CN-121998945-A - Method and system for detecting surface defects of insulating layer of polyethylene insulated cable

CN121998945ACN 121998945 ACN121998945 ACN 121998945ACN-121998945-A

Abstract

The invention relates to the technical field of image processing, in particular to a method and a system for detecting surface defects of an insulating layer of a polyethylene insulated cable, comprising the steps of obtaining a surface image of the insulating layer; and calculating a defect response index of each pixel in the surface image, wherein the defect response index is positively correlated with the product of the gradient amplitude values of each pixel in the corresponding local window, and is positively correlated with the absolute value of the difference value of the gray average value of all pixels in the local window. According to the invention, a defect response index is constructed by introducing a gradient direction angle variance, the two are distinguished by utilizing the physical characteristics that the gradient direction of a highlight region is consistent and the defect region is disordered, and a defect significance enhancement factor is constructed according to the physical characteristics, so that the Gaussian kernel scale is dynamically adjusted. According to the method, large-kernel smooth interference is automatically adopted in a highlight region, and small kernels are adopted in a defect region to keep details, so that the limitation of a traditional fixed parameter algorithm is effectively overcome, and the defect detection precision under a complex reflection background is remarkably improved.

Inventors

LIN KUNDA
ZHANG QIURUI
WU SHUOXIN

Assignees

广州市珠江电线厂有限公司

Dates

Publication Date: 20260508
Application Date: 20260127

Claims (10)

1. The method for detecting the surface defects of the insulating layer of the polyethylene insulated cable is characterized by comprising the steps of obtaining a surface image of the insulating layer, calculating a defect response index of each pixel in the surface image, wherein the defect response index is positively correlated with the product of gradient amplitude values of each pixel in a corresponding local window, is positively correlated with the absolute value of the difference value of the gray average value of all pixels in the local window, and is positively correlated with the variance of the gradient direction angle of all pixels in the local window; Constructing defect significance enhancement factors of all pixels, wherein the defect significance enhancement factors are positively correlated with the absolute values of the differences of the defect response indexes and the global average value, and are positively correlated with the absolute values of the differences of the defect response indexes and the global minimum value; The defect significance enhancement factors are used as control parameters to construct self-adaptive kernel scales, the self-adaptive kernel scales are in negative correlation with the defect significance enhancement factors, a local binary fitting model containing the self-adaptive kernel scales is adopted to segment the preprocessed image, and the area in the outline is extracted to serve as a surface defect.
2. The method for detecting surface defects of an insulating layer of a polyethylene insulated cable according to claim 1, wherein obtaining a local window corresponding to each pixel comprises: For any pixel, a window with the size of W is built by taking the pixel as the center and is used as a local window corresponding to the pixel.
3. The method for detecting surface defects of an insulating layer of a polyethylene insulated cable according to claim 1, wherein extracting a region within a contour as a surface defect comprises: After the Gaussian kernel scale parameter corresponding to each pixel point of the full graph is obtained, the Gaussian kernel scale parameter is substituted into the full graph And iterating in the model energy functional, and continuously updating the level set function along with the increase of the iteration times until the profile curve does not change significantly or reaches the preset maximum iteration times, and finally extracting the zero level set profile of the level set function, wherein the inner area surrounded by the profile is the surface defect of the insulating layer.
4. The method for detecting surface defects of an insulation layer of a polyethylene insulated cable according to claim 1, wherein the defect significance enhancement factor has a calculation formula as follows: ; in the formula, For image coordinates A defect significance enhancement factor at the corresponding local window; For image coordinates A defect response index at the corresponding local window; Is the average value of all defect response indexes in the whole graph; Is the minimum value of all defect response indexes in the whole graph; The parameter is adjusted for a sensitivity.
5. The method for detecting surface defects of an insulating layer of a polyethylene insulated cable according to claim 4, wherein the value of the sensitivity adjustment parameter is obtained by performing adaptive threshold segmentation on a set of all defect response indexes by using an oxford method.
6. The method for detecting surface defects of an insulating layer of a polyethylene insulated cable according to claim 1, further comprising, after acquiring a surface image of the insulating layer: the surface image is preprocessed, wherein the preprocessing comprises the steps of removing background and noise points by adopting morphological-based opening and closing operation, and smoothing the image by adopting median filtering.
7. The method for detecting surface defects of an insulation layer of a polyethylene insulated cable according to claim 6, wherein the preprocessing further comprises gamma correction of the surface image to enhance contrast of an image edge region.
8. The method for detecting surface defects of an insulating layer of a polyethylene insulated cable according to claim 1, wherein acquiring a surface image of the insulating layer comprises: and synchronously shooting the surface of the insulating layer by adopting a plurality of linear array CCD cameras which are arranged around the circumference of the polyethylene cable.
9. The method for detecting surface defects of an insulating layer of a polyethylene insulated cable according to claim 1, wherein the gradient magnitude of each pixel is calculated by using a Sobel operator.
10. An insulation layer surface defect detection system of a polyethylene insulated cable, comprising a processor and a memory, characterized in that the memory stores a computer program, the processor executing the computer program to implement the insulation layer surface defect detection method of a polyethylene insulated cable according to any one of claims 1-9.

Description

Method and system for detecting surface defects of insulating layer of polyethylene insulated cable Technical Field The invention relates to the technical field of image processing. More particularly, the invention relates to a method and a system for detecting surface defects of an insulating layer of a polyethylene insulated cable. Background The integrity and surface quality of the insulation layer of the polyethylene insulated cable directly determine the electrical performance, operation safety and service life of the cable as a core component in the power transmission system. In modern cable manufacturing processes, strict surface defect detection of cables on a production line is often required in order to ensure product quality. In an actual detection scene, as the cable on the production line is in a high-speed continuous motion state, and the surface of the polyethylene insulating layer is of a smooth cylinder structure, the material has strong light reflection property. When an industrial camera is used in conjunction with illumination equipment for image acquisition, the cable surface inevitably forms specular reflection bands (high light areas) of high brightness. This highlights phenomenon not only occupies a significant position in the image, but also tends to mask the true surface texture. However, existing surface defect detection techniques, such as conventional level set (e.g., LBF model) based image segmentation methods, typically rely on globally fixed kernel scale parameters for evolution. When processing such cable images with both strong highlight interference and weak defect characteristics, the method for fixing parameters faces the practical contradiction that if a larger kernel scale is selected for suppressing interference and background noise of a highlight region, the edges of weak texture defects such as extremely tiny scratches or shallow bubbles on the surface of an insulating layer are excessively smoothed to cause omission, otherwise, if a smaller kernel scale is selected for capturing the weak defect details, an algorithm is extremely easy to generate anaphylactic reaction on the highlight edges and noise points, and the highlight reflection band is mistakenly identified as the defect. Disclosure of Invention The invention provides a method and a system for detecting surface defects of an insulating layer of a polyethylene insulated cable, and aims to solve the problem of how to accurately separate weak defects from a complex background while inhibiting highlight interference in the related art. In a first aspect, the invention provides an insulation layer surface defect detection method of a polyethylene insulation cable, which comprises the steps of obtaining a surface image of the insulation layer, calculating a defect response index of each pixel in the surface image, wherein the defect response index is in positive correlation with a product of gradient magnitudes of each pixel in a corresponding local window, in positive correlation with an absolute value of a difference value of a gray average value of all pixels in the local window and in positive correlation with a variance of a gradient direction angle of all pixels in the local window, constructing a defect saliency enhancement factor of each pixel, in positive correlation with the difference absolute value of the defect response index and a global average value, in positive correlation with the absolute value of the difference value of the defect response index and a global minimum value, using the defect saliency enhancement factor as a control parameter, constructing a self-adaptive kernel scale, in negative correlation with the defect saliency enhancement factor, segmenting the preprocessed image by adopting a local binary model containing the self-adaptive kernel scale, and extracting a region in the contour as a surface defect. By constructing the defect response index which merges the characteristics of gradient direction angle variance and the like, a highlight region (consistent gradient direction) and a real defect (disordered gradient direction) are effectively distinguished, and an LBF model with a self-adaptive kernel scale is built according to the defect response index, the contradiction that the existing fixed kernel scale technology cannot balance between inhibiting highlight interference and retaining weak defect details is solved, and the accurate segmentation and detection of surface defects such as micro scratches and bubbles under a strong light reflection background are realized. Further, obtaining the local window corresponding to each pixel comprises constructing a window with a size of W by taking the pixel as a center for any pixel, and taking the window as the local window corresponding to the pixel. The spatial range of local feature calculation is defined, gray scale and gradient statistical information in the pixel neighborhood can be fully utilized when the defect response index