A real-time detection method and system for conductor defects

Abstract

The present application relates to the technical field of conductor defect detection, and specifically provides a method and system for real-time detection of conductor surface defects, the method comprising the steps of: obtaining insulation layer transparency information based on optical features of an optical image with an optical wavelength of a preset wavelength, and then determining parameters of a region recognition algorithm and parameters of a first defect detection algorithm based on the insulation layer transparency information; using the region recognition algorithm to perform surface print mark recognition on an optical image with an optical wavelength within a preset band to obtain a surface print identification area; performing feature extraction on areas other than the surface print identification area in all optical images to obtain a first feature map reflecting information at different levels; using a first defect detection algorithm to perform defect recognition on the first feature map, and then determining the location of the surface defect based on the feature differences of the same surface defect in different optical images; this method can effectively improve the accuracy, reliability and robustness of defect detection.

Description

Technical Field

[0001] The present application relates to the technical field of conductor defect detection, and in particular to a method and system for real-time detection of conductor defects. Background Art

[0002] When producing translucent insulated conductors (such as insulated wires with a translucent insulation layer), the integrity and consistency of the insulation layer are directly related to the electrical performance and safety of the product. Existing technology uses an automated inspection system based on machine vision to detect defects in translucent insulated conductors.

[0003] However, the translucent nature of insulating wires presents numerous challenges for existing automated inspection systems. First, the translucency of the insulation material allows the camera to capture not only the features of the outer surface of the insulation, but also, to a certain extent, the interior of the insulation and even the surface of the conductor. This means that the inspection system must simultaneously monitor surface defects (such as scratches or surface roughness), internal defects (such as bubbles or impurities), and surface anomalies that may be visible through the insulation. These defects at different levels may overlap or interfere with each other in the image, making them difficult to accurately distinguish. Consequently, existing automated inspection systems lack the accuracy and reliability to detect defects in translucent insulated conductors.

[0004] Secondly, the transparency of the insulation material may fluctuate due to factors such as raw material batch differences, extrusion temperature and cooling rate. Changes in the transparency of the insulation layer will directly affect the visibility of the underlying conductor or internal features. Existing machine vision algorithms usually use fixed parameters for defect detection. Fixed parameter defect detection is difficult to adapt to the impact of such transparency fluctuations, resulting in reduced robustness and accuracy of defect detection. For example, when the transparency is high, the internal and conductor features are clearer, but it is also easier to misjudge the normal texture of the conductor surface as a defect; when the transparency is low, internal and conductor defects may become blurred or even invisible, resulting in missed detection.

[0005] In addition, in order to identify product information, the outer surface of the insulation layer is usually marked with inkjet printing or laser printing (such as model, specifications, production date, etc.). These printed marks will interfere with defect detection. Therefore, the existing technology also has the problem of misjudging the printed marks as defects due to the interference of the printed marks with defect detection.

[0006] There is no effective technical solution to the above problems. It should be noted that the above information disclosed in this section is only used to understand the background of the present invention, and therefore may contain information that does not constitute prior art. Summary of the Invention

[0007] The purpose of this application is to provide a method and system for real-time detection of conductor defects, which can effectively improve the accuracy, reliability and robustness of defect detection.

[0008] In a first aspect, the present application provides a method for real-time detection of conductor defects, which is used to detect defects in a translucent insulated conductor, and comprises the following steps:

[0009] S1. Acquire a multispectral image sequence of a semi-transparent insulating conductor, where the multispectral image sequence includes optical images at different optical wavelengths;

[0010] S2. Obtaining insulation layer transparency information based on optical features of an optical image having a preset wavelength, and then determining parameters of a region recognition algorithm and a first defect detection algorithm based on the insulation layer transparency information;

[0011] S3, using a region recognition algorithm to perform surface printing mark recognition on an optical image whose optical wavelength is within a preset wavelength band to obtain a surface printing mark area;

[0012] S4, performing feature extraction on the areas except the surface printed mark area in all optical images to obtain a first feature map reflecting information at different levels;

[0013] S5. Use the first defect detection algorithm to identify defects in the first feature image. Then, if a defect exists, determine the location of the defect based on feature differences in optical images corresponding to different optical wavelengths for the same defect.

[0014] The present application provides a real-time conductor defect detection method that can adapt the region recognition algorithm and the first defect detection algorithm to the current transparency of the insulation layer by determining the parameters of the region recognition algorithm and the first defect detection algorithm based on the transparency information of the insulation layer. Therefore, the present application can effectively solve the problem of reduced robustness and accuracy of defect detection due to the difficulty of defect detection with fixed parameters to adapt to the influence of such transparency fluctuations. In addition, the present application can eliminate the interference of printed marks on defect detection by first using the region recognition algorithm to identify the surface printed mark area and then performing defect detection on the area other than the surface printed mark area in the optical image. Therefore, the present application can effectively solve the problem of misjudging printed marks as defects due to the interference of printed marks in defect detection. Moreover, since the present application can realize defect identification and positioning at different levels by performing multi-spectral feature extraction and defect identification on non-printed mark areas and determining the position of the same defect by utilizing the feature differences in optical images corresponding to different optical wavelengths, the present application can also effectively solve the problem of insufficient defect detection accuracy and reliability of translucent insulated conductors due to the fact that defects at different levels may overlap or interfere with each other in the image and are difficult to distinguish accurately. That is, the present application can effectively improve the accuracy, reliability and robustness of defect detection.

[0015] In a second aspect, the present application further provides a conductor defect real-time detection system for detecting defects in a translucent insulated conductor, comprising:

[0016] An image sequence acquisition module is used to acquire a multispectral image sequence of the semi-transparent insulating conductor, wherein the multispectral image sequence includes optical images of different optical wavelengths;

[0017] an algorithm parameter confirmation module, configured to obtain insulation layer transparency information based on optical features of an optical image having a preset wavelength, and then determine parameters of a region recognition algorithm and parameters of a first defect detection algorithm based on the insulation layer transparency information;

[0018] A print mark area recognition module is used to use an area recognition algorithm to perform surface print mark recognition on an optical image whose optical wavelength is within a preset wavelength band to obtain a surface print mark area;

[0019] A feature extraction module is used to extract features from all areas of the optical image except the surface printed mark area to obtain a first feature map reflecting information at different levels;

[0020] The defect detection module is used to identify defects on the first feature map using a first defect detection algorithm, and then determine the location of the defect based on the feature differences of the same defect in optical images corresponding to different optical wavelengths when the defect exists.

[0021] The present application provides a real-time conductor defect detection system that can adapt the region recognition algorithm and the first defect detection algorithm to the current transparency of the insulation layer by determining the parameters of the region recognition algorithm and the first defect detection algorithm based on the transparency information of the insulation layer. Therefore, the present application can effectively solve the problem of reduced robustness and accuracy of defect detection caused by the difficulty of defect detection with fixed parameters to adapt to the influence of such transparency fluctuations. In addition, the present application can eliminate the interference of printed marks on defect detection by first using the region recognition algorithm to identify the surface printed mark area and then performing defect detection on the area other than the surface printed mark area in the optical image. Therefore, the present application can effectively solve the problem of misjudging printed marks as defects due to the interference of printed marks in defect detection. Moreover, since the present application can realize defect identification and positioning at different levels by performing multi-spectral feature extraction and defect identification on non-printed mark areas and determining the position of the same defect by utilizing the feature differences in optical images corresponding to different optical wavelengths, the present application can also effectively solve the problem of insufficient defect detection accuracy and reliability of translucent insulated conductors due to the fact that defects at different levels may overlap or interfere with each other in the image and are difficult to distinguish accurately. That is, the present application can effectively improve the accuracy, reliability and robustness of defect detection.

[0022] From the above, it can be seen that the present application provides a method and system for real-time detection of conductor defects, which can make the region recognition algorithm and the first defect detection algorithm suitable for the current transparency of the insulation layer by determining the parameters of the region recognition algorithm and the parameters of the first defect detection algorithm according to the transparency information of the insulation layer. Therefore, the present application can effectively solve the problem that the robustness and accuracy of defect detection are reduced due to the fact that defect detection with fixed parameters is difficult to adapt to the influence of such transparency fluctuations. Since the present application can eliminate the influence of printed marks on defect detection by first using the region recognition algorithm to identify the surface printed mark area, and then performing defect detection on the area other than the surface printed mark area in the optical image, Therefore, the present application can effectively solve the problem of misjudging printed marks as defects due to the interference of printed marks on defect detection, and because the present application can realize defect recognition and positioning at different levels by performing multi-spectral feature extraction and defect recognition on non-printed mark areas and determining the position of the same defect by utilizing the feature differences in optical images corresponding to different optical wavelengths, the present application can also effectively solve the problem of insufficient defect detection accuracy and reliability of translucent insulated conductors due to the fact that defects at different levels may overlap or interfere with each other in the image and are difficult to distinguish accurately, that is, the present application can effectively improve the accuracy, reliability and robustness of defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1A flowchart of a method for real-time detection of conductor defects provided in an embodiment of the present application.

[0024] Figure 2 A schematic structural diagram of a real-time conductor defect detection system provided in an embodiment of the present application.

[0025] Figure numerals: 1. Image sequence acquisition module; 2. Algorithm parameter confirmation module; 3. Print mark area recognition module; 4. Feature extraction module; 5. Defect detection module. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work fall within the scope of protection of the present application.

[0027] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0028] Traditional automated inspection systems for defect detection of translucent insulated conductors suffer from insufficient accuracy, robustness, and reliability. The translucency of the insulation material causes the image captured by the camera to include information about the outer surface, the interior of the insulation layer, and the conductor surface. Defects at different levels overlap or interfere with each other in the image, making accurate distinction difficult. Furthermore, the transparency of the insulation material can fluctuate. Existing machine vision algorithms typically use fixed parameters for defect detection, making it difficult to adapt to the effects of such transparency fluctuations, resulting in reduced robustness and accuracy. Printed markings on the outer surface of the insulation layer can also interfere with defect detection, causing printed markings to be misidentified as defects.

[0029] For example, consider a high-speed production line using a camera-based inspection system to perform in-line defect detection on translucent insulated wires. Due to the translucency of the insulation layer, not only scratches on the insulation surface are visible in the image, but also bubbles within the insulation layer or texture on the conductor surface. If the transparency of the insulation layer increases due to variations in production batches or process parameters, the normal surface texture of the conductor becomes more prominent in the image, potentially leading to it being mistakenly identified as a defect. Conversely, if the transparency decreases, bubbles within the insulation layer may become obscured, causing them to be missed. Furthermore, the model printed markings on the wire surface create specific patterns in the image. These patterns may resemble the image characteristics of certain defects, causing the inspection system to misidentify the printed markings as defects. These issues are particularly prominent in high-speed production environments, requiring the inspection system to handle them quickly and accurately.

[0030] If these issues are not addressed, defect detection systems for translucent insulated conductors will experience high rates of false positives and missed negatives. False positives increase subsequent manual re-inspection costs and reduce production efficiency, while missed negatives can lead to defective products entering the market, compromising product quality, user safety, and damaging corporate reputation. This directly impacts the effectiveness of automated inspection systems in quality control of translucent insulated conductor production.

[0031] In this regard, firstly, Figure 1 As shown, the present application proposes a real-time conductor defect detection method for detecting defects in a translucent insulating conductor, which includes the following steps:

[0032] S1. Acquire a multispectral image sequence of a semi-transparent insulating conductor, where the multispectral image sequence includes optical images at different optical wavelengths;

[0033] S2. Obtaining insulation layer transparency information based on optical features of an optical image having a preset wavelength, and then determining parameters of a region recognition algorithm and a first defect detection algorithm based on the insulation layer transparency information;

[0034] S3, using a region recognition algorithm to perform surface printing mark recognition on an optical image whose optical wavelength is within a preset wavelength band to obtain a surface printing mark area;

[0035] S4, performing feature extraction on the areas except the surface printed mark area in all optical images to obtain a first feature map reflecting information at different levels;

[0036] S5. Use the first defect detection algorithm to identify defects in the first feature image. Then, if a defect exists, determine the location of the defect based on feature differences in optical images corresponding to different optical wavelengths for the same defect.

[0037] A semi-transparent insulated conductor is one covered by an insulating layer with a certain degree of light transmittance. This characteristic allows simultaneous image capture of surface, interior, and conductor-level information during image acquisition, posing a challenge for defect detection. A multispectral image sequence is a series of optical images acquired at different optical wavelengths. Its purpose is to capture characteristic information reflecting different layers of the conductor (surface, interior, and conductor) by exploiting the differences in material penetration and reflectivity at different wavelengths. This embodiment can utilize a multispectral camera or a filter set in conjunction with the camera to acquire a multispectral image sequence.

[0038] The optical characteristics of an optical image with an optical wavelength of a preset wavelength refer to the optical properties exhibited by the optical image with an optical wavelength of a preset wavelength in a multispectral image sequence (equivalent to the optical properties exhibited by an image collected at a specific selected optical wavelength). The optical characteristics can be pixel values, color values, and texture characteristics of pixels. Since the optical characteristics of the optical image are associated with the transparency of the insulating layer, for example, if the transparency of the insulating layer is higher, the image will be brighter; conversely, if the transparency is lower, the image will be darker. Therefore, this embodiment can obtain the transparency information of the insulating layer based on the optical characteristics of the optical image with an optical wavelength of a preset wavelength. It should be understood that since the insulating layer of a translucent insulating conductor has different absorption and scattering characteristics for light of different wavelengths, and at a specific wavelength, the absorption coefficient or scattering coefficient of the insulating layer will change with slight changes in transparency. That is, even if the transparency of the insulating layer fluctuates slightly, the optical characteristics (such as pixel grayscale value, brightness, contrast, etc.) of the optical image collected at the specific wavelength will also change significantly accordingly. Therefore, the preset wavelength of this embodiment is preferably a wavelength that is more sensitive to changes in the transparency of the insulating layer. Step S2 can obtain the transparency information of the insulating layer by querying a pre-constructed mapping relationship table of optical characteristics and transparency according to the optical characteristics of the optical image with the preset wavelength. Step S2 can also obtain the transparency information of the insulating layer by inputting the optical characteristics of the optical image with the preset wavelength into a pre-trained transparency evaluation model. The transparency evaluation model is used to evaluate the transparency of the insulating layer according to the input optical characteristics, and generate and output the transparency information of the insulating layer according to the evaluation results. The insulation layer transparency information of this embodiment refers to a quantitative description of the light transmittance of the insulation layer. Since step S2 determines the parameters of the region recognition algorithm and the first defect detection algorithm based on the insulation layer transparency information, that is, step S2 uses the insulation layer transparency information as confirmation of the parameters of the region recognition algorithm and the first defect detection algorithm, step S2 is equivalent to adaptively adjusting the parameters of the region recognition algorithm and the first defect detection algorithm based on the insulation layer transparency information, so that the region recognition algorithm and the first defect detection algorithm can adapt to the current transparency of the insulation layer. That is, even if the transparency of the insulation layer material fluctuates, this embodiment can accurately identify the surface printed mark area and defects. The parameters of the region recognition algorithm and the first defect detection algorithm refer to configuration values ​​or settings used to control the behavior and performance of the region recognition algorithm and the first defect detection algorithm. Their purpose is to enable the algorithm to adaptively adjust its processing method based on the insulation layer transparency information to improve the robustness of detection.

[0039] Surface printed mark recognition in step S3 involves automatically identifying and locating printed marks on the surface of a translucent insulated conductor using a region recognition algorithm after parameter confirmation. The goal is to exclude these printed marks from defect detection to avoid misidentifying them as defects. Specifically, the region recognition algorithm in this embodiment can employ template matching, feature point detection, or deep learning methods to identify surface printed marks in optical images with an optical wavelength within a preset wavelength band. It should be understood that the preset wavelength is a specific wavelength value, and the preset wavelength band is a wavelength range. Therefore, the surface printed mark recognition target in step S3 is essentially an optical image with a wavelength corresponding to the preset wavelength within the preset wavelength band.

[0040] Because the multispectral image sequence includes optical images of different optical wavelengths, and light of different wavelengths interacts differently with the translucent insulating conductor, optical images of different optical wavelengths can reflect information at different levels of the conductor. For example, optical images of shorter wavelengths (such as ultraviolet light) primarily reflect information on the surface of the insulation layer (such as scratches or contaminants on the insulation layer surface), while optical images of longer wavelengths (such as infrared light, which can penetrate the insulation layer) can reflect information within the insulation layer or on the surface of the conductor (such as internal bubbles, impurities, or texture on the conductor surface). Therefore, step S4 can extract features from these optical images of different wavelengths and fuse the extracted features to obtain a comprehensive feature representation (first feature map) that can reflect information at different levels of the conductor. This can be achieved by extracting features from all areas of the optical images except the surface printed mark area to obtain the first feature map reflecting information at different levels. Step S4 can use edge detection, texture analysis, color space conversion, or deep learning feature extraction to extract features from all areas of the optical images except the surface printed mark area.

[0041] Defect identification in step S5 refers to automatically identifying defects in the first feature map using the first defect detection algorithm after parameter confirmation. It should be understood that if no defect is identified (i.e., no defect exists), step S5 terminates. Because the same defect may appear differently in optical images corresponding to different optical wavelengths (i.e., the features of the same defect in optical images corresponding to different optical wavelengths differ, i.e., the same defect may have different appearance, contrast, shape, or visibility in optical images captured at different optical wavelengths), for example, defects on the surface of an insulating layer are more obvious in short-wavelength images, while defects within the insulating layer and defects on the surface of a conductor are more obvious in long-wavelength images. Therefore, when a defect exists, step S5 determines the location of the defect based on the feature differences between the optical images corresponding to different optical wavelengths, thereby achieving layered defect location. Because step S5 determines the defect location based on the feature differences between the optical images corresponding to different optical wavelengths, this embodiment effectively avoids the situation where defect location misjudgment occurs due to using only single image information for defect location (e.g., misjudging a defect actually located on the surface of an insulating layer as located within the insulating layer or on the surface of a conductor), thereby effectively improving the accuracy and reliability of defect location.

[0042] The core innovation of this application lies in obtaining a multispectral image sequence of a translucent insulating conductor and obtaining the transparency information of the insulation layer based on the optical characteristics of the image at a specific wavelength, thereby adaptively adjusting the parameters of the region recognition algorithm and the first defect detection algorithm according to the transparency information. At the same time, the region recognition algorithm is used to exclude the surface printed mark area, thereby solving the problem of insufficient detection accuracy caused by fluctuations in the transparency of the insulation layer and interference from printed marks, and achieving the effect of improving detection robustness and accuracy.

[0043] The solution of the present application provides multi-dimensional information for subsequent analysis by acquiring a multi-spectral image sequence of a translucent insulating conductor. Then, the transparency information of the insulating layer is obtained based on the optical characteristics of the image at a specific optical wavelength, and the parameters of the region recognition algorithm and the first defect detection algorithm are dynamically adjusted using this information so that the detection process can adapt to fluctuations in the transparency of the insulating layer. Subsequently, the adjusted region recognition algorithm is used to identify and exclude the surface printed mark area on the specific band image, thereby avoiding the interference of the printed mark on the defect detection. Then, feature extraction is performed on the non-printed mark area in all optical images to obtain a first feature map reflecting information at different levels. Finally, the first feature map is used to identify defects using the adjusted first defect detection algorithm, and when a defect is detected, its position is determined based on the feature difference of the same defect in the optical images corresponding to different optical wavelengths, thereby achieving accurate identification and positioning of defects in the translucent insulating conductor.

[0044] In some preferred embodiments, a multispectral image sequence can be acquired by equipping multiple narrowband filters or using a multispectral camera, for example, to capture images in the visible and near-infrared bands. The preset wavelength can be a specific wavelength sensitive to changes in the transparency of the insulating layer material, such as a near-infrared wavelength. In this embodiment, transparency can be characterized by calculating the average pixel value in the optical image at this preset wavelength as an optical feature. This embodiment determines parameters for the region recognition algorithm and the first defect detection algorithm by querying a pre-established mapping table. This mapping table maps these average pixel values ​​to insulation layer transparency levels, and then maps the insulation layer transparency levels to specific parameter sets (such as thresholds, filter parameters, or model weights) for the region recognition algorithm (e.g., based on template matching or feature point detection) and the first defect detection algorithm (e.g., based on image segmentation or classification). For example, when the optical feature is the average grayscale value, a grayscale value in the range of 0-50 can be set to indicate low transparency; a grayscale value in the range of 51-150 can be set to indicate medium transparency; and a grayscale value in the range of 151-255 can be set to indicate high transparency. When the optical feature is a color feature, reddish colors can be assigned a low transparency level, greenish colors can be assigned a medium transparency level, and bluish colors can be assigned a high transparency level. Surface printed mark recognition can be performed on visible light images, using a region recognition algorithm to identify the printed mark area. Feature extraction can be performed using a multi-channel convolutional neural network (CNN) for joint feature extraction of visible and near-infrared images. This CNN can be an existing multi-channel weighted convolutional neural network (MC-WCNN), a multi-channel one-dimensional convolutional network (MC-1DCNN), a multi-channel convolutional attention network (MC-CAN), or a depthwise separable multi-channel convolutional network (such as MobileNet and ShuffleNet). Defect identification can utilize a first defect detection algorithm to analyze the extracted feature maps. This first defect detection algorithm can be obtained by training a convolutional neural network using a pre-constructed dataset containing various defective and non-defective samples, and continuously adjusting the network parameters using a back-propagation algorithm during the training process to enable the algorithm to learn effective features for distinguishing defects from non-defects. This first defect detection algorithm can also be obtained by fine-tuning a pre-trained model (preferably a convolutional neural network model pre-trained on a large dataset such as ImageNet, such as ResNet, VGGNet, or EfficientNet) using a pre-constructed dataset containing a small number of defective and non-defective samples. When a defect is detected, the precise location of the defect can be determined by comparing the contrast or morphological differences of the defect in visible light and near-infrared images.

[0045] The present application provides a real-time conductor defect detection method that can adapt the region recognition algorithm and the first defect detection algorithm to the current transparency of the insulation layer by determining the parameters of the region recognition algorithm and the first defect detection algorithm based on the transparency information of the insulation layer. Therefore, the present application can effectively solve the problem of reduced robustness and accuracy of defect detection due to the difficulty of defect detection with fixed parameters to adapt to the influence of such transparency fluctuations. In addition, the present application can eliminate the interference of printed marks on defect detection by first using the region recognition algorithm to identify the surface printed mark area and then performing defect detection on the area other than the surface printed mark area in the optical image. Therefore, the present application can effectively solve the problem of misjudging printed marks as defects due to the interference of printed marks in defect detection. Moreover, since the present application can realize defect identification and positioning at different levels by performing multi-spectral feature extraction and defect identification on non-printed mark areas and determining the position of the same defect by utilizing the feature differences in optical images corresponding to different optical wavelengths, the present application can also effectively solve the problem of insufficient defect detection accuracy and reliability of translucent insulated conductors due to the fact that defects at different levels may overlap or interfere with each other in the image and are difficult to distinguish accurately. That is, the present application can effectively improve the accuracy, reliability and robustness of defect detection.

[0046] Since there may be local differences in the transparency of the insulation layer, if only globally unified parameters are used for region identification and defect detection, the accuracy of region identification and defect detection may be reduced. For example, in areas with low transparency, the printed mark may be blurred, resulting in failure of printed mark recognition; in areas with high transparency, normal textures inside the insulation layer may be mistakenly identified as defects.

[0047] In order to solve this technical problem, in some preferred embodiments, step S2 includes:

[0048] S21, obtaining a pixel value of each pixel point in an optical image having an optical wavelength of a preset wavelength;

[0049] S22, using a clustering algorithm to divide the optical image having a preset optical wavelength into a plurality of transparency analysis regions according to all pixel values;

[0050] S23. For each transparency analysis area, query a pre-built mapping relationship table between optical features and transparency based on the optical features of the area in the optical image having a preset optical wavelength to obtain transparency information of the insulating layer;

[0051] S24. For each transparency analysis region, query a pre-built mapping relationship table of insulation layer transparency, region recognition algorithm parameters, and defect detection parameters based on the insulation layer transparency information to obtain corresponding parameters of the region recognition algorithm and the first defect detection algorithm;

[0052] Step S3 includes:

[0053] S31, for each transparency analysis area, using a corresponding area recognition algorithm to perform surface printing mark recognition on the transparency analysis area to obtain a surface printing mark area;

[0054] Step S5 includes:

[0055] S51. For each transparency analysis area, use the corresponding first defect detection algorithm to identify defects in the area of ​​the transparency analysis area except the surface printed mark area. Then, when a defect exists, determine the location of the defect based on the characteristic differences of the same defect in optical images corresponding to different optical wavelengths.

[0056] The clustering algorithm refers to an unsupervised learning algorithm used to divide a dataset into several clusters based on the similarity between samples, such that samples within the same cluster have a high degree of similarity, while samples between different clusters have a low degree of similarity. The clustering algorithm of this embodiment can adopt a K-means algorithm, a mean-shift clustering algorithm, or a spectral clustering algorithm, among others. This embodiment can divide an optical image having a preset optical wavelength into multiple transparency analysis regions based on the optical characteristics reflected by the pixel values ​​by utilizing the clustering algorithm to divide the optical image into regions with relatively consistent pixel values. Since transparency is associated with pixel values, this embodiment is equivalent to dividing the optical image into regions with relatively consistent transparency. The transparency analysis region refers to an image subregion obtained by the clustering algorithm. The transparency of the insulating layer within each transparency analysis region has a certain degree of consistency. The transparency analysis region is the basic unit for performing local transparency analysis and parameter adjustment. The mapping relationship table between optical features and transparency in this embodiment refers to a lookup table that stores the relationship between multiple optical features and corresponding insulation layer transparency values. The mapping relationship table can be established by performing image acquisition and feature analysis on insulation layer samples with known transparency. Its purpose is to quickly obtain the transparency information of the insulation layer based on the optical features of the optical image. The mapping relationship table of the insulating layer transparency, region recognition algorithm parameters and defect detection parameters of this embodiment refers to a lookup table that stores the relationship between different insulating layer transparency values ​​and the region recognition algorithm and first defect detection algorithm parameter set optimized for the transparency. For example, for low-transparency areas, the similarity threshold of the region recognition algorithm (such as template matching based on correlation coefficient) can be set to a lower value, and the edge strength threshold of the first defect detection algorithm (such as detection based on edge strength) can be set to a lower value; for high-transparency areas, the similarity threshold of the region recognition algorithm can be set to a higher value, the edge strength threshold of the first defect detection algorithm can be set to a higher value, and the texture filtering parameters can be enabled; for medium-transparency areas, the parameters are set between the two. The mapping relationship table can be established by testing and adjusting the algorithm parameters on different transparency samples. Its purpose is to determine the algorithm parameters applicable to the transparency based on the insulating layer transparency information.

[0057] This embodiment achieves adaptive segmentation of an optical image with a preset wavelength by utilizing a clustering algorithm to divide the optical image with a preset wavelength into multiple transparency analysis areas based on all pixel values, so that the transparency of the insulating layer in each transparency analysis area is relatively consistent and overcomes the challenges brought by local differences in the transparency of the insulating layer. This embodiment can adaptively adjust the parameters of the region recognition algorithm and the first defect detection algorithm according to the local differences in the transparency of the insulating layer by first determining the transparency information of the insulating layer based on its optical characteristics for each transparency analysis area, and then determining the parameters of the region recognition algorithm and the parameters of the first defect detection algorithm based on the transparency information of the insulating layer. For example, in areas with lower transparency, parameters that are more sensitive to fuzzy features can be used for printed mark recognition and defect detection to reduce missed detections; in areas with higher transparency, parameters that are insensitive to internal texture can be used to reduce misjudgments. Therefore, this embodiment can effectively eliminate the influence of local differences in the transparency of the insulating layer on the region recognition accuracy and defect detection accuracy, thereby effectively improving the accuracy and reliability of printed mark recognition and defect detection.

[0058] Since the texture difference on the surface of the insulating layer will also affect the judgment of transparency, for example, areas with rough surfaces may scatter light, resulting in underestimated transparency, while areas with smooth surfaces may cause overestimation of transparency. This error will further affect the parameter selection of the subsequent area recognition algorithm and the first defect detection algorithm, reducing the accuracy of printed mark recognition and defect detection. Therefore, the above embodiment has the problem that the accuracy of the insulation layer transparency information is low and the accuracy of printed mark recognition and defect detection is reduced due to not considering the influence of the surface texture of the insulating layer on the transparency of the insulating layer.

[0059] In order to solve this technical problem, in some preferred embodiments, step S23 includes:

[0060] S231. For each transparency analysis area, query a pre-built mapping relationship table between optical features and transparency based on the optical features of the area in the optical image having a preset optical wavelength to obtain preliminary transparency information.

[0061] S232: For each transparency analysis area, calculate the actual grayscale variance used to characterize the surface texture based on the grayscale values ​​of all pixels contained therein, and then query a pre-built mapping relationship table between grayscale variance and transparency compensation amount based on the actual grayscale variance to obtain a first transparency compensation amount.

[0062] S233 . For each transparency analysis area, calculate the insulation layer transparency information according to the corresponding preliminary transparency information and the first transparency compensation amount.

[0063] Among them, the actual grayscale variance of this embodiment refers to the statistical variance of the grayscale values ​​of all pixels in the transparency analysis area, which can be obtained by calculating the average value of the square of the difference between the grayscale values ​​of all pixels in the transparency analysis area and their average grayscale value. Its purpose is to quantify the degree of discreteness of the grayscale values ​​of pixels in the area. When the surface texture is relatively smooth, the reflection or transmission of light in the transparency analysis area is relatively uniform, and the difference in the grayscale values ​​of each pixel is small. At this time, the actual grayscale variance calculated is also small; when the surface texture is relatively rough, the reflection or transmission of light in the transparency analysis area will be scattered due to the unevenness of the surface, and the grayscale values ​​of each pixel are relatively different. At this time, the actual grayscale variance calculated is large, so the actual grayscale variance of this embodiment can characterize the surface texture. In this embodiment, the grayscale variance-to-transparency compensation mapping table stores transparency compensation amounts corresponding to different grayscale variances. For example, when the grayscale variance is 0-10, the transparency compensation amount is 0; when the grayscale variance is 10-20, the transparency compensation amount is 0.02; when the grayscale variance is 20-30, the transparency compensation amount is 0.04; when the grayscale variance is 30-40, the transparency compensation amount is 0.05, etc. Step S233 can calculate the insulation layer transparency information by directly summing or weighted summing the preliminary transparency information corresponding to the transparency analysis area and the first transparency compensation amount. This embodiment is equivalent to not only considering the optical characteristics of the insulating layer material itself when obtaining the transparency information of the insulating layer, but also introducing the quantification and correction of the influence of the surface texture, so as to obtain more accurate insulation layer transparency information and provide a more reliable parameter basis for subsequent area recognition and defect detection algorithms. Therefore, this embodiment can effectively solve the problem of low accuracy of insulation layer transparency information and reduced accuracy of printed mark recognition and defect detection due to failure to consider the influence of the surface texture of the insulating layer on the transparency of the insulating layer, thereby further improving the accuracy of printed mark recognition and defect detection.

[0064] In practical applications, the intensity of ambient light will also affect the brightness and contrast of the image, thereby affecting the accuracy of transparency analysis. If the influence of ambient light is ignored, it may lead to inaccurate transparency assessment, ultimately affecting the reliability of defect detection.

[0065] In order to solve this technical problem, in some preferred implementations, step S233 includes:

[0066] A1. Acquire ambient light intensity information, and then query a pre-built mapping relationship table between ambient light intensity and transparency compensation amount based on the ambient light intensity information to obtain a second transparency compensation amount;

[0067] A2. For each transparency analysis area, calculate the insulation layer transparency information according to the second transparency compensation amount, the corresponding preliminary transparency information, and the first transparency compensation amount.

[0068] The ambient light intensity information refers to a value or signal reflecting the intensity of the detected ambient light, which can be obtained using a light sensor. The mapping table for ambient light intensity and transparency compensation in this embodiment refers to a data table established through experiments or calibration. This mapping table records the required transparency compensation amounts under different ambient light intensities. For example, under low light intensity conditions, such as 100 Lux, the image is generally dark and transparency is underestimated. In this case, the mapping table may recommend a positive transparency compensation amount, such as +0.05, to improve the transparency evaluation value. Under moderate light intensity conditions, such as 500 Lux, the image brightness is moderate and the transparency evaluation is relatively accurate. In this case, the mapping table may recommend a transparency compensation amount close to zero, such as 0 or +0.01, for fine-tuning. Under high light intensity conditions, such as 1000 Lux, the image is generally bright and transparency is overestimated. In this case, the mapping table may recommend a negative transparency compensation amount, such as -0.03, to reduce the transparency evaluation value. The second transparency compensation amount in this embodiment refers to a correction value used to correct the impact of ambient light on transparency evaluation. Step A1 can calculate the insulation layer transparency information by directly or weightedly summing the preliminary transparency information, the first transparency compensation amount, and the second transparency compensation amount corresponding to the transparency analysis area. This embodiment is equivalent to introducing ambient light intensity information and compensating the preliminary transparency information based on the ambient light intensity information, so that the calculated insulation layer transparency information can effectively eliminate the impact of ambient light fluctuations. This avoids the problem of not considering the impact of ambient light intensity on insulation layer transparency, which leads to a decrease in the accuracy of insulation layer transparency information, printed mark recognition, and defect detection. In other words, this embodiment can further improve the accuracy of printed mark recognition and defect detection.

[0069] Since different insulating layer materials have different sensitivities to light of different wavelengths, that is, some wavelengths of light may be more likely to penetrate certain insulating layer materials, while other wavelengths of light may penetrate certain insulating layer materials. If a fixed preset wavelength is used, the pixel value of each pixel point in the optical image with the preset wavelength cannot accurately reflect the transparency information of different materials, thereby affecting the accuracy of subsequent defect detection.

[0070] In order to solve this technical problem, in some preferred embodiments, step S21 includes:

[0071] S211, obtaining the material type information of the insulating layer, and then querying a pre-built mapping relationship table of material types and transparency change sensitive wavelengths according to the insulating layer material type information to obtain a preset wavelength;

[0072] S212: Obtain a pixel value of each pixel point in the optical image having an optical wavelength equal to a preset wavelength.

[0073] The insulation layer material type information refers to the specific material type used in the insulation layer of the semi-transparent insulated conductor, such as polyethylene, polyvinyl chloride, or polyurethane. This information can be obtained by manual input, reading product identification, or using a material identification sensor. Its purpose is to distinguish the optical properties of different insulation layer materials. The pre-constructed mapping relationship table between material type and transparency change sensitive wavelength refers to a data table that stores the correspondence between different insulation layer material types and the optical wavelengths most sensitive to their transparency changes. For example, for polyethylene (PE) material, the transparency change sensitive wavelength is 650nm; for polyvinyl chloride (PVC) material, the transparency change sensitive wavelength is 520nm; and for polypropylene (PP) material, the transparency change sensitive wavelength is 1450nm. This mapping relationship table can be determined by experimentally measuring the transmittance or absorptivity change characteristics of different materials at different wavelengths. Its purpose is to select the most suitable optical wavelength for transparency analysis for different insulation layer materials. The preset wavelength in this embodiment refers to the optical wavelength that is most sensitive to changes in the transparency of the current insulating layer material, obtained by querying the mapping relationship table based on the insulating layer material type information. This preset wavelength is a specific wavelength used to capture the optical image required for subsequent transparency analysis, and its purpose is to ensure that the acquired image data accurately reflects the transparency characteristics of the current insulating layer material. This embodiment is equivalent to dynamically selecting the optical wavelength that is most sensitive to changes in the transparency of the insulating layer material type to acquire image pixel values ​​and optical characteristics. In other words, this embodiment can more accurately capture and reflect the true transparency characteristics of different insulating layer materials, providing a more reliable data foundation for subsequent acquisition of insulating layer transparency information, regional division, and transparency-based algorithm parameter adjustment. This significantly improves the adaptability of printed mark recognition and defect detection for different insulating layer materials, thereby enhancing the accuracy and reliability of defect detection and effectively avoiding misjudgments and missed detections due to material differences.

[0074] In some preferred embodiments, the method for real-time detection of conductor defects further includes the following steps performed after step S3:

[0075] S6. Querying a pre-built mapping relationship table between printed mark image features and defect detection algorithm parameters based on the image features of the surface printed mark area to determine parameters of the second defect detection algorithm;

[0076] S7. Extract features of the surface printed mark areas in all optical images to obtain a second feature map that can reflect information at different levels, and then use a second defect detection algorithm to identify defects in the second feature map.

[0077] The image features of the surface printed mark area refer to data used to describe the visual attributes of the surface printed mark area, which can be represented by color distribution, texture features, shape features, or a combination thereof. The pre-constructed mapping relationship table of printed mark image features and defect detection algorithm parameters refers to a data table that stores the association between different printed mark image features and corresponding preferred defect detection algorithm parameters. For example, when the printed mark image feature is a dark blue print mark with a relatively simple texture, the defect detection algorithm parameters are to select a defect detection algorithm based on color threshold segmentation and the color threshold is a value close to dark blue; when the printed mark image feature is a light gray print mark with a relatively complex texture, the defect detection algorithm parameters are to select a defect detection algorithm based on texture analysis and use a Gabor filter to extract texture features. The parameters of the second defect detection algorithm refer to the settings used to configure the second defect detection algorithm to adapt it to the image features of a specific printed mark area, which can include thresholds, filter parameters, model weights, or algorithm process control parameters. Feature extraction of the surface printed identification area in all optical images refers to extracting information that can characterize the characteristics of the area from the image data corresponding to the surface printed identification area in optical images of different optical wavelengths. It can be achieved by methods based on filtering, transformation, statistical analysis or deep learning. The second feature map of this embodiment refers to a data representation of the visual or physical characteristics of the surface printed identification area at different optical wavelengths obtained by feature extraction, which can be a feature vector or feature image composed of multiple channels or dimensions. The second defect detection algorithm refers to an algorithm specifically used to detect defects in the surface printed identification area, which can be achieved by methods based on rules, statistics, machine learning or deep learning. Defect recognition refers to the process of using the second defect detection algorithm to determine whether there are defects in the surface printed identification area based on the second feature map. This embodiment can first determine the parameters of the second defect detection algorithm based on the image features of the surface printed mark area, then extract features of the surface printed mark area in all optical images to obtain a second feature map that can reflect information at different levels, and finally use the second defect detection algorithm to perform defect identification on the second feature map to achieve defect detection in the surface printed mark area while overcoming the interference of the printed mark on the defect detection. Therefore, this embodiment can identify defects in the surface printed mark area without misjudging the printed mark as a defect, thereby effectively avoiding the situation where defects in the surface printed mark area are missed due to the lack of defect detection on the surface printed mark area.

[0078] In some preferred embodiments, step S1 includes:

[0079] S11, acquiring a multispectral image sequence of a semi-transparent insulating conductor;

[0080] S12. Preprocess the multispectral image sequence.

[0081] The preprocessing of this embodiment refers to various operations performed on the multispectral image sequence after the multispectral image sequence is acquired, which may specifically include but are not limited to image denoising, contrast enhancement, and brightness correction. This embodiment can improve the image quality of the multispectral image sequence by preprocessing the multispectral image sequence, thereby providing more accurate and reliable input data for subsequent algorithm parameter determination, surface print mark recognition, feature extraction of information at different levels, and the final defect recognition process, thereby effectively avoiding error accumulation caused by original image quality issues, and effectively improving the accuracy and robustness of the real-time conductor defect detection method.

[0082] In some preferred embodiments, preprocessing includes image registration and distortion correction. Image registration in this embodiment refers to the process of spatially aligning multiple images of the same scene acquired at different times, using different sensors, from different viewing angles, or in different wavelengths through spatial transformation. This can be achieved using either feature point-based or region-based registration methods. This embodiment can address potential misalignment issues between different optical images in a multispectral image sequence by performing image registration on the multispectral image sequence. Distortion correction in this embodiment refers to the process of eliminating or reducing image geometric distortion caused by factors such as the optical system or sensor during image acquisition by using a distortion correction model or a distortion correction algorithm. This embodiment can first obtain camera intrinsic parameters and distortion coefficients using a camera calibration method, and then apply these parameters to perform distortion identification and correction on each optical image to achieve distortion correction for a multispectral image sequence. This embodiment can compensate for curvature distortion caused by changes in wire diameter or thickness by performing distortion correction on the multispectral image sequence, thereby effectively eliminating or suppressing curvature distortion caused by changes in wire diameter or thickness under the influence of factors such as the characteristics of the extrusion process and the flexibility of the conductor itself, thereby affecting printed mark recognition and defect detection.

[0083] From the above, it can be seen that the present application provides a real-time detection method for conductor defects, which can make the region recognition algorithm and the first defect detection algorithm suitable for the current transparency of the insulation layer by determining the parameters of the region recognition algorithm and the parameters of the first defect detection algorithm according to the transparency information of the insulation layer. Therefore, the present application can effectively solve the problem that the robustness and accuracy of defect detection are reduced due to the fact that defect detection with fixed parameters is difficult to adapt to the influence of such transparency fluctuations. Since the present application can eliminate the influence of printed marks on defect detection by first using the region recognition algorithm to identify the surface printed mark area, and then performing defect detection on the area other than the surface printed mark area in the optical image, Interference, therefore, the present application can effectively solve the problem of misjudging printed marks as defects due to the interference of printed marks with defect detection, and since the present application can realize defect recognition and positioning at different levels by performing multi-spectral feature extraction and defect recognition on non-printed mark areas and determining the position of the same defect by utilizing the feature differences in optical images corresponding to different optical wavelengths, the present application can also effectively solve the problem of insufficient defect detection accuracy and reliability of translucent insulated conductors due to the fact that defects at different levels may overlap or interfere with each other in the image and are difficult to distinguish accurately, that is, the present application can effectively improve the accuracy, reliability and robustness of defect detection.

[0084] Second, as Figure 2 As shown, the present application also provides a conductor defect real-time detection system for detecting defects in a translucent insulated conductor, which includes:

[0085] An image sequence acquisition module 1 is used to acquire a multispectral image sequence of a semi-transparent insulating conductor, wherein the multispectral image sequence includes optical images of different optical wavelengths;

[0086] Algorithm parameter confirmation module 2, used to obtain insulation layer transparency information based on the optical characteristics of the optical image with a preset wavelength, and then determine parameters of the region recognition algorithm and parameters of the first defect detection algorithm based on the insulation layer transparency information;

[0087] The print mark area recognition module 3 is used to use an area recognition algorithm to perform surface print mark recognition on an optical image whose optical wavelength is within a preset wavelength band to obtain a surface print mark area;

[0088] The feature extraction module 4 is used to extract features from all areas of the optical image except the surface printed mark area to obtain a first feature map reflecting information at different levels;

[0089] The defect detection module 5 is used to identify defects on the first feature image using a first defect detection algorithm, and then determine the location of the defect based on the feature differences of the same defect in optical images corresponding to different optical wavelengths when the defect exists.

[0090] A real-time conductor defect detection system provided in the present application includes an image sequence acquisition module 1, an algorithm parameter confirmation module 2, a print mark area recognition module 3, a feature extraction module 4 and a defect detection module 5. The real-time conductor defect detection system provided in this embodiment is used to execute the steps in the real-time conductor defect detection method provided in the first aspect above. The principle of the real-time conductor defect detection system provided in this embodiment is the same as the principle of the real-time conductor defect detection method provided in the first aspect above, and will not be discussed in detail here.

[0091] In some preferred embodiments, the process of obtaining the insulation layer transparency information based on the optical characteristics of the optical image having a preset wavelength, and then determining the parameters of the region recognition algorithm and the parameters of the first defect detection algorithm based on the insulation layer transparency information includes:

[0092] B1. Obtaining the pixel value of each pixel point in the optical image having an optical wavelength of a preset wavelength;

[0093] B2. using a clustering algorithm to divide the optical image having a preset optical wavelength into a plurality of transparency analysis regions according to all pixel values;

[0094] B3. For each transparency analysis area, query a pre-built mapping relationship table between optical features and transparency based on its optical features in the optical image with a preset optical wavelength to obtain the transparency information of the insulating layer;

[0095] B4. For each transparency analysis area, query a pre-built mapping relationship table of insulation layer transparency, area recognition algorithm parameters, and defect detection parameters based on the insulation layer transparency information to obtain corresponding area recognition algorithm and first defect detection algorithm parameters;

[0096] The process of using a region recognition algorithm to identify surface printed marks on an optical image having an optical wavelength within a preset wavelength band to obtain a surface printed marking region includes:

[0097] For each transparency analysis area, a corresponding area recognition algorithm is used to identify the surface printing mark of the transparency analysis area to obtain a surface printing mark area;

[0098] The process of using the first defect detection algorithm to identify defects in the first feature image and then, when a defect exists, determining the location of the defect based on feature differences in optical images corresponding to different optical wavelengths for the same defect includes:

[0099] For each transparency analysis area, the corresponding first defect detection algorithm is used to identify defects in the area of ​​the transparency analysis area except the surface printed mark area. Then, when a defect exists, the location of the defect is determined based on the characteristic differences of the same defect in the optical images corresponding to different optical wavelengths.

[0100] From the above, it can be seen that the present application provides a method and system for real-time detection of conductor defects, which can make the region recognition algorithm and the first defect detection algorithm suitable for the current transparency of the insulation layer by determining the parameters of the region recognition algorithm and the parameters of the first defect detection algorithm according to the transparency information of the insulation layer. Therefore, the present application can effectively solve the problem that the robustness and accuracy of defect detection are reduced due to the fact that defect detection with fixed parameters is difficult to adapt to the influence of such transparency fluctuations. Since the present application can eliminate the influence of printed marks on defect detection by first using the region recognition algorithm to identify the surface printed mark area, and then performing defect detection on the area other than the surface printed mark area in the optical image, Therefore, the present application can effectively solve the problem of misjudging printed marks as defects due to the interference of printed marks on defect detection, and because the present application can realize defect recognition and positioning at different levels by performing multi-spectral feature extraction and defect recognition on non-printed mark areas and determining the position of the same defect by utilizing the feature differences in optical images corresponding to different optical wavelengths, the present application can also effectively solve the problem of insufficient defect detection accuracy and reliability of translucent insulated conductors due to the fact that defects at different levels may overlap or interfere with each other in the image and are difficult to distinguish accurately, that is, the present application can effectively improve the accuracy, reliability and robustness of defect detection.

[0101] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another robot, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interface, the indirect coupling or communication connection of the device or unit can be electrical, mechanical or other forms.

[0102] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0103] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any actual relationship or order between these entities or operations.

[0104] The above are merely examples of the present application and are not intended to limit the scope of protection of the present application. Those skilled in the art will appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A method for real-time detection of conductor defects, characterized in that: For detecting defects in a translucent insulated conductor, the method for real-time detection of conductor defects comprises the following steps: S1. Acquire a multispectral image sequence of the translucent insulating conductor, wherein the multispectral image sequence includes optical images of different optical wavelengths; S2. Obtaining insulation layer transparency information based on the optical features of the optical image having the preset optical wavelength, and then determining parameters of a region recognition algorithm and a first defect detection algorithm based on the insulation layer transparency information; S3, using the region recognition algorithm to perform surface printing mark recognition on the optical image whose optical wavelength is within the preset wavelength band to obtain a surface printing mark area; S4, performing feature extraction on all regions of the optical image except the surface printed mark region to obtain a first feature map reflecting information at different levels; S5. Use the first defect detection algorithm to identify defects in the first feature image. Then, when a defect exists, determine the location of the defect based on feature differences in optical images corresponding to different optical wavelengths for the same defect.

2. The method for real-time detection of conductor defects according to claim 1, characterized in that: Step S2 includes: S21, obtaining a pixel value of each pixel point in the optical image having the optical wavelength being a preset wavelength; S22, using a clustering algorithm to divide the optical image having the preset optical wavelength into a plurality of transparency analysis areas according to all the pixel values; S23. For each transparency analysis area, query a pre-built mapping relationship table between optical features and transparency based on the optical features of the area in the optical image having the preset optical wavelength, so as to obtain the transparency information of the insulating layer; S24. For each transparency analysis region, query a pre-built mapping relationship table of insulation layer transparency, region recognition algorithm parameters, and defect detection parameters according to the insulation layer transparency information to obtain corresponding parameters of the region recognition algorithm and the first defect detection algorithm; Step S3 includes: S31, for each transparency analysis area, using a corresponding area recognition algorithm to perform surface printing mark recognition on the transparency analysis area to obtain a surface printing identification area; Step S5 includes: S51. For each transparency analysis area, use the corresponding first defect detection algorithm to identify defects in the area of ​​the transparency analysis area except the surface printed mark area. Then, when a defect exists, determine the location of the defect based on the characteristic differences of the same defect in the optical images corresponding to different optical wavelengths.

3. The method for real-time detection of conductor defects according to claim 2, wherein: Step S23 includes: S231. For each transparency analysis area, query a pre-built mapping relationship table between optical features and transparency based on the optical features of the area in the optical image having the preset optical wavelength to obtain preliminary transparency information. S232: For each transparency analysis area, calculate an actual grayscale variance for characterizing surface texture based on the grayscale values ​​of all pixels contained therein, and then query a pre-built mapping relationship table between grayscale variance and transparency compensation amount based on the actual grayscale variance to obtain a first transparency compensation amount. S233 . For each transparency analysis area, calculate the insulation layer transparency information according to the corresponding preliminary transparency information and the first transparency compensation amount.

4. The method for real-time detection of conductor defects according to claim 3, characterized in that: Step S233 includes: A1. Acquire ambient light intensity information, and then query a pre-built mapping relationship table between ambient light intensity and transparency compensation amount based on the ambient light intensity information to obtain a second transparency compensation amount; A2. For each transparency analysis area, calculate the insulation layer transparency information according to the second transparency compensation amount, the corresponding preliminary transparency information, and the first transparency compensation amount.

5. The method for real-time detection of conductor defects according to claim 2, characterized in that: Step S21 includes: S211. Obtaining material type information of the insulating layer, and then querying a pre-built mapping relationship table between material type and transparency change sensitive wavelength according to the insulating layer material type information to obtain a preset wavelength; S212: Obtain a pixel value of each pixel point in the optical image whose optical wavelength is the preset wavelength.

6. The method for real-time detection of conductor defects according to claim 1, characterized in that: The method for real-time detection of conductor defects further includes the following steps performed after step S3: S6. Querying a pre-built mapping relationship table between printed mark image features and defect detection algorithm parameters based on the image features of the surface printed mark area to determine parameters of a second defect detection algorithm; S7. Extract features of the surface printed mark areas in all the optical images to obtain a second feature map that can reflect information at different levels, and then use the second defect detection algorithm to perform defect recognition on the second feature map.

7. The method for real-time detection of conductor defects according to claim 1, characterized in that: Step S1 includes: S11, acquiring a multispectral image sequence of the translucent insulating conductor; S12. Preprocessing the multispectral image sequence.

8. The method for real-time detection of conductor defects according to claim 7, characterized in that: The preprocessing includes image registration and distortion correction.

9. A real-time conductor defect detection system, characterized in that: Used to detect defects in translucent insulated conductors, the conductor defect real-time detection system includes: an image sequence acquisition module, configured to acquire a multispectral image sequence of the translucent insulating conductor, wherein the multispectral image sequence includes optical images of different optical wavelengths; an algorithm parameter confirmation module, configured to obtain insulation layer transparency information based on the optical characteristics of the optical image having the preset optical wavelength, and then determine parameters of the region recognition algorithm and parameters of the first defect detection algorithm based on the insulation layer transparency information; a print mark area recognition module, configured to use the area recognition algorithm to perform surface print mark recognition on the optical image whose optical wavelength is within a preset wavelength band to obtain a surface print mark area; a feature extraction module, configured to extract features from all regions of the optical image except the surface printed mark region, to obtain a first feature map reflecting information at different levels; The defect detection module is used to use the first defect detection algorithm to identify defects in the first feature map, and then determine the location of the defect based on the feature differences in the optical images corresponding to the same defect at different optical wavelengths when the defect exists.

10. The conductor defect real-time detection system according to claim 9, characterized in that: The process of obtaining the insulation layer transparency information according to the optical features of the optical image having the preset wavelength, and then determining the parameters of the region recognition algorithm and the parameters of the first defect detection algorithm according to the insulation layer transparency information includes: B1. Obtaining the pixel value of each pixel point in the optical image having the optical wavelength being a preset wavelength; B2. using a clustering algorithm to divide the optical image having the preset optical wavelength into a plurality of transparency analysis regions according to all the pixel values; B3. For each transparency analysis area, query a pre-built mapping relationship table between optical features and transparency based on the optical features of the area in the optical image having the preset optical wavelength, so as to obtain the transparency information of the insulating layer; B4. For each transparency analysis region, query a pre-built mapping relationship table of insulation layer transparency, region recognition algorithm parameters, and defect detection parameters based on the insulation layer transparency information to obtain corresponding parameters of the region recognition algorithm and the first defect detection algorithm; The process of using the region recognition algorithm to perform surface printing mark recognition on the optical image whose optical wavelength is within the preset wavelength band to obtain the surface printing mark area includes: For each transparency analysis area, using a corresponding area recognition algorithm to perform surface printing mark recognition on the transparency analysis area to obtain a surface printing identification area; The process of using the first defect detection algorithm to identify defects on the first feature image and then, when a defect exists, determining the location of the defect based on feature differences in optical images corresponding to different optical wavelengths for the same defect includes: For each transparency analysis area, the corresponding first defect detection algorithm is used to identify defects in the transparency analysis area. Then, when a defect exists, the location of the defect is determined based on the characteristic differences in the optical images corresponding to different optical wavelengths of the same defect.

Images