Defect detection system for permanent magnet circuit breaker and detection method thereof

Abstract

The invention discloses a defect detection system for a permanent magnet circuit breaker and a detection method of the defect detection system, and belongs to the field of permanent magnet circuit breaker detection. The defect detection system comprises a rotary platform, a feeding module arranged on the rotary platform, and a detection module and a material taking module which are arranged around the rotary platform; the detection module is located at the position between the feeding module and the material taking module. The detection module comprises a stand column, and a power-on detection assembly and a camera detection assembly which are arranged on the stand column; according to the defect detection system for the permanent magnet circuit breaker and the detection method of the defect detection system, the electrical data and the image data of the workpiece to be detected are collected in real time through the arranged detection module, intelligent fusion of multi-modal data is achieved through fusion of the electrical data and the image data, single-modal misjudgment is avoided, and the detection accuracy is improved. And the detection rate of composite defects is improved, and the detection efficiency and the detection precision are greatly improved, so that intelligent detection of the permanent magnet circuit breaker is realized.

Description

Technical Field

[0001] This invention belongs to the field of permanent magnet circuit breaker testing, specifically relating to a defect detection system and method for permanent magnet circuit breakers. Background Technology

[0002] A permanent magnet circuit breaker is a circuit breaker structure that uses a permanent magnet mechanism as its operating mechanism. It has advantages such as simple structure, fewer parts, long mechanical life, and maintenance-free operation. Compared with traditional spring and electromagnetic mechanisms, the permanent magnet mechanism uses permanent magnets and opening / closing control coils, solving the problem of requiring high power energy during closing and greatly simplifying the mechanical structure. Therefore, it has been widely used in power systems.

[0003] With the increasing automation and intelligence of power systems, the requirements for the accuracy and real-time performance of permanent magnet circuit breaker defect detection are also becoming higher. Traditional detection methods may not be able to meet the needs of modern power grids for rapid fault response. Currently, the detection of permanent magnet circuit breakers is generally divided into two main directions: external appearance inspection and power-on inspection. Power-on inspection is generally used to determine the electrical performance of permanent magnet circuit breakers, while external appearance inspection is generally used to determine surface defects of insulators. In current technology, both methods are usually carried out manually or using simple testing agencies. However, current detection solutions, whether conducted manually or using simple testing agencies, suffer from problems such as low detection efficiency and poor detection accuracy.

[0004] Therefore, we introduce a defect detection system and method for permanent magnet circuit breakers to solve the above problems. Summary of the Invention

[0005] In response to one or more of the above-mentioned defects or improvement needs in the existing technology, the present invention provides a defect detection system and method for permanent magnet circuit breakers. The system collects electrical data and image data of the workpiece under test in real time through a set detection module, and achieves intelligent fusion of multi-modal data by fusing electrical data and image data. This avoids misjudgment of single modes and improves the detection rate of composite defects, thereby greatly improving detection efficiency and accuracy, and realizing intelligent detection of permanent magnet circuit breakers.

[0006] To achieve the above objectives, the present invention provides a defect detection system for permanent magnet circuit breakers, comprising a rotary platform and a feeding module disposed on the rotary platform, as well as a detection module and a material handling module disposed around the rotary platform; the detection module is located between the feeding module and the material handling module; The loading module includes a linear guide rail and a support assembly mounted on the linear guide rail for carrying the workpiece to be tested; the unloading module includes a gantry frame and an unloading robot, the unloading robot being mounted on the gantry frame for removing the tested workpiece from the support assembly. The detection module includes a column and an electrical detection component and a camera detection component mounted on the column. The electrical detection component includes a fixed plate mounted on the side of the column near the center of the rotating platform, and a movable plate connected to the bottom of the fixed plate via a first telescopic member. A conductive component is mounted on the bottom of the movable plate, and a sensor is mounted on the conductive component. When the conductive component comes into contact with the workpiece to be tested, it can perform an electrical detection operation and collect electrical data during the electrical detection operation. The camera detection assembly includes a first detection camera mounted on a fixed plate, with its shooting direction set vertically downward, for capturing top view image data of the workpiece to be tested; A lateral displacement module is also provided on the side of the column near the rotary platform. Two vertical displacement modules are spaced apart on the lateral displacement module. A second telescopic component is provided on the vertical displacement module. A second detection camera is provided at the telescopic end of the second telescopic component for capturing the peripheral image data of the workpiece to be tested. The camera detection assembly also includes a third detection camera, which is positioned to shoot towards the back of the workpiece to be tested; It also includes a data processing module, which is communicatively connected to the power-on detection component and the camera detection component, respectively. The data processing module is used to receive electrical data and image data of the surface of the workpiece to be tested collected by the sensor, fuse the electrical data and image data, and compare them with reference data to obtain the defects on the surface of the workpiece to be tested.

[0007] As a further improvement of the present invention, the support assembly includes a base plate, a top plate, and hydraulic rods; the top plate and the base plate are connected by at least two of the hydraulic rods. The top plate is provided with a clamping assembly for holding the insulator.

[0008] As a further improvement of the present invention, there are two vertical displacement modules, and the two vertical displacement modules are arranged at a distance from each other in the horizontal direction; At least one of the vertical displacement modules can move along the width direction of the column.

[0009] As a further improvement of the present invention, an alarm module is also included, which is communicatively connected to the data processing module.

[0010] As a further improvement of the present invention, a conveyor belt is also provided corresponding to the gantry frame for transporting the workpiece to be tested held by the robot arm.

[0011] As a further improvement of the present invention, an aperture component is provided corresponding to the first detection camera and / or the second detection camera.

[0012] Based on this, the present invention also provides a detection method for permanent magnet circuit breakers, which utilizes the aforementioned defect detection system. The method includes the following steps: S1. Place the workpiece to be tested onto the carrier assembly and control the rotary platform to rotate, moving the workpiece to be tested to the bottom of the detection module; S2. Control the movement of the conductive component to make it contact the contact point on the workpiece to perform energization, and detect the electrical data of the workpiece during the energization process through the sensor. S3. Control the first detection camera to capture image data of the top view of the workpiece to be tested, and extend the second detection camera between two adjacent insulators in sequence to capture image data of the outer periphery of the insulator; S4. Process the image data of the outer periphery of the insulator and the image data of the top view to obtain complete image data of each insulator; S5. Next, align the processed image data and electrical data in the time series, and then fuse the processed image data and electrical data. S6. Finally, the fused data is compared with the standard data to obtain the defect type of the insulator on the workpiece to be tested. As a further improvement of the present invention, step S2 also includes the following steps: preprocessing the collected electrical data, and then extracting the time domain features and frequency domain features of the electrical parameters. Step S4 also includes the following steps: The image data is preprocessed by first denoising the images from different sides, then augmenting the images, and finally stitching the images from different sides together and performing image segmentation to obtain complete image data.

[0013] As a further improvement of the present invention, step S5 specifically includes the following steps: Based on complete insulator image data, extract geometric and texture features of the insulator surface; and extract time-domain and frequency-domain features of electrical parameters based on electrical parameter data. The surface image features and electrical parameter features of the insulator are fused to obtain the fused feature vector.

[0014] As a further improvement to the present invention, step S6 further includes the following step: Construct a classifier for defect classification and train it using the fused feature vectors of known defect types as the training set; The fused feature vector is input into the classifier to obtain the defect features on the surface of the permanent magnet circuit breaker.

[0015] In summary, the beneficial effects of the above-described technical solutions conceived by this invention compared with the prior art include: The present invention relates to a defect detection system and method for permanent magnet circuit breakers. The system collects electrical and image data of the workpiece under test in real time through a set detection module. By fusing the electrical and image data, it achieves intelligent fusion of multi-modal data, which avoids misjudgment of single modes and improves the detection rate of composite defects, thus greatly improving detection efficiency and accuracy, thereby realizing intelligent detection of permanent magnet circuit breakers. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall three-dimensional structure of the defect detection system during material loading in an embodiment of the present invention; Figure 2 This is a three-dimensional structural diagram of the defect detection system after material loading in an embodiment of the present invention; Figure 3 This is a three-dimensional structural diagram of the detection module in the defect detection system of this invention. Figure 4 This is a three-dimensional structural diagram of the feeding module in the defect detection system of this invention. Figure 5 This is a three-dimensional structural diagram of the material picking module in the defect detection system of this invention. Figure 6 This is a three-dimensional structural diagram of the detection module in the defect detection system of this invention from another perspective; Figure 7 This is a system composition diagram of the defect detection system in an embodiment of the present invention; Figure 8 This is a flowchart of the defect detection method in an embodiment of the present invention; In all the accompanying drawings, the same reference numerals denote the same technical features, specifically: 100. Slewing platform; 200. Feeding module; 201. Linear guide rail; 202. Support assembly; 2021. Base plate; 2022. Top plate; 2023. Hydraulic rod; 203. Clamping assembly; 204. Bracket; 205. Third inspection camera; 300. The workpiece to be tested; 400. Detection module; 401. Column; 402. Power-on detection assembly; 4021. Movable plate; 4022. Telescopic component; 4023. Conductive component; 403. Camera detection module; 4031. First detection camera; 4032. Lateral displacement module; 4033. Vertical displacement module; 4034. Lateral telescopic component; 4035. Second detection camera; 500. Material handling module; 501. Gantry crane; 502. Material handling robot; 503. Conveyor belt; 600. Control module.

[0017] 700. Data processing module; 800. Alarm module. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0019] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, the technical features involved in the various embodiments of the invention described below can be combined with each other as long as they do not conflict with each other.

[0020] Example: Please see Figures 1-6The defect detection system in a preferred embodiment of the present invention includes a rotary platform 100 and a loading module 200 disposed on the rotary platform 100, as well as a detection module 400 and a picking module 500 disposed around the rotary platform 100. The detection module 400 is located between the loading module 200 and the picking module 500. The detection module 400 includes a column 401 and an electrical detection component 402 and a camera detection component 403 disposed on the column 401. The electrical detection component 402 includes a fixed plate disposed on the side of the column 401 near the center of the rotary platform 100, and a movable plate 4021 connected to the bottom of the fixed plate via a first telescopic member 4022. A conductive component 4023 is disposed at the bottom of the movable plate 4021, and a sensor is disposed corresponding to the conductive component 4023. When the conductive component 4023 comes into contact with the workpiece 300 to be tested, it can detect the sensor. The camera detection assembly 403 includes a first detection camera 4031 mounted on a fixed plate, with its shooting direction vertically downward, used to capture top view image data of the workpiece 300 under test; a lateral displacement module 4032 is also provided on the side of the column 401 near the rotary platform 100, with two vertical displacement modules 4033 spaced apart on the lateral displacement module 4032, a second telescopic member 4034 is provided on the vertical displacement module 4033, and a second detection camera 4035 is provided at the telescopic end of the second telescopic member 4034, used to capture image data of the outer periphery of the workpiece 300 under test; the camera detection assembly 403 also includes a third detection camera 205, which is mounted on the loading module 200 through a bracket 204, with its shooting direction facing the back of the workpiece 300 under test.

[0021] In practical use, this method can simultaneously perform electrical and image-based inspections on permanent magnet circuit breakers, thereby compensating for the errors caused by traditional single-mode electrical or image-based inspections. By aligning data from different modes in a time series, they complement each other, significantly improving the accuracy of the defect detection system.

[0022] More specifically, the defect detection system in the preferred embodiment of the present invention further includes a data processing module 700, which is communicatively connected to the power-on detection component 402 and the camera detection component 403, respectively, for receiving electrical data collected by the sensor and image data of the surface of the workpiece 300 to be tested, fusing the electrical data and image data, and comparing them with reference data to obtain defects on the surface of the workpiece 300 to be tested.

[0023] In addition, in a preferred embodiment of the present invention, a control module 600 is also included, which is communicatively connected to the rotary platform 100, the power-on detection component 402 and the camera detection component 403, and is used to control the start and stop operations of each device.

[0024] Furthermore, such as Figure 1 and Figure 2 As shown in the figure, in a preferred embodiment of the present invention, the feeding module 200 includes a linear guide rail 201 and a support component 202 disposed on the linear guide rail 201 for bearing the workpiece 300 to be tested; the picking module 500 includes a gantry frame 501 and a picking robot 502, the picking robot 502 being mounted on the gantry frame 501 for removing the tested workpiece 300 from the support component 202.

[0025] This arrangement significantly improves the testing efficiency of the workpiece 300, i.e., the permanent magnet circuit breaker, facilitating large-scale testing operations. Simultaneously, the unloading method using the gantry 501 and the robotic arm makes unloading operations more convenient, while the support assembly 202 for loading effectively clamps and secures the permanent magnet circuit breaker.

[0026] More specifically, the support assembly 202 includes a base plate 2021, a top plate 2022, and hydraulic rods 2023; the top plate 2022 and the base plate 2021 are connected by at least two hydraulic rods 2023; the top plate 2022 is provided with a clamping assembly 203 for clamping the insulator. That is, when the operator places the permanent magnet circuit breaker onto the support assembly 202, it can be quickly secured by the clamping assembly 203 to prevent the permanent magnet circuit breaker from falling off during rotation.

[0027] It is worth noting that, in the preferred embodiment of the present invention, such as Figure 1 and Figure 2 As shown, a gap is left between the detection module 400 and the material handling module 500, which is a reserved manual unloading position. Typically, after the defect detection module 400 detects a defect in the permanent magnet circuit breaker, the unloading operation can be performed through this reserved position. Then, the rotary platform 100 drives the support assembly 202 back to the loading position. Figure 1 The initial position in.

[0028] At the same time, such as Figure 1 As shown, a linear module is provided for the support component 202, so that when the support component 202 is loading or unloading, the linear module can drive the entire support component 202 and the workpiece 300 to be tested to outside the space of the rotary platform 100, preventing operators or robots from intruding into the entire operating space when loading or unloading, and reducing the occurrence of safety problems.

[0029] In addition, a conveyor belt 503 is provided for the corresponding gantry 501 to transport the workpiece 300 to be tested held by the robot arm.

[0030] Furthermore, such as Figure 4 As shown, in the preferred embodiment of the present invention, there are two vertical displacement modules 4033, which are arranged at a distance from each other in the horizontal direction; at least one vertical displacement module 4033 can move along the width direction of the column 401.

[0031] In actual operation, the two second detection cameras 4035 can sequentially penetrate into the gap between the two insulator surfaces of the permanent magnet circuit breaker to capture image data. To accommodate permanent magnet circuit breakers of different sizes, the lateral spacing between the two second detection cameras 4035 can be fine-tuned to improve the overall practicality of the detection system.

[0032] Of course, in actual setup, four cameras can be simultaneously inserted into the two gaps between the three insulators; however, considering overall cost and other factors, two cameras are preferred, with the two cameras shooting in opposite directions. Figure 4 As shown, one of them is shooting to the left and the other to the right.

[0033] Furthermore, in order to improve the accuracy of image data during acquisition and facilitate subsequent defect judgment, an aperture component is provided for the first detection camera 4031 and / or the second detection camera 4035.

[0034] Furthermore, the defect detection module in the preferred embodiment of the present invention also includes an alarm module 800, which is communicatively connected to the data processing module 700 and is used to perform alarm processing after the detection module 400 detects defects on the surface of the insulator.

[0035] Based on this, such as Figure 7 and Figure 8 As shown, the present invention also provides a detection method for permanent magnet circuit breakers, which utilizes the aforementioned defect detection system. The method includes the following steps: S1. Place the workpiece to be tested onto the carrier assembly and control the rotary platform to rotate, moving the workpiece to be tested to the bottom of the detection module; S2. Control the movement of the conductive component to make it contact the contact point on the workpiece to perform energization, and detect the electrical data of the workpiece during the energization process through the sensor. Step S2 also includes the following steps: preprocessing the collected electrical data, firstly detecting outliers in the electrical parameters, and then processing the missing values ​​to exclude outliers and supplement missing data; then normalizing the data after the first processing to map the electrical parameters to the interval {0, 1}.

[0036] Next, the time-domain and frequency-domain characteristics of the electrical parameters are extracted. First, the mean values ​​of parameters such as current and voltage over a period of time are calculated to reflect their average level. Then, their variances are calculated to measure the degree of fluctuation of the parameters, thereby obtaining the time-domain characteristics of the electrical parameters. Then, the parameters such as current and voltage are processed according to Fourier transform to obtain their spectral characteristics, extracting features such as dominant frequency and harmonic content. Finally, by calculating the power spectral density, the energy distribution of the parameters at different frequencies is analyzed to obtain the frequency-domain characteristics of the electrical parameters.

[0037] S3. Control the first detection camera to capture image data of the top view of the workpiece to be tested, and extend the second detection camera between two adjacent insulators in sequence to capture image data of the outer periphery of the insulator.

[0038] S4. Process the image data of the outer periphery of the insulator and the image data of the top view to obtain complete image data of each insulator; Step S4 also includes the following steps: preprocessing the image data, first denoising the image data of different surfaces, that is, using Gaussian filters and median filters to denoise the data.

[0039] Then, the image contrast is enhanced by adjusting the gray-level histogram. The specific steps are as follows: calculate the image's gray-level histogram, then calculate the cumulative distribution function, and finally map the image's gray-level values ​​according to the cumulative distribution function; then, use the Laplacian operator to sharpen the image, highlighting the edges and details of the insulators, thereby enhancing the image data.

[0040] Finally, the images from different sides are stitched together. The scale-invariant feature transform algorithm is used to extract feature points from the images captured by each camera. Then, Euclidean distance is used to match the feature points of different images. Using the matched point pairs, the random sampling consensus algorithm is used to calculate the transformation matrix between the images. Based on the calculated transformation matrix, the images captured by different cameras are fused. Finally, a threshold segmentation method is used to separate the insulator from the background to achieve image segmentation processing and obtain complete and clear image data.

[0041] Furthermore, step S4 also includes the following step: extracting the geometric and texture features of the insulator surface based on the complete insulator image data; The extraction of geometric features from the insulator surface in the above steps includes the following steps: determining the crack contour using edge detection and contour tracking algorithms, then calculating the crack length and width. Simultaneously, pixel statistics are performed on the segmented contaminated areas to obtain the contaminated area, thus representing the geometric features of the insulator surface.

[0042] The extraction of texture features includes the following steps: calculating the GLCM of the image and extracting texture features such as contrast, correlation, energy, and homogeneity; then comparing each pixel in the image with its neighboring pixels to generate LBP codes; and finally, calculating the histogram of the LBP codes as texture features. S5. Next, align the processed image data and electrical data in the time series, and then fuse the processed image data and electrical data. More specifically, step S5 includes the following steps: concatenating the feature vector extracted from the image data and the feature vector extracted from the electrical parameters into a new feature vector; Image features are: The electrical parameters are characterized as follows: The fused feature vector: Then, PCA dimensionality reduction is performed on the concatenated feature vectors to remove redundant information and retain the main features.

[0043] Next, classifiers were built based on image data and electrical parameters to perform preliminary defect assessments. For image data, a convolutional neural network (CNN) classifier was used. The processed image was input into the trained CNN model, and the model output the probability of different defect types. The result with the highest probability was selected as the preliminary assessment result for the image data. For electrical parameters, a support vector machine (SVM) classifier was used. The extracted electrical features were input into the SVM model to obtain the preliminary assessment result for the electrical parameters.

[0044] Next, define the defect type set. In the formula, This indicates that it is normal. Indicates a crack defect. This indicates stains, defects, etc.; however, if the image data judgment result is... The electrical parameter judgment result is ;when i=j The final judgment result is that ,like i and jThey are not the same; count the number of votes for each defect type and select the defect type with the most votes as the final result.

[0045] Different combinations of feature vectors are intrinsically linked to specific defect types. For example, when the crack length and width feature values ​​in the image feature vector are large, and the insulation resistance and leakage current in the electrical feature vector are significantly reduced and increased, it is highly likely to correspond to a crack defect. If the stain area feature value in the image feature vector is large, and the voltage distribution in the electrical parameters shows abnormal fluctuations, it may indicate a stain defect.

[0046] S6. Finally, the fused data is compared with the standard data to obtain the defect type of the insulator on the workpiece to be tested. Step S6 also includes the following steps: constructing a classifier for defect classification and training it using the fused feature vectors of known defect types as a training set; Historical data was divided into 80% training and 20% validation sets. The training set included images of known defect types and electrical fusion feature vectors.

[0047] Then calculate the prior probability. P(C j ) Statistically analyze each defect type in the training set C j Frequency of occurrence, such as in the training set N In each sample n 1 Second-rate; ; Then calculate its conditional probability, in the defect type C j Below, statistical fusion feature vectors F f Frequency of occurrence, if C j Down F f Appear m j Number of samples of this type n j ,but; The fused feature vector F f The input is fed into the classifier to calculate the posterior probability. Choose posterior probability P ( C j | F fThe largest defect type is used as the final judgment result to obtain the defect characteristics on the surface of the permanent magnet circuit breaker.

[0048] The present invention relates to a defect detection system and method for permanent magnet circuit breakers. The system collects electrical and image data of the workpiece under test in real time through a set detection module. By fusing the electrical and image data and through dynamic weight allocation and hierarchical decision-making logic, it achieves intelligent fusion of multi-modal data, which avoids misjudgment of single modes and improves the detection rate of composite defects, thus greatly improving detection efficiency and accuracy, thereby realizing intelligent detection of permanent magnet circuit breakers.

[0049] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A defect detection system for permanent magnet circuit breakers, characterized in that, It includes a rotary platform (100) and a loading module (200) disposed on the rotary platform (100), as well as a detection module (400) and a picking module (500) disposed around the rotary platform (100); the detection module (400) is located between the loading module (200) and the picking module (500); The loading module (200) includes a linear guide rail (201) and a support assembly (202) mounted on the linear guide rail (201) for carrying the workpiece (300) to be tested; the unloading module (500) includes a gantry frame (501) and an unloading robot (502), the unloading robot (502) being mounted on the gantry frame (501) for removing the tested workpiece (300) from the support assembly (202); The detection module (400) includes a column (401) and an electrical detection component (402) and a camera detection component (403) disposed on the column (401); wherein, the electrical detection component (402) includes a fixed plate disposed on the side of the column (401) near the center of the rotary platform (100), the bottom of the fixed plate is connected to a movable plate (4021) through a first telescopic member (4022), and a conductive member (4023) is disposed at the bottom of the movable plate (4021), and a sensor is disposed corresponding to the conductive member (4023), so that after the conductive member (4023) comes into contact with the workpiece (300) to be tested, it can perform an electrical detection operation and collect electrical data during the electrical detection operation. The camera detection assembly (403) includes a first detection camera (4031) mounted on a fixed plate, with its shooting direction set vertically downward, for capturing top view image data of the workpiece (300) to be tested; A lateral displacement module (4032) is also provided on the side of the column (401) near the rotary platform (100). Two vertical displacement modules (4033) are spaced apart on the lateral displacement module (4032). A second telescopic member (4034) is provided on the vertical displacement module (4033), and a second detection camera (4035) is provided at the telescopic end of the second telescopic member (4034) for capturing the peripheral image data of the workpiece (300) to be tested. The camera detection assembly (403) also includes a third detection camera (205) whose shooting direction is set towards the back of the workpiece (300) to be tested; It also includes a data processing module (700), which is communicatively connected to the power-on detection component (402) and the camera detection component (403) respectively. It is used to receive electrical data collected by the sensor and image data of the surface of the workpiece (300) under test, fuse the electrical data and image data, and compare them with reference data to obtain the defects on the surface of the workpiece (300) under test.

2. The defect detection system for permanent magnet circuit breakers according to claim 1, characterized in that, The support assembly (202) includes a base plate (2021), a top plate (2022), and hydraulic rods (2023); the top plate (2022) and the base plate (2021) are connected by at least two of the hydraulic rods (2023); The top plate (2022) is provided with a clamping assembly (203) for clamping the insulator.

3. The defect detection system for permanent magnet circuit breakers according to claim 1, characterized in that, There are two vertical displacement modules (4033), and the two vertical displacement modules (4033) are arranged at a horizontal interval; At least one of the vertical displacement modules (4033) can move along the width direction of the column (401).

4. The defect detection system for permanent magnet circuit breakers according to claim 1, characterized in that, It also includes an alarm module (800), which is communicatively connected to the data processing module (700).

5. The defect detection system for permanent magnet circuit breakers according to claim 1, characterized in that, A conveyor belt (503) is also provided corresponding to the gantry (501) for transporting the workpiece (300) to be tested held by the robot arm.

6. The defect detection system for permanent magnet circuit breakers according to claim 1, characterized in that, An aperture assembly is provided for the first detection camera (4031) and / or the second detection camera (4035).

7. A testing method for permanent magnet circuit breakers, characterized in that, It is implemented using the defect detection system according to any one of claims 1 to 6, and the method includes the following steps: S1. Place the workpiece (300) to be tested on the carrier assembly and control the rotary platform (100) to rotate, moving the workpiece (300) to be tested below the detection module (400); S2. Control the movement of the conductive component (4023) to make it contact the contact point on the workpiece (300) to perform energization operation, and detect the electrical data of the workpiece (300) during the energization process through the sensor; S3. Control the first detection camera (4031) to capture image data of the top view of the workpiece (300) to be tested, and extend the second detection camera (4035) between two adjacent insulators in sequence to capture image data of the outer periphery of the insulator; S4. Process the image data of the outer periphery of the insulator and the image data of the top view to obtain complete image data of each insulator; S5. Next, align the processed image data and electrical data in the time series, and then fuse the processed image data and electrical data. S6. Finally, the fused data is compared with the standard data to obtain the defect type of the insulator on the workpiece (300) to be tested.

8. The detection method for permanent magnet circuit breakers according to claim 7, characterized in that, Step S2 also includes the following steps: The collected electrical data is preprocessed, and then the electrical parameters are extracted using time-domain and frequency-domain features. Step S4 also includes the following steps: The image data is preprocessed by first denoising the images from different sides, then augmenting the images, and finally stitching the images from different sides together and performing image segmentation to obtain complete image data.

9. The detection method for permanent magnet circuit breakers according to claim 7, characterized in that, Step S5 specifically includes the following steps: Based on complete insulator image data, extract geometric and texture features of the insulator surface; and extract time-domain and frequency-domain features of electrical parameters based on electrical parameter data. The surface image features and electrical parameter features of the insulator are fused to obtain the fused feature vector.

10. The detection method for permanent magnet circuit breakers according to claim 7, further comprising the following step S6: Construct a classifier for defect classification and train it using the fused feature vectors of known defect types as the training set; The fused feature vector is input into a classifier to obtain the defect features on the surface of the permanent magnet circuit breaker.

Images