Power transmission line channel hidden danger target positioning method and system based on image segmentation

Abstract

The invention belongs to the technical field of image processing, and provides a power transmission line channel hidden danger target positioning method and system based on image segmentation. The method comprises the following steps: acquiring two-dimensional image data and three-dimensional point cloud data of a power transmission line channel scene; extracting a multi-scale global feature map of the two-dimensional image data of the power transmission line channel scene, and fusing all different-scale global feature maps based on a coordinate attention mechanism to obtain all hidden danger target segmentation regions in the two-dimensional image data of the power transmission line channel scene; based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the power transmission line channel scene, screening out the three-dimensional point cloud data projected in each hidden danger target segmentation region; and classifying the three-dimensional point cloud data of each hidden danger target segmentation region to obtain point cloud data corresponding to each hidden danger target, and finally realizing positioning of each hidden danger target. According to the method, hidden danger targets with different scales under a complex background can be realized, and the hidden danger target segmentation performance is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and in particular relates to a method and system for locating hidden danger targets in a power transmission line channel based on image segmentation. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In complex scenarios such as engineering operations near transmission lines, it is very easy for dangerous situations to occur where engineering vehicles may touch the transmission lines due to improper operation, which may cause damage to the transmission lines and casualties, seriously affecting the normal operation of the power grid.

[0004] The prior art provides a method of using two-dimensional image data of a scene to perform target segmentation processing on the scene image based on a segmentation neural network. However, for complex scenes such as construction near transmission lines, there are large variations in the size of hidden danger targets, complex background environments, and the influence of occlusion by other objects. The existing image segmentation method is not suitable for complex transmission line scenes, resulting in poor segmentation accuracy and difficulty in accurately locating hidden danger targets. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method and system for locating hidden danger targets in a transmission line channel based on image segmentation, which can realize hidden danger targets of different scales under complex backgrounds and significantly improve the hidden danger target segmentation performance.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a method for locating hidden danger targets in a power transmission line channel based on image segmentation.

[0008] In one or more embodiments, a method for locating hidden danger targets in a power transmission line channel based on image segmentation is provided, comprising:

[0009] Obtain 2D image data and 3D point cloud data of the transmission line channel scene;

[0010] Extract multi-scale global feature maps of the two-dimensional image data of the transmission line channel scene, fuse all global feature maps of different scales based on the coordinate attention mechanism, and then obtain all hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area;

[0011] Based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the transmission line channel scene, the three-dimensional point cloud data projected on each hidden danger target segmentation area is screened out;

[0012] The three-dimensional point cloud data of each hidden danger target segmentation area is classified to obtain the point cloud data corresponding to each hidden danger target, and finally the positioning of each hidden danger target is achieved.

[0013] As an implementation method, the process of fusing all global feature maps of different scales based on the coordinate attention mechanism includes:

[0014] Assign a channel to each global feature map of different scales;

[0015] For each channel, the global feature map of the corresponding scale is encoded along the horizontal coordinate and the vertical coordinate respectively to obtain the corresponding horizontal coordinate and vertical coordinate encoding features;

[0016] For each channel, the corresponding horizontal coordinate and vertical coordinate encoding features are spliced ​​along the spatial dimension, and the spliced ​​features are regularized and activated through normalization and nonlinear activation functions. Then, the activation function and normalization weighting are used to obtain the fused feature map of each channel.

[0017] As an implementation method, the expression of the fusion feature map of each channel is:

[0018]

[0019] Among them, δ is the regularization and nonlinear activation operation, σ represents the activation function operation, Conv is the convolution operation, and Y v (i, j) is the fusion feature map of the vth channel; (i, j) are the horizontal coordinate point and the vertical coordinate point respectively; is the horizontal coordinate encoding feature, is the vertical coordinate encoding feature; m is the maximum coordinate value of the fused feature map along the horizontal direction; n is the maximum coordinate value of the fused feature map along the vertical direction; x, y are the horizontal and vertical coordinates respectively.

[0020] As an implementation method, for each channel, the process of encoding the corresponding scale global feature map along the horizontal coordinate is:

[0021] One-dimensional mean pooling is performed on the global feature map of the corresponding scale along the horizontal coordinate direction. Starting from the leftmost side of the global feature map of the corresponding scale, the window size is set and slid in sequence along the width direction. The pixel values ​​covered in the window are averaged to obtain the horizontal coordinate encoding features.

[0022] As an implementation method, for each channel, the process of encoding the corresponding scale global feature map along the vertical coordinate is specifically as follows:

[0023] One-dimensional mean pooling is performed on the global feature map of the corresponding scale along the vertical coordinate direction. Starting from the bottom of the global feature map of the corresponding scale, the window size is set and slid in sequence along the height direction. The pixel values ​​covered in the window are averaged to obtain the vertical coordinate encoding features.

[0024] As an implementation method, the two-dimensional image data of the transmission line channel scene is processed using an image segmentation model to obtain all the hidden danger target segmentation areas therein;

[0025] The image segmentation model includes a feature extraction network, a cross-level feature fusion network and a small target segmentation head; the feature extraction network is used to extract a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene; the cross-level feature fusion network is used to fuse all global feature maps of different scales based on a coordinate attention mechanism to obtain a fused feature map; the small target segmentation head is used to obtain all hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area.

[0026] A second aspect of the present invention provides a transmission line channel hidden danger target positioning system based on image segmentation.

[0027] In one or more embodiments, a transmission line channel hidden danger target positioning system based on image segmentation includes:

[0028] A data acquisition module, which is used to acquire two-dimensional image data and three-dimensional point cloud data of the transmission line channel scene;

[0029] An image segmentation module is used to extract a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene, fuse all the global feature maps of different scales based on the coordinate attention mechanism, and then obtain all the hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area;

[0030] Point cloud projection module, which is used to screen out the three-dimensional point cloud data projected on each hidden danger target segmentation area based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the transmission line channel scene;

[0031] The target positioning module is used to classify the three-dimensional point cloud data of each hidden danger target segmentation area, obtain the point cloud data corresponding to each hidden danger target, and finally realize the positioning of each hidden danger target.

[0032] A third aspect of the present invention provides a computer-readable storage medium.

[0033] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in the method for locating hidden danger targets in a power transmission line channel based on image segmentation as described above.

[0034] A fourth aspect of the present invention provides a computer program product.

[0035] A computer program product includes a computer program / instruction, which, when executed by a processor, implements the steps in the above-mentioned method for locating hidden danger targets in a power transmission line channel based on image segmentation.

[0036] A fifth aspect of the present invention provides an electronic device.

[0037] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for locating hidden danger targets in a power transmission line channel based on image segmentation as described above are implemented.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] (1) The present invention segments the two-dimensional image data of the transmission line channel scene to determine all hidden danger target segmentation areas, and then screens out the three-dimensional point cloud data projected on each hidden danger target segmentation area based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data. The three-dimensional point cloud data of each hidden danger target segmentation area is then classified, thereby achieving accurate positioning of hidden danger targets in transmission line channels of different scales under complex backgrounds, and significantly improving the hidden danger target segmentation performance.

[0040] (2) The present invention targets the multi-scale global feature map of two-dimensional image data of the power transmission line channel scene, performs feature fusion based on the coordinate attention mechanism fusion, introduces grid coordinate information, and expands the attention mechanism from the pixel level to the image coordinate level, which can capture the global spatial structure of the image, pay more attention to the position information of the image, and enhance the understanding and representation capabilities of the image semantic information; the coordinate attention mechanism models the correlation between features in the channel dimension, and captures the relationship between different positions in the spatial dimension, thereby improving the understanding and modeling capabilities of image features; when processing large-size images, the coordinate attention mechanism has good scalability and can effectively handle complex scenes and image contexts. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1It is a flow chart of a method for locating hidden danger targets in a power transmission line channel based on image segmentation according to an embodiment of the present invention;

[0043] Figure 2 It is a schematic diagram of the structure of a transmission line channel hidden danger target positioning system based on image segmentation according to an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0046] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0047] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0048] Figure 1 FIG. 1 is a flow chart of a method for locating hidden danger targets in a power transmission line channel based on image segmentation in an embodiment of the present invention. Figure 1 The method for locating hidden danger targets in a power transmission line channel based on image segmentation in the present embodiment shown may include:

[0049] S101, acquiring two-dimensional image data and three-dimensional point cloud data of a transmission line channel scene;

[0050] S102, extracting a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene, fusing all the global feature maps of different scales based on a coordinate attention mechanism, and then obtaining all the hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on a mapping relationship between the fused feature map and the hidden danger target segmentation area;

[0051] S103, based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the transmission line channel scene, screening out the three-dimensional point cloud data projected on each hidden danger target segmentation area;

[0052] S104, classifying the three-dimensional point cloud data of each hidden danger target segmentation area, obtaining the point cloud data corresponding to each hidden danger target, and finally realizing the positioning of each hidden danger target.

[0053] This embodiment segments the two-dimensional image data of the transmission line channel scene to determine all hidden danger target segmentation areas, and then screens out the three-dimensional point cloud data projected on each hidden danger target segmentation area based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data. The three-dimensional point cloud data of each hidden danger target segmentation area is then classified, thereby achieving accurate positioning of hidden danger targets in transmission line channels of different scales under complex backgrounds, and significantly improving the hidden danger target segmentation performance.

[0054] In step S101, two-dimensional image data of the transmission line channel scene may be acquired using an image acquisition device, such as a camera mounted on an unmanned vehicle.

[0055] The three-dimensional point cloud data of the transmission line channel scene can be obtained using equipment such as laser radar.

[0056] In step S102, the multi-scale global feature map of the extracted two-dimensional image data of the power transmission line channel scene includes but is not limited to shape features, color features, spatial position features, and the like.

[0057] In step S102, the process of fusing all global feature maps of different scales based on the coordinate attention mechanism includes:

[0058] Step S1021: assigning a channel to each global feature map of different scales;

[0059] Step S1022: For each channel, encode the corresponding scale global feature map along the horizontal coordinate and the vertical coordinate respectively to obtain the corresponding horizontal coordinate and vertical coordinate encoding features;

[0060] In step S1022, for each channel, the process of encoding the corresponding scale global feature map along the horizontal coordinate is:

[0061] One-dimensional mean pooling is performed on the global feature map of the corresponding scale along the horizontal coordinate direction. Starting from the leftmost side of the global feature map of the corresponding scale, it slides along the width direction with a set window size (for example, the window size is 3) in sequence, and the pixel values ​​covered in the window are averaged to obtain the horizontal coordinate encoding feature.

[0062] This embodiment integrates the pixel information dispersed in the horizontal direction through the mean pooling operation. After the pixels at different positions are averaged, they are connected with each other, so that the network can capture the long-range dependency relationship in the horizontal direction.

[0063] Among them, for each channel, the process of encoding the corresponding scale global feature map along the vertical coordinate is specifically as follows:

[0064] One-dimensional mean pooling is performed on the global feature map of the corresponding scale along the vertical coordinate direction. Starting from the bottom of the global feature map of the corresponding scale, the window size is set and slid in sequence along the height direction. The pixel values ​​covered in the window are averaged to obtain the vertical coordinate encoding features.

[0065] This embodiment integrates the pixel information dispersed in the vertical direction through the mean pooling operation. After the pixels at different positions are averaged, they are connected with each other, so that the network can capture the long-range dependency in the vertical direction.

[0066] Step S1023: For each channel, the corresponding horizontal coordinate and vertical coordinate encoding features are spliced ​​along the spatial dimension, and the spliced ​​features are regularized and activated through normalization and non-linear activation functions, and then weighted by activation function and normalization to obtain a fused feature map of each channel.

[0067] Specifically, the expression of the fusion feature map of each channel in step S1023 is:

[0068]

[0069] Among them, δ is the regularization and nonlinear activation operation, σ represents the activation function operation, Conv is the convolution operation, and Y v (i, j) is the fusion feature map of the vth channel; (i, j) are the horizontal coordinate point and the vertical coordinate point respectively; is the horizontal coordinate encoding feature, is the vertical coordinate encoding feature; m is the maximum coordinate value of the fused feature map along the horizontal direction; n is the maximum coordinate value of the fused feature map along the vertical direction; x, y are the horizontal and vertical coordinates respectively.

[0070] This embodiment targets a multi-scale global feature map of two-dimensional image data of a transmission line channel scene, performs feature fusion based on a coordinate attention mechanism, introduces grid coordinate information, and expands the attention mechanism from the pixel level to the image coordinate level, which can capture the global spatial structure of the image, pay more attention to the position information of the image, and enhance the ability to understand and represent the semantic information of the image; the coordinate attention mechanism models the correlation between features in the channel dimension, and captures the relationship between different positions in the spatial dimension, thereby improving the ability to understand and model image features; when processing large-size images, the coordinate attention mechanism has good scalability and can effectively handle complex scenes and image contexts.

[0071] In step S102, the two-dimensional image data of the transmission line channel scene is processed using an image segmentation model to obtain all hidden danger target segmentation areas therein;

[0072] The image segmentation model includes a feature extraction network, a cross-level feature fusion network and a small target segmentation head; the feature extraction network is used to extract a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene; the cross-level feature fusion network is used to fuse all global feature maps of different scales based on a coordinate attention mechanism to obtain a fused feature map; the small target segmentation head is used to obtain all hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area.

[0073] For example, the image segmentation model can be implemented using the YOLOv series of algorithms. In this way, the C2f_Faster module can be used in the feature extraction network and the cross-level feature fusion network to improve the model performance while focusing on the simplicity and interpretability of the model, making it more suitable for hidden danger target image segmentation and positioning tasks in complex scenarios.

[0074] It is understandable that in other embodiments, the image segmentation model may also be implemented using other segmentation algorithms, which will not be described in detail here.

[0075] In step S103, the two-dimensional image data and three-dimensional point cloud data of the power transmission line channel scene may be pre-calibrated according to the coordinate relationship of the corresponding acquisition device.

[0076] In step S104, the algorithm for classifying the three-dimensional point cloud data of each hidden danger target segmentation area may adopt the SVM algorithm, etc. It should be noted here that the classification algorithm is used for fine segmentation of point cloud, and the type of classification algorithm does not affect the essence of this solution.

[0077] The types of three-dimensional point cloud data here include but are not limited to transmission line point cloud data, ground point cloud data, vegetation point cloud data, building point cloud data and hidden danger target point cloud data.

[0078] Figure 2 is a schematic diagram of a transmission line channel hidden danger target positioning system based on image segmentation in an embodiment of the present invention. Figure 1 The method of locating hidden danger targets in power transmission line channels based on image segmentation is corresponding to that in Figure 2 As shown, the transmission line channel hidden danger target positioning system based on image segmentation in this embodiment may include:

[0079] A data acquisition module 201 is used to acquire two-dimensional image data and three-dimensional point cloud data of a transmission line channel scene;

[0080] An image segmentation module 202 is used to extract a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene, fuse all the global feature maps of different scales based on the coordinate attention mechanism, and then obtain all the hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area;

[0081] The point cloud projection module 203 is used to screen out the three-dimensional point cloud data projected on each hidden danger target segmentation area based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the transmission line channel scene;

[0082] The target positioning module 204 is used to classify the three-dimensional point cloud data of each hidden danger target segmentation area, obtain the point cloud data corresponding to each hidden danger target, and finally realize the positioning of each hidden danger target.

[0083] It should be noted here that Figure 2 The various modules in the transmission line channel hidden danger target positioning system based on image segmentation are Figure 1 The various steps in the method for locating hidden danger targets in transmission line channels based on image segmentation correspond one to another, and the specific implementation process is the same, which will not be repeated here.

[0084] This embodiment segments the two-dimensional image data of the transmission line channel scene to determine all hidden danger target segmentation areas, and then screens out the three-dimensional point cloud data projected on each hidden danger target segmentation area based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data. The three-dimensional point cloud data of each hidden danger target segmentation area is then classified, thereby achieving accurate positioning of hidden danger targets in transmission line channels of different scales under complex backgrounds, and significantly improving the hidden danger target segmentation performance.

[0085] Reference Figure 3 , a schematic diagram of an electronic device is given. It should be noted that, Figure 3 The electronic device 300 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.

[0086] like Figure 3 As shown, electronic device 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage part 308 into a random access memory (RAM) 303. In RAM 303, various programs and data required for system operation are also stored. Central processing unit 301, ROM 302 and RAM 303 are connected to each other via a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0087] The following components are connected to the I / O interface 305: an input section 306 including a keyboard, a mouse, etc.; an output section 307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a local area network (LAN) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0088] When the central processing unit 301 in the electronic device of this embodiment executes the program, the following is achieved: Figure 1 The steps in the method for locating hidden danger targets in transmission line channels based on image segmentation are shown.

[0089] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, the computer program including a computer program for executing Figure 1 In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit 301, various functions defined in the apparatus of the present application are executed.

[0090] in, Figure 1 The computer program instructions corresponding to the method shown may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0092] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for locating hidden danger targets in power transmission line channels based on image segmentation, characterized in that: include: Obtain 2D image data and 3D point cloud data of the transmission line channel scene; Extract multi-scale global feature maps of the two-dimensional image data of the transmission line channel scene, fuse all global feature maps of different scales based on the coordinate attention mechanism, and then obtain all hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area; Based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the transmission line channel scene, the three-dimensional point cloud data projected on each hidden danger target segmentation area is screened out; The three-dimensional point cloud data of each hidden danger target segmentation area is classified to obtain the point cloud data corresponding to each hidden danger target, and finally the positioning of each hidden danger target is achieved.

2. The method for locating hidden danger targets in power transmission line channels based on image segmentation according to claim 1, characterized in that: The process of fusing all global feature maps of different scales based on the coordinate attention mechanism includes: Assign a channel to each global feature map of different scales; For each channel, the global feature map of the corresponding scale is encoded along the horizontal coordinate and the vertical coordinate respectively to obtain the corresponding horizontal coordinate and vertical coordinate encoding features; For each channel, the corresponding horizontal coordinate and vertical coordinate encoding features are spliced ​​along the spatial dimension, and the spliced ​​features are regularized and activated through normalization and nonlinear activation functions. Then, the activation function and normalization weighting are used to obtain the fused feature map of each channel.

3. The method for locating hidden danger targets in power transmission line channels based on image segmentation according to claim 2, characterized in that: The expression of the fused feature map of each channel is: Among them, δ is the regularization and nonlinear activation operation, σ represents the activation function operation, Conv is the convolution operation, and Y v (i, j) is the fusion feature map of the vth channel; (i, j) are the horizontal coordinate point and the vertical coordinate point respectively; is the horizontal coordinate encoding feature, is the vertical coordinate encoding feature; m is the maximum coordinate value of the fused feature map along the horizontal direction; n is the maximum coordinate value of the fused feature map along the vertical direction; x, y are the horizontal and vertical coordinates respectively.

4. The method for locating hidden danger targets in power transmission line channels based on image segmentation according to claim 2, characterized in that: For each channel, the process of encoding the corresponding scale global feature map along the horizontal coordinate is: One-dimensional mean pooling is performed on the global feature map of the corresponding scale along the horizontal coordinate direction. Starting from the leftmost side of the global feature map of the corresponding scale, the window size is set and slid in sequence along the width direction. The pixel values ​​covered in the window are averaged to obtain the horizontal coordinate encoding features.

5. The method for locating hidden danger targets in power transmission line channels based on image segmentation according to claim 2, characterized in that: For each channel, the process of encoding the corresponding scale global feature map along the vertical coordinate is as follows: One-dimensional mean pooling is performed on the global feature map of the corresponding scale along the vertical coordinate direction. Starting from the bottom of the global feature map of the corresponding scale, the window size is set and slid in sequence along the height direction. The pixel values ​​covered in the window are averaged to obtain the vertical coordinate encoding features.

6. The method for locating hidden danger targets in power transmission line channels based on image segmentation according to claim 1, characterized in that: The image segmentation model is used to process the two-dimensional image data of the transmission line channel scene to obtain all the hidden danger target segmentation areas therein; The image segmentation model includes a feature extraction network, a cross-level feature fusion network and a small target segmentation head; the feature extraction network is used to extract a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene; the cross-level feature fusion network is used to fuse all global feature maps of different scales based on a coordinate attention mechanism to obtain a fused feature map; the small target segmentation head is used to obtain all hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area.

7. A transmission line channel hidden danger target positioning system based on image segmentation, characterized in that: include: A data acquisition module, which is used to acquire two-dimensional image data and three-dimensional point cloud data of the transmission line channel scene; An image segmentation module is used to extract a multi-scale global feature map of the two-dimensional image data of the transmission line channel scene, fuse all the global feature maps of different scales based on the coordinate attention mechanism, and then obtain all the hidden danger target segmentation areas in the two-dimensional image data of the transmission line channel scene based on the mapping relationship between the fused feature map and the hidden danger target segmentation area; Point cloud projection module, which is used to screen out the three-dimensional point cloud data projected on each hidden danger target segmentation area based on the calibration relationship between the two-dimensional image data and the three-dimensional point cloud data of the transmission line channel scene; The target positioning module is used to classify the three-dimensional point cloud data of each hidden danger target segmentation area, obtain the point cloud data corresponding to each hidden danger target, and finally realize the positioning of each hidden danger target.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for locating hidden danger targets in a transmission line channel based on image segmentation as described in any one of claims 1 to 6 are implemented.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps in the method for locating hidden danger targets in a transmission line channel based on image segmentation as described in any one of claims 1 to 6 are implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the method for locating hidden danger targets in a power transmission line channel based on image segmentation as described in any one of claims 1 to 6 are implemented.

Images