Distribution line detection method and device based on airborne platform

Abstract

The invention discloses a distribution line detection method based on an airborne platform. The method comprises the following steps: step 1, carrying detection equipment on the airborne platform; step 2, the airborne platform flies according to a predetermined route; 3, collecting a visible light image, an array sound wave signal and an ultraviolet light image; step 4, matching the ultraviolet light image and the array sound wave signal; 5, fusing the amplitude vector, the ultraviolet light descriptor and the visible light descriptor; and step 6, a classification module is established based on a support vector machine, the trained classification model is used for detecting the fusion vector of the distribution line in real time, and the discharge type is output. The device is simple in structure and reasonable in design, synchronous acquisition of sound wave signals, ultraviolet light images and visible light images can be realized, and data matching can be carried out more easily; the sound wave signal and the ultraviolet light image are matched, the interference of noise can be reduced by utilizing complementarity between the sound wave signal and the ultraviolet light image, and more comprehensive monitoring information can be obtained.

Description

Technical Field

[0001] The present invention belongs to the technical field of power distribution line detection, and in particular relates to a power distribution line detection method and device based on an airborne platform. Background Art

[0002] Partial discharge is an early manifestation of insulation degradation and a common fault phenomenon in power equipment and lines. If not detected and handled in time, it may cause serious power accidents and affect the stability and safety of power supply. It is a daily task in the power industry to monitor and locate partial discharge, detect insulation faults in time, and perform targeted maintenance and care on equipment to prevent accidents.

[0003] Based on a series of physical phenomena caused by partial discharge, such as electromagnetic waves, ultrasound, light, heat, etc., it is a common practice to use corresponding detection technologies to monitor and identify partial discharge of distribution lines. However, this approach has the following technical difficulties:

[0004] First, the ultraviolet light signal generated by partial discharge is often very weak and short-lived, making it difficult to capture. In addition, ultraviolet images are easily affected by weather conditions, such as haze, rain and snow, which results in a decrease in image quality.

[0005] Second, the data collection of acoustic signals, ultraviolet images and visible light images is often not synchronized, and data synchronization processing is required. How to effectively fuse the features from different sensors to improve the accuracy and reliability of partial discharge identification is a key issue.

[0006] Therefore, there are many technical difficulties in detecting partial discharge of distribution lines based on acoustic wave signals, ultraviolet light images and visible light image recognition. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide a distribution line detection method based on an airborne platform in view of the deficiencies in the above-mentioned prior art. The method has a simple structure and a reasonable design, can realize the synchronous acquisition of acoustic wave signals, ultraviolet light images and visible light images, can simplify the data processing flow, reduce the workload of data integration, cleaning and conversion, and can more easily perform data matching; by matching acoustic wave signals and ultraviolet light images, the complementarity between them can be utilized to reduce noise interference, and more comprehensive monitoring information can be obtained.

[0008] In order to solve the above technical problems, the technical solution adopted by the present invention is: a power distribution line detection method based on an airborne platform, characterized in that it includes the following steps:

[0009] Step 1: Carrying a detection device on the airborne platform: the detection device includes at least a camera, a wireless acoustic imager and an ultraviolet imager, and the wireless acoustic imager has a microphone array;

[0010] Step 2: The airborne platform flies along a predetermined route;

[0011] Step 3: The camera collects visible light images of the distribution line area to be detected, the wireless acoustic imager collects array acoustic wave signals of the distribution line area to be detected; the ultraviolet imager collects ultraviolet images of the distribution line area to be detected;

[0012] Step 4: Matching the ultraviolet light image and the array acoustic wave signal;

[0013] Step 5: Fusing the amplitude vector, the ultraviolet light descriptor and the visible light descriptor corresponding to the ultraviolet light descriptor to obtain a fused vector;

[0014] Step 6: Establish a classification module based on the support vector machine, use the labeled fusion vector to train the classification module, and use the trained classification model to detect the fusion vector of the distribution line in real time and output the discharge type.

[0015] The above-mentioned distribution line detection method based on the airborne platform is characterized in that: in step six, the marked fusion vectors have at least 5 categories, namely internal discharge, surface discharge, corona discharge, suspension discharge and air gap discharge.

[0016] The above-mentioned distribution line detection method based on the airborne platform is characterized in that: the specific method of "matching the ultraviolet light image and the array acoustic wave signal" in step 4 is:

[0017] Step 401, obtaining a visible light descriptor of a visible light image, and obtaining an ultraviolet light descriptor of an ultraviolet light image;

[0018] Step 402: extracting features from the array acoustic wave signal, where the extracted features are amplitude vectors;

[0019] Step 403: Calculate the distance D between the amplitude vector and all ultraviolet light descriptors in the same context. i , D i represents the distance between the amplitude vector and the i-th ultraviolet light descriptor;

[0020] Step 404: min(D i ) is the most similar matching vector to the ultraviolet light descriptor.

[0021] The above-mentioned distribution line detection method based on the airborne platform is characterized in that: the specific method of "obtaining the visible light descriptor of the visible light image and obtaining the ultraviolet light descriptor of the ultraviolet light image" in step 401 is:

[0022] Step 401a: performing target frame recognition on the visible light image and the ultraviolet light image in the same context respectively, and scaling the ultraviolet light image based on the target frame angle of the visible light image;

[0023] Step 401b: extracting target features from the image data in the target frame of the visible light image and the target frame of the ultraviolet light image respectively;

[0024] Step 401c: converting the visible light feature into a visible light descriptor, and converting the ultraviolet light feature into an ultraviolet light descriptor, wherein the visible light descriptor and the ultraviolet light descriptor are represented in vector form.

[0025] The above-mentioned distribution line detection method based on the airborne platform is characterized in that the specific method of "converting the visible light feature into a visible light descriptor" is:

[0026] Select a feature point in the visible light image, assign a main direction to each feature point, select multiple pairs of sampling points in a circular area around the feature point, and for any pair of sampling points, if the gray value of the first point is greater than the gray value of the second point, the binary comparison result of the corresponding position is 1; otherwise, it is 0; combine all binary comparison results into a binary string as the visible light descriptor of the feature point.

[0027] The above-mentioned distribution line detection method based on the airborne platform is characterized in that: the specific method of "fusing the amplitude vector, the ultraviolet light descriptor and the visible light descriptor corresponding to the ultraviolet light descriptor" in step five is: fusing based on the multi-scale feature fusion model of SwinTransformer, and the number of channels of the multi-scale feature fusion model is 3.

[0028] The above-mentioned distribution line detection method based on the airborne platform is characterized in that: the distance D i is the Euclidean distance.

[0029] The power distribution line detection device based on the airborne platform of the present invention is characterized by:

[0030] These include drones designed to carry cameras, wireless acoustic imagers, and ultraviolet imagers;

[0031] A wireless acoustic imager, carried on an airborne platform, is used to transmit high-frequency sound waves to a target area of ​​a power distribution line, and the wireless acoustic imager has a microphone array, and the microphone array is used to receive array sound wave signals from different positions in space;

[0032] A camera, mounted on the airborne platform, is used to collect visible light images of the distribution line area to be inspected in real time;

[0033] The UV imager is carried on the airborne platform and is used to collect UV images of the distribution line area to be inspected;

[0034] The UAV airport is wirelessly connected to the airborne platform to receive, process and store data from the airborne platform;

[0035] The ground terminal analyzes and processes the collected data to obtain the discharge type.

[0036] The above-mentioned distribution line detection device based on the airborne platform is characterized in that: the camera is located directly in front of the airborne platform; the wireless acoustic imager and the ultraviolet imager are installed at the bottom of the airborne platform through a mounting box and a mounting plate, and the wireless acoustic imager is installed on the mounting plate through a telescopic rod. The wireless acoustic imager and the ultraviolet imager are oriented in the same direction, and the ultraviolet imager is located on the side close to the nose of the airborne platform.

[0037] The above-mentioned power distribution line detection device based on the airborne platform is characterized in that a low-power heating type pitot tube, an ultraviolet imager and a wireless acoustic imager are sequentially arranged along the length direction of the mounting plate.

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

[0039] 1. The invention has a simple structure, reasonable design, and is easy to implement and operate.

[0040] 2. The present invention matches the acoustic wave signal and the ultraviolet light image, and can utilize the complementarity between them to reduce the interference of noise, so as to obtain more comprehensive monitoring information, improve the signal-to-noise ratio of the signal, enhance the image quality, and improve the robustness of the system, thereby improving the recognition and diagnosis accuracy of the target object.

[0041] 3. In the present invention, the camera and the wireless acoustic imager have the same orientation and collect data synchronously. By taking the target frame of the visible light image as a reference and scaling the target frame of the ultraviolet light image, corresponding image data can be obtained, and the use effect is good.

[0042] 4. The present invention can realize the synchronous acquisition of acoustic wave signals, ultraviolet light images and visible light images. Synchronous acquisition can simplify the data processing process, reduce the workload of data integration, cleaning and conversion, make data matching easier, and provide a more accurate and reliable basis, thereby avoiding data inconsistency caused by time differences, thereby speeding up data processing and improving overall work efficiency.

[0043] In summary, the present invention has a simple structure and a reasonable design, can realize the synchronous acquisition of acoustic wave signals, ultraviolet light images and visible light images, can simplify the data processing process, reduce the workload of data integration, cleaning and conversion, and can make data matching easier; by matching acoustic wave signals and ultraviolet light images, the complementarity between them can be used to reduce noise interference, and more comprehensive monitoring information can be obtained.

[0044] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 The present invention is a flow chart of the method.

[0046] Figure 2 The present invention is a flow chart of the method for matching ultraviolet light images and array acoustic wave signals.

[0047] Figure 3 It is a schematic diagram of the structure of the present invention.

[0048] Figure 4 It is a structural schematic diagram of the airborne platform equipped with detection equipment of the present invention.

[0049] Figure 5 The figure is a schematic diagram of the installation structure of the wireless acoustic imager and the ultraviolet imager in the present invention.

[0050] Figure 6 It is a structural schematic diagram of the distribution line detection device in the third embodiment of the present invention.

[0051] Description of reference numerals:

[0052] 1—Wireless acoustic imager; 2—Airborne platform; 3—UAV airport;

[0053] 4—ground terminal; 5—ultraviolet imager; 6—camera;

[0054] 7—Mounting plate; 8—Mounting box; 9—Low power consumption heated pitot tube. DETAILED DESCRIPTION

[0055] The method of the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments of the present invention.

[0056] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0057] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. 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 "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0058] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0059] For ease of description, spatially relative terms such as "above", "above", "on the upper surface of", "above", etc. may be used here to describe the spatial positional relationship between a device or feature and other devices or features as shown in the figure. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figure. For example, if the device in the accompanying drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below other devices or structures". Thus, the exemplary term "above" can include both "above" and "below". The device can also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used here are interpreted accordingly.

[0060] like Figure 1 As shown, the power distribution line detection method based on the airborne platform of the present invention comprises:

[0061] Step 1: Carrying a detection device on the airborne platform 2: The detection device at least includes a camera 6, a wireless acoustic imager 1 and an ultraviolet imager 5, and the wireless acoustic imager 1 has a microphone array.

[0062] In a possible embodiment, the camera 6 is located in front of the airborne platform 2; the wireless acoustic imager 1 and the ultraviolet imager 5 are installed at the bottom of the airborne platform 2 through the installation box 8 and the installation plate 7, the wireless acoustic imager 1 and the ultraviolet imager 5 are oriented in the same direction, and the ultraviolet imager 5 is located on the side close to the nose of the airborne platform 2. The wireless acoustic imager 1 is installed on the installation plate 7 through a telescopic rod.

[0063] The airborne platform 2 is equipped with a camera 6, a wireless acoustic imager 1 and an ultraviolet imager 5 at the same time, and the layout positions of the camera 6, the wireless acoustic imager 1 and the ultraviolet imager 5 do not interfere with each other, so the synchronous acquisition of acoustic wave signals, ultraviolet images and visible light images can be achieved. Synchronous acquisition can simplify the data processing process, reduce the workload of data integration, cleaning and conversion, make data matching easier, and provide a more accurate and reliable basis, thereby avoiding the inconsistency of data caused by time differences, thereby speeding up data processing and improving overall work efficiency.

[0064] Step 2: The airborne platform flies along a predetermined route: The UAV airport 3 works in coordination with the flight control system of the airborne platform 2 to plan and control the flight trajectory of the airborne platform 2, and the airborne platform 2 flies along the predetermined route.

[0065] Step 3: During the flight of the airborne platform 2, the camera 6 collects visible light images of the distribution line area to be detected, the wireless acoustic imager 1 collects array acoustic wave signals of the distribution line area to be detected; and the ultraviolet imager 5 collects ultraviolet images of the distribution line area to be detected.

[0066] Visible light images can reflect the appearance and structural information of distribution lines, and help identify abnormal conditions of lines. The ultraviolet imager 5 uses ultraviolet imaging technology to obtain ultraviolet images. Discharge points usually appear as bright spots or spots in ultraviolet images, and their position and intensity can reflect the severity and distribution range of partial discharge.

[0067] Step 4: Matching the ultraviolet light image and the array acoustic wave signal;

[0068] like Figure 2 As shown, the specific method of "matching the ultraviolet light image and the array acoustic wave signal" in step 4 is:

[0069] Step 401: Obtain a visible light descriptor of a visible light image, and obtain an ultraviolet light descriptor of an ultraviolet light image.

[0070] The specific method of "obtaining the visible light descriptor of the visible light image and obtaining the ultraviolet light descriptor of the ultraviolet light image" in step 401 is:

[0071] Step 401a: performing target frame recognition on the visible light image and the ultraviolet light image in the same context respectively, and scaling the ultraviolet light image based on the target frame angle of the visible light image.

[0072] The target box recognition uses the CornerNet network model, which predicts the corner points of the target, which are usually the upper left corner and the lower right corner, and then constructs the target box through the upper left corner and the lower right corner.

[0073] Since the camera 6 and the wireless acoustic imager 1 are both facing one side of the nose of the airborne platform 2, and the visible light image and the ultraviolet light image are collected synchronously, there is no need to rotate the target frame of the ultraviolet light image. It is only necessary to scale the target frame of the ultraviolet light image to adapt it to the target frame of the visible light image, which is easy to operate.

[0074] Step 401b: extract target features from the target frame of the visible light image and the target frame of the ultraviolet light image respectively. Target feature extraction is to obtain feature points of the visible light image and the ultraviolet light image. In a possible embodiment, a SuperPoint network model is used, which is specially designed for feature point detection.

[0075] Step 401c: converting the visible light feature into a visible light descriptor, and converting the ultraviolet light feature into an ultraviolet light descriptor, wherein the visible light descriptor and the ultraviolet light descriptor are represented in vector form.

[0076] The method for converting visible light features into visible light descriptors is the same as that for converting UV light features into UV light descriptors.

[0077] The specific method of "converting visible light features into visible light descriptors" is:

[0078] Select a feature point in the visible light image, assign a main direction to each feature point, select multiple pairs of sampling points in a circular area around the feature point, and for any pair of sampling points, if the gray value of the first point is greater than the gray value of the second point, the binary comparison result of the corresponding position is 1; otherwise, it is 0; combine all binary comparison results into a binary string as the visible light descriptor of the feature point.

[0079] In a possible embodiment, the SURF algorithm is used to extract feature points and descriptors thereof in a visible light image, and the SURF algorithm is also used to extract feature points and descriptors thereof in an ultraviolet light image.

[0080] Step 402: Extract features from the array acoustic wave signal, where the extracted features are amplitude vectors. In one possible embodiment, for each frame of the array acoustic wave signal, the amplitude is calculated by taking the absolute value or square root of the signal, and the amplitude values ​​of each frame are arranged in order to form an amplitude vector.

[0081] Step 403: Calculate the distance D between the amplitude vector and all ultraviolet light descriptors in the same context. i , D i Represents the distance between the amplitude vector and the i-th ultraviolet light descriptor.

[0082] By calculating the distance between the features, it is determined whether the amplitude vector matches the ultraviolet light descriptor. In one possible embodiment, the match is determined by calculating the Euclidean distance. Matching features should have a smaller distance or a higher similarity. Distance D i is the Euclidean distance.

[0083] Step 404: min(D i ) is the most similar matching vector to the ultraviolet light descriptor.

[0084] Acoustic wave signals and ultraviolet light images provide different physical information respectively; acoustic wave signals can reflect the vibration, movement of objects and the propagation characteristics of the medium, while ultraviolet light images can capture phenomena such as radiation and discharge on the surface of objects.

[0085] The ultraviolet light signal generated by partial discharge is often very weak, and the ultraviolet image may be interfered by noise during transmission and reception. By matching the acoustic signal and the ultraviolet image, the complementarity between them can be used to reduce the interference of noise, obtain more comprehensive monitoring information, improve the signal-to-noise ratio, enhance the image quality, and improve the robustness of the system, thereby improving the recognition and diagnosis accuracy of the target object.

[0086] At the same time, visible light images provide a realistic view of the distribution lines and surroundings, such as the arrangement of wires and the appearance of insulators, while ultraviolet images can capture ultraviolet radiation generated by discharge phenomena. These radiations usually appear in the form of light spots and indicate the specific location and intensity of the discharge. Combining the two can accurately correspond the discharge point to the actual location of the distribution line, giving full play to their respective advantages and achieving technical complementarity.

[0087] Step 5: Fuse the amplitude vector, the ultraviolet light descriptor and the visible light descriptor corresponding to the ultraviolet light descriptor to obtain a fused vector.

[0088] It should be noted that the visible light descriptor corresponding to the ultraviolet light descriptor is the visible light descriptor of the reference visible light image in step 401a.

[0089] The fusion is performed based on the multi-scale feature fusion model of Swin Transformer, and the number of channels of the multi-scale feature fusion model is 3.

[0090] Step 6: Establish a classification module based on the support vector machine, use the labeled fusion vector set to train the classification module, and use the trained classification model to detect the fusion vector of the distribution line in real time and output the discharge type. Establish a classification module based on the support vector machine, and the classification module can automatically learn the association and rules between features to improve the accuracy and efficiency of feature recognition.

[0091] The output layer of the classification module contains neurons equal to the number of discharge types, which are used to output the prediction results of the discharge type.

[0092] It should be noted that, based on historical data and experimental results, a database containing known discharge features is established. The feature library contains fusion vectors of various discharge types. The fusion vectors are labeled, and a label representing the discharge type is assigned to each fusion vector. The discharge type is also the label of the fusion vector, and there are at least five categories, namely internal discharge, surface discharge, corona discharge, suspension discharge and air gap discharge.

[0093] Use the labeled fusion vector set to train the classification module; during the training process, adjust the weight and bias of the classification module through the back propagation algorithm to minimize the loss function. The loss function uses the cross entropy loss to calculate the accuracy of the classification module, and use the accuracy to evaluate the recognition effect of the classification model; a classification model that can recognize different discharge types can be trained.

[0094] The trained classification model is deployed in practical applications for real-time detection of partial discharge in distribution lines and improving the safety and reliability of the power system.

[0095] In one possible embodiment, the discharge type and discharge location are output in the form of a report, which is stored in the ground terminal for subsequent tracking and comparison, providing data support for the long-term operation and maintenance of the distribution line. The report is easy for the operation and maintenance personnel to view, understand and process; the operation and maintenance personnel perform maintenance according to the report, such as replacing damaged insulators and repairing damaged wires, so as to eliminate the hidden dangers of partial discharge and provide strong guarantee for the safe and stable operation of the power system. The discharge location is calculated based on the array acoustic wave signal.

[0096] Embodiment 2

[0097] like Figure 3 , Figure 4 and Figure 5 As shown, the power distribution line detection device based on the airborne platform of the present invention includes a drone 2, which is used to carry a camera 6, a wireless acoustic imager 1 and an ultraviolet imager 5;

[0098] A wireless acoustic imager 1 is carried on an airborne platform 2 and is used to transmit high-frequency sound waves to a target area of ​​a power distribution line. The wireless acoustic imager 1 has a microphone array, and the microphone array is used to receive array sound wave signals from different positions in space.

[0099] The camera 6 is mounted on the airborne platform 2 and is used to collect visible light images of the distribution line area to be inspected in real time;

[0100] The ultraviolet imager 5 is mounted on the airborne platform 2 and is used to collect ultraviolet images of the distribution line area to be inspected;

[0101] The UAV airport 3 is wirelessly connected to the airborne platform 2 to receive, process and store data from the airborne platform 2;

[0102] The ground terminal 4 analyzes and processes the collected data to obtain the discharge type.

[0103] In this embodiment, the camera 6 is located directly in front of the airborne platform 2; the wireless acoustic imager 1 and the ultraviolet imager 5 are installed at the bottom of the airborne platform 2 through the mounting box 8 and the mounting plate 7, the wireless acoustic imager 1 and the ultraviolet imager 5 have the same orientation, and the ultraviolet imager 5 is located on the side close to the nose of the airborne platform 2.

[0104] The wireless acoustic imager 1 is mounted on the mounting plate 7 via a telescopic rod. It should be noted that this arrangement not only ensures flexible deployment of the equipment, but also facilitates height adjustment according to different operating environments and requirements.

[0105] The ultraviolet imager 5 is positioned at the front end of the wireless acoustic imager 1 so that the acoustic data and the on-site image can be synchronized in real time, which greatly improves the accuracy and efficiency of detection and diagnosis.

[0106] In actual operation, the operator can adjust the length of the telescopic rod easily according to the on-site situation to slowly and steadily lower the wireless acoustic imager 1 to the target detection area. In this process, the microphone array of the wireless acoustic imager 1 will not be blocked by the ultraviolet imager 5, avoiding signal interference or data acquisition errors that may be caused by physical blocking, thereby ensuring high definition and accuracy of acoustic imaging.

[0107] The airborne platform 2 is equipped with a camera 6, a wireless acoustic imager 1 and an ultraviolet imager 5 at the same time, and the layout positions of the camera 6, the wireless acoustic imager 1 and the ultraviolet imager 5 do not interfere with each other, so the synchronous acquisition of acoustic wave signals, ultraviolet images and visible light images can be achieved. Synchronous acquisition can simplify the data processing process, reduce the workload of data integration, cleaning and conversion, make data matching easier, and provide a more accurate and reliable basis, thereby avoiding the inconsistency of data caused by time differences, thereby speeding up data processing and improving overall work efficiency.

[0108] Embodiment 3

[0109] like Figure 6 As shown, in a possible embodiment, a low-power heated pitot tube 9, an ultraviolet imager 5 and a wireless acoustic imager 1 are sequentially arranged along the length direction of the mounting plate 7. The low-power heated pitot tube 9 is used to compensate for the ambient temperature of the ultraviolet imager 5, so that the ultraviolet imager 5 is kept within the optimal operating temperature range.

[0110] The above description is only an embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A power distribution line detection method based on an airborne platform, characterized in that: The following steps are involved: Step 1: Mounting a detection device on the airborne platform (2): the detection device at least comprises a camera (6), a wireless acoustic imager (1) and an ultraviolet imager (5), wherein the wireless acoustic imager (1) has a microphone array; Step 2: The airborne platform flies along a predetermined route; Step 3: The camera (6) collects visible light images of the distribution line area to be detected, the wireless acoustic imager (1) collects array acoustic wave signals of the distribution line area to be detected; and the ultraviolet imager (5) collects ultraviolet images of the distribution line area to be detected; Step 4: Matching the ultraviolet light image and the array acoustic wave signal; Step 5: Fusing the amplitude vector, the ultraviolet light descriptor and the visible light descriptor corresponding to the ultraviolet light descriptor to obtain a fused vector; Step 6: Establish a classification module based on the support vector machine, use the labeled fusion vector to train the classification module, and use the trained classification model to detect the fusion vector of the distribution line in real time and output the discharge type.

2. The power distribution line detection method based on an airborne platform according to claim 1, characterized in that: In step six, the labeled fusion vectors have at least five categories, namely internal discharge, surface discharge, corona discharge, suspension discharge and air gap discharge.

3. The power distribution line detection method based on an airborne platform according to claim 1, characterized in that: The specific method of "matching ultraviolet light image and array acoustic wave signal" in step 4 is: Step 401, obtaining a visible light descriptor of a visible light image, and obtaining an ultraviolet light descriptor of an ultraviolet light image; Step 402: extracting features from the array acoustic wave signal, where the extracted features are amplitude vectors; Step 403: Calculate the distance D between the amplitude vector and all ultraviolet light descriptors in the same context. i , D i represents the distance between the amplitude vector and the i-th ultraviolet light descriptor; Step 404: min(D i ) is the most similar matching vector to the ultraviolet light descriptor.

4. The power distribution line detection method based on an airborne platform according to claim 3 is characterized in that: The specific method of "obtaining the visible light descriptor of the visible light image and obtaining the ultraviolet light descriptor of the ultraviolet light image" in step 401 is: Step 401a: performing target frame recognition on the visible light image and the ultraviolet light image in the same context respectively, and scaling the ultraviolet light image based on the target frame angle of the visible light image; Step 401b: extracting target features from the image data in the target frame of the visible light image and the target frame of the ultraviolet light image respectively; Step 401c: converting the visible light feature into a visible light descriptor, and converting the ultraviolet light feature into an ultraviolet light descriptor, wherein the visible light descriptor and the ultraviolet light descriptor are represented in vector form.

5. The power distribution line detection method based on an airborne platform according to claim 1, characterized in that: The specific method of "converting visible light features into visible light descriptors" is: Select a feature point in the visible light image, assign a main direction to each feature point, select multiple pairs of sampling points in a circular area around the feature point, and for any pair of sampling points, if the gray value of the first point is greater than the gray value of the second point, the binary comparison result of the corresponding position is 1; otherwise, it is 0; combine all binary comparison results into a binary string as the visible light descriptor of the feature point.

6. The power distribution line detection method based on an airborne platform according to claim 1, characterized in that: In step 5, the specific method of "fusing the amplitude vector, the ultraviolet light descriptor and the visible light descriptor corresponding to the ultraviolet light descriptor" is: fusing based on the multi-scale feature fusion model of Swin Transformer, and the number of channels of the multi-scale feature fusion model is 3.

7. The power distribution line detection method based on an airborne platform according to claim 1, characterized in that: Distance D i is the Euclidean distance.

8. A power distribution line detection device based on an airborne platform, characterized in that: It includes a drone (2) for carrying a camera (6), a wireless acoustic imager (1) and an ultraviolet imager (5); A wireless acoustic imager (1) is mounted on an airborne platform (2) and is used to transmit high-frequency sound waves to a target area of ​​a power distribution line. The wireless acoustic imager (1) has a microphone array, and the microphone array is used to receive array sound wave signals from different positions in space. A camera (6), mounted on the airborne platform (2), is used to collect visible light images of the distribution line area to be inspected in real time; An ultraviolet imager (5), mounted on the airborne platform (2), is used to collect ultraviolet images of the distribution line area to be inspected; The UAV airport (3) is wirelessly connected to the airborne platform (2) to receive, process and store data from the airborne platform (2); The ground terminal (4) analyzes and processes the collected data to obtain the discharge type.

9. The power distribution line detection device based on an airborne platform according to claim 8, characterized in that: The camera (6) is located in front of the airborne platform (2); the wireless acoustic imager (1) and the ultraviolet imager (5) are installed at the bottom of the airborne platform (2) via a mounting box (8) and a mounting plate (7); the wireless acoustic imager (1) is installed on the mounting plate (7) via a telescopic rod; the wireless acoustic imager (1) and the ultraviolet imager (5) are oriented in the same direction; the ultraviolet imager (5) is located on a side close to the nose of the airborne platform (2).

10. The power distribution line detection device based on an airborne platform according to claim 8, characterized in that: A low-power-consumption heating type pitot tube (9), an ultraviolet imager (5) and a wireless acoustic imager (1) are sequentially arranged along the length direction of the mounting plate (7).

Images