CN112614130A - Unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and YOLOv3 - Google Patents
Unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and YOLOv3 Download PDFInfo
- Publication number
- CN112614130A CN112614130A CN202110002537.1A CN202110002537A CN112614130A CN 112614130 A CN112614130 A CN 112614130A CN 202110002537 A CN202110002537 A CN 202110002537A CN 112614130 A CN112614130 A CN 112614130A
- Authority
- CN
- China
- Prior art keywords
- insulator
- transmission line
- unmanned aerial
- aerial vehicle
- yolov3
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Locating Faults (AREA)
Abstract
The invention discloses an unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and improved YOLOv3, belonging to the field related to sensors and comprising the following steps: the invention carries out defect detection and fault location on the insulator of the power transmission line by establishing an improved YOLOv3 cascade network structure, thereby greatly improving the defect detection accuracy and realizing the accurate location of the fault insulator block. Aiming at the problems of low transmission speed and unstable signals of a communication scheme in a remote area, the method aims to adopt a 5G unmanned aerial vehicle base station technology to transmit video data of a transmission line area to be patrolled, which is shot by an unmanned aerial vehicle, so that the real-time performance of data transmission is improved, and high-resolution video frame images are provided for subsequent insulator detection, positioning and fault classification.
Description
Technical Field
The invention relates to a transmission line insulator fault detection method based on 5G transmission and improved YOLOv3, and belongs to the field related to sensors.
Background
Unmanned aerial vehicle power inspection is a novel inspection mode for maintaining an overhead transmission line by using unmanned aerial vehicle autonomous operation. The unmanned aerial vehicle acquires high-definition aerial insulating sub-images through image acquisition equipment and a remote terminal control system which are carried by the unmanned aerial vehicle body, and performs fault analysis. Different from traditional manual inspection modes, unmanned aerial vehicle power inspection has the advantages of strong adaptability to complex terrains, high safety factor, small risk, low cost, accuracy, reliability and the like, and becomes one of key development directions of transmission line operation and maintenance technology. However, due to the complexity of the aerial images of the unmanned aerial vehicle and the limitation of the existing visual recognition algorithm technology, the precision of small fault detection is not high, the calculation complexity is high, the real-time performance is difficult to realize, and the application depth and the application range of the power patrol inspection on the unmanned aerial vehicle are limited.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: due to the complexity of aerial images of the unmanned aerial vehicle and the limitation of the existing visual recognition algorithm technology, the precision of small fault detection is not high, the calculation complexity is high, the real-time performance is difficult to realize, and the application depth and the application range of the power patrol inspection on the unmanned aerial vehicle are limited.
In order to solve the technical problem, the technical scheme of the invention is to provide an unmanned aerial vehicle transmission line insulator fault detection method based on 5G transmission and YOLOv3, which is characterized by comprising the following steps:
utilize unmanned aerial vehicle to carry image acquisition equipment to obtain high definition's transmission line insulator video image, adopt 5G unmanned aerial vehicle basic station technique with transmission line insulator video image data transmission to control center, control center will receive transmission line insulator video image input modified YOLOv3 cascade network structure, cascade network structure by modified YOLOv3 and detect the transmission line insulator, judge whether there is the trouble to output transmission line insulator, wherein:
the improved YOLOv3 cascade network structure comprises ResNet-50, wherein video images of the power transmission line insulators are input into ResNet-50 for processing, an insulator detection network module utilizes the output of ResNet-50 to construct feature pyramids with different scales, the feature pyramids with different scales are input into IoU prediction branches, the IoU prediction branches pass through IoU of a prediction target candidate bounding box and a real target label, and products of classification scores are used as detection confidence degrees to be used for NMS calculation.
Preferably, when the improved YOLOv3 cascade network structure is trained, an unmanned aerial vehicle acquires various fault high-definition mixed photos of insulators in a real environment, an insulation sub data set is constructed, a GAN generation type countermeasure network is used for preprocessing images in the insulation sub data set, defective insulation sub images are fused with various complex backgrounds, the defective insulation sub data set is expanded, and then the insulation sub image data set is sent into the improved YOLOv3 cascade network structure for training.
Preferably, the insulator detection network module is of a top-down network structure with lateral connection and is used for constructing feature maps with different sizes and high-level semantic information.
Preferably, the IoU predicted branch multiplies the predicted IoU by the classification probability and objective score to compute the final detection confidence, which is more relevant to localization accuracy, and then uses the final detection confidence as input for subsequent NMS.
Preferably, after the video image of the insulator of the power transmission line is obtained, the accurate positioning block diagram of the insulator of the power transmission line is obtained, then the accurate positioning block diagram of the insulator of the power transmission line is cut to obtain an effective cutting block diagram of the insulator, and the effective cutting block diagram of the insulator is input into the improved YOLOv3 cascade network structure.
The invention provides a transmission line insulator fault detection method based on 5G transmission and improved YOLOv3, which adopts an unmanned aerial vehicle image autonomous insulator fault detection method based on 5G network transmission and improved YOLOv3, fully exerts respective advantages of a 5G network and a YOLOv3 detection network, solves the problems of low small fault detection precision and insufficient signal fidelity and high transmission delay in the transmission process of video data of a transmission line patrol area of a transmission line in the traditional algorithm, and accurately positions the insulator fault appearing in an input image acquired from an actual detection environment.
Compared with the prior art, the invention has the beneficial effects that: the respective advantage of full play 5G network and YOLOv3 network can improve unmanned aerial vehicle and patrol and examine discernment processing speed and recognition efficiency, realizes that electric power patrols and examines the intellectuality, reduces electric power and patrols and examines personnel's work load, and the advantage is patrolled and examined to better performance unmanned aerial vehicle electric power.
Drawings
Fig. 1 is a flow chart of a transmission line insulator fault detection method based on improved YOLOv 3;
FIG. 2 is a modified YOLOv3 network structure;
fig. 3 is a structure diagram of a video transmission scheme of an unmanned aerial vehicle based on the 5G technology.
Detailed Description
The invention will be further illustrated with reference to the following specific examples. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Further, it should be understood that various changes or modifications of the present invention may be made by those skilled in the art after reading the teaching of the present invention, and such equivalents may fall within the scope of the present invention as defined in the appended claims.
As shown in fig. 1, in order to improve a flow chart of a method for detecting faults of insulators of a YOLOv3 transmission line, an unmanned aerial vehicle is used for collecting various high-definition mixed photos of insulators in a real environment to construct an insulation subdata set, a GAN generation countermeasure network is used for preprocessing images due to scarcity of the defect insulation subdata set, defect insulation subdata images are fused with various complex backgrounds to expand the defect insulation subdata set, then the insulator image data set is sent into an improved YOLOv3 detection network model for training, after model weights are obtained, the model weights are applied to actual detection of insulators of the transmission line, after an accurate positioning block diagram of the insulators is obtained, the insulators are effectively cut, image parameters are reduced while main features of the images are kept, and detection speed is improved. And sending the insulator fault classification to the improved YOLOv3 detection network again to obtain insulator fault classification, and automatically capturing the space coordinates of the fault insulator in the current actual detection environment by the ground control center and informing inspection personnel of removing the fault.
In fig. 1, a confidence mechanism is added to the YOLOv3 network structure by the detection module, so that the insulator in the line patrol area can be better positioned. And the second detection module is used for carrying out fault classification after cutting the positioned insulator image, so that redundant information in the image is greatly reduced, and the detection speed can be accelerated.
Inspired by a residual error network ResNet and a characteristic pyramid method, the method is supposed to provide a transmission line insulator fault detection method based on 5G transmission and improved YOLOv 3. The backbone network used in the method is different from the DarkNet-53 selected in the initial YOLOv3, and a more extensive and better optimized ResNet-50 is selected instead. An Insulator Detection Network (IDN) module is introduced to construct feature pyramids of different scales and IoU prediction branches to measure positioning accuracy.
In order to solve the problem of multi-scale change caused by size difference of objects, a top-down network structure with lateral connection is introduced to construct feature maps with different sizes and high-level semantic information, an IDN module is shown in FIG. 2, the network structure takes a feature pyramid as a basic structure, predicts the feature maps of each level respectively, can furthest reserve the semantic information of the feature maps in the process of up-sampling, fuses the feature maps with strong low-resolution semantic information and the feature maps with weak semantic information but rich spatial information with high-resolution semantic information on the premise of increasing less calculation amount, thereby obtaining the characteristic diagram with good spatial information and strong semantic information.
Bounding box regression is a key step in target detection. In YOLOv3, bounding box regression employed an L1 loss function. In the past models, the classification score is often used as the confidence of inaccurate target positioning, and based on the confidence, a large number of candidate target bounding boxes NMS (network management system) are discovered by more and more work, and the high classification score cannot ensure high positioning precision. Unlike YOLOv4, the L1 Loss is not directly replaced by the IoU Loss, but rather a branch is added to calculate the IoU Loss. IoU predict branches As shown in FIG. 2, the branch is calculated by predicting IoU the target candidate bounding box and the true target label, and based on the product of this classification score as the detection confidence for NMS computation. During the training process, IoU perceptual loss is adopted to train IoU predicted branches; in testing, the predicted IoU is multiplied by the classification probability and the objective score to calculate the final detection confidence, which is more relevant to localization accuracy, and then used as input for subsequent NMS. The positioning accuracy of the insulator fault can be greatly improved by introducing IoU prediction branches.
At present, 5G is used as a new generation mobile communication technology, and provides bandwidth of more than 10Gbit/s, millisecond-level time delay and ultrahigh-density connection by using a brand-new network architecture, so that new jump of network performance is realized. According to the method, based on scientific problems, the 5G base stations carried by multiple unmanned aerial vehicles are used for constructing a 5G core network when the method is creatively provided in a remote area, as shown in figure 3, the unmanned aerial vehicles acquire line inspection video data through the carried high-definition cameras, transmit the acquired video data to a ground control center through a 5G terminal module in real time, and provide effective video data for subsequent image detection and fault location. The characteristics of the 5G network such as large bandwidth, low time delay, high reliability and the like can effectively meet the requirements of high-definition video data transmission and remote real-time control of the unmanned aerial vehicle.
Claims (5)
1. The unmanned aerial vehicle transmission line insulator fault detection method based on 5G transmission and YOLOv3 is characterized by comprising the following steps:
utilize unmanned aerial vehicle to carry image acquisition equipment to obtain high definition's transmission line insulator video image, adopt 5G unmanned aerial vehicle basic station technique with transmission line insulator video image data transmission to control center, control center will receive transmission line insulator video image input modified YOLOv3 cascade network structure, cascade network structure by modified YOLOv3 and detect the transmission line insulator, judge whether there is the trouble to output transmission line insulator, wherein:
the improved YOLOv3 cascade network structure comprises ResNet-50, wherein video images of the power transmission line insulators are input into ResNet-50 for processing, an insulator detection network module utilizes the output of ResNet-50 to construct feature pyramids with different scales, the feature pyramids with different scales are input into IoU prediction branches, the IoU prediction branches pass through IoU of a prediction target candidate bounding box and a real target label, and products of classification scores are used as detection confidence degrees to be used for NMS calculation.
2. The method for detecting insulator faults of the electric transmission line of the unmanned aerial vehicle based on 5G transmission and YOLOv3 is characterized in that when the improved YOLOv3 cascade network structure is trained, the unmanned aerial vehicle acquires various fault high-definition mixed photos of insulators in a real environment, an insulator data set is constructed, a GAN generation type countermeasure network is used for preprocessing images in the insulator data set, a defect insulator sub-image is fused with various complex backgrounds, the defect insulator sub-image data set is expanded, and then the insulator sub-image data set is sent into the improved YOLOv3 cascade network structure for training.
3. The unmanned aerial vehicle transmission line insulator fault detection method based on 5G transmission and YOLOv3 of claim 1, wherein the insulator detection network module is a top-down network structure with lateral connections and is used for constructing feature maps with different sizes and high-level semantic information.
4. The unmanned aerial vehicle transmission line insulator fault detection method based on 5G transmission and YOLOv3 as claimed in claim 1, wherein the IoU prediction branch multiplies the predicted IoU by classification probability and objective score to calculate final detection confidence, the confidence has a greater correlation with positioning accuracy, and then the final detection confidence is used as an input of a subsequent NMS.
5. The unmanned aerial vehicle transmission line insulator fault detection method based on 5G transmission and YOLOv3, according to claim 1, is characterized in that after a transmission line insulator video image is obtained, a transmission line insulator accurate positioning block diagram is obtained, then the transmission line insulator accurate positioning block diagram is cut to obtain an insulator effective cutting block diagram, and the insulator effective cutting block diagram is input into the improved YOLOv3 cascade network structure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110002537.1A CN112614130A (en) | 2021-01-04 | 2021-01-04 | Unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and YOLOv3 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110002537.1A CN112614130A (en) | 2021-01-04 | 2021-01-04 | Unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and YOLOv3 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112614130A true CN112614130A (en) | 2021-04-06 |
Family
ID=75253375
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110002537.1A Pending CN112614130A (en) | 2021-01-04 | 2021-01-04 | Unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and YOLOv3 |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112614130A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113192071A (en) * | 2021-06-29 | 2021-07-30 | 南方电网数字电网研究院有限公司 | Insulator binding wire missing identification method and device and computer equipment |
| CN113297996A (en) * | 2021-05-31 | 2021-08-24 | 贵州电网有限责任公司 | Unmanned aerial vehicle aerial photographing insulator target detection method based on YoloV3 |
| CN113536938A (en) * | 2021-06-18 | 2021-10-22 | 云南电网有限责任公司 | 5G-fused intelligent early warning method and system for forest fire of power transmission line |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110033453A (en) * | 2019-04-18 | 2019-07-19 | 国网山西省电力公司电力科学研究院 | Based on the power transmission and transformation line insulator Aerial Images fault detection method for improving YOLOv3 |
-
2021
- 2021-01-04 CN CN202110002537.1A patent/CN112614130A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110033453A (en) * | 2019-04-18 | 2019-07-19 | 国网山西省电力公司电力科学研究院 | Based on the power transmission and transformation line insulator Aerial Images fault detection method for improving YOLOv3 |
Non-Patent Citations (4)
| Title |
|---|
| XIANG LONG ET AL.: "PP-YOLO: An Effective and Efficient Implementation of Object Detector", 《ARXIV:2007.12099V3》 * |
| 唐小煜等: "基于U－net和YOLOv4的绝缘子图像分割与缺陷检测", 《华南师范大学学报(自然科学版)》 * |
| 王海文等: "一种基于生成式对抗网络的图像数据扩充方法", 《计算机技术与发展》 * |
| 赵宏大等: "5G通信技术在泛在电力物联网的应用", 《南方电网技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113297996A (en) * | 2021-05-31 | 2021-08-24 | 贵州电网有限责任公司 | Unmanned aerial vehicle aerial photographing insulator target detection method based on YoloV3 |
| CN113536938A (en) * | 2021-06-18 | 2021-10-22 | 云南电网有限责任公司 | 5G-fused intelligent early warning method and system for forest fire of power transmission line |
| CN113192071A (en) * | 2021-06-29 | 2021-07-30 | 南方电网数字电网研究院有限公司 | Insulator binding wire missing identification method and device and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112614130A (en) | Unmanned aerial vehicle power transmission line insulator fault detection method based on 5G transmission and YOLOv3 | |
| CN110033453A (en) | Based on the power transmission and transformation line insulator Aerial Images fault detection method for improving YOLOv3 | |
| CN113205039B (en) | Power equipment fault image recognition disaster investigation system and method based on multiple DCNN networks | |
| CN111431986A (en) | Industrial intelligent quality inspection system based on 5G and AI cloud edge cooperation | |
| WO2023216721A1 (en) | Concrete dam defect time sequence image intelligent identification method | |
| CN112528979B (en) | Transformer substation inspection robot obstacle distinguishing method and system | |
| CN110910440B (en) | Power transmission line length determination method and system based on power image data | |
| CN113589837A (en) | Electric power real-time inspection method based on edge cloud | |
| CN112906654A (en) | Anti-vibration hammer detection method based on deep learning algorithm | |
| CN113252701B (en) | Cloud edge cooperation-based power transmission line insulator self-explosion defect detection system and method | |
| CN113099242B (en) | Power transmission line video monitoring data processing method and system | |
| CN112906620A (en) | Unmanned aerial vehicle-assisted insulator fault detection method and device and electronic equipment | |
| CN115909092A (en) | Light-weight power transmission channel hidden danger distance measuring method and hidden danger early warning device | |
| CN114863118A (en) | Self-learning identification system and method based on external hidden danger of power transmission line | |
| CN115082813A (en) | Detection method, unmanned aerial vehicle, detection system and medium | |
| CN114119528A (en) | Defect detection method and device for distribution line insulator | |
| CN113962973A (en) | Power transmission line unmanned aerial vehicle intelligent inspection system and method based on satellite technology | |
| CN112734732A (en) | Railway tunnel leaky cable clamp detection method based on improved SSD algorithm | |
| CN115457411B (en) | Unmanned inspection method and device for aviation oil pipeline and aviation oil safety operation and maintenance system | |
| CN116866512A (en) | Photovoltaic power station inspection system and operation method thereof | |
| CN115765848A (en) | 5G network signal strength detection system and use method | |
| CN115147591A (en) | Transformer equipment infrared image voltage heating type defect diagnosis method and system | |
| Sheng et al. | A YOLOX-Based Detection Method of Triple-Cascade Feature Level Fusion for Power System External Defects | |
| CN116909320B (en) | Electric power collaborative inspection strategy analysis method based on ant colony algorithm | |
| CN112948459B (en) | Railway infrastructure inspection device, method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210406 |
|
| WD01 | Invention patent application deemed withdrawn after publication |