A connector quality detection system and method based on image processing

Abstract

The application provides a connector quality detection system and method based on image processing. The method collects thermal imaging images of a connector through an Internet of Things sensor to obtain a thermal imaging image sequence of the connector. The method extracts differential features from the thermal imaging image sequence of the connector to obtain an electrical detection image of the thermal imaging image sequence. The method divides a contact area through a gradient differential image to obtain a contact area sub-image of the gradient differential image. The method extracts contact features from the contact area sub-image to obtain an image contact loss feature of the connector. The method extracts an image apparent and hidden diffusion feature of the electrical detection image based on a contrast image and a plurality of key detection points. The method analyzes the quality of the connector based on the image contact loss feature and the image apparent and hidden diffusion feature. The method can generate an electrical detection image of the connector through a thermal imaging image set and realize non-contact electrical performance detection of the connector based on the electrical detection image.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and more specifically, to a connector quality inspection system and method based on image processing, such as terminal equipment, chips, computer storage media, etc. Background Technology

[0002] Connectors, also known as plugs, connectors, or sockets, are electrical connection components used in electrical equipment. Their basic function is to achieve electrical connection between two active devices, thereby enabling the transmission of current or signals. Structurally, connectors typically consist of male and female terminals, which establish a conductive relationship through mutual contact, thus completing the transmission of information or electrical energy. Therefore, connectors are not only key basic components in electrical systems, but also important supporting components for the modular design, ease of installation, maintenance, and upgrade of electronic products.

[0003] In the existing technology, defect detection of connectors mainly focuses on surface defect detection of connectors using visible light. This method can only identify obvious surface defects and is difficult to extract hidden defects inside the connector. At the same time, traditional detection methods cannot effectively evaluate the electrical performance of connectors, which may lead to potential safety risks in actual use of connectors, greatly restricting the application and development of connector quality inspection. Summary of the Invention

[0004] This application provides a connector quality inspection system and method based on image processing, which can generate electrical inspection images of connectors from a set of thermal imaging images, and realize non-contact electrical performance inspection of connectors based on the electrical inspection images.

[0005] In a first aspect, this application provides a connector quality inspection method based on image processing. This method can be executed by a network device, or by a chip configured in the network device, and this application does not limit the execution of such method.

[0006] Specifically, the method includes:

[0007] During the use of the connector, thermal imaging images of the connector are acquired through IoT sensors to obtain a sequence of thermal imaging images of the connector.

[0008] Differential feature extraction is performed on the thermal imaging image sequence of the connector to obtain the electrical inspection image of the thermal imaging image sequence;

[0009] The contact region is divided by the gradient difference image to obtain a contact region sub-image of the gradient difference image. Contact features are extracted based on the contact region sub-image to obtain the image contact loss features of the connector.

[0010] Obtain a comparison image of the connector and a number of preset key detection points, and extract the image manifestation and diffusion features of the electrical test image based on the comparison image and the number of key detection points;

[0011] Connector quality analysis is performed based on the image contact loss characteristics and the image visibility diffusion characteristics.

[0012] In conjunction with the first aspect, in certain implementations of the first aspect, obtaining the electrical inspection image of the thermal imaging image sequence by performing differential feature extraction on the thermal imaging image sequence specifically includes:

[0013] Each thermal imaging image in the thermal imaging image sequence is acquired, and differential image extraction is performed on each thermal imaging image using the adjacent frame difference method to obtain multiple differential images;

[0014] The average gray level is extracted based on the co-position pixels of each difference image to obtain the average difference gray level corresponding to each pixel coordinate.

[0015] The average gradient is extracted based on the co-position pixels of each difference image to obtain the average difference gradient corresponding to each pixel coordinate.

[0016] Image reconstruction is performed based on the average differential gradient and average differential gradient corresponding to each pixel coordinate to obtain the electrical detection image of the thermal imaging image sequence.

[0017] In conjunction with the first aspect, in certain implementations of the first aspect, the process of dividing the contact region using the gradient difference image to obtain the contact region sub-image of the gradient difference image specifically includes:

[0018] Obtain a standard image of the connector, and perform image alignment on the gradient difference image based on the standard image;

[0019] The gradient difference image is segmented into a sub-image of the contact region based on the boundary points of the contact region of the standard image.

[0020] In conjunction with the first aspect, in certain implementations of the first aspect, extracting contact features based on the contact area sub-image to obtain the image contact loss features of the connector specifically includes:

[0021] Based on the grayscale feature statistics of the contact area sub-image, the contact area feature vector of the contact area sub-image is obtained;

[0022] The feature vectors of the contact area are subjected to feature clustering, and the corresponding image contact loss features are determined based on the clustering results.

[0023] In conjunction with the first aspect, in some implementations of the first aspect, before performing differential feature extraction using the thermal imaging image sequence of the connector, the method further includes: performing image preprocessing on the thermal imaging image sequence of the connector.

[0024] In conjunction with the first aspect, in some implementations of the first aspect, a thermal imaging camera is used as an Internet of Things sensor to acquire thermal imaging images of the connector.

[0025] In conjunction with the first aspect, in certain implementations of the first aspect, extracting the image manifestation and concealment diffusion features of the electrical detection image based on the comparison image and multiple key detection points specifically includes:

[0026] Acquire comparison images and multiple key detection points;

[0027] For any key detection point, obtain the position coordinates of the key detection point, extract the image grayscale value of the electrical detection image at the position coordinates, and perform differential detection with the corresponding pixel points of the comparison image to obtain the thermal diffusion differential value.

[0028] When the thermal diffusion difference value is higher than a preset threshold, a standard image of the connector is acquired, and the proportion of visible defects and the proportion of hidden defects corresponding to the key detection point are determined based on the image grayscale value of the standard image of the connector at the position coordinates of the key detection point.

[0029] Image explicit and implicit diffusion features are constructed based on the thermal diffusion difference value corresponding to each key detection point, the proportion of explicit defects, and the proportion of implicit defects.

[0030] Secondly, this application provides a connector quality inspection system based on image processing, which includes an image processing unit, the image processing unit comprising:

[0031] The image acquisition module is used to acquire thermal imaging images of the connector through an Internet of Things sensor during the use of the connector, and obtain a thermal imaging image sequence of the connector.

[0032] An image processing module is used to perform differential feature extraction on the thermal imaging image sequence of the connector to obtain an electrical inspection image of the thermal imaging image sequence;

[0033] The image processing module is further configured to divide the contact area using the gradient difference image to obtain a contact area sub-image of the gradient difference image, and extract contact features based on the contact area sub-image to obtain the image contact loss features of the connector.

[0034] The image processing module is also used to acquire a comparison image of the connector and a number of preset key detection points, and to extract the image visibility and diffusion features of the electrical detection image based on the comparison image and the number of key detection points.

[0035] The image feature analysis module is used to perform connector quality analysis based on the image contact loss characteristics and the image visible diffusion characteristics.

[0036] Thirdly, this application provides a computer terminal device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described image processing-based connector quality detection method.

[0037] Fourthly, this application provides a computer-readable storage medium storing at least one computer program, which is loaded and executed by a processor to perform the operations described above in the image processing-based connector quality inspection method.

[0038] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0039] This application provides a connector quality inspection system and method based on image processing. First, thermal imaging images of the connector are acquired using an IoT sensor to obtain a sequence of thermal imaging images. Differential feature extraction is performed on the thermal imaging image sequence to obtain an electrical inspection image of the thermal imaging image sequence. The contact area is divided using gradient differential images to obtain contact area sub-images of the gradient differential images. Contact feature extraction is performed based on the contact area sub-images to obtain the contact loss features of the connector. A comparison image of the connector and several preset key detection points are acquired. Image manifestation and diffusion features of the electrical inspection image are extracted based on the comparison image and the multiple key detection points. Finally, connector quality analysis is performed based on the image contact loss features and the image manifestation and diffusion features.

[0040] Therefore, this application considers that the connector contact area will generate local Joule heating due to the passage of current. By analyzing the heat intensity and temperature gradient of different areas, the contact resistance of each contact point can be indirectly reflected, thereby generating an electrical test image reflecting the electrical performance. Then, the contact area is divided by gradient difference image, and the regional features are extracted and the contact loss characteristics of the image are identified based on the contact area. By comparing the preset key detection points with the standard image, the thermal diffusion characteristics can be extracted, and abnormal thermal diffusion caused by potential defects (obvious defects and latent defects) on the surface and inside of the connector can be detected, and the image explicit and implicit diffusion characteristics are generated, corresponding to the decline in the electrical performance of the connector. This makes thermal imaging images not only usable for surface inspection, but also reflect the internal electrical state, realizing non-contact inspection.

[0041] In summary, this application can generate electrical test images of connectors from a set of thermal imaging images, and realize non-contact electrical performance testing of connectors based on the electrical test images. Attached Figure Description

[0042] Figure 1 This is an exemplary flowchart of an image processing-based connector quality inspection method according to some embodiments of this application;

[0043] Figure 2 This is a schematic diagram of the structure of an image processing unit according to some embodiments of this application;

[0044] Figure 3 This is a schematic diagram of the structure of a computer terminal device that implements an image processing-based connector quality inspection method according to some embodiments of this application. Detailed Implementation

[0045] This application acquires thermal imaging images of a connector using an IoT sensor to obtain a sequence of thermal imaging images of the connector; it then extracts differential features from the thermal imaging image sequence to obtain an electrical inspection image of the thermal imaging image sequence; it divides the contact area using gradient differential images to obtain contact area sub-images of the gradient differential images, and extracts contact features based on the contact area sub-images to obtain the contact loss features of the connector; it acquires a comparison image of the connector and multiple preset key detection points, and extracts the image manifestation and diffusion features of the electrical inspection image based on the comparison image and the multiple key detection points; it performs connector quality analysis based on the image contact loss features and the image manifestation and diffusion features, and can generate an electrical inspection image of the connector from a set of thermal imaging images, thereby realizing non-contact electrical performance testing of the connector based on the electrical inspection image.

[0046] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific implementation methods. (Reference) Figure 1 The figure is an exemplary flowchart of an image processing-based connector quality inspection method according to some embodiments of this application. The image processing-based connector quality inspection method 100 mainly includes the following steps:

[0047] In step S101, during the use of the connector, thermal imaging images of the connector are acquired by an IoT sensor to obtain a thermal imaging image sequence of the connector.

[0048] Preferably, in some embodiments, a thermal imaging camera is used as an IoT sensor to acquire thermal images of the connector. In other embodiments, other devices or equipment capable of acquiring thermal images, such as an edge gateway with an integrated infrared imaging module, can also be used to acquire thermal images of the connector. This application does not limit this.

[0049] Optionally, in some embodiments, the acquisition interval of the thermal imaging images can be set to a fixed interval based on empirical values ​​or operational requirements, such as 200ms, 500ms or 1s, and after acquiring a preset number of thermal imaging images, such as 50 frames, the thermal imaging image sequence is constructed according to the acquisition time sequence of the images.

[0050] In specific implementation, when the connector is first used, the temperature can be detected by an external temperature sensor arranged on or near its surface. When the surface temperature of the connector reaches a preset threshold, such as 35°C, 40°C, or a calculated value based on the rated current, the connector is determined to be in working state, and this moment is set as the initial acquisition time point of the thermal imaging image. Thereafter, the thermal imaging camera automatically starts the acquisition process at preset intervals and combines the timestamp to form the thermal imaging image sequence.

[0051] In step S102, differential feature extraction is performed on the thermal imaging image sequence of the connector to obtain the electrical detection image of the thermal imaging image sequence.

[0052] Preferably, in some embodiments, before performing differential feature extraction using the thermal imaging image sequence of the connector, the method further includes: image preprocessing of the thermal imaging image sequence of the connector.

[0053] In a specific implementation, the process of image preprocessing for the thermal imaging image sequence of the connector includes: performing grayscale processing on each thermal imaging image of the connector through color space mapping. In some other embodiments, performing grayscale processing on each thermal imaging image of the connector may also include: performing median filtering on each thermal imaging image of the connector.

[0054] Preferably, in some embodiments, obtaining the electrical inspection image of the thermal imaging image sequence by performing differential feature extraction on the thermal imaging image sequence specifically includes:

[0055] Each thermal imaging image in the thermal imaging image sequence is acquired, and differential image extraction is performed on each thermal imaging image using the adjacent frame difference method to obtain multiple differential images;

[0056] The average gray level is extracted based on the co-position pixels of each difference image to obtain the average difference gray level corresponding to each pixel coordinate.

[0057] The average gradient is extracted based on the co-position pixels of each difference image to obtain the average difference gradient corresponding to each pixel coordinate.

[0058] Image reconstruction is performed based on the average differential gradient and average differential gradient corresponding to each pixel coordinate to obtain the electrical detection image of the thermal imaging image sequence.

[0059] In a specific implementation, grayscale difference can be performed on n adjacent frames in the thermal imaging image sequence to obtain n-1 difference images. The grayscale values ​​of the difference images are used to reflect the dynamic change of temperature over time. The average grayscale value of all difference images is taken at each pixel position to determine the average difference grayscale corresponding to each pixel coordinate. The Sobel operator is used at each pixel position to extract the gradient value corresponding to each difference image, and the gradient values ​​corresponding to each difference image are averaged at the same pixel position to obtain the average difference gradient corresponding to each pixel coordinate.

[0060] In some embodiments, during the image reconstruction stage, the average difference gray level and average difference gradient corresponding to each pixel coordinate are weighted and fused according to preset weight coefficients, and the fusion result is rounded down to obtain the gray level value, thus completing the reconstruction of the electrical inspection image. In some optional embodiments, the reconstructed electrical inspection image can be binarized or thresholded to highlight abnormal heating areas or abnormal contact points of the connector. In addition, the above difference and gradient calculations can be completed in an edge processing unit or a local IoT node, or uploaded to a cloud server for centralized processing. The values ​​of the weight coefficients can be set or dynamically optimized according to the connector material, operating current, temperature rise rate, or historical operating data, thereby improving the accuracy and reliability of the electrical inspection image under different operating conditions.

[0061] In step S103, the contact area is divided using the gradient difference image to obtain a contact area sub-image of the gradient difference image. Contact features are extracted based on the contact area sub-image to obtain the image contact loss features of the connector.

[0062] It should be noted that the contact area refers to the portion of the image area in the connector where there is contact resistance and Joule heat is continuously generated during operation. Specifically, the contact resistance of the connector through electrical connection will generate Joule heat when energized, which is the ohmic loss of the connector. The greater the temperature characteristic change in this area, the greater the ohmic loss of the connector during contact operation. Therefore, the temperature change in this area can reflect the actual contact resistance and contact performance of the connector.

[0063] Preferably, in some embodiments, the process of dividing the contact region using the gradient difference image to obtain the contact region sub-image of the gradient difference image specifically includes:

[0064] Obtain a standard image of the connector, and perform image alignment on the gradient difference image based on the standard image;

[0065] The gradient difference image is segmented into a sub-image of the contact region based on the boundary points of the contact region of the standard image.

[0066] In specific implementation, a standard image of the connector is first obtained. This standard image is a visible light image of the connector. To facilitate image alignment, in some embodiments, the standard image is acquired by an industrial camera coaxial with a thermal imaging image acquisition device. This standard image records the geometric structure and contact area distribution of the connector under normal conditions. In some preferred embodiments, the standard image can be acquired by laser scanning, structured light imaging, etc. This application will not elaborate on this.

[0067] Optionally, in some embodiments, during the image alignment operation on the gradient difference image based on the standard image, the operation may include image geometric transformation, affine transformation, or keypoint-based registration algorithm, so that the gradient difference image and the standard image are spatially consistent. Then, based on the boundary points of the contact area in the standard image, the gradient difference image is segmented into regions, and the pixels corresponding to the contact area in the gradient difference image are extracted to obtain the contact area sub-image. The boundary points of the contact area in the standard image can be set through multiple tests. The region segmentation can employ various image segmentation methods, such as threshold segmentation, edge-based segmentation, or mask-based ROI extraction methods, so that the contact area sub-image can accurately cover the part of the connector that actually generates Joule heat.

[0068] Preferably, in some embodiments, extracting contact features based on the contact area sub-image to obtain the image contact loss features of the connector specifically includes:

[0069] Based on the grayscale feature statistics of the contact area sub-image, the contact area feature vector of the contact area sub-image is obtained;

[0070] The feature vectors of the contact area are subjected to feature clustering, and the corresponding image contact loss features are determined based on the clustering results.

[0071] In specific implementation, the process of extracting contact features based on the contact area sub-image to obtain the image contact loss features of the connector includes the following steps: First, perform grayscale feature statistics on the contact area sub-image. This process may include calculating statistical indicators such as the average grayscale value, grayscale variance, grayscale histogram distribution, and local grayscale gradient of each pixel in the contact area sub-image, thereby forming a multidimensional vector describing the thermal distribution features of the contact area.

[0072] Preferably, in some embodiments, the K-means clustering algorithm can be used to cluster the feature vectors of the contact area. Below is a specific embodiment of this application using the K-means clustering algorithm to cluster the feature vectors of the contact area:

[0073] The gray-level feature vector of the contact area sub-image is input into a K-means clustering algorithm. Different dimensions of the feature vector include pixel gray-level mean, gray-level variance, local gray-level gradient, thermal distribution uniformity index, and local Joule thermal peak value. First, the K-means clustering algorithm initializes multiple cluster centers, divides the contact area feature vector according to its distance from each cluster center, and iteratively updates the cluster centers until convergence, outputting the final two-dimensional clustering result. Based on the first clustering result, each cluster can be mapped to a contact loss level, such as normal contact area, slightly abnormal contact area, and severely abnormal contact area. The statistical features corresponding to each category are mapped to image contact loss features, thereby reflecting the contact performance status of the connector during use.

[0074] In the secondary clustering, K-means clustering is performed again using the same contact area feature vectors to independently generate another set of clustering results, which are used to map the connection stability level of the connector. Through two clustering analyses, independent evaluation indicators for contact loss level and connection stability level can be obtained respectively, providing multi-dimensional data support for the overall quality analysis of the connector.

[0075] During the model initialization phase, multiple contact region sub-graph feature samples and corresponding manually labeled contact loss and connectivity stability levels can be prepared in advance. These samples are then input into a K-means clustering algorithm for training or cluster center initialization. By adjusting the number of clusters and the iteration threshold, the clustering results can be made to highly match the manually labeled contact loss and connectivity stability levels. When the correlation between the clustering results and the manually labeled results falls below a preset threshold, the clustering process can be adjusted by re-initializing cluster centers or optimizing feature weights until the correlation between the clustering results and the manually labeled contact loss and connectivity stability levels reaches a preset standard, and the mapping relationship between the clustering results and the contact loss and connectivity stability levels is obtained.

[0076] In step S104, a comparison image of the connector and a number of preset key detection points are obtained, and the image visibility and diffusion features of the electrical detection image are extracted based on the comparison image and the number of key detection points.

[0077] It should be noted that the comparison image of the connector is an electrical inspection image extracted from a qualified connector. This image is used to compare the thermal diffusion characteristics with the electrical inspection image of the connector to be inspected, thereby extracting features of latent defects in the connector based on abnormal thermal diffusion. The comparison image reflects the typical heat distribution pattern of the connector during operation under normal conditions, including the thermal diffusion characteristics of the contact area, non-contact area, and edge area, providing a reliable reference benchmark for the connector to be inspected.

[0078] In practice, the comparison image also includes multiple key detection points for refined analysis of thermal diffusion characteristics. These key detection points can be distributed at the center, edges, and expected heat concentration areas of the contact region. By collecting grayscale values ​​or temperature change information from these key detection points, the difference between the thermal diffusion of the connector under test and the standard comparison image can be quantitatively analyzed. In practice, the number and location of the key detection points can be preset based on the connector type, contact structure, and operating current distribution, or optimized by combining historical testing data or simulated thermal distribution results. By comparing the thermal diffusion characteristics of the key detection points, areas of localized abnormal heating or uneven thermal diffusion can be effectively identified, thereby further extracting latent defect characteristics of the connector and providing data support for subsequent quality analysis and fault early warning.

[0079] Preferably, in some embodiments, extracting the image manifestation / latent diffusion features of the electrical detection image based on the comparison image and multiple key detection points specifically includes:

[0080] Acquire comparison images and multiple key detection points;

[0081] For any key detection point, obtain the position coordinates of the key detection point, extract the image grayscale value of the electrical detection image at the position coordinates, and perform differential detection with the corresponding pixel points of the comparison image to obtain the thermal diffusion differential value.

[0082] When the thermal diffusion difference value is higher than a preset threshold, a standard image of the connector is acquired, and the proportion of visible defects and the proportion of hidden defects corresponding to the key detection point are determined based on the image grayscale value of the standard image of the connector at the position coordinates of the key detection point.

[0083] Image explicit and implicit diffusion features are constructed based on the thermal diffusion difference value corresponding to each key detection point, the proportion of explicit defects, and the proportion of implicit defects.

[0084] In specific implementation, the comparison image and multiple preset key detection points are acquired, each key detection point corresponding to a specific location coordinate in the connector electrical inspection image. For any key detection point, the corresponding pixel grayscale value is extracted from the electrical inspection image to be inspected using its location coordinates, and the grayscale value is differentially calculated with the pixel grayscale value at the same location in the comparison image to obtain the thermal diffusion difference value of the key detection point. The thermal diffusion difference value is used to quantify the thermal diffusion deviation of the connector to be inspected at that location, and then the obtained thermal diffusion difference value is compared with a preset threshold. When the difference value is higher than the threshold, a standard image of the connector can be further obtained. Based on the pixel grayscale value at the position coordinates of the corresponding key detection point in the standard image, combined with the features of the comparison image and the electrical inspection image, the proportion of visible defects and the proportion of hidden defects corresponding to the key detection point are determined. In a specific implementation, the sum of the proportion of visible defects and the proportion of hidden defects is 1, which is used to fully describe the distribution characteristics of the thermal diffusion anomaly at the key detection point. Preferably, in some embodiments, the proportion of visible defects and the proportion of hidden defects can be determined by comparing the visible light image of the connector with the visible light image of the qualified connector. The specific operation includes: extracting the pixel gradient information of the visible light image at the key detection point, calculating the pixel gradient difference between the connector to be tested and the standard qualified connector at the same key detection point, thereby quantifying the degree of thermal anomaly caused by surface defects, and determining the corresponding proportion of visible defects by performing interval mapping on the pixel gradient difference value according to a preset mapping table. The proportion of visible defects reflects the abnormal thermal diffusion caused by the surface defects of the connector, while the proportion of hidden defects represents the abnormal thermal diffusion caused by internal defects of the connector.

[0085] Optionally, in some embodiments, during the process of constructing the image visible and hidden diffusion features of the electrical inspection image based on the thermal diffusion difference values ​​of each key detection point and the corresponding proportions of visible and hidden defects, the features can be represented in vector form to describe the degree and distribution of thermal diffusion anomalies of the connector at multiple key detection points, thereby achieving accurate extraction and quantitative analysis of hidden defects of the connector.

[0086] In step S105, connector quality analysis is performed based on the image contact loss characteristics and the image visibility diffusion characteristics.

[0087] Preferably, in some embodiments, connector quality analysis based on the image contact loss characteristics and the image visible / hidden diffusion characteristics specifically includes: first, obtaining the image contact loss characteristics and image visible / hidden diffusion characteristics of the connector; the image contact loss characteristics are used to reflect the contact performance of the connector, and may include the Joule thermal intensity, contact stability, and contact loss level of each contact area; the image visible / hidden diffusion characteristics are used to reflect the defect distribution on the surface and inside of the connector, including the thermal diffusion difference value of key detection points, the proportion of visible defects, and the proportion of hidden defects.

[0088] Subsequently, the two types of features are judged based on preset detection thresholds. Specifically, for image contact loss features, a contact loss threshold is set. When the loss feature of a certain contact area exceeds the threshold, it is determined that the contact area has contact abnormalities or excessive loss. For image visible and hidden diffusion features, visible defect thresholds and hidden defect thresholds are set respectively. When the thermal diffusion difference value of a key detection point is higher than the visible defect threshold, it is determined that there is a visible defect on the surface. When the thermal diffusion difference value of a key detection point is higher than the hidden defect threshold and is not covered by visible defects, it is determined that there is a hidden defect inside.

[0089] After the threshold judgment is completed, the abnormal judgment results of each contact area and key detection point are summarized to generate an overall quality evaluation of the connector. The connector can be divided into different health levels according to the number, degree and distribution of abnormal areas, such as "good contact and no defects", "normal contact but with minor defects", "abnormal contact or obvious defects", etc. Furthermore, the location of abnormal areas can be marked to provide location information of obvious and latent defects for subsequent maintenance, repair or production quality control.

[0090] Preferably, in some embodiments, time series analysis can be performed on the contact loss characteristics and latent diffusion characteristics of consecutive frames of images, and the dynamic change trend of key areas can be tracked by threshold judgment, thereby realizing dynamic monitoring of the long-term operating status and potential latent defects of connectors, and improving the accuracy and reliability of quality analysis.

[0091] Furthermore, in another aspect of this application, in some embodiments, this application provides an image processing-based connector quality inspection system, which includes an image processing unit, with reference to... Figure 2 The figure is a schematic diagram of the exemplary hardware and / or software structure of an image processing unit according to some embodiments of this application. The image processing unit 200 includes: an image acquisition module 201, an image processing module 202, and an image feature analysis module 203, which are described below:

[0092] The image acquisition module 201 is used to acquire thermal imaging images of the connector through an Internet of Things sensor during the use of the connector, and obtain a thermal imaging image sequence of the connector.

[0093] Image processing module 202 is used to perform differential feature extraction on the thermal imaging image sequence of the connector to obtain an electrical detection image of the thermal imaging image sequence;

[0094] The image processing module 202 is further configured to divide the contact area using the gradient difference image to obtain a contact area sub-image of the gradient difference image, and extract contact features based on the contact area sub-image to obtain the image contact loss features of the connector.

[0095] The image processing module 202 is also used to acquire a comparison image of the connector and a number of preset key detection points, and extract the image visibility and diffusion features of the electrical detection image based on the comparison image and the number of key detection points;

[0096] The image feature analysis module 203 is used to perform connector quality analysis based on the image contact loss characteristics and the image visible diffusion characteristics.

[0097] The foregoing has provided a detailed example of a connector quality inspection system and method based on image processing provided in the embodiments of this application. It is understood that the corresponding device includes hardware structures and / or software modules for performing each function in order to achieve the above functions.

[0098] Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a certain function in the application is executed in a manner that drives hardware or computer software depends on the specific application and design constraints of the technical solution. Therefore, those skilled in the art can use different methods to implement the described function for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0099] In addition, this application also provides a computer terminal device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described image processing-based connector quality detection method.

[0100] In some embodiments, reference Figure 3 The figure is a schematic diagram of the structure of a computer terminal device implementing an image processing-based connector quality inspection method according to some embodiments of this application. The image processing-based connector quality inspection method in the above embodiments can... Figure 3 The computer terminal device 300 shown is used to implement this, and the computer terminal device 300 includes at least one communication bus 301, communication interface 302, processor 303 and memory 304.

[0101] The processor 303 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more devices used to control the execution of an image processing-based connector quality inspection method as described in this application.

[0102] The communication bus 301 may include a path for transmitting information between the aforementioned components.

[0103] Memory 304 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 304 may exist independently and be connected to processor 303 via communication bus 301. Memory 304 may also be integrated with processor 303.

[0104] The memory 304 stores program code for executing the scheme of this application, and its execution is controlled by the processor 303. The processor 303 executes the program code stored in the memory 304. The program code may include one or more software modules. In the above embodiments, the determination of image contact loss characteristics can be achieved by the processor 303 and one or more software modules in the program code in the memory 304.

[0105] Communication interface 302 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0106] Optionally, the computer terminal device 300 may also include a power supply 305 for providing power to various devices or circuits in the real-time computer terminal device.

[0107] In a specific implementation, as one example, a computer terminal device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used for processing data (e.g., computer program instructions).

[0108] The aforementioned computer terminal device can be a general-purpose computer terminal device or a dedicated computer terminal device. In specific implementations, the computer terminal device can be a desktop computer, a portable computer, a network server, a handheld computer (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer terminal device.

[0109] In addition, other aspects of this application provide a computer-readable storage medium storing at least one computer program that is loaded and executed by a processor to perform the operations described above in the image processing-based connector quality inspection method.

[0110] In summary, the connector quality inspection system and method based on image processing disclosed in this application first acquires thermal imaging images of the connector using an IoT sensor to obtain a thermal imaging image sequence of the connector; differential feature extraction is performed on the thermal imaging image sequence of the connector to obtain an electrical inspection image of the thermal imaging image sequence; the contact area is divided using gradient differential images to obtain contact area sub-images of the gradient differential images; contact features are extracted based on the contact area sub-images to obtain the image contact loss features of the connector; a comparison image of the connector and multiple preset key detection points are acquired; image manifestation and diffusion features of the electrical inspection image are extracted based on the comparison image and multiple key detection points; connector quality analysis is performed based on the image contact loss features and image manifestation and diffusion features, which can generate an electrical inspection image of the connector from a set of thermal imaging images, and realize non-contact electrical performance inspection of the connector based on the electrical inspection image.

[0111] The above descriptions are merely embodiments of this application, and common knowledge such as specific technical solutions or characteristics in the solutions are not described in detail here. It should be noted that those skilled in the art can make several modifications and improvements without departing from the technical solutions of this application, and these should also be considered within the scope of protection of this application. None of the above will affect the effectiveness of the implementation of this application or the practicality of the patent.

[0112] The scope of protection claimed in this application shall be determined by the content of its claims. The specific embodiments described in the specification can be used to interpret the content of the claims. Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of the invention. Therefore, if the above-mentioned modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include the above-mentioned modifications and variations.

Claims

1. A connector quality inspection method based on image processing, characterized in that, include: During the use of the connector, thermal imaging images of the connector are acquired through IoT sensors to obtain a sequence of thermal imaging images of the connector. Differential feature extraction is performed on the thermal imaging image sequence of the connector to obtain the gradient difference image of the thermal imaging image sequence; The contact region is divided by the gradient difference image to obtain a contact region sub-image of the gradient difference image. Contact features are extracted based on the contact region sub-image to obtain the image contact loss features of the connector. Obtain a comparison image of the connector and several preset key detection points. Based on the comparison image and the multiple key detection points, extract the image manifestation and concealment diffusion features of the gradient difference image, specifically including: Acquire comparison images and multiple key detection points; For any key detection point, obtain the position coordinates of the key detection point, extract the image gray value of the gradient difference image under the position coordinates, and perform differential detection with the corresponding pixel of the comparison image to obtain the thermal diffusion difference value. When the thermal diffusion difference value is higher than a preset threshold, a standard image of the connector is acquired, and the proportion of visible defects and the proportion of hidden defects corresponding to the key detection point are determined based on the image grayscale value of the standard image of the connector at the position coordinates of the key detection point. Image explicit and implicit diffusion features are constructed based on the thermal diffusion difference value corresponding to each key detection point, the proportion of explicit defects, and the proportion of implicit defects. Connector quality analysis is performed based on the image contact loss characteristics and the image visibility diffusion characteristics.

2. The method as described in claim 1, characterized in that, The gradient difference image of the thermal imaging image sequence obtained by performing differential feature extraction on the thermal imaging image sequence of the connector specifically includes: Each thermal imaging image in the thermal imaging image sequence is acquired, and differential image extraction is performed on each thermal imaging image using the adjacent frame difference method to obtain multiple differential images; The average gray level is extracted based on the co-position pixels of each difference image to obtain the average difference gray level corresponding to each pixel coordinate. The average gradient is extracted based on the co-position pixels of each difference image to obtain the average difference gradient corresponding to each pixel coordinate. Image reconstruction is performed based on the average difference gray level and average difference gradient corresponding to each pixel coordinate to obtain the gradient difference image of the thermal imaging image sequence.

3. The method as described in claim 1, characterized in that, The contact region sub-map obtained by dividing the gradient difference image into contact regions specifically includes: Obtain a standard image of the connector, and perform image alignment on the gradient difference image based on the standard image; The gradient difference image is segmented into a sub-image of the contact region based on the boundary points of the contact region of the standard image.

4. The method as described in claim 1, characterized in that, The contact feature extraction based on the contact area sub-image specifically yields the image contact loss features of the connector, including: Based on the grayscale feature statistics of the contact area sub-image, the contact area feature vector of the contact area sub-image is obtained; The feature vectors of the contact area are subjected to feature clustering, and the corresponding image contact loss features are determined based on the clustering results.

5. The method as described in claim 1, characterized in that, Before performing differential feature extraction from the thermal imaging image sequence of the connector, the method further includes: image preprocessing of the thermal imaging image sequence of the connector.

6. The method as described in claim 1, characterized in that, A thermal imaging camera is used as an IoT sensor to acquire thermal images of the connector.

7. A connector quality inspection system based on image processing, comprising an image processing unit, wherein the image processing unit is used to execute the connector quality inspection method based on image processing according to any one of claims 1 to 6, characterized in that, The image processing unit includes: The image acquisition module is used to acquire thermal imaging images of the connector through an Internet of Things sensor during the use of the connector, and obtain a thermal imaging image sequence of the connector. An image processing module is used to extract differential features from the thermal imaging image sequence of the connector to obtain a gradient difference image of the thermal imaging image sequence. The image processing module is further configured to divide the contact area using the gradient difference image to obtain a contact area sub-image of the gradient difference image, and extract contact features based on the contact area sub-image to obtain the image contact loss features of the connector. The image processing module is also used to acquire a comparison image of the connector and a number of preset key detection points, and to extract the image manifestation and concealment diffusion features of the gradient difference image based on the comparison image and the number of key detection points. The image feature analysis module is used to perform connector quality analysis based on the image contact loss characteristics and the image visible diffusion characteristics.

8. A computer terminal device, characterized in that, The computer terminal device includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute a connector quality inspection method based on image processing as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing at least one computer program, characterized in that, The computer program is loaded and executed by a processor to perform the operations described in any one of claims 1 to 6 of the image processing-based connector quality inspection method.

Images