may use different terms to refer to the same component. This specification and claims do not use differences in names as a way to distinguish components, but use differences in the functions of components as the criteria for distinction. For example, "including" mentioned throughout the specification and claims is an open term, so it should be interpreted as "including but not limited to". "Approximately" means that within an acceptable error range, those skilled in the art can solve technical problems within a certain error range and basically achieve technical effects.

[0031] It should be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a product or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such product or system. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the product or system including the elements.

[0032] The above description shows and describes several preferred embodiments of the present invention, but as before, it should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the invention concept of this article through the above teachings or the technology or knowledge of the relevant field. And the changes and modifications made by those skilled in the art do not depart from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.

Claims

1. An image-based mechanical terminal crack detection method, characterized in that: The following steps are included: S100, taking a first image of the mechanical terminal by a visual acquisition device, determining an incident component and a reflected component according to the positions of the visual acquisition device and the mechanical terminal, and processing noise by Fourier transform to obtain a second image; S200, performing wavelet decomposition, morphological analysis and wavelet transform on the second image to obtain an image showing cracks, and segmenting the image according to the number of cracks to obtain a third image; S300: judging the damage degree of the crack by using the fractal dimension of the third image.

2. The image-based mechanical terminal crack detection method according to claim 1, characterized in that: The steps of obtaining the second image by processing the noise through the illumination reflection model and Fourier transform in step S100 are as follows: S1001, performing logarithmic transformation on the first image; S1002, transforming the logarithmically transformed image from the spatial domain to the frequency domain through Fourier transformation; S1003, constructing a filter K (p, b) to filter and obtain an enhanced visual image; S1004, using Fourier transform again to convert the visual image in step S1003 from the frequency domain to the spatial domain; S1005. Perform exponential transformation on the image in S1004 to obtain a second image.

3. The image-based mechanical terminal crack detection method according to claim 2, characterized in that: S1001, when performing logarithmic transformation on the first image, the formula is as follows: lng(x,y)=lnj(x,y)+ lne(x,y); Where: g(x,y) is the light intensity function, j(x,y) is the incident component, e(x,y) is the reflected component, and 0<j (x,y)<∞, 0<e(x,y)<1; When the logarithmically transformed image is converted from the spatial domain to the frequency domain by Fourier transform in S1002, the following formula is used: G[ln g(x,y)]=G[ln j(x,y)]+G[ln e(x,y)]; Convert the above formula to: G(p,b) = J(p,b) + E(p,b); Where: J(p,b), E(p,b) and G(p,b) are the Fourier transforms of ln j(x,y), ln e(x,y) and ln g(x,y); E(p,b) represents the spectrum function, which is concentrated in the high frequency band; J(p,b) represents the spectrum function, which is concentrated in the low frequency band; In S1003, the high-frequency spectrum function E(p, b) is enhanced by the filter K(p, b) to obtain an enhanced visual image A(p, b): A(p,b) =K(p,b)G(p,b) =K(p,b)J(p,b) + K(p,b)E(p,b); When Fourier transform is used in S1004 to transform the image from the frequency domain to the spatial domain, k ( x , y ) is set to A ( p , b ) is the inverse Fourier transform of: k ( x , y ) = G -1 [ A ( p , b )]= kj ( x , y ) + ke ( x , y )； Where kj(x,y) and ke(x,y) represent the transformation of the filter in the incident component and the reflected component; G -1 stands for Inverse Fourier Transform; In S1005, the above formula is exponentially modified to obtain the processed second image: h(x,y) = exp| k(x,y) | .

4. The image-based mechanical terminal crack detection method according to claim 1, characterized in that: When the second image is subjected to wavelet decomposition, the high-frequency detail component is obtained, and the formula is as follows: I1 (x,y) = I(x,y) + I H [h(x,y)]+ I N [h(x,y)]+ I D [h(x,y)]; In the formula, I H represents the high-frequency horizontal component; I D represents the high-frequency diagonal component; I N represents the high-frequency vertical component; H , D and N Represent the horizontal, diagonal and vertical directions respectively; I 1 ( x , y ) represents the high-frequency detail component.

5. The image-based mechanical terminal crack detection method according to claim 4, characterized in that: The high-frequency component is subjected to morphological analysis. The high-frequency detail component is corroded by the structural element to obtain the crack boundary. After the crack boundary is obtained, the internal area of ​​the obtained crack boundary is filled in the form of expansion. When obtaining the crack boundary, the high-frequency detail component is subtracted from the corrosion to obtain the boundary. The formula is as follows: I(x,y)= I1 (x,y)- [I1 (x,y)ΔS]; In the above formula, S Represents structural elements; △ represents the difference operation; The formula for expansion is: I i O (x,y) = (I i O-1 (x,y) ⊙ S) ∩ I e’ ′ (x,y); where I i (x,y) ∈ I(x,y); In the above formula, I e’ ′ (x,y) represents the complement of the high-frequency component; O represents the number of iterations ; ⊙ represents the XOR operation.

6. The image-based mechanical terminal crack detection method according to claim 5, characterized in that: After dilation, the image is operated to obtain connected components, and the formula is as follows: I i O (x,y)= (I i O-1 (x,y)⊙S)∩I e’ (x,y),I e’ (x,y) is the high frequency component, when I i O (x,y) = I i O (x,y)-1, get the connected component: Y=OI i O (x, y), and then refine the image through the structural element, the formula is: Y'=Y-((I e’ ′ (x,y) △S1)∩(I e’ (x,y) △S2)), S1, S2∈S; where S1 is a set of elements associated with the object; S2 is a set of elements related to the corresponding background; repeat the above process until the refined image of Y' no longer changes.

7. The image-based mechanical terminal crack detection method according to claim 6, characterized in that: The morphological skeleton features of the cracks are extracted from the refined image.

8. The image-based mechanical terminal crack detection method according to claim 1, characterized in that: The damage degree of the third image is judged by the fractal dimension, and the method is as follows: Step 1: Establish an a×b matrix A; Step 2: Fill the edge pixel values ​​of the matrix A with a crack image of size (a+l) ×(b+l), select a pixel point containing a×b from the image of size (a+l) ×(b+l) and select a rectangular area as an area in the image; Step 3: Divide each rectangular area into s×s areas again, and s is selected as 2 1 , 2 2 and 2 3 , and calculate r at different scales, r = s / a; Step 4: At each scale r, calculate n(i,j) = q-v+1 for each grid, where q and v are the box numbers of the maximum and minimum grayscale values ​​of the grid, respectively, and then move the grid; Step 5: After calculating n(i,j) for each grid, find b= ; Step 6: Use the least squares method to fit -log(r) and log(b) to calculate the slope, which is used as the fractal dimension D of the crack image in this area; Step 7: Traverse the pixel points in A row by row, repeat steps 2 to 6, calculate the fractal dimension D of each rectangular area, calculate the average value of the fractal dimension D of each rectangular area as the fractal dimension of the crack image, and judge the degree of damage according to the size of the fractal dimension.

9. A computer device, characterized in that: The device includes a memory and a processor, the memory is used to store computer instructions, and the processor executes the image-based mechanical terminal crack detection method described in any one of claims 1-8 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the image-based mechanical terminal crack detection method described in any one of claims 1-8.

Images