CN113239838A - Unmanned aerial vehicle intelligent inspection identification method for power transmission tower - Google Patents
Unmanned aerial vehicle intelligent inspection identification method for power transmission tower Download PDFInfo
- Publication number
- CN113239838A CN113239838A CN202110562089.0A CN202110562089A CN113239838A CN 113239838 A CN113239838 A CN 113239838A CN 202110562089 A CN202110562089 A CN 202110562089A CN 113239838 A CN113239838 A CN 113239838A
- Authority
- CN
- China
- Prior art keywords
- module
- image
- insulator
- feature
- aerial vehicle
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000007689 inspection Methods 0.000 title claims abstract description 22
- 230000005540 biological transmission Effects 0.000 title claims abstract description 19
- 239000012212 insulator Substances 0.000 claims abstract description 32
- 238000000605 extraction Methods 0.000 claims abstract description 21
- 230000004927 fusion Effects 0.000 claims abstract description 20
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 230000001537 neural effect Effects 0.000 claims abstract 2
- 238000001514 detection method Methods 0.000 claims description 11
- 238000010586 diagram Methods 0.000 claims description 7
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 230000003993 interaction Effects 0.000 claims description 2
- 238000011176 pooling Methods 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 abstract description 3
- 230000007547 defect Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 230000009467 reduction Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- RZVHIXYEVGDQDX-UHFFFAOYSA-N 9,10-anthraquinone Chemical compound C1=CC=C2C(=O)C3=CC=CC=C3C(=O)C2=C1 RZVHIXYEVGDQDX-UHFFFAOYSA-N 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention relates to an intelligent inspection identification method of an unmanned aerial vehicle for a power transmission tower, which is used for inspecting an insulator by adopting an intelligent inspection system, wherein the intelligent inspection system comprises an image acquisition module, a preprocessing module, a feature extraction module, a feature fusion module, a prediction judgment module and a fault reporting module which are carried by the unmanned aerial vehicle; the identification method specifically comprises the following steps: 1) shooting an image of the insulator and transmitting the image to a preprocessing module; 2) carrying out non-local average denoising processing on the image and inputting the image into a feature extraction module; 3) extracting shape characteristics and fault characteristics of the insulator by adopting a neural characteristic extraction network; 4) performing feature fusion on the extracted high-dimensional features; 5) judging the characteristics through a prediction regression network, and judging whether the shot insulator has a fault or not; the method and the device quickly and accurately realize real-time identification of the insulator fault of the transmission tower and improve the maintenance efficiency of the insulator.
Description
Technical Field
The invention relates to the technical field of power transmission tower fault inspection and maintenance, in particular to an unmanned aerial vehicle intelligent inspection and identification method for a power transmission tower.
Background
The modern times of artificial intelligence and big data are entered, and deep learning and neural networks are rapidly developed. The power grid occupies a very important position in China, the maintenance of the power grid is important, and the intelligent power grid is built and continuously paid attention. An overhead transmission line is a steel frame structure facility for erecting a lead to keep the lead away from the ground. The insulator plays the effect of connecting wire and gold utensil at overhead transmission line, and whether the insulator has trouble hidden danger, has very big influence to the steady operation of electric wire netting, will lead to the insulator fracture seriously, and wire ground connection arouses huge economic loss and casualties, consequently appears very important early the trouble of insulator to, and the insulator is huge, and the tradition is wasted time and energy through the mode of artifical patrolling and examining, and can not be fine judgement insulator has trouble or defect.
Disclosure of Invention
The invention aims to solve the problems that the traditional manual inspection mode wastes time and labor and can not well judge whether the insulator has faults or defects, and provides an unmanned aerial vehicle intelligent inspection identification method for a power transmission tower.
The specific scheme of the invention is as follows: an unmanned aerial vehicle image intelligent inspection system comprises an image acquisition module, an image preprocessing module, a feature extraction module, a feature fusion module, a prediction judgment module and a fault reporting module.
The image acquisition module adopt unmanned aerial vehicle to carry SHD10T3 cloud platform of making a video recording and shoot the insulator visible light image, the image database is unified to be stored to the image of shooing, the follow-up module of being convenient for is handled.
The image preprocessing module reduces Gaussian noise generated in the image transmission process by adopting a fast non-local mean denoising method, and the idea is to denoise by using a redundant part in an image, wherein the redundant part refers to a repeated or similar neighborhood structure in the image; the time complexity is relatively high, which results in overlong calculation time, so that the integral image is accelerated, and the Euclidean distance replaces the Gaussian weighted Euclidean distance, thereby obtaining the rapid filtering and noise reduction.
The characteristic extraction module is based on a YOLO-V4tiny network structure, an improved network is provided, a Resnet _ Y module is used for replacing an original CSPblock network, the Resnet _ Y network is composed of two branches, one input characteristic is firstly subjected to 1 multiplied by 1 convolution to adjust the number of channels, then a 3 multiplied by 3 convolution module is connected for characteristic extraction, the other branch is used for carrying out maximum pooling processing on the input characteristic, input layer down sampling is carried out, redundant information is removed, the characteristic is compressed, and the receptive field is expanded; and then connecting a 1 x 1 convolution module, integrating channel information, performing cross-channel information interaction, enhancing the nonlinear characteristic of the network, and finally adding and outputting the characteristic information of the two branches. An auxiliary feature network is added in the main feature network, the detection precision is improved, a channel attention and space attention algorithm is added, and the channel attention algorithm enhances the network learning global information to selectively strengthen the features containing useful information and inhibit useless features; the spatial attention algorithm is used for obtaining weight by positioning a target, enhancing feature extraction on an interested target region in an image, finally obtaining a plurality of high-dimensional feature maps, and changing the size of an output feature layer to enable the feature to adapt to the shape feature of the insulator. MishReLU is used for replacing the original LeakyReLU activation function, so that the model has good generalization capability, and the quality is optimized.
The feature fusion module uses a BiFPN feature fusion structure, and on the FPN feature fusion structure, connection among cross units is increased, so that superior features are linked with fused features, the upper feature diagram and the lower feature diagram have closer relevance, and the feature fusion result is more excellent.
The prediction judgment module removes the detection head with the maximum prediction resolution by using a YOLO prediction structure, and improves the detection rate by using two detection heads; and a prediction structure with large-size resolution is increased, and the prior frame size suitable for the insulator is selected by using k-mean clustering, so that the judgment efficiency is improved.
And the fault reporting module uploads the acquired fault image and fault information to the client for alarming, dials up the network by using the SIM7600G-H module, and sends the fault information and the fault image.
The invention has the following beneficial effects: 1. the image data set transmitted by an unmanned aerial vehicle image acquisition system is adopted, the image data set is preprocessed, and the Gaussian noise of the image is reduced by using a rapid non-local average denoising method to help the feature extraction of the image; the image identification module adopts end-to-end mode identification, and improves a feature extraction network based on a YOLO-V4tiny network structure, so that the method is faster and more accurate, and can realize real-time identification of the insulator fault of the transmission tower;
2. the image recognition system can be used for defect recognition, and a lightweight feature extraction network is designed, so that edge rapid calculation can be realized;
3. after the defect characteristics are extracted, fault information is identified, a 4G network is used for transmitting a fault image and an alarm signal back to the information management system, the fault image is confirmed through alarm information, a maintainer can be helped to arrive at the site in time, insulators are replaced in time, and large-area power failure of a power grid is prevented.
Drawings
FIG. 1 is a schematic diagram of the intelligent inspection system of the present invention;
FIG. 2 is a schematic diagram of a feature extraction network architecture of the present invention;
FIG. 3 is a schematic diagram of a feature fusion network architecture of the present invention;
fig. 4 is an insulator image taken with the present invention for inspection with defects.
Detailed Description
The unmanned aerial vehicle image intelligent inspection system comprises an image acquisition module, an image preprocessing module, a feature extraction module, a feature fusion module, a prediction judgment module and a fault reporting module.
Referring to fig. 1, in the image acquisition module of this embodiment, an unmanned aerial vehicle is used to carry an SHD10T3 camera pan-tilt, the pan-tilt can acquire an optical zoom visible light image and an infrared image, the pan-tilt captures an insulator visible light image, the insulator visible light image is acquired every 3 seconds, and an HDMI real-time signal is simultaneously output, and the captured images are uniformly stored in an image database in a TF card, so as to facilitate subsequent module processing; one end of the HDMI real-time signal is connected with the sky end of the Dajiang flight control, the video stream signal is transmitted back to the flight controller to display a real-time picture, and the other end of the HDMI real-time signal is connected with the data conversion module and is input into the detection module as a video.
The image preprocessing module of the embodiment reduces the gaussian noise generated in the image transmission process by adopting a fast non-local mean denoising method, and the idea is to denoise by using a redundant part in an image, wherein the redundant part refers to a repeated or similar neighborhood structure in the image, the similarity degree between neighborhoods is firstly measured by Euclidean distance, then the similar neighborhoods are given with larger weight by weight distribution, and finally, the image after denoising is obtained by weighted averaging. The time complexity is relatively high, the calculation time is too long, the integral image is accelerated, and the Euclidean distance replaces the Gaussian weighted Euclidean distance to obtain the rapid filtering noise reduction.
Referring to fig. 2, the feature extraction module described in this embodiment is based on a YOLO-V4tiny network structure, and provides an improved network, and performs downsampling on an input image twice, replaces an original CSPblock network with a Resnet _ Y module, increases a detection rate, adds an auxiliary feature network, adds a channel attention and spatial attention algorithm, enhances feature extraction on an interested target area in the image, so as to compensate for accuracy reduction caused by speed increase, finally obtains a plurality of high-dimensional feature maps, and changes the size of an output feature layer to adapt to the shape feature of an insulator. MishReLU is used for replacing the original LeakyReLU activation function, so that the model has good generalization capability, and the quality is optimized.
Referring to fig. 3, the feature fusion module described in this embodiment uses a BiFPN feature fusion structure, adds a connection edge between units, spans the connection between the units, and associates a superior feature with a fused feature, so that the upper and lower feature maps have closer association, and the result of feature fusion is more excellent; compared with an FPN characteristic pyramid structure, downsampling is added, so that the connection between characteristics with different scales is more, and two detection branches are output finally.
Referring to fig. 4, the prediction determining module of the present embodiment includes the following steps:
adjusting the number of channels after feature fusion by using a convolution block to be Bx (C +5), wherein B is the number of predicted prior frames of each grid, and C is the number of categories;
generating all prior frames on the whole feature map by using the set prior frames, wherein the number of the prior frames in the algorithm is 6; dividing all the prior frames into positive, negative and neglected samples according to the IOU of the real frame and the distributed network;
random sampling is carried out, and a part of prior frames are selected for loss calculation; the real frame is coded into the same form of network output, so that the loss is convenient to calculate; and calculating the classification, the confidence coefficient, the position of the matrix box and the loss of the width and the height, and weighting and summing the final output for calculating the gradient and back propagation.
And finally, obtaining a detection result, and detecting the self-explosion fault of the insulator and the umbrella skirt of the insulator.
In this embodiment, the fault reporting module uploads the acquired fault image and fault information to the database, and the client reads data from the database to alarm, and uses the SIM7600G-H module to dial up to the internet to send the fault information and the fault image.
Claims (6)
1. An unmanned aerial vehicle intelligent inspection identification method for a power transmission tower is characterized by comprising the following steps: the insulator inspection method comprises the following steps that an intelligent inspection system is adopted to inspect the insulator, wherein the intelligent inspection system comprises an image acquisition module, a preprocessing module, a feature extraction module, a feature fusion module, a prediction judgment module and a fault reporting module which are carried by an unmanned aerial vehicle; the identification method specifically comprises the following steps: 1) an image acquisition module carried by the unmanned aerial vehicle shoots an image of the insulator and transmits the image to the preprocessing module; 2) the preprocessing module carries out non-local average denoising processing on the image and inputs the image into the feature extraction module; 3) the characteristic extraction module adopts a neural characteristic extraction network to extract the shape characteristic and the fault characteristic of the insulator; 4) the feature fusion module performs feature fusion on the extracted high-dimensional features; 5) the prediction judging module judges the characteristics through a prediction regression network and judges whether the shot insulator has a fault or not; 6) and the fault reporting module uploads the insulator image judged as the fault and the fault information to the client and gives an alarm.
2. The intelligent unmanned aerial vehicle inspection identification method for the transmission tower according to claim 1, wherein the method comprises the following steps: the image preprocessing module in the step 2) reduces Gaussian noise generated in the image transmission process by adopting a fast non-local mean denoising method, denoising is carried out by utilizing a repeated or similar neighborhood structure in the image, the similarity degree between neighborhoods is measured through Euclidean distance, then a weight value is distributed to give a larger weight to the similar neighborhoods, and finally weighted averaging is carried out to obtain the denoised image.
3. The intelligent unmanned aerial vehicle inspection identification method for the transmission tower according to claim 1, wherein the method comprises the following steps: the feature extraction module in the step 3) adopts a YOLO-V4tiny network structure as a basis, a Resnet _ Y network is used for replacing a CSPblock network in the YOLO-V4tiny network structure, the Resnet _ Y network consists of two branches, one input feature is firstly subjected to 1 × 1 convolution to adjust the number of channels, and then a 3 × 3 convolution module is connected for feature extraction; the other branch performs maximum pooling on the input features, performs layer down-sampling on the input features, removes redundant information, compresses the features and enlarges the receptive field; and then connecting a 1 x 1 convolution module, integrating channel information, performing cross-channel information interaction, enhancing the nonlinear characteristic of the network, and finally adding and outputting the characteristic information of the two branches.
4. The intelligent unmanned aerial vehicle inspection identification method for the transmission tower according to claim 3, wherein the method comprises the following steps: a channel attention and space attention algorithm is added into the YOLO-V4tiny network structure, and the channel attention selectively strengthens the characteristics containing useful information and inhibits useless characteristics; and (3) performing feature extraction on an interested target region in the image by a spatial attention algorithm, finally obtaining a plurality of high-dimensional feature maps, and changing the size of an output feature layer to adapt to the shape feature of the insulator.
5. The intelligent unmanned aerial vehicle inspection identification method for the transmission tower according to claim 1, wherein the method comprises the following steps: the feature fusion module in the step 4) uses a BiFPN feature fusion structure, and the connection between cross units is added on the FPN feature fusion structure to enable the superior features to be linked with the fused features, so that the upper feature diagram and the lower feature diagram have closer relevance, and the feature fusion result is more excellent.
6. The intelligent unmanned aerial vehicle inspection identification method for the transmission tower according to claim 1, wherein the method comprises the following steps: the prediction judging module in the step 5) uses a YOLO prediction structure, removes the detection head with the maximum prediction resolution and uses two detection heads to improve the detection rate; and increasing a prediction structure with large-size resolution, and selecting the prior frame size suitable for the insulator by using k-mean clustering to improve the judgment efficiency.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110562089.0A CN113239838B (en) | 2021-05-24 | 2021-05-24 | Unmanned aerial vehicle intelligent inspection identification method for power transmission tower |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110562089.0A CN113239838B (en) | 2021-05-24 | 2021-05-24 | Unmanned aerial vehicle intelligent inspection identification method for power transmission tower |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113239838A true CN113239838A (en) | 2021-08-10 |
| CN113239838B CN113239838B (en) | 2022-09-20 |
Family
ID=77138377
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110562089.0A Active CN113239838B (en) | 2021-05-24 | 2021-05-24 | Unmanned aerial vehicle intelligent inspection identification method for power transmission tower |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113239838B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114332697A (en) * | 2021-12-19 | 2022-04-12 | 西安科技大学 | Method, system, equipment and medium for detecting faults of multiple types of targets in power transmission line |
| CN116721095A (en) * | 2023-08-04 | 2023-09-08 | 杭州瑞琦信息技术有限公司 | Aerial photographing road illumination fault detection method and device |
Citations (5)
| 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 |
| CN110636715A (en) * | 2019-08-27 | 2019-12-31 | 杭州电子科技大学 | Self-learning-based automatic welding and defect detection method |
| CN110674866A (en) * | 2019-09-23 | 2020-01-10 | 兰州理工大学 | Method for detecting X-ray breast lesion images by using transfer learning characteristic pyramid network |
| CN112183667A (en) * | 2020-10-31 | 2021-01-05 | 哈尔滨理工大学 | Insulator fault detection method in cooperation with deep learning |
| CN112464910A (en) * | 2020-12-18 | 2021-03-09 | 杭州电子科技大学 | Traffic sign identification method based on YOLO v4-tiny |
-
2021
- 2021-05-24 CN CN202110562089.0A patent/CN113239838B/en active Active
Patent Citations (5)
| 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 |
| CN110636715A (en) * | 2019-08-27 | 2019-12-31 | 杭州电子科技大学 | Self-learning-based automatic welding and defect detection method |
| CN110674866A (en) * | 2019-09-23 | 2020-01-10 | 兰州理工大学 | Method for detecting X-ray breast lesion images by using transfer learning characteristic pyramid network |
| CN112183667A (en) * | 2020-10-31 | 2021-01-05 | 哈尔滨理工大学 | Insulator fault detection method in cooperation with deep learning |
| CN112464910A (en) * | 2020-12-18 | 2021-03-09 | 杭州电子科技大学 | Traffic sign identification method based on YOLO v4-tiny |
Non-Patent Citations (2)
| Title |
|---|
| ZICONG JIANG，ETC.: ""Real-time object detection method for embedded devices"", 《COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 赵庆平，等: ""一种改进权重的非局部均值图像去噪算法"", 《电子测量与仪器学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114332697A (en) * | 2021-12-19 | 2022-04-12 | 西安科技大学 | Method, system, equipment and medium for detecting faults of multiple types of targets in power transmission line |
| CN116721095A (en) * | 2023-08-04 | 2023-09-08 | 杭州瑞琦信息技术有限公司 | Aerial photographing road illumination fault detection method and device |
| CN116721095B (en) * | 2023-08-04 | 2023-11-03 | 杭州瑞琦信息技术有限公司 | Aerial photographing road illumination fault detection method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113239838B (en) | 2022-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108109385B (en) | System and method for identifying and judging dangerous behaviors of power transmission line anti-external damage vehicle | |
| CN112381784B (en) | Equipment detecting system based on multispectral image | |
| CN106356757B (en) | A kind of power circuit unmanned plane method for inspecting based on human-eye visual characteristic | |
| CN113239838B (en) | Unmanned aerial vehicle intelligent inspection identification method for power transmission tower | |
| CN111986240A (en) | Drowning person detection method and system based on visible light and thermal imaging data fusion | |
| CN113963315A (en) | Real-time video multi-user behavior recognition method and system in complex scene | |
| CN112183313A (en) | SlowFast-based power operation field action identification method | |
| CN112083437A (en) | Marine laser radar and video combined target capturing system and method | |
| CN103729620B (en) | A kind of multi-view pedestrian detection method based on multi-view Bayesian network | |
| CN103090946B (en) | Method and system for measuring single fruit tree yield | |
| CN116823800A (en) | Bridge concrete crack detection method based on deep learning under complex background | |
| CN116681979A (en) | Power equipment target detection method under complex environment | |
| CN114973032A (en) | Photovoltaic panel hot spot detection method and device based on deep convolutional neural network | |
| CN111582069B (en) | Track obstacle zero sample classification method and device for air-based monitoring platform | |
| CN113902698A (en) | Unmanned aerial vehicle holder progressive target focusing method based on intelligent visual control | |
| Wang et al. | Improvements of YoloV3 for road damage detection | |
| Yuan et al. | Identification method of typical defects in transmission lines based on YOLOv5 object detection algorithm | |
| CN116740419A (en) | Target detection method based on graph regulation network | |
| CN109359545B (en) | Cooperative monitoring method and device under complex low-altitude environment | |
| CN114359167A (en) | Insulator defect detection method based on lightweight YOLOv4 in complex scene | |
| CN111915565B (en) | Method for analyzing cracks of porcelain insulator of power transmission and transformation line in real time based on YOLACT algorithm | |
| CN111354028B (en) | Binocular vision-based power transmission channel hidden danger identification and tracking method | |
| CN116994161A (en) | Insulator defect detection method based on improved YOLOv5 | |
| CN114549432A (en) | Intelligent unmanned aerial vehicle image identification method | |
| CN113034598B (en) | Unmanned aerial vehicle power line inspection method based on deep learning |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |