CN-121998982-A - Cable interlayer structure detection method and system based on image processing

CN121998982ACN 121998982 ACN121998982 ACN 121998982ACN-121998982-A

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a cable interlayer structure detection method and system based on image processing, wherein the method comprises the steps of calculating the boundary credibility score of pixel points based on multi-dimensional characteristics of cable images so as to accurately identify boundaries; the method comprises the steps of generating an internal constraint factor by combining the shape and the quality of a boundary band, dynamically adjusting parameters of an active contour model according to the internal constraint factor to accurately extract a closed contour line, and finally evaluating thickness uniformity of an insulating layer based on the contour line. The invention solves the problem of limitation of the fixed parameter model in processing fuzzy or broken boundaries by dynamically adjusting the model parameters, and improves the accuracy and the robustness of the detection of the cable interlayer structure.

Inventors

GUO XIONGTAO
YAO WANG
WANG ZHANSHENG
LIU BINGJI

Assignees

淳化昆仑优佳电缆有限公司

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. An image processing-based cable interlayer structure detection method is characterized by comprising the following steps: carrying out circular detection on the target image to obtain an ROI region; obtaining boundary credibility scores of pixel points based on gray level change intensity, texture complexity degree and spatial coherence characteristics of edge directions of the pixel points in the ROI area, obtaining suspicious insulating layer boundaries according to the relation between the boundary credibility scores and a preset threshold value, dividing the suspicious insulating layer boundaries to obtain a plurality of interlayer boundary bands, and further combining the discrete degree of the pixel points on each interlayer boundary band with the boundary credibility scores to obtain internal constraint factors of the interlayer boundary bands; According to the internal constraint factors and preset basic model parameters, dynamic model parameters of the movable contour model are obtained; The method comprises the steps of obtaining a closed contour line group based on the obtained centroid distance between the closed contour lines, obtaining the quality index of an insulating layer according to continuous thickness data of the closed contour line group under continuous angles, comparing the quality index with a quality threshold value, and evaluating thickness uniformity of the insulating layer.
2. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein the acquisition of the ROI area comprises: A hough circle detection is used on the target image to detect a circular contour, with the largest circle being the ROI area.
3. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein the boundary confidence score of the pixel point satisfies a relation: ; Wherein, the Is a target pixel point positioned in the region of the ROI Boundary trust scores for (2); is the target pixel point in the ROI area Is a gradient magnitude of (a); is the target pixel point in the ROI area Is a local neighborhood entropy of (1); is the target pixel point in the ROI area Is used for normalizing the gradient vector; is the target pixel point in the ROI area The average value of all pixel gradient vectors in the local neighborhood is normalized to obtain a unit vector; Is a preset tiny value; is a normalization function; is a maximum function; Is the euclidean norm symbol of the vector.
4. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein the obtaining of the boundary of the suspected insulating layer comprises: Setting a threshold value, and binarizing the boundary credibility scores of all the pixel points; setting the pixel points with the boundary credibility score larger than a preset threshold value as 1 on the ROI area, and setting the rest pixel points as 0; And extracting all pixel points with the value of 1 to form a suspected insulating layer boundary.
5. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein the acquisition of the interlayer boundary strip comprises: and (3) carrying out region segmentation on the boundary of the suspected insulating layer by adopting a watershed algorithm, and dividing the boundary into interlayer boundary bands with preset numbers.
6. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein an internal constraint factor of the interlayer boundary band satisfies a relation: ; Wherein, the Is the first Internal constraint factors of the strip interlayer boundary bands; Is the first The boundary between the stripe layers has the total area covered by pixel points; Is to the first The pixel length of the skeleton line is obtained after the skeletonizing treatment of the strip interlayer boundary band; Is the maximum value of the square ratio of the total area covered by the pixel points and the length of the skeleton line in all interlayer boundary bands; Is the first Boundary credible fractional average value of all pixel points on the boundary band between the strips; a standard threshold value of a preset standard boundary credibility score; is a minimum function.
7. The method for detecting a cable interlayer structure based on image processing according to claim 1, wherein the acquisition of the dynamic model parameters of the active contour model comprises: Calculating the product of a preset basic external force weight and a difference value obtained by subtracting the internal constraint factor from 1 to obtain a dynamic external force weight of the movable contour model; calculating the product of a preset basic rigidity coefficient and the sum of 1 and the internal constraint factor to obtain a dynamic rigidity coefficient of the movable contour model; And calculating the difference value of the product result subtracted from the 1 and multiplying the difference value with a preset basic elastic coefficient to obtain a dynamic elastic coefficient.
8. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein the acquisition of the closed contour line group comprises: Calculating the mass center of each closed contour line, and carrying out ascending arrangement according to the distance from the mass center of each closed contour line to the center of the ROI area to obtain a mass center sequence; And taking two closed contour lines corresponding to two adjacent centroids in the centroid sequence as a closed contour line group.
9. The method for detecting an interlayer structure of a cable based on image processing according to claim 1, wherein the obtaining of the quality index of the insulating layer comprises: taking the average value of two centroids of the closed contour line group as a reference center; Emitting radial rays along a reference center to obtain intersection points with the closed contour line group, and obtaining continuous thickness data of the insulating layer; Obtaining a maximum thickness value and a minimum thickness value of each insulating layer under all angles from the continuous thickness data; and adding the difference value of the maximum thickness value and the minimum thickness value of the continuous thickness data of each insulating layer to obtain the quality index of the insulating layer.
10. An image processing-based cable interlayer structure detection system comprising a processor and a memory, the memory storing computer program instructions which, when executed by the processor, implement an image processing-based cable interlayer structure detection method according to any one of claims 1 to 9.

Description

Cable interlayer structure detection method and system based on image processing Technical Field The invention relates to the technical field of image processing. More particularly, the invention relates to a cable interlayer structure detection method and system based on image processing. Background The cable is used as a key carrier for power and information transmission, and the manufacturing quality of the cable is related to the safety of the system. In the cable interlayer structure, the thickness and uniformity of the insulating layer are core quality indicators. The insulation layer is partially insufficient in thickness, so that the insulation layer becomes a defect of electric field concentration, breakdown is easy to occur under high voltage, faults are caused, meanwhile, the mechanical strength and ageing resistance of the cable are reduced due to uneven thickness, and the service life of the cable is shortened. In the links of cable production quality control and later nondestructive testing, the cross-section images of the cable are usually required to be analyzed, and the thickness uniformity of the insulating layer is ensured. In order to realize automatic detection of the cable interlayer structure, a visual analysis method based on image segmentation and contour extraction is commonly adopted at present. In the related art, for example, chinese patent CN120182257B discloses a method and a system for detecting a cross section of a cable core of a power supply and distribution cable for a building, which discloses that an inner contour and an outer contour of a conductor are obtained by cutting points, and an outer contour is estimated by combining with a convolutional neural network iteration, so as to evaluate the roundness and the quantity of the conductor to determine the quality. However, when the method is applied to insulating layer detection, the method relies on a preset geometric model and deep learning estimation, and is difficult to directly apply to weak contrast boundary extraction of non-conductor materials such as insulating layers. When the image is blurred or the gray level difference between layers is weak, adjacent insulating interfaces cannot be effectively distinguished, and boundary mismatching or missed detection is caused. In order to further improve the boundary positioning accuracy, the prior art also commonly adopts a movable contour model to carry out closed contour evolution. However, as shown in the above CN120182257B patent, the whole frame still converges with the fixed parameter driving profile, and when there is a blur or fracture in the cable section image, the movable profile model with the fixed parameter is prone to leakage due to insufficient driving force or loss of details due to excessive smoothness, which results in deviation of the extracted boundary from the real physical interface, and affects the accuracy of uniform evaluation of the thickness of the insulating layer. Disclosure of Invention In order to solve the technical problem of the above-described evaluation accuracy, the present invention provides the following aspects. In a first aspect, the present invention provides a method for detecting an interlayer structure of a cable based on image processing, including: The method comprises the steps of preprocessing a cable cross section image to obtain a target image, carrying out circular detection on the target image to obtain an ROI region, obtaining boundary credibility scores of pixel points based on pixel point gray level change intensity, texture complexity and spatial coherence characteristics of edge directions in the ROI region, obtaining suspicious insulating layer boundaries according to the relation between the boundary credibility scores and a preset threshold value, dividing the suspicious insulating layer boundaries to obtain a plurality of interlayer boundary bands, further combining the discrete degree of the pixel points on each interlayer boundary band and the boundary credibility scores to obtain internal constraint factors of the interlayer boundary bands, obtaining dynamic model parameters of a movable contour model according to the internal constraint factors and preset basic model parameters, inputting the dynamic model parameters into the movable contour model to obtain closed contour lines of the interlayer boundary bands, obtaining a closed contour line group based on centroid distances among the obtained closed contour lines, obtaining quality indexes of the insulating layer according to continuous thickness data of the closed contour line group under continuous angles, comparing the quality indexes with the quality threshold value, and evaluating thickness uniformity of the insulating layer. The method comprises the steps of carrying out multistage processing and analysis on the cable cross-section image, including pixel point boundary credibility score calculation, inter-layer boundary zone internal constraint f