A crack image enhancement method, device, equipment and storage medium

Abstract

The application discloses a crack image enhancement method and device, equipment and storage medium, apply to image enhancement field, this method comprises: obtaining the crack image and calculating the gradient image of the crack image;The crack edge pixels in each gradient image are subjected to interpolation processing to obtain a strong gradient contour set;Call a deep learning algorithm to update the parameters of the strong gradient contour set;Call the strong gradient contour set after updating the parameters to process the crack image to be processed image enhancement. The method constructs a strong gradient contour set by interpolating the gradient image, improves the spatial continuity of the discrete pixels of the crack image, realizes sub-pixel level enhancement of the crack image, and avoids the problem of low image measurement accuracy caused by only considering enhancing the crack gradient feature to improve the image recognition rate in the existing crack enhancement technology without considering the spatial continuity of the discrete pixels.

Description

Technical Field

[0001] This invention relates to the field of image enhancement, and in particular to a method, apparatus, device, and computer-readable storage medium for enhancing cracked images. Background Technology

[0002] With the continuous growth of my country's expressway mileage, the pressure on expressway use is increasing daily, and the density of expressway maintenance is rising year by year. Asphalt concrete pavement is a major type of expressway pavement and one of the key maintenance targets. The demand for asphalt concrete pavement distress detection is increasing annually, with cracks being a common type of distress that significantly impacts the performance of expressways. Timely identification, measurement, and repair of cracks can reduce traffic safety hazards on expressways, extend the safe service life of expressways, and significantly reduce maintenance costs. Crack image enhancement can improve the accuracy of crack identification and increase the efficiency of crack measurement and repair. Therefore, image enhancement of expressway pavement cracks is of significant practical importance to the development of expressway maintenance in my country.

[0003] Current methods for enhancing crack images improve the crack image recognition rate by enhancing the gradient features of the crack image and increasing the pixel value of crack pixels. Although this method improves the crack image recognition rate, it does not take into account the spatial continuity of discrete pixels in the crack image, resulting in the problem of low crack measurement accuracy in the enhanced image. Summary of the Invention

[0004] The purpose of this invention is to provide a crack image enhancement method, apparatus, device, and storage medium, which are applied in the field of image enhancement. This method constructs a strong gradient contour set by interpolating gradient images, thereby improving the spatial continuity of discrete pixels in crack images and achieving sub-pixel-level enhancement of crack images. This avoids the problem of low image measurement accuracy caused by existing crack enhancement technologies that only consider enhancing crack gradient features to improve image recognition rate without considering the spatial continuity of discrete pixels.

[0005] To solve the above-mentioned computational problems, the present invention provides a crack image enhancement method, comprising:

[0006] Acquire crack images and calculate the gradient image of the crack images;

[0007] Interpolation processing is performed on the crack edge pixels in each gradient image to obtain a strong gradient contour set;

[0008] The parameters of the strong gradient contour set are updated by invoking a deep learning algorithm;

[0009] The updated strong gradient contour set is used to perform image enhancement on the crack image to be processed.

[0010] Optionally, the step of interpolating the crack edge pixels in each gradient image to obtain a strong gradient contour set includes:

[0011] The strong gradient contour set is obtained by interpolating the crack edge pixels in each gradient image using bilinear interpolation.

[0012] Optionally, acquiring the crack image and calculating the gradient image of the crack image includes:

[0013] Obtain the crack image;

[0014] The mapping transformation coefficients are determined based on the line-of-sight and tilt angle parameters in the camera parameters.

[0015] The mapping transformation coefficients and the tilt angle parameter are input into the first model to normalize the crack image, resulting in a normalized image. The expression of the first model is:

[0016]

[0017] In the formula, s is the mapping change coefficient, and α i For the tilt angle parameter, (C x C y ) represents the pixel coordinates of the crack image, (I x I y () represents the normalized image pixel coordinates;

[0018] Calculate the gradient image of the normalized image.

[0019] Optionally, acquiring the crack image and calculating the gradient image of the crack image includes:

[0020] The pixel values ​​of the neighboring pixels of each pixel in the crack image are input into the second model to calculate the gradient image of the crack image. The expression of the second model is:

[0021]

[0022] In the formula, Let q1 and q2 be the pixel values ​​of the adjacent pixels, and G(x,y;σ) be a Gaussian kernel function with a standard deviation of σ.

[0023] Optionally, the step of calling the updated strong gradient contour set to perform image enhancement on the crack image to be processed includes:

[0024] Calculate the gradient image to be processed from the crack image to be processed;

[0025] The gradient image to be processed is determined as an iterative gradient image, and the iterative strong gradient contour of the iterative gradient image is calculated.

[0026] The minimum transformation rate parameter is determined based on the iterative strong gradient profile and the set of strong gradient profiles after updating the parameters.

[0027] Calculate the enhanced gradient image based on the minimum transformation rate parameter and the iterative gradient image;

[0028] Calculate the energy value based on the enhanced gradient image and the gradient image to be processed, and determine whether the energy value meets the preset conditions;

[0029] If not, the enhanced gradient image is identified as the iterative gradient image and the step of calculating the minimum transformation rate parameter is restarted;

[0030] If so, the crack image corresponding to the enhanced gradient image is determined as the enhanced crack image.

[0031] Optionally, calculating the enhanced gradient image based on the minimum transformation rate parameter and the gradient image to be processed includes:

[0032] The minimum transformation rate parameter and the gradient image to be processed are input into the third model to calculate the enhanced gradient image. The expression of the third model is:

[0033]

[0034] In the formula, For the enhanced gradient image, Let r(x) be the iterative gradient image, and r(x) be the minimum transformation parameter.

[0035] Optionally, calculating the energy value based on the enhanced gradient image and the gradient image to be processed includes:

[0036] The enhanced gradient image and the gradient image to be processed are input into the fourth model to calculate the energy value. The expression of the fourth model is:

[0037]

[0038] In the formula, For the enhanced gradient image, The gradient image to be processed, The energy value is mentioned above.

[0039] To address the aforementioned technical problems, the present invention also provides a crack image enhancement device, comprising:

[0040] An acquisition module is used to acquire crack images and calculate the gradient image of the crack images;

[0041] An interpolation module is used to interpolate the crack edge pixels in each gradient image to obtain a strong gradient contour set.

[0042] The training module is used to call deep learning algorithms to update the parameters of the strong gradient contour set;

[0043] The enhancement module is used to call the strong gradient contour set after updating the parameters to perform image enhancement on the crack image to be processed.

[0044] To address the aforementioned technical problems, the present invention also provides a crack image enhancement device, comprising:

[0045] Memory, used to store computer programs;

[0046] A processor, configured to implement any of the aforementioned crack image enhancement methods when executing the computer program.

[0047] To address the aforementioned technical problems, the present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement any of the aforementioned crack image enhancement methods.

[0048] As can be seen, the method of this invention acquires a crack image and calculates its gradient image; interpolates the crack edge pixels in each gradient image to obtain a strong gradient contour set; updates the parameters of the strong gradient contour set using a deep learning algorithm; and then uses the updated strong gradient contour set to enhance the crack image to be processed. By constructing a strong gradient contour set through interpolation of the gradient image, the spatial continuity of discrete pixels in the crack image is improved, achieving sub-pixel-level enhancement of the crack image. This avoids the problem of low image measurement accuracy caused by existing crack enhancement techniques that only consider enhancing crack gradient features to improve image recognition rate without considering the spatial continuity of discrete pixels. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0050] Figure 1 A flowchart of a crack image enhancement method provided in an embodiment of the present invention;

[0051] Figure 2This is a schematic diagram of an image acquisition process provided in an embodiment of the present invention;

[0052] Figure 3 This is a schematic diagram of a strong gradient profile of a crack provided in an embodiment of the present invention;

[0053] Figure 4 A specific embodiment of a crack image enhancement method provided in this invention is shown in the figure.

[0054] Figure 5 This is a structural block diagram of a crack image enhancement device provided in an embodiment of the present invention. Detailed Implementation

[0055] 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 embodiments of the present invention, and not all embodiments. 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.

[0056] The following combination Figure 1 , Figure 1 A flowchart of a crack image enhancement method provided in an embodiment of the present invention, the method may include:

[0057] S101: Obtain the crack image and calculate the gradient image of the crack image.

[0058] This embodiment does not limit the method of acquiring crack images. Currently, highway pavement crack images can be divided into two types according to different acquisition methods: mobile and fixed. Mobile methods involve periodically acquiring crack images using mobile inspection vehicles to perform crack image enhancement, identification, and measurement. Fixed methods involve real-time acquisition of crack images using fixed PTZ (Pan Tilt Zoom) cameras, followed by crack image enhancement for crack identification and measurement. Because mobile methods affect traffic flow, the acquired crack images are temporally and spatially discrete, which is detrimental to subsequent highway maintenance work. Furthermore, highway pavement maintenance based on mobile inspection vehicles is time-consuming and costly. This embodiment acquires crack images based on fixed PTZ cameras acquired under different line-of-sight and lighting conditions, such as… Figure 2 As shown, Figure 2 This is a schematic diagram of an image acquisition process provided by an embodiment of the present invention. A PTZ camera and related equipment are fixed on a column. O represents the optical center. The PTZ camera can operate at different line-of-sight parameters d. i Acquired samples of corresponding crack p under different illumination parameters iThe image is shown in (x), where p represents the projection surface of the road surface onto the fixed PTZ camera, and the projection parameter s is... i This refers to the actual physical distance from the center of the virtual projection surface corresponding to the road surface to the intersection line of the virtual projection surface and the road surface when the fixed PTZ camera is imaging.

[0059] In this embodiment, the gradient image of the crack image acquired by the PTZ camera can be calculated. This embodiment does not limit the acquisition method; each crack image can be acquired at the focal length that maximizes the image clarity. The method for determining clarity is not limited; it can be determined manually or by machine. This embodiment does not limit the specific calculation method; the gradient image can be directly calculated for general crack images. However, because the fixed PTZ camera is at a certain angle to the road surface when acquiring each crack image, the crack image is not captured from the front, resulting in some error and affecting crack measurement in subsequent highway maintenance. This embodiment can normalize the crack images acquired by the fixed PTZ camera to obtain a normalized image corresponding to each crack image. This embodiment can determine the mapping transformation coefficient of the crack image based on the line-of-sight and tilt angle parameters in the camera parameters corresponding to each crack image acquisition. Inputting the mapping transformation coefficient and tilt angle parameters into the first model can normalize the crack image to obtain a normalized image. The expression of the first model is:

[0060]

[0061] In the formula, s is the mapping change coefficient, and α i For the tilt angle parameter, (C x C y ) represents the pixel coordinates of the crack image, (I x I y () represents the normalized image pixel coordinates.

[0062] After normalizing the crack image, the gradient image of the normalized image can be calculated. The method for calculating the gradient image of the crack image is the same as the method for calculating the gradient image of the normalized image. The pixel values ​​of the neighboring pixels of each pixel in the normalized image are input into the second model to calculate the normalized gradient image. The expression for the second model is:

[0063]

[0064] In the formula, Let q1 and q2 be the pixel values ​​of adjacent pixels, and G(x,y;σ) be a Gaussian kernel function with a standard deviation of σ.

[0065] S102: Interpolate the crack edge pixels in each gradient image to obtain a strong gradient contour set.

[0066] After calculating the gradient images of the crack image, interpolation can be performed on the crack edge pixels in each gradient image to improve the spatial continuity of discrete pixels in the crack image and enhance the accuracy of subsequent crack measurements. This embodiment does not limit the specific method of interpolation; it can be performed on the gradient images... The strong gradient contour of the crack edge pixels is calculated, and a one-dimensional gradient contour is searched along the gradient direction of any crack edge pixel for interpolation. In this embodiment, the points on the strong gradient contour are generated using pixel bilinear interpolation, such as... Figure 3 As shown, Figure 3 This is a schematic diagram of a strong gradient contour of a crack provided in an embodiment of the present invention. The real pixels in the gradient image can be x0, x1, and x2. The pixel value of the pixel x to the left of the real pixel x0 can be generated by its adjacent gradient values ​​a, b, c, and d. When the gradient value no longer decreases, a strong gradient contour is generated. A strong gradient contour g(x; σ, λ) of the crack image related to the gradient contour intensity λ is established. The gradient contour intensity λ is a parameter representing the number or density of pixels. For each strong gradient contour corresponding to a crack image, this parameter can be the same or different. σ is the standard deviation of the pixel values ​​on the gradient contour.

[0067] Bilinear interpolation was performed on all crack images acquired under different viewing distance and illumination parameters to obtain a two-dimensional strong gradient contour set g of illumination and viewing distance parameters. c (x;σ,λ,d i ), g1(x;σ,λ,d1), can be the strong gradient contour corresponding to the crack image with a viewing distance parameter of d1 under the first illumination parameter, c represents the number of illumination parameters, and i represents the number of viewing distance parameters.

[0068] S103: Call the deep learning algorithm to update the parameters of the strong gradient contour set.

[0069] S104: Use the updated strong gradient contour set to perform image enhancement on the crack image to be processed.

[0070] After obtaining the strong gradient contour set, a deep learning algorithm can be called to update the parameters of the strong gradient contour set, solve and optimize the strong gradient contour set, and obtain an accurate strong gradient contour set of the crack image.

[0071] In this embodiment, image enhancement can be performed on the image to be processed using the strong gradient contour set after training. This embodiment does not limit the specific method of image enhancement. Generally, the gradient image to be processed of the crack image to be processed can be calculated; the gradient image to be processed can be determined as the iterative gradient image, and the iterative strong gradient contour of the iterative gradient image can be calculated; the minimum transformation rate parameter can be determined based on the iterative strong gradient contour and the strong gradient contour set after parameter update; the enhanced gradient image can be calculated based on the minimum transformation rate parameter and the iterative gradient image; the energy value can be calculated based on the enhanced gradient image and the gradient image to be processed, and it can be determined whether the energy value meets the preset condition; if the energy value does not meet the preset condition, the enhanced gradient image can be determined as the iterative gradient image and the step of calculating the minimum transformation rate parameter can be restarted; if the energy value meets the preset condition, the crack image corresponding to the enhanced gradient image can be determined as the enhanced crack image.

[0072] The formula for calculating the minimum conversion rate parameter in this embodiment can be as follows:

[0073]

[0074] In the formula, g s (x;σ,λ,d r ) represents the iterative strong gradient contour of the iterative gradient image, g c (x;σ,λ,d i ) represents the set of strong gradient contours after parameter updates, and r(x) represents the minimum transformation rate parameter.

[0075] After the minimum transformation rate parameter is calculated, the minimum transformation rate parameter and the gradient image to be processed are input into the third model to calculate the enhanced gradient image. The expression of the third model is:

[0076]

[0077] In the formula, To enhance gradient images, For the iterative gradient image, r(x) is the minimum transformation parameter.

[0078] The enhanced gradient image and the gradient image to be processed are input into the fourth model to calculate the energy value. The expression of the fourth model is:

[0079]

[0080] In the formula, To enhance gradient images, The gradient image to be processed. This represents the energy value.

[0081] This embodiment does not limit the specific content of the preset conditions, which can be set according to the actual application. In this embodiment, when the energy value does not meet the preset conditions, the enhanced gradient image is determined as the iterative gradient image and the step of calculating the minimum transformation rate parameter is restarted; when the energy value meets the preset conditions, the crack image corresponding to the enhanced gradient image is determined as the enhanced crack image, and the enhancement of the image to be processed is completed.

[0082] This invention constructs a strong gradient contour set by interpolating gradient images, which improves the spatial continuity of discrete pixels in crack images and achieves sub-pixel-level enhancement of crack images. This avoids the problem of low image measurement accuracy caused by existing crack enhancement technologies that only consider enhancing crack gradient features to improve image recognition rate without considering the spatial continuity of discrete pixels.

[0083] The following combination Figure 4 , Figure 4 This is a specific embodiment of a crack image enhancement method provided by the present invention. This specific embodiment may include:

[0084] 1. Acquire crack images acquired by a fixed PTZ camera under different line-of-sight and illumination parameters.

[0085] 2. Normalize each crack image to obtain the corresponding normalized image.

[0086] 3. Calculate the gradient image of the normalized image.

[0087] 4. Interpolate the crack edge pixels in each gradient image to obtain a strong gradient contour set.

[0088] 5. Call the deep learning algorithm to update the parameters of the strong gradient contour set.

[0089] 6. Use the updated strong gradient contour set to perform image enhancement on the crack image to be processed.

[0090] The following combination Figure 5 , Figure 5 This is a structural block diagram of a crack image enhancement device provided in an embodiment of the present invention. The device may include:

[0091] The acquisition module 100 is used to acquire crack images and calculate the gradient image of the crack images;

[0092] Interpolation module 200 is used to interpolate the crack edge pixels in each gradient image to obtain a strong gradient contour set.

[0093] Training module 300 is used to call a deep learning algorithm to update the parameters of the strong gradient contour set;

[0094] The enhancement module 400 is used to call the strong gradient contour set after updating the parameters to perform image enhancement on the crack image to be processed.

[0095] Based on the above embodiments, the present invention constructs a strong gradient contour set by interpolating the gradient image, which improves the spatial continuity of discrete pixels in the crack image and realizes sub-pixel-level enhancement of the crack image. This avoids the problem of low image measurement accuracy caused by existing crack enhancement technologies that only consider enhancing crack gradient features to improve image recognition rate without considering the spatial continuity of discrete pixels.

[0096] Based on the above embodiments, the interpolation module 200 may include:

[0097] A bilinear interpolation unit is used to interpolate the crack edge pixels in each gradient image using bilinear interpolation to obtain the strong gradient contour set.

[0098] Based on the above embodiments, the acquisition module 100 may include:

[0099] Acquisition unit, used to acquire the crack image;

[0100] The coefficient unit determines the mapping transformation coefficients based on the line-of-sight and tilt angle parameters in the camera parameters.

[0101] The normalization unit is used to input the mapping transformation coefficients and the tilt angle parameter into the first model to normalize the crack image and obtain a normalized image. The expression of the first model is:

[0102]

[0103] In the formula, s is the mapping change coefficient, and α i For the tilt angle parameter, (C x C y ) represents the pixel coordinates of the crack image, (I x I y () represents the normalized image pixel coordinates;

[0104] A first gradient unit is used to calculate the gradient image of the normalized image.

[0105] Based on the above embodiments, the acquisition module 100 may include:

[0106] The second gradient unit is used to input the pixel values ​​of the neighboring pixels of each pixel in the crack image into the second model to calculate the gradient image of the crack image. The expression of the second model is:

[0107]

[0108] In the formula, Let q1 and q2 be the pixel values ​​of the adjacent pixels, and G(x,y;σ) be a Gaussian kernel function with a standard deviation of σ.

[0109] Based on the above embodiments, the enhancement module 400 may include:

[0110] The processing unit is used to calculate the gradient image to be processed of the crack image to be processed;

[0111] The determining unit is used to determine the gradient image to be processed as an iterative gradient image and to calculate the iterative strong gradient contour of the iterative gradient image.

[0112] A minimum transformation unit is used to determine a minimum transformation rate parameter based on the iterative strong gradient profile and the set of strong gradient profiles after updating the parameters.

[0113] An enhanced gradient unit is used to calculate an enhanced gradient image based on the minimum transformation rate parameter and the iterative gradient image;

[0114] An energy value unit is used to calculate an energy value based on the enhanced gradient image and the gradient image to be processed, and to determine whether the energy value meets a preset condition.

[0115] The first execution unit is configured to, if not, determine the enhanced gradient image as the iterative gradient image and restart the step of calculating the minimum transformation rate parameter;

[0116] The second execution unit is used to determine the crack image corresponding to the enhanced gradient image as the enhanced crack image if the condition is met.

[0117] Based on the above embodiments, the enhanced gradient unit may include:

[0118] An enhanced gradient subunit is used to input the minimum transformation rate parameter and the gradient image to be processed into a third model to calculate the enhanced gradient image. The expression of the third model is:

[0119]

[0120] In the formula, For the enhanced gradient image, Let r(x) be the iterative gradient image, and r(x) be the minimum transformation parameter.

[0121] Based on the above embodiments, the energy value unit may include:

[0122] The energy value subunit is used to input the enhanced gradient image and the gradient image to be processed into the fourth model to calculate the energy value. The expression of the fourth model is:

[0123]

[0124] In the formula, For the enhanced gradient image, The gradient image to be processed, The energy value is mentioned above.

[0125] Based on the above embodiments, the present invention also provides a crack image enhancement device, which may include a memory and a processor. The memory stores a computer program, and when the processor calls the computer program in the memory, it can implement the steps provided in the above embodiments. Of course, the device may also include various necessary network interfaces, a power supply, and other components.

[0126] The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by an execution terminal or processor, it can implement the crack image enhancement method provided in the embodiments of the present invention. The storage medium may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0127] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0128] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0129] The present invention provides a detailed description of a crack image enhancement method, apparatus, device, and storage medium. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for enhancing crack images, characterized in that, include: Acquire crack images and calculate the gradient image of the crack images; Interpolation is performed on the crack edge pixels in each gradient image to obtain a strong gradient contour set; The parameters of the strong gradient contour set are updated by invoking a deep learning algorithm; The updated strong gradient contour set is used to perform image enhancement on the crack image to be processed. The process of calling and updating the strong gradient contour set after the parameters are called to perform image enhancement on the crack image to be processed includes: Calculate the gradient image to be processed from the crack image to be processed; The gradient image to be processed is determined as an iterative gradient image, and the iterative strong gradient contour of the iterative gradient image is calculated. The minimum transformation rate parameter is determined based on the iterative strong gradient profile and the set of strong gradient profiles after updating the parameters. Calculate the enhanced gradient image based on the minimum transformation rate parameter and the gradient image to be processed; Calculate the energy value based on the enhanced gradient image and the gradient image to be processed, and determine whether the energy value meets the preset conditions; If not, the enhanced gradient image is identified as the iterative gradient image and the step of calculating the minimum transformation rate parameter is restarted; If so, the crack image corresponding to the enhanced gradient image is determined as the enhanced crack image; The step of calculating the enhanced gradient image based on the minimum transformation rate parameter and the gradient image to be processed includes: The minimum transformation rate parameter and the gradient image to be processed are input into the third model to calculate the enhanced gradient image. The expression of the third model is: ； In the formula, For the enhanced gradient image, For the iterative gradient image, The minimum conversion rate parameter; The step of calculating the energy value based on the enhanced gradient image and the gradient image to be processed includes: The enhanced gradient image and the gradient image to be processed are input into the fourth model to calculate the energy value. The expression of the fourth model is: ； In the formula, For the enhanced gradient image, The gradient image to be processed, The energy value; The formula for calculating the minimum conversion rate parameter is: ； In the formula, For iterative strong gradient contours of the iterative gradient image, For the set of strong gradient contours after parameter updates, The minimum conversion rate parameter, For pixels, For gradient profile intensity, Let be the standard deviation of the pixel values ​​of the pixels on the gradient contour. Indicates the number of lighting parameters. For sight distance parameters, The number of the sight distance parameters.

2. The crack image enhancement method according to claim 1, characterized in that, The step of interpolating the crack edge pixels in each gradient image to obtain a strong gradient contour set includes: The strong gradient contour set is obtained by interpolating the crack edge pixels in each gradient image using bilinear interpolation.

3. The crack image enhancement method according to claim 1, characterized in that, The process of acquiring the crack image and calculating the gradient image of the crack image includes: Obtain the crack image; The mapping transformation coefficients are determined based on the line-of-sight and tilt angle parameters in the camera parameters. The mapping transformation coefficients and the tilt angle parameter are input into the first model to normalize the crack image, resulting in a normalized image. The expression of the first model is: ； In the formula, A is the mapping change coefficient. The tilt angle parameter, The pixel coordinates of the crack image. Normalized image pixel coordinates; Calculate the gradient image of the normalized image.

4. The crack image enhancement method according to claim 1, characterized in that, The process of acquiring the crack image and calculating the gradient image of the crack image includes: The pixel values ​​of the neighboring pixels of each pixel in the crack image are input into the second model to calculate the gradient image of the crack image. The expression of the second model is: ； In the formula, The gradient image, , The pixel values ​​of the adjacent pixels. The standard deviation is The Gaussian kernel function.

5. A crack image enhancement device, characterized in that, include: An acquisition module is used to acquire crack images and calculate the gradient image of the crack images; An interpolation module is used to interpolate the crack edge pixels in each gradient image to obtain a strong gradient contour set. The training module is used to call deep learning algorithms to update the parameters of the strong gradient contour set; The enhancement module is used to call the strong gradient contour set after updating the parameters to perform image enhancement on the crack image to be processed; The process of calling and updating the strong gradient contour set after the parameters are called to perform image enhancement on the crack image to be processed includes: Calculate the gradient image to be processed from the crack image to be processed; The gradient image to be processed is determined as an iterative gradient image, and the iterative strong gradient contour of the iterative gradient image is calculated. The minimum transformation rate parameter is determined based on the iterative strong gradient profile and the set of strong gradient profiles after updating the parameters. Calculate the enhanced gradient image based on the minimum transformation rate parameter and the gradient image to be processed; Calculate the energy value based on the enhanced gradient image and the gradient image to be processed, and determine whether the energy value meets the preset conditions; If not, the enhanced gradient image is identified as the iterative gradient image and the step of calculating the minimum transformation rate parameter is restarted; If so, the crack image corresponding to the enhanced gradient image is determined as the enhanced crack image; The step of calculating the enhanced gradient image based on the minimum transformation rate parameter and the gradient image to be processed includes: The minimum transformation rate parameter and the gradient image to be processed are input into the third model to calculate the enhanced gradient image. The expression of the third model is: ； In the formula, For the enhanced gradient image, For the iterative gradient image, The minimum conversion rate parameter; The step of calculating the energy value based on the enhanced gradient image and the gradient image to be processed includes: The enhanced gradient image and the gradient image to be processed are input into the fourth model to calculate the energy value. The expression of the fourth model is: ； In the formula, For the enhanced gradient image, The gradient image to be processed, The energy value; The formula for calculating the minimum conversion rate parameter is: ； In the formula, For iterative strong gradient contours of the iterative gradient image, For the set of strong gradient contours after parameter updates, The minimum conversion rate parameter, For pixels, For gradient profile intensity, Let be the standard deviation of the pixel values ​​of the pixels on the gradient contour. Indicates the number of lighting parameters. For sight distance parameters, The number of the sight distance parameters.

6. A crack image enhancement device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the crack image enhancement method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the crack image enhancement method as described in any one of claims 1 to 4.

Images