Homomorphic filtering enhancement method for cable X-ray image

Abstract

The invention relates to the technical field of image filtering, and provides a cable X-ray image homomorphic filtering enhancement method comprising the following steps: collecting a cable X-ray image, setting an initial value of a sharpening coefficient, and obtaining a cable X-ray enhanced image; obtaining a significance cluster and a length cluster, and calculating the gray level variation degree, the image richness variation degree and the sharpening coefficient suppression degree of the cable X-ray enhanced image; and according to the gray level change degree, the image richness change degree and the sharpening coefficient suppression degree, determining the value of a sharpening coefficient of homomorphic filtering, performing image enhancement on the cable X-ray image by using homomorphic filtering, and obtaining an image enhancement result of the homomorphic filtering. According to the invention, the processing effect of homomorphic filtering image enhancement can be improved.

Description

Technical Field

[0001] This application relates to the field of image filtering technology, specifically to a homomorphic filtering enhancement method for cable X-ray images. Background Technology

[0002] High-voltage cables have complex internal structures, making X-ray imaging a core method for cable quality inspection. However, different material regions show relatively small grayscale differences in cable X-ray images, and system noise and scattering make minute defects difficult to distinguish. This results in generally insufficient contrast and strong noise interference in cable X-ray images, directly affecting the defect identification rate of high-voltage cables. Therefore, homomorphic filtering is needed to enhance cable X-ray images, improving high-frequency details, compressing the low-frequency dynamic range, and enhancing image contrast and detail, thereby improving image quality.

[0003] Homomorphic filtering uses the sigmoid function as the transfer function, reducing the number of parameters that need to be adjusted. However, homomorphic filtering operates globally in the frequency domain based on the X-ray image. When amplifying high-frequency information, it is easy to amplify noise simultaneously, which can lead to weak high-frequency enhancement and poor low-frequency suppression in the image after image enhancement processing. Summary of the Invention

[0004] This application provides a homomorphic filtering enhancement method for cable X-ray images to solve the problem that homomorphic filtering images often exhibit weak high-frequency enhancement and poor low-frequency suppression. The specific technical solution adopted is as follows: One embodiment of this application provides a method for homomorphic filtering enhancement of cable X-ray images, the method comprising the following steps: Acquire cable X-ray images, set initial values ​​for sharpening coefficients, and use homomorphic filtering to enhance the cable X-ray images to obtain enhanced cable X-ray images; The corresponding pixels in the cable X-ray image and the cable X-ray enhanced image are recorded as the corresponding matching pixels. The gray-scale change of the cable X-ray enhanced image is calculated based on the difference in gray-scale values ​​between the corresponding matching pixels and their adjacent pixels. Clustering is performed on the saliency analysis results of cable X-ray images and cable X-ray enhanced images to obtain saliency clusters. Based on the differences in saliency analysis results within all saliency clusters corresponding to cable X-ray images and cable X-ray enhanced images, the image richness variation of cable X-ray enhanced images is calculated. Edge detection is performed on cable X-ray images and cable X-ray enhanced images to obtain length clusters of edges. Based on the differences in the mean and number of all lengths within the same cluster, the edge enhancement effect of the cable X-ray enhanced image is calculated. The sharpening coefficient suppression degree of the cable X-ray enhanced image is calculated by combining the differences in the number of edge pixels contained in the neighborhood of the corresponding matching pixels that are all edge pixels. Based on the grayscale variation, image richness variation, and sharpening coefficient suppression of the enhanced cable X-ray image, the value of the sharpening coefficient for homomorphic filtering is determined. Homomorphic filtering is then used to enhance the cable X-ray image, and the image enhancement result of the cable X-ray image after homomorphic filtering is obtained.

[0005] Furthermore, the specific method for obtaining the grayscale variation of the cable X-ray enhanced image is as follows: The sum of the differences between the gray values ​​of the matched pixel and all pixels in its eight neighboring regions is recorded as the adjacent gray value difference of the matched pixel. The grayscale variation of the cable X-ray enhanced image is calculated based on the grayscale difference between adjacent matching pixels in the cable X-ray enhanced image and the cable X-ray image.

[0006] Furthermore, the specific steps for calculating the grayscale change of the cable X-ray enhanced image based on the adjacent grayscale differences of all corresponding matching pixels in the cable X-ray enhanced image and the cable X-ray image are as follows: The normalized value of the sum of the differences between adjacent gray levels of all corresponding matching pixels in the cable X-ray enhanced image and the cable X-ray image is denoted as the gray level variation of the cable X-ray enhanced image.

[0007] Furthermore, the specific method for obtaining the image richness variation of the enhanced cable X-ray image is as follows: Based on the saliency analysis results of cable X-ray images and cable X-ray enhanced images, the saliency eigenvalues ​​of each saliency cluster are calculated. The difference between the number of pixels contained in the salient cluster with the largest salient feature value in the cable X-ray enhanced image and the number of pixels contained in the salient cluster with the largest salient feature value in the cable X-ray image is denoted as the number of salient differences in the cable X-ray enhanced image. The positive correlation between the saliency eigenvalues ​​and the number of significant differences of all saliency clusters corresponding to the cable X-ray enhanced image is denoted as the image richness variation degree of the cable X-ray enhanced image.

[0008] Furthermore, the specific calculation steps for the saliency eigenvalue are as follows: Calculate the saliency of each pixel in cable X-ray images and cable X-ray enhanced images; The mean of all saliencies within the same saliency cluster is denoted as the saliency eigenvalue of the same saliency cluster.

[0009] Furthermore, the specific calculation method for the image edge enhancement effect of the cable X-ray enhanced image is as follows: Calculate the length feature value of the length cluster based on all lengths within the length cluster; The difference between the number of pixels contained in the length cluster with the largest length feature value in the cable X-ray enhanced image and the number of lengths contained in the length cluster with the largest length feature value in the cable X-ray image is denoted as the length difference number of the cable X-ray enhanced image. The positive correlation between the length feature values ​​of all length clusters corresponding to the cable X-ray enhanced image and the number of length differences is denoted as the image edge enhancement effect of the cable X-ray enhanced image.

[0010] Furthermore, the specific method for obtaining the length feature value is as follows: The mean of all lengths within a cluster of the same length is denoted as the length feature value of the cluster of the same length.

[0011] Furthermore, the specific calculation steps for the sharpening coefficient suppression degree of the cable X-ray enhanced image are as follows: Select the corresponding matching pixels that are all edge pixels. The difference between the number of edge pixels contained in the eight neighborhood of the selected matching pixels in the cable X-ray enhanced image and the cable X-ray image is recorded as the edge number difference of the selected matching pixels. The normalized value of the mean of the edge number differences of all selected matching pixels is recorded as the edge enhancement degree of the cable X-ray enhanced image. The normalized value of the sum of the edge enhancement effect and the edge enhancement degree of the cable X-ray enhanced image is denoted as the edge feature value of the cable X-ray enhanced image. The difference between the number 1 and the edge feature value of the cable X-ray enhanced image is denoted as the sharpening coefficient suppression degree of the cable X-ray enhanced image.

[0012] Furthermore, the specific method for determining the sharpening coefficient of the homomorphic filter based on the grayscale variation, image richness variation, and sharpening coefficient suppression of the enhanced cable X-ray image includes: The normalized value of the difference between the product of the grayscale variation and the image richness variation of the cable X-ray enhanced image and the sharpening coefficient suppression is denoted as the correction coefficient of the cable X-ray enhanced image; the enhancement coefficient of the cable X-ray enhanced image is calculated based on the correction coefficient of the cable X-ray enhanced image. When the correction coefficient of the cable X-ray enhanced image is less than or equal to the correction coefficient threshold, the enhancement coefficient of the cable X-ray enhanced image is used as the value of the sharpening coefficient. When the correction factor of the cable X-ray enhanced image is greater than the correction factor threshold, the initial value of the sharpening factor is used as the value of the sharpening factor.

[0013] Furthermore, the enhancement coefficient of the cable X-ray enhanced image is specifically as follows: The enhancement coefficient of the cable X-ray enhanced image is denoted as the product of the sum of the correction coefficient and the number 1 and the initial value of the sharpening coefficient.

[0014] The beneficial effects of this application are: When using homomorphic filtering to enhance cable X-ray images, the sharpening coefficient can adjust the enhancement intensity, control the smoothness of the transition, and affect the visual effect of the image. First, homomorphic filtering is applied to the cable X-ray image based on the initial sharpening coefficient value to obtain an enhanced cable X-ray image. The image enhancement effect is evaluated based on the differences in high-frequency and detail information between the enhanced and original cable X-ray images. The grayscale variation and image richness variation of the enhanced cable X-ray image are calculated. Furthermore, since higher-quality image enhancement can improve the richness of details in the image, image enhancement of cable X-ray images can connect edges that were originally single but disconnected, making the edge information more complete. It can also highlight edges that were not originally present due to noise interference or environmental factors. This study analyzes the edge detection results of cable X-ray images and enhanced cable X-ray images, evaluates the degree of change in edge detail information in the enhanced cable X-ray images, and calculates the sharpening coefficient suppression degree of the enhanced cable X-ray images. The smaller the sharpening coefficient suppression degree, the better the image enhancement effect, and the lower the necessity for suppressing the image enhancement effect. Finally, based on the grayscale variation, image richness variation, and sharpening coefficient suppression degree of the enhanced cable X-ray images, the value of the sharpening coefficient for homomorphic filtering is determined. Homomorphic filtering is then used to enhance the cable X-ray images, obtaining the image enhancement results after homomorphic filtering. This addresses the problem that homomorphic filtering often results in weak high-frequency enhancement and poor low-frequency suppression, improving the processing effect of homomorphic filtering image enhancement and achieving image quality improvement. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application 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 some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1This is a schematic flowchart of a homomorphic filtering enhancement method for cable X-ray images provided in one embodiment of this application; Figure 2 This is a flowchart illustrating the grayscale variation acquisition process provided in one embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] Please see Figure 1 The diagram illustrates a flowchart of a homomorphic filtering enhancement method for cable X-ray images according to an embodiment of this application. The method includes the following steps: Step S001: Acquire cable X-ray images, set the initial value of the sharpening coefficient, and use homomorphic filtering to perform image enhancement processing on the cable X-ray images to obtain enhanced cable X-ray images.

[0019] An industrial X-ray digital imaging inspection system is used to acquire X-ray images of cables.

[0020] It should be noted that, for ease of processing, the cable X-ray images in this embodiment of the invention have undergone denoising preprocessing to eliminate the impact of excessive noise on subsequent image enhancement. Image denoising is a technique well-known to those skilled in the art and is not limited thereto.

[0021] When processing cable X-ray images using homomorphic filtering, this embodiment uses a sigmoid function as the transfer function, which contains a parameter k. Parameter k is the sharpening coefficient. The core function of the sharpening coefficient is to control the steepness of the transition from low to high frequencies in the filter. It can adjust the enhancement intensity, control the smoothness of the transition, and affect the visual effect of the image. Specifically: the larger the sharpening coefficient value, the steeper the slope of the filter function, and the more obvious the enhancement effect of high-frequency components relative to low-frequency components in the image. The value of the sharpening coefficient determines the behavior of the filter near the cutoff frequency; a smaller sharpening coefficient value produces a smoother transition, while a larger sharpening coefficient value produces a sharper transition. Increasing the sharpening coefficient value can further enhance the contrast and edge sharpness of the image, but excessive increase may lead to noise amplification or artifacts. Here, high-frequency components in the image correspond to image details and edges, while low-frequency components correspond to the lighting background.

[0022] The initial value of the sharpening coefficient is set to the first preset parameter, and homomorphic filtering is used to enhance the cable X-ray image to obtain the enhanced cable X-ray image.

[0023] In this embodiment, the first preset parameter is set to 0.5; the use of state filtering to enhance cable X-ray images is a well-known technique and will not be described in detail here.

[0024] At this point, cable X-ray images and cable X-ray enhanced images have been obtained.

[0025] Step S002: Record the corresponding pixels in the cable X-ray image and the cable X-ray enhanced image as the corresponding matching pixels. Calculate the grayscale change of the cable X-ray enhanced image based on the difference in grayscale values ​​between the corresponding matching pixels and their adjacent pixels.

[0026] High-frequency information in cable X-ray images and enhanced cable X-ray images corresponds to detail and edge information, such as cable edge defects and voids in the internal plastic layer. The pixels corresponding to this high-frequency information exhibit significant grayscale variations with adjacent grayscale values ​​in both the original and enhanced cable X-ray images. In other words, the more pixels with significant grayscale variations in the enhanced cable X-ray image compared to the original, the more high-frequency information it contains, resulting in better image enhancement. Therefore, the image enhancement effect can be evaluated based on the differences in high-frequency information between the original and enhanced cable X-ray images.

[0027] All corresponding pixels in the cable X-ray image and the cable X-ray enhanced image are recorded as corresponding matching pixels. The sum of the differences between the gray values ​​of the matching pixels and all pixels in their eight neighborhoods is recorded as the adjacent gray value difference of the matching pixels. The normalized value of the sum of the differences between the adjacent gray values ​​of the cable X-ray enhanced image and all corresponding matching pixels in the cable X-ray image is recorded as the gray value change of the cable X-ray enhanced image.

[0028] In this embodiment, the sigmoid function is used to calculate the normalized value. The sigmoid function is a well-known technique and will not be described in detail here. As other implementation methods, implementers can use other methods of the prior art, such as the tanh function.

[0029] The greater the grayscale variation in the cable X-ray enhanced image, the more prominent the high-frequency information in the cable X-ray enhanced image is compared to the cable X-ray image, and the more obvious the image enhancement effect of the cable X-ray enhanced image is compared to the cable X-ray image.

[0030] At this point, the grayscale variation of the enhanced X-ray image of the cable has been obtained. The flowchart for obtaining the grayscale variation is as follows: Figure 2 As shown.

[0031] Step S003: Cluster the saliency analysis results of the cable X-ray image and the cable X-ray enhanced image to obtain saliency clusters. Based on the differences in saliency analysis results within all saliency clusters corresponding to the cable X-ray image and the cable X-ray enhanced image, calculate the image richness variation of the cable X-ray enhanced image.

[0032] Cable defects in X-ray images often correspond to image details and high-frequency information, with the corresponding pixels having prominent gray values ​​and high local contrast.

[0033] Saliency analysis was performed on both the cable X-ray image and the enhanced cable X-ray image to obtain the saliency of each pixel. Saliency clusters were then formed for all pixels in both images. The mean of all saliencies within the same saliency cluster was denoted as the saliency eigenvalue of that cluster. The difference between the number of pixels in the saliency cluster with the highest saliency eigenvalue in the enhanced cable X-ray image and the number of pixels in the same cluster was denoted as the number of saliency differences in the enhanced cable X-ray image. The positive correlation between the saliency eigenvalues ​​and the number of saliency differences for all saliency clusters corresponding to the enhanced cable X-ray image was denoted as the image richness variation of the enhanced cable X-ray image.

[0034] In this embodiment, the CA saliency analysis algorithm is used for saliency analysis, and the DBSCAN algorithm is used for saliency clustering. Saliency analysis of images is a well-known technique and will not be described in detail here.

[0035] It is understood that a positive correlation is applied to the salient eigenvalues ​​and the number of significant differences of all salient clusters corresponding to the cable X-ray enhanced image, ensuring that the salient eigenvalues ​​and the number of significant differences of all salient clusters corresponding to the cable X-ray enhanced image are positively correlated with the image richness variation of the cable X-ray enhanced image. It is understood that the positive correlation in this application refers to the relationship between the independent and dependent variables. The independent variable is the salient eigenvalues ​​and the number of significant differences of all salient clusters corresponding to the cable X-ray enhanced image, and the dependent variable is the image richness variation of the cable X-ray enhanced image. The positive correlation means that the dependent variable increases (decreases) as the independent variable increases (decreases), and can be an additive or multiplicative relationship.

[0036] Preferably, as an embodiment of this application, the normalized value of the sum of the saliency feature values ​​of all saliency clusters corresponding to the cable X-ray enhanced image and the sum of the number of saliency differences is denoted as the image richness variation degree of the cable X-ray enhanced image.

[0037] The greater the difference between the number of pixels contained in the salient cluster with the largest salient feature value in the cable X-ray enhanced image and the number of pixels contained in the salient cluster with the largest salient feature value in the cable X-ray enhanced image, and the greater the salient feature value of all salient clusters corresponding to the cable X-ray enhanced image, the more pixels with richer details will be in the cable X-ray enhanced image after image enhancement. At this time, the greater the variation in image richness of the cable X-ray enhanced image, and the more obvious the image enhancement effect of the cable X-ray enhanced image compared to the original cable X-ray image.

[0038] At this point, the image richness variation of the enhanced cable X-ray image has been obtained.

[0039] Step S004: Perform edge detection on the cable X-ray image and the cable X-ray enhanced image to obtain the length clusters of the edges, obtain length clusters, calculate the image edge enhancement effect of the cable X-ray enhanced image based on the difference in the mean and number of all lengths in all the same clusters, and calculate the sharpening coefficient suppression degree of the cable X-ray enhanced image by combining the difference in the number of edge pixels contained in the neighborhood of the corresponding matching pixels that are all edge pixels.

[0040] Furthermore, to avoid excessive amplification of noise information in the enhanced cable X-ray image obtained by image enhancement of the cable X-ray image, the cable X-ray image and the enhanced cable X-ray image are further analyzed by combining noise characteristics.

[0041] Edge detection was performed on the cable X-ray image and the cable X-ray enhanced image respectively to obtain the cable X-ray edge image and the cable X-ray enhanced edge image.

[0042] Cable defects in cable X-ray images and cable X-ray enhanced images generally have corresponding edge information. After image enhancement of cable X-ray images, edges that were originally a single but unconnected edge line can be connected to each other, making the edge information more complete. Edges that were not originally present due to noise interference or shooting environment interference can also be highlighted. Therefore, higher quality image enhancement can improve the richness of details in the image.

[0043] Clustering is performed on the lengths of all edges contained in the cable X-ray edge image and the cable X-ray enhanced edge image to obtain length clusters. The mean of all lengths within the same length cluster is denoted as the length feature value of that cluster. The difference between the number of pixels contained in the length cluster with the largest length feature value in the cable X-ray enhanced image and the number of lengths contained in the same length cluster is denoted as the length difference quantity of the cable X-ray enhanced image. The positive correlation between the length feature values ​​and the length difference quantity of all length clusters corresponding to the cable X-ray enhanced image is denoted as the image edge enhancement effect of the cable X-ray enhanced image.

[0044] Preferably, as an embodiment of this application, the normalized value of the sum of the length feature values ​​of all length clusters corresponding to the cable X-ray enhanced image and the sum of the number of length differences is denoted as the image edge enhancement effect of the cable X-ray enhanced image.

[0045] In the process of calculating the image edge enhancement effect of cable X-ray enhanced images, the Z-score normalization algorithm is used to calculate the normalization value, so that the value range of the image edge enhancement effect of cable X-ray enhanced images is greater than or equal to -1 and less than or equal to 1.

[0046] When the edge enhancement effect of the cable X-ray image is greater than 0, more detailed edge information is added when the cable X-ray image is enhanced, making the edge information more coherent and rich. When the edge enhancement effect of the cable X-ray image is less than 0, the image noise is increased when the cable X-ray image is enhanced, and the image quality deteriorates.

[0047] However, excessive edge enhancement in cable X-ray images can lead to an overabundance of pixels at the banana-shaped edge and increased high-frequency noise. Therefore, further evaluation is needed by combining the cable X-ray enhanced image with overlapping edge pixels in the cable X-ray image to determine whether image enhancement reduces edge blur. If edge blur is reduced, then the image enhancement is excessive.

[0048] Matching pixels that are all edge pixels are selected. The difference between the number of edge pixels contained in the eight neighborhoods of the selected matching pixels in the cable X-ray enhanced image and the cable X-ray image is recorded as the edge number difference of the selected matching pixels. The normalized value of the mean of the edge number differences of all selected matching pixels is recorded as the edge enhancement degree of the cable X-ray enhanced image. The normalized value of the sum of the image edge enhancement effect and the edge enhancement degree of the cable X-ray enhanced image is recorded as the edge feature value of the cable X-ray enhanced image. The difference between the number 1 and the edge feature value of the cable X-ray enhanced image is recorded as the sharpening coefficient suppression degree of the cable X-ray enhanced image.

[0049] In calculating the edge enhancement degree and edge feature value of the cable X-ray enhanced image, the Z-score normalization algorithm is used to calculate the normalized value, so that the range of the edge enhancement degree and edge feature value of the cable X-ray enhanced image is greater than or equal to -1 and less than or equal to 1. When the edge enhancement degree of the cable X-ray enhanced image is less than 0, the edge information richness of the cable X-ray enhanced image obtained by image enhancement is small. When the edge enhancement degree of the cable X-ray enhanced image is greater than 0, the edge information richness of the cable X-ray enhanced image obtained by image enhancement is large. In this case, the image enhancement is more likely to have the problem of over-enhancement, resulting in excessive edge information, thereby increasing the difficulty of edge information recognition.

[0050] It is understandable that the smaller the sharpening factor suppression of the cable X-ray enhanced image, the more obvious the change in edge detail information in the cable X-ray enhanced image after image enhancement, the better the image enhancement effect, and the lower the necessity to suppress the image enhancement effect.

[0051] At this point, the sharpening factor suppression degree of the cable X-ray enhanced image is obtained.

[0052] Step S005: Based on the grayscale variation, image richness variation, and sharpening coefficient suppression of the cable X-ray enhanced image, determine the value of the sharpening coefficient for homomorphic filtering, and use homomorphic filtering to perform image enhancement processing on the cable X-ray image to obtain the image enhancement result of the cable X-ray image after homomorphic filtering.

[0053] The enhancement coefficient of the cable X-ray enhanced image is calculated based on the grayscale variation, image richness variation, and sharpness suppression of the enhanced image.

[0054] The normalized value of the difference between the product of the grayscale variation and the image richness variation of the cable X-ray enhanced image and the sharpening coefficient suppression is denoted as the correction coefficient of the cable X-ray enhanced image. The product of the correction coefficient of the cable X-ray enhanced image and the number 1 and the initial value of the sharpening coefficient is denoted as the enhancement coefficient of the cable X-ray enhanced image.

[0055] In the process of calculating the correction coefficient of the cable X-ray enhanced image, the Z-score normalization algorithm is used to calculate the normalized value, so that the value range of the correction coefficient of the cable X-ray enhanced image is greater than or equal to -1 and less than or equal to 1.

[0056] Experts in the field selected a total of 100 cable X-ray images and image-enhanced X-ray images from a historical database. Among them, the image-enhanced X-ray images were selected from those that were deemed to have qualified image enhancement effects by experts in the field. The mean of the correction coefficients of all image-enhanced X-ray images was calculated and recorded as the correction coefficient threshold.

[0057] When the correction coefficient of the cable X-ray enhanced image is less than or equal to the correction coefficient threshold, the enhancement coefficient of the cable X-ray enhanced image is used as the sharpening coefficient value, and homomorphic filtering is applied to the cable X-ray image for image enhancement processing to obtain the image enhancement result after homomorphic filtering. When the correction coefficient of the cable X-ray enhanced image is greater than the correction coefficient threshold, the cable X-ray enhanced image is taken as the image enhancement result after homomorphic filtering. It can be understood that when the correction coefficient of the cable X-ray enhanced image is greater than the correction coefficient threshold, the sharpening coefficient value used for image enhancement with homomorphic filtering is the initial value of the sharpening coefficient.

[0058] At this point, the image enhancement result of the cable X-ray image after homomorphic filtering is obtained.

[0059] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for homomorphic filtering enhancement of cable X-ray images, characterized in that, The method includes the following steps: Acquire cable X-ray images, set initial values ​​for sharpening coefficients, and use homomorphic filtering to enhance the cable X-ray images to obtain enhanced cable X-ray images; The corresponding pixels in the cable X-ray image and the cable X-ray enhanced image are recorded as the corresponding matching pixels. The gray-scale change of the cable X-ray enhanced image is calculated based on the difference in gray-scale values ​​between the corresponding matching pixels and their adjacent pixels. Clustering is performed on the saliency analysis results of cable X-ray images and cable X-ray enhanced images to obtain saliency clusters. Based on the differences in saliency analysis results within all saliency clusters corresponding to cable X-ray images and cable X-ray enhanced images, the image richness variation of cable X-ray enhanced images is calculated. Edge detection is performed on cable X-ray images and cable X-ray enhanced images to obtain length clusters of edges. Based on the differences in the mean and number of all lengths within the same cluster, the edge enhancement effect of the cable X-ray enhanced image is calculated. The sharpening coefficient suppression degree of the cable X-ray enhanced image is calculated by combining the differences in the number of edge pixels contained in the neighborhood of the corresponding matching pixels that are all edge pixels. Based on the grayscale variation, image richness variation, and sharpening coefficient suppression of the enhanced cable X-ray image, the value of the sharpening coefficient for homomorphic filtering is determined. Homomorphic filtering is then used to enhance the cable X-ray image, and the image enhancement result of the cable X-ray image after homomorphic filtering is obtained.

2. The method for homomorphic filtering enhancement of cable X-ray images according to claim 1, characterized in that, The specific method for obtaining the grayscale variation of the cable X-ray enhanced image is as follows: The sum of the differences between the gray values ​​of the matched pixel and all pixels in its eight neighboring regions is recorded as the adjacent gray value difference of the matched pixel. The grayscale variation of the cable X-ray enhanced image is calculated based on the grayscale difference between adjacent matching pixels in the cable X-ray enhanced image and the cable X-ray image.

3. The method for homomorphic filtering enhancement of cable X-ray images according to claim 2, characterized in that, The specific steps for calculating the grayscale variation of the cable X-ray enhanced image based on the adjacent grayscale differences of all corresponding matching pixels in the cable X-ray enhanced image and the cable X-ray image are as follows: The normalized value of the sum of the differences between adjacent gray levels of all corresponding matching pixels in the cable X-ray enhanced image and the cable X-ray image is denoted as the gray level change of the cable X-ray enhanced image.

4. The method for homomorphic filtering enhancement of cable X-ray images according to claim 1, characterized in that, The specific method for obtaining the image richness variation of the cable X-ray enhanced image is as follows: Based on the saliency analysis results of cable X-ray images and cable X-ray enhanced images, the saliency eigenvalues ​​of each saliency cluster are calculated respectively; The difference between the number of pixels contained in the salient cluster with the largest salient feature value in the cable X-ray enhanced image and the number of pixels contained in the salient cluster with the largest salient feature value in the cable X-ray image is denoted as the number of salient differences in the cable X-ray enhanced image. The positive correlation between the saliency eigenvalues ​​of all saliency clusters corresponding to the cable X-ray enhanced image and the number of saliency differences is denoted as the image richness variation degree of the cable X-ray enhanced image.

5. The method for homomorphic filtering enhancement of cable X-ray images according to claim 4, characterized in that, The specific calculation steps for the significant eigenvalue are as follows: Calculate the saliency of each pixel in cable X-ray images and cable X-ray enhanced images; The mean of all saliencies within the same saliency cluster is denoted as the saliency eigenvalue of the same saliency cluster.

6. The method for homomorphic filtering enhancement of cable X-ray images according to claim 1, characterized in that, The specific calculation method for the image edge enhancement effect of the cable X-ray enhanced image is as follows: Calculate the length feature value of the length cluster based on all lengths within the length cluster; The difference between the number of pixels contained in the length cluster with the largest length feature value in the cable X-ray enhanced image and the number of lengths contained in the length cluster with the largest length feature value in the cable X-ray image is denoted as the length difference number of the cable X-ray enhanced image. The positive correlation between the length feature values ​​of all length clusters corresponding to the cable X-ray enhanced image and the number of length differences is denoted as the image edge enhancement effect of the cable X-ray enhanced image.

7. The method for homomorphic filtering enhancement of cable X-ray images according to claim 6, characterized in that, The specific method for obtaining the length feature value is as follows: The mean of all lengths within a cluster of the same length is denoted as the length feature value of the cluster of the same length.

8. The method for homomorphic filtering enhancement of cable X-ray images according to claim 1, characterized in that, The specific calculation steps for the sharpening coefficient suppression of the cable X-ray enhanced image are as follows: Select the corresponding matching pixels that are all edge pixels. The difference between the number of edge pixels contained in the eight neighborhood of the selected matching pixels in the cable X-ray enhanced image and the cable X-ray image is recorded as the edge number difference of the selected matching pixels. The normalized value of the mean of the edge number differences of all selected matching pixels is recorded as the edge enhancement degree of the cable X-ray enhanced image. The normalized value of the sum of the edge enhancement effect and the edge enhancement degree of the cable X-ray enhanced image is denoted as the edge feature value of the cable X-ray enhanced image. The difference between the number 1 and the edge feature value of the cable X-ray enhanced image is denoted as the sharpening coefficient suppression degree of the cable X-ray enhanced image.

9. The method for homomorphic filtering enhancement of cable X-ray images according to claim 1, characterized in that, The method for determining the sharpening coefficient of homomorphic filtering based on the grayscale variation, image richness variation, and sharpening coefficient suppression of the enhanced cable X-ray image includes the following specific methods: The normalized value of the difference between the product of the grayscale variation and the image richness variation of the cable X-ray enhanced image and the sharpening coefficient suppression is denoted as the correction coefficient of the cable X-ray enhanced image. Calculate the enhancement coefficient of the cable X-ray enhanced image based on the correction coefficient of the cable X-ray enhanced image; When the correction coefficient of the cable X-ray enhanced image is less than or equal to the correction coefficient threshold, the enhancement coefficient of the cable X-ray enhanced image is used as the value of the sharpening coefficient. When the correction factor of the cable X-ray enhanced image is greater than the correction factor threshold, the initial value of the sharpening factor is used as the value of the sharpening factor.

10. The method for homomorphic filtering enhancement of cable X-ray images according to claim 9, characterized in that, The enhancement coefficient of the cable X-ray enhanced image is specifically as follows: The enhancement coefficient of the cable X-ray enhanced image is denoted as the product of the sum of the correction coefficient and the number 1 and the initial value of the sharpening coefficient.

Images