Insulator defect detection method and system based on 3D attention focusing and pure convolution

Abstract

The invention relates to the technical field of power grid operation and maintenance, and discloses an insulator defect detection method and system based on 3D attention focusing and pure convolution, by introducing a 3D attention mechanism and pure convolution, the model can enhance the extraction and focusing capability of insulator defect features, and reduce the interference of irrelevant background features; besides, through an improved mixed spatial pyramid pooling module and a lightweight upsampling operator, more detail information can be reserved, and the detection precision of a small defect target is remarkably improved; the technical innovation ensures that the model can realize high-precision and real-time insulator defect detection in a complex environment, and provides powerful guarantee for stable operation of a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid operation and maintenance, and in particular to an insulator defect detection method and system based on 3D attention focusing and pure convolution. Background Art

[0002] At present, the defect detection technology of power line insulators is facing a series of severe challenges in dealing with the complex and ever-changing actual environment. These challenges are not only due to the limitations of the technology itself, but also deeply affected by the complexity and uncertainty of the external environment.

[0003] First, the performance of traditional detection methods, especially those based on vision, is largely restricted by environmental factors such as complex background and lighting conditions. In practical applications, complex and changing backgrounds often cause serious interference to the detection results, resulting in a significant reduction in the stability and accuracy of the detection results. At the same time, changes in lighting conditions will also have a significant impact on the performance of visual detection methods, making it difficult to ensure detection efficiency and even more difficult to meet real-time requirements.

[0004] In addition, although deep learning methods have made significant progress in the fields of image processing and computer vision in recent years, providing new possibilities for insulator defect detection, existing deep learning methods still have problems that cannot be ignored when facing the small target of insulator defects. Specifically, when dealing with the specific task of insulator defects, existing methods often find it difficult to fully extract and utilize the feature information in the image, especially those detailed features that are critical for defect detection. This leads to a significant increase in the difficulty of accurately detecting defects in practical applications when the color of the insulator is similar to the background or is in a complex and changing scene, which may even lead to missed detection or false detection.

[0005] In summary, the existing power line insulator defect detection technology faces multiple challenges in a complex environment. These challenges not only stem from the limitations of the technology itself, but are also deeply affected by the complexity and uncertainty of the external environment.

[0006] Therefore, in order to improve the accuracy and stability of insulator defect detection technology, it is necessary to continuously explore new detection methods and means to better meet these challenges.

[0007] The above information is presented as background information only to assist with understanding the present disclosure and no determination or admission is made as to whether any of the above may be used as prior art with respect to the present disclosure. Summary of the invention

[0008] The present invention provides an insulator defect detection method and system based on 3D attention focusing and pure convolution to solve the problems existing in the prior art.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] In a first aspect, the present invention provides an insulator defect detection method based on 3D attention focusing and pure convolution, the method comprising:

[0011] Acquire an insulator image;

[0012] The insulator image is input into the CMYOLOv7 model to obtain the insulator defect detection result; the CMYOLOv7 model is based on the YOLOv7 model and is constructed by integrating the ConvSimCB pure convolution feature extraction module, the MIXPCSPC mixed space pyramid pooling module and the CARAFE lightweight upsampling operator module.

[0013] Furthermore, in the insulator defect detection method based on 3D attention focusing and pure convolution, the integration position of the ConvSimCB module is the position of the first ELAN module and the fourth ELAN module in the original backbone network.

[0014] Furthermore, in the insulator defect detection method based on 3D attention focusing and pure convolution, the ConvSimCB pure convolution feature extraction module is integrated by the ConvNeXtBlock pure convolution module, the SimAM attention module and the CBS convolution batch normalization module;

[0015] The ConvNeXtBlock pure convolution module is based on excellent convolution technology;

[0016] The SimAM attention module guides the model to focus on and pay attention to the target features through 3D attention weights;

[0017] The CBS convolution batch normalization module further enhances the feature extraction function of the ConvSimCB module.

[0018] Further, in the insulator defect detection method based on 3D attention focusing and pure convolution, the MIXPCSPC mixed spatial pyramid pooling module includes a mixed pooling layer;

[0019] The CARAFE lightweight upsampling operator module includes a prediction part of the upsampling kernel and a content-based reconstruction part.

[0020] In a second aspect, the present invention provides an insulator defect detection system based on 3D attention focusing and pure convolution, the system comprising:

[0021] An image acquisition module, used for acquiring an insulator image;

[0022] A defect detection module is used to input the insulator image into a CMYOLOv7 model to obtain an insulator defect detection result; the CMYOLOv7 model is based on a YOLOv7 model and is constructed by integrating a ConvSimCB pure convolution feature extraction module, a MIXPCSPC mixed space pyramid pooling module, and a CARAFE lightweight upsampling operator module.

[0023] Furthermore, in the insulator defect detection system based on 3D attention focusing and pure convolution, the integration position of the ConvSimCB module is the position of the first ELAN module and the fourth ELAN module in the original backbone network.

[0024] Furthermore, in the insulator defect detection system based on 3D attention focusing and pure convolution, the ConvSimCB pure convolution feature extraction module is integrated by the ConvNeXtBlock pure convolution module, the SimAM attention module and the CBS convolution batch normalization module;

[0025] The ConvNeXtBlock pure convolution module is based on excellent convolution technology;

[0026] The SimAM attention module guides the model to focus on and pay attention to the target features through 3D attention weights;

[0027] The CBS convolution batch normalization module further enhances the feature extraction function of the ConvSimCB module.

[0028] Further, in the insulator defect detection system based on 3D attention focusing and pure convolution, the MIXPCSPC mixed spatial pyramid pooling module includes a mixed pooling layer;

[0029] The CARAFE lightweight upsampling operator module includes a prediction part of the upsampling kernel and a content-based reconstruction part.

[0030] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the insulator defect detection method based on 3D attention focusing and pure convolution as provided in the first aspect above is implemented.

[0031] In a fourth aspect, the present invention provides a computer-readable storage medium having computer-executable instructions stored thereon, wherein the computer-executable instructions are executed by a computer processor to implement the insulator defect detection method based on 3D attention focusing and pure convolution as provided in the first aspect above.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention provides an insulator defect detection method and system based on 3D attention focusing and pure convolution. By introducing the 3D attention mechanism and pure convolution, the model can enhance the ability to extract and focus on insulator defect features while reducing the interference of irrelevant background features; in addition, through the improved hybrid spatial pyramid pooling module and lightweight upsampling operator, the present invention can retain more detail information and significantly improve the detection accuracy of small defect targets; these technical innovations ensure that the model can achieve high-precision and real-time insulator defect detection in complex environments, providing a strong guarantee for the stable operation of the power system.

[0034] The present invention has other features and advantages, which will be apparent from or will be described in detail in the accompanying drawings and the following detailed description incorporated herein, which together serve to explain the specific principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 It is a flowchart of an insulator defect detection method based on 3D attention focusing and pure convolution provided in the first embodiment of the present invention;

[0037] Figure 2 It is a CMYOLOv7 model structure diagram mentioned in the first embodiment of the present invention;

[0038] Figure 3 It is a structural diagram of the ConvSimCB pure convolution feature extraction module mentioned in the first embodiment of the present invention;

[0039] Figure 4 is a structural diagram of ConvNeXt mentioned in the first embodiment of the present invention;

[0040] Figure 5 It is the SimAM attention mechanism diagram mentioned in the first embodiment of the present invention;

[0041] Figure 6 It is a structural diagram of the MIXPCSPC hybrid spatial pyramid pooling module mentioned in the first embodiment of the present invention;

[0042] Figure 74 is a structural diagram of the CARAFE lightweight upsampling operator module mentioned in the first embodiment of the present invention;

[0043] Figure 8 is a training loss iteration curve diagram of the comparison algorithm mentioned in the first embodiment of the present invention;

[0044] Fig. 9 It is a functional module schematic diagram of an insulator defect detection system based on 3D attention focusing and pure convolution provided in the second embodiment of the present invention;

[0045] Fig.10 It is a structural diagram of a computer device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0046] In order to explain in detail the possible application scenarios, technical principles, specific schemes that can be implemented, and the purposes and effects that can be achieved, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0047] Reference to "embodiment" herein means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The term "embodiment" appearing in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or association with other embodiments. In principle, in the present application, as long as there is no technical contradiction or conflict, the various technical features mentioned in the embodiments can be combined in any way to form a corresponding implementable technical solution.

[0048] Unless otherwise defined, the technical terms used in this document have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms in this document is only for describing specific embodiments and is not intended to limit this application.

[0049] In the description of this application, the term "and / or" is an expression used to describe the logical relationship between objects, indicating that three relationships may exist, for example, A and / or B, which means: A exists, B exists, and A and B exist at the same time. In addition, the character " / " in this article generally indicates that the objects before and after are in an "or" logical relationship.

[0050] In the present application, terms such as “first” and “second” are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship of quantity, priority or sequence between these entities or operations.

[0051] Without further limitations, in this application, the words "include", "comprises", "has" or other similar expressions used in the sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those limited elements, but also other elements not explicitly listed, or also include elements inherent to such process, method or product.

[0052] In this application, expressions such as "greater than", "less than", "exceed" and the like are understood to exclude the number itself; expressions such as "above", "below", "within" and the like are understood to include the number itself. In addition, in the description of the embodiments of this application, "multiple" means more than two (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups", "multiple times", etc., unless otherwise clearly and specifically limited.

[0053] In the description of the embodiments of the present application, space-related expressions used, such as "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "vertical", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicate the orientation or position relationship based on the orientation or position relationship shown in the specific embodiments or drawings, and are only for the convenience of describing the specific embodiments of the present application or facilitating the reader's understanding, and do not indicate or imply that the referred device or component must have a specific position, a specific orientation, or be constructed or operated in a specific orientation, and therefore cannot be understood as a limitation on the embodiments of the present application.

[0054] Unless otherwise expressly specified or limited, in the description of the embodiments of the present application, the terms such as "install", "connect", "connect", "fix", "set", etc. used should be understood in a broad sense. For example, the "connection" can be a fixed connection, a detachable connection, or an integrated setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For technicians in the technical field to which the present application belongs, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0055] Embodiment 1

[0056] Please refer to Figure 1 , is a flow chart of an insulator defect detection method based on 3D attention focusing and pure convolution provided in Embodiment 1 of the present invention, wherein the method performs defect detection on insulators of transmission lines. The method specifically comprises the following steps:

[0057] S101. Acquire an insulator image.

[0058] It should be noted that appropriate image acquisition equipment (such as drones, inspection robots, cameras, etc.) is used to capture images of insulators on transmission lines.

[0059] It is understandable that some preprocessing may be performed on the collected insulator images, including denoising, enhancement, normalization and other operations, to improve the accuracy of subsequent processing.

[0060] S102, inputting the insulator image into the CMYOLOv7 model to obtain the insulator defect detection result; the CMYOLOv7 model is based on the YOLOv7 model and is constructed by integrating the ConvSimCB pure convolution feature extraction module, the MIXPCSPC mixed space pyramid pooling module and the CARAFE lightweight upsampling operator module.

[0061] It should be noted that the detected insulator fault information (such as fault location, type, severity, etc.) is output in the form of visualization or data for further processing by the user or system.

[0062] The CMYOLOv7 model is Figure 2 As shown, Figure 2 (a) is the structure diagram of the YOLOv7 model, and (b) is the structure diagram of the CMYOLOv7 model.

[0063] exist Figure 2 In (b), the integration position of the ConvSimCB module is the position of the first ELAN module and the fourth ELAN module in the original backbone network. This improved strategy enhances the feature extraction capability of tiny target insulator defects, significantly attenuates feature interference irrelevant to the detection target, and enhances the focus on insulator defect features, thereby greatly improving the detection accuracy of the model proposed in this paper.

[0064] The structure of the ConvSimCB pure convolution feature extraction module is as follows: Figure 3 As shown, it is integrated by ConvNeXtBlock pure convolution module, SimAM attention module and CBS convolution batch normalization module;

[0065] The ConvNeXtBlock pure convolution module is based on excellent convolution technology;

[0066] The SimAM attention module guides the model to focus on and pay attention to the target features through 3D attention weights;

[0067] The CBS convolution batch normalization module further enhances the feature extraction function of the ConvSimCB module.

[0068] The structure of ConvNeXt is as follows Figure 4 As shown in the figure, ResNet adopts a "high-dimensional-low-dimensional-high-dimensional" bottleneck structure to reduce the amount of calculation. In contrast, MobileNetV2 proposes an inverse bottleneck structure, using a "low-dimensional-high-dimensional-low-dimensional" format to avoid excessive information loss when compressing dimensions.

[0069] The ConvNeXtBlock module of the ConvNeXt network is introduced as a feature extraction structure in ConvSimCB to enhance the ability to extract insulator defect features. The ConvNeXtBlock module efficiently extracts global and local features by dividing the input feature map into multiple sub-paths. In addition, the parameter sharing mechanism of the sub-paths greatly reduces the model parameters and improves the training efficiency. The ConvNeXtBlock module adaptively aggregates features based on the attention weight of each sub-path, thereby improving model performance and optimizing the training process.

[0070] The present invention introduces the SimAM attention mechanism in ConvSimCB. The SimAM attention mechanism is as follows Figure 5 As shown. The SimAM attention mechanism distinguishes the difference between the target neuron and other neurons by defining a linearly separable minimum energy function to evaluate the importance of each neuron in the network. The minimum energy function is:

[0071] ……(1);

[0072] ……(2);

[0073] ……(3);

[0074] Among them, λ is a hyperparameter; M = H × W is the number of neurons in each channel; t is the target neuron; is the i-th neuron of the input feature map on a single channel; is the average of all neurons on a single channel; is the variance of all neurons on a single channel; and Used to reduce the amount of calculation.

[0075] The minimum energy function of equation (1) shows that The lower it is, the greater the difference between the target neuron and its neighboring neurons, indicating that the target neuron is more critical. The weight of each neuron on the feature map can be obtained by Found. The final output is

[0076] ……(4);

[0077] Where E is the total number of channels and spatial dimensions. grouping.

[0078] A CBS module with a convolution kernel size of 3×3 and a stride of 1 is added next to the SimAM attention mechanism to extract deeper semantic information.

[0079] In order to solve the problem of over-smoothing and loss of important edge detail information in images, the present invention proposes a MIXPCSPC mixed spatial pyramid pooling module (abbreviated as MIXPCSPC module) to solve the above problems. The structure of the MIXPSPC module is as follows Figure 6 . In the MIXPCSPC module, a mixed pooling layer is introduced instead of a maximum pooling layer. This improved strategy can reduce the resolution of the feature map while effectively achieving the model's translation invariance and rotation invariance, more effectively retain rich image edge detail information, and significantly improve the accuracy of the detection model. The introduction of mixed pooling balances the size of the receptive field and the accuracy of feature representation, thereby providing the model with a comprehensive understanding of key information in the image, and has significant advantages in dealing with detection tasks or problems in various complex scenes. Mixed pooling is a pooling strategy that is derived by simultaneously utilizing the advantages of maximum pooling and average pooling.

[0080] YOLOv7 uses the nearest neighbor interpolation upsampling method to enhance the resolution of the feature map. This method is simple, computationally inefficient and fast. However, this method may produce obvious jagged structures when processing edge details, thereby affecting the recognition of detail features. The jagged effect in the edge area may cause the edge contour of the target object to be not smooth enough, thereby reducing the ability to accurately perceive the target boundary. In order to solve this problem, the present invention proposes upsampling through the CARAFE lightweight upsampling operator (hereinafter referred to as the CARAFE operator). This method adaptively perceives the receptive field of the feature map through convolution operations, effectively improves the resolution of the feature map, and improves the fidelity of image details. CARAFE outperforms traditional upsampling methods in terms of details and accuracy.

[0081] The structure of the upsampling operator CARAFE is as follows Figure 7 As shown in Figure 2, the CARAFE upsampling method consists of two parts: prediction of upsampling kernels and content-based reconstruction. First, the upsampling kernel is predicted based on the content of each object position. The size is The feature map channels are compressed to The upsampling factor used in this study is σ, which is set to 2. The size of the upsampling kernel is . With content encoding, the number of channels becomes , and then expand the channel in the spatial dimension, and normalize the predicted upsampled kernel by applying the softmax() function to generate a convolution kernel with a weight sum of 1. Finally, the content-based reconstruction part maps each position of the output feature map to the corresponding area of ​​the input feature map through a dot product operation; this process is achieved by the dot product operation of the predicted upsampled kernel and the original feature map area, and the final output size is feature map.

[0082] By predicting the upsampling kernel and completing feature reorganization based on the upsampling kernel, the upsampling process can make full use of the perceptual information of the input feature map. The CARAFE upsampling method is used to replace the nearest neighbor interpolation upsampling in the original network, which significantly restores the detailed features of the insulator defects and significantly improves the detection accuracy. This optimization strategy has a significant effect on the reconstruction of detailed features and helps to improve the accuracy and stability of insulator defect detection.

[0083] Experimental results:

[0084] In order to verify the performance of the proposed CMYOLOv7, it is compared with advanced detection algorithms such as YOLOv7, YOLOv5s, YOLOv4, YOLOv3, SSD, EfficientDet and RetinaNet. The training loss iteration curves of the compared algorithms are shown in Figure 2. Figure 8 As shown. Table 1 shows the precision, recall, MAP, F1 score and model size of the compared algorithms. Table 1 shows that CMYOLOv7 has better performance than the original YOLOv7. The precision, recall, MAP and F1 score of CMYOLOv7 are 0.61%, 8.64%, 5.41% and 5% higher than those of the YOLOv7 algorithm, respectively. The MAP of the CMYOLOv7 algorithm is 0.3% lower than that of the SSD algorithm, but higher than that of other advanced comparison algorithms, which verifies that the performance of CMYOLOv7 is better than that of the original YOLOv7, YOLOv5s, YOLOv4, YOLOv3, EfficientDet and RetinaNet; although CMYOLOv7 and SSD are similar in all evaluation indicators, the overall performance of CMYOLOv7 is higher than that of SSD. Among the compared algorithms, EfficientDet has the smallest model size, and its indicators reflect performance much lower than other algorithms.

[0085] Table 1:

[0086]

[0087] An analysis of the results is as follows.

[0088] YOLOv3 misidentifies one defect as two. The defect detection algorithms of YOLOv4 and YOLOv3 cannot identify defects well when the color of the insulator is similar to the background color. However, the proposed CMYOLOv7 insulator defect detection method has a low false detection rate and is able to achieve accurate detection of insulator edge defects. CMYOLOv7 is able to accurately detect defects even when the color features of the insulator are similar to the background color features. This is attributed to the proposed ConvSimCB module, which enhances the feature extraction capability of insulator defects in complex environments and improves the attention and focus on defects.

[0089] The EfficientDet and RetinaNet methods have obvious under-detection of tiny insulator defects in complex environments and cannot accurately capture defects at the edge. The MIXPCSPC module proposed in this study and the CARAFE upsampling method introduced can retain more detailed features and enhance the fused feature representation, thereby improving the effectiveness of the model in detecting tiny insulator defects in complex environments. Therefore, CMYOLOv7 can accurately capture defects at the edge.

[0090] The detection effect of SSD is similar to that of CMYOLOv7. However, the overall indicators of CMYOLOv7 are more balanced, which further verifies that the proposed CMYOLOv7 is superior to SSD.

[0091] The present invention explains to users the credibility of the CMYOLOv7 model in the insulator defect detection task through actual detection results, helps users understand the difference between the model output results and the actual results, and effectively applies the CMYOLOv7 model to the actual transmission line insulator defect detection. In addition, the insulators in the detection results are located in lakes, grasslands, the sky, and various complex environmental backgrounds similar to the color of the insulators, and include different lighting conditions such as sunny and cloudy days. The CMYOLOv7 model can still accurately detect the defective position of the insulator, further verifying the robustness of the CMYOLOv7 model.

[0092] Therefore, the CMYOLOv7 model proposed in this patent is verified to be advanced. The interaction between the three improvement strategies (i.e., ConvSimCB, MIXPCSPC, and CARAFE) enables the CMYOLOv7 insulator defect detection method to achieve high-precision insulator defect detection in complex environmental backgrounds.

[0093] Although the terms such as model and detection are used more frequently in this application, the possibility of using other terms is not excluded. These terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional restrictions is contrary to the spirit of the present invention.

[0094] An embodiment of the present invention provides an insulator defect detection method based on 3D attention focusing and pure convolution. By introducing the 3D attention mechanism and pure convolution, the model can enhance the ability to extract and focus on insulator defect features while reducing the interference of irrelevant background features; in addition, through the improved hybrid spatial pyramid pooling module and lightweight upsampling operator, the present invention can retain more detail information and significantly improve the detection accuracy of small defect targets; these technical innovations ensure that the model can achieve high-precision and real-time insulator defect detection in complex environments, providing a strong guarantee for the stable operation of the power system.

[0095] Embodiment 2

[0096] Please refer to Fig. 9 , Fig. 9 A functional module diagram of an insulator defect detection system based on 3D attention focusing and pure convolution provided in the second embodiment of the present invention, which is suitable for executing the insulator defect detection method based on 3D attention focusing and pure convolution provided in the embodiment of the present invention. The system specifically includes the following modules:

[0097] An image acquisition module 201 is used to acquire an insulator image;

[0098] The defect detection module 202 is used to input the insulator image into the CMYOLOv7 model to obtain the insulator defect detection result; the CMYOLOv7 model is based on the YOLOv7 model and is constructed by integrating the ConvSimCB pure convolution feature extraction module, the MIXPCSPC mixed space pyramid pooling module and the CARAFE lightweight upsampling operator module.

[0099] Furthermore, the integration position of the ConvSimCB module is the position of the first ELAN module and the fourth ELAN module in the original backbone network.

[0100] Furthermore, the ConvSimCB pure convolution feature extraction module is integrated by the ConvNeXtBlock pure convolution module, the SimAM attention module and the CBS convolution batch normalization module;

[0101] The ConvNeXtBlock pure convolution module is based on excellent convolution technology;

[0102] The SimAM attention module guides the model to focus on and pay attention to the target features through 3D attention weights;

[0103] The CBS convolution batch normalization module further enhances the feature extraction function of the ConvSimCB module.

[0104] Further, the MIXPCSPC mixed spatial pyramid pooling module includes a mixed pooling layer;

[0105] The CARAFE lightweight upsampling operator module includes a prediction part of the upsampling kernel and a content-based reconstruction part.

[0106] An embodiment of the present invention provides an insulator defect detection system based on 3D attention focusing and pure convolution. By introducing the 3D attention mechanism and pure convolution, the model can enhance the ability to extract and focus on insulator defect features while reducing the interference of irrelevant background features; in addition, through the improved hybrid spatial pyramid pooling module and lightweight upsampling operator, the present invention can retain more detail information and significantly improve the detection accuracy of small defect targets; these technical innovations ensure that the model can achieve high-precision and real-time insulator defect detection in complex environments, providing a strong guarantee for the stable operation of the power system.

[0107] The above system can execute the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0108] Embodiment 3

[0109] Fig.10 A schematic diagram of the structure of a computer device provided in Embodiment 3 of the present invention. Fig.10 A block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention is shown. Fig.10 The computer device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0110] like Fig.10 As shown, the computer device 12 is in the form of a general-purpose computing device. The components of the computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).

[0111] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0112] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including volatile and non-volatile media, removable and non-removable media.

[0113] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Fig.10 not shown, usually called a "hard drive"). Although Fig.10 Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0114] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in the memory 28, such program modules 42 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0115] The computer device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), one or more devices that enable a user to interact with the computer device 12, and / or any device that enables the computer device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface 22. In addition, the computer device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with the other modules of the computer device 12 via a bus 18. It should be understood that although Fig.10Not shown, other hardware and / or software modules may be used in conjunction with computer device 12, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0116] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, such as implementing the insulator defect detection method based on 3D attention focusing and pure convolution provided in an embodiment of the present invention.

[0117] Embodiment 4

[0118] Embodiment 4 of the present invention provides a computer-readable storage medium having computer-executable instructions stored thereon. When the instructions are executed by a processor, an insulator defect detection method based on 3D attention focusing and pure convolution as provided in all the embodiments of the invention of this application is implemented.

[0119] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, device, or device.

[0120] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0121] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0122] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0123] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of patent protection of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concept of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of patent protection of this application.

Claims

1. An insulator defect detection method based on 3D attention focusing and pure convolution, characterized in that: The method comprises: Acquire an insulator image; The insulator image is input into the CMYOLOv7 model to obtain the insulator defect detection result; the CMYOLOv7 model is based on the YOLOv7 model and is constructed by integrating the ConvSimCB pure convolution feature extraction module, the MIXPCSPC mixed space pyramid pooling module and the CARAFE lightweight upsampling operator module.

2. The insulator defect detection method based on 3D attention focusing and pure convolution according to claim 1 is characterized in that: The integration position of the ConvSimCB module is the position of the first ELAN module and the fourth ELAN module in the original backbone network.

3. The insulator defect detection method based on 3D attention focusing and pure convolution according to claim 2 is characterized in that: The ConvSimCB pure convolution feature extraction module is integrated by the ConvNeXtBlock pure convolution module, the SimAM attention module and the CBS convolution batch normalization module; The ConvNeXtBlock pure convolution module is based on excellent convolution technology; The SimAM attention module guides the model to focus on and pay attention to the target features through 3D attention weights; The CBS convolution batch normalization module further enhances the feature extraction function of the ConvSimCB module.

4. The insulator defect detection method based on 3D attention focusing and pure convolution according to claim 1 is characterized in that: The MIXPCSPC mixed spatial pyramid pooling module includes a mixed pooling layer; The CARAFE lightweight upsampling operator module includes a prediction part of the upsampling kernel and a content-based reconstruction part.

5. An insulator defect detection system based on 3D attention focusing and pure convolution, characterized in that: The system comprises: An image acquisition module, used for acquiring an insulator image; A defect detection module is used to input the insulator image into a CMYOLOv7 model to obtain an insulator defect detection result; the CMYOLOv7 model is based on a YOLOv7 model and is constructed by integrating a ConvSimCB pure convolution feature extraction module, a MIXPCSPC mixed space pyramid pooling module, and a CARAFE lightweight upsampling operator module.

6. The insulator defect detection system based on 3D attention focusing and pure convolution according to claim 5, characterized in that: The integration position of the ConvSimCB module is the position of the first ELAN module and the fourth ELAN module in the original backbone network.

7. The insulator defect detection system based on 3D attention focusing and pure convolution according to claim 6, characterized in that: The ConvSimCB pure convolution feature extraction module is integrated by the ConvNeXtBlock pure convolution module, the SimAM attention module and the CBS convolution batch normalization module; The ConvNeXtBlock pure convolution module is based on excellent convolution technology; The SimAM attention module guides the model to focus on and pay attention to the target features through 3D attention weights; The CBS convolution batch normalization module further enhances the feature extraction function of the ConvSimCB module.

8. The insulator defect detection system based on 3D attention focusing and pure convolution according to claim 5, characterized in that: The MIXPCSPC mixed spatial pyramid pooling module includes a mixed pooling layer; The CARAFE lightweight upsampling operator module includes a prediction part of the upsampling kernel and a content-based reconstruction part.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the insulator defect detection method based on 3D attention focusing and pure convolution as described in any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: The computer executable instructions are executed by a computer processor to implement the insulator defect detection method based on 3D attention focusing and pure convolution as described in any one of claims 1 to 4.

Images