CN113392701A - YN-Net convolution neural network-based power transmission line obstacle detection method - Google Patents
YN-Net convolution neural network-based power transmission line obstacle detection method Download PDFInfo
- Publication number
- CN113392701A CN113392701A CN202110504527.8A CN202110504527A CN113392701A CN 113392701 A CN113392701 A CN 113392701A CN 202110504527 A CN202110504527 A CN 202110504527A CN 113392701 A CN113392701 A CN 113392701A
- Authority
- CN
- China
- Prior art keywords
- neural network
- transmission line
- convolutional neural
- power transmission
- image
- 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
- 230000005540 biological transmission Effects 0.000 title claims abstract description 68
- 238000013528 artificial neural network Methods 0.000 title claims description 5
- 238000001514 detection method Methods 0.000 title abstract description 10
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 51
- 238000000034 method Methods 0.000 claims abstract description 31
- 238000007689 inspection Methods 0.000 claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 230000004888 barrier function Effects 0.000 claims abstract description 4
- 238000000605 extraction Methods 0.000 claims description 12
- 230000009467 reduction Effects 0.000 claims description 9
- 230000003416 augmentation Effects 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims 1
- 238000011161 development Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 230000010354 integration Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a power transmission line obstacle detection method based on an YN-Net convolutional neural network, which comprises the following steps of: acquiring a barrier image of the power transmission line based on a ZED binocular camera carried by the power transmission line inspection robot; preprocessing the acquired images of the obstacles of the power transmission line and making the images into a data set; building an YN-Net convolutional neural network, and fully training the YN-Net convolutional neural network to obtain a trained YN-Net convolutional neural network model; and classifying the acquired real-time images of the obstacles of the power transmission line by using an YN-Net convolutional neural network model, and judging whether the obstacles exist in front of the power transmission line inspection robot or not according to the probability value of the obstacles existing in the input images and the obstacles not existing in the input images. The method can improve the real-time performance and reliability of the convolutional neural network, solves the problems of low accuracy, poor real-time performance and the like of the existing power transmission line obstacle detection algorithm, has high transportability, can be applied to an embedded development platform, and has wide application prospect.
Description
Technical Field
The invention relates to the technical field of binocular camera image acquisition and deep learning, in particular to a power transmission line obstacle detection method based on an YN-Net convolutional neural network.
Background
Along with the enlargement of urban construction scale, the urban and rural integration speed is accelerated, the requirement of users on the reliability of the power grid is higher and higher, and the operation and inspection work of the power transmission line is more and more important for the normal operation of the power grid. However, the number of professional teams engaged in power transmission line inspection work is basically unchanged, and the contradiction between the increasing expansion of the power grid scale and the structural shortage of operators is increasingly prominent. The traditional manual operation and inspection mode is difficult to meet the operation and maintenance requirements of modern power grids, and a novel intelligent device capable of replacing manpower is urgently needed. Under the background that the robot and the artificial intelligence technology are mature gradually, the research and the application of the power transmission line inspection robot are important ways for changing the operating mode of line operation inspection and improving the working quality. The key and difficulty of the power transmission line inspection robot research mainly focuses on the detection task of the power transmission line obstacle.
The existing detection algorithm for the obstacles of the power transmission line has the problems of low accuracy, poor real-time performance and the like, and a new technical scheme is needed to solve the problems.
Disclosure of Invention
The purpose of the invention is as follows: in order to overcome the defects in the prior art, the method for detecting the obstacles of the power transmission line based on the YN-Net convolutional neural network is provided, the accuracy and the real-time performance of the detection algorithm of the obstacles of the power transmission line are improved, and various obstacles of the power transmission line can be reliably detected.
The technical scheme is as follows: in order to achieve the aim, the invention provides a method for detecting obstacles of a power transmission line based on an YN-Net convolutional neural network, which comprises the following steps:
s1: acquiring a barrier image of the power transmission line based on a ZED binocular camera carried by the power transmission line inspection robot;
s2: preprocessing the acquired images of the obstacles of the power transmission line and making the images into a data set;
s3: building an YN-Net convolutional neural network, and fully training the YN-Net convolutional neural network by using the data set of the step S2 to obtain a trained YN-Net convolutional neural network model;
s4: and classifying the acquired real-time images of the transmission line obstacles by using the trained YN-Net convolutional neural network model, and judging whether the obstacles exist in front of the transmission line inspection robot or not according to the probability value of the existence and nonexistence of the obstacles in the input images.
Further, in step S2, the collected images of the obstacles in the power transmission line are preprocessed by image screening and image augmentation.
Further, the data set obtaining method in step S2 is as follows: and according to whether the obstacles exist, performing labeling operation on the preprocessed images of the obstacles of the power transmission line to obtain a labeled data set.
Further, the preprocessing operation in step S2 includes:
a1: screening the acquired image set, and deleting the images with fuzzy shooting;
a2: image augmentation is performed on the acquired image set, and the operation comprises the following steps: horizontally turning, vertically turning, adding Gaussian noise, and adjusting the saturation, contrast and brightness of the image;
a3: the method comprises the steps of dividing a sensitive area of an acquired image set, dividing the sensitive area by taking the middle point of the image as the center, wherein the longitudinal length of the sensitive area is the same as that of the image, the transverse length of the sensitive area is one third of that of the image, cutting the image in the image set according to the sensitive area, and only keeping the image in the sensitive area.
Further, the method for building the YN-Net convolutional neural network in step S3 includes:
b1: designing a novel convolution structure unit named as Reduce-Group convolution, wherein the Reduce-Group convolution firstly uses 1 × 1 standard convolution to Reduce the dimension of an input feature map, then uses 2 different convolution channels to extract the features of the feature map after dimension reduction, uses channel 1 to perform 3 × 3 grouped convolution feature extraction once on the feature map after dimension reduction, uses channel 2 to perform 3 × 3 grouped convolution feature extraction twice on the feature map after dimension reduction, and finally superposes the results of the two channels and uses 1 × 1 grouped convolution to perform feature extraction to obtain a final output feature map;
b2: and constructing an YN-Net convolution neural network based on the Reduce-Group convolution.
Further, the method for building the YN-Net convolutional neural network based on Reduce-Group convolution in the step B2 includes:
firstly, the size of an input image is adjusted to 256 × 384 × 3 (pixels), a layer of 3 × 3 standard convolution layer is used for carrying out feature extraction on the adjusted input image to obtain a preliminary feature map, then, five layers of 3 × 3 Reduce-Group convolution layers are used for carrying out further feature extraction on the preliminary feature map to obtain a depth feature map, feature fusion is carried out on the depth feature map by two layers of 1 × 1 standard convolution layers to obtain a final feature map of an YN-Net convolutional neural network, the final feature map is processed by using global mean pooling to obtain an output vector out with the length of 2, and the composition of the out is shown in formula 1, wherein a and b respectively represent probability values of the existence of obstacles and the nonexistence of obstacles in the current input image.
out=[a,b] (1)
Further, the method for training the YN-Net convolutional neural network model in step S3 includes:
and dividing the Obs-R data set into a training set and a testing set according to the ratio of 4:1, and training the built YN-Net convolutional neural network model, wherein the loss function uses a cross entropy loss function, and the network optimizer adopts a adam optimizer to carry out 300 times of iterative training on the network, so that the trained YN-Net convolutional neural network model is obtained.
Further, the transmission line obstacle includes 6 kinds: the inspection robot comprises a vibration damper, a wire clamp, a bridge fitting, a plastic bag, a kite and a balloon, the transmission line inspection robot is enabled to walk on an acquisition line in a reciprocating mode on sunny days, cloudy days and rainy days, a ZED binocular camera carried by the transmission line inspection robot acquires images of 6 obstacles respectively, visible light images acquired by a lens on the left side in the ZED binocular camera are obtained and stored until stored images of the obstacles of each transmission line are larger than 2000, images without the obstacles are larger than 4000, and an image set is obtained through integration.
Has the advantages that: compared with the prior art, the method has the advantages that the distribution characteristics of the obstacles of the power transmission line are fully considered, the YN-Net convolutional neural network is designed, the interference of a complex background in the image is avoided based on the image sensitive area, the Reduce-Group novel convolutional structure is designed, the real-time performance and the reliability of the convolutional neural network can be effectively improved, the problems of low accuracy, poor real-time performance and the like of the existing obstacle detection algorithm of the power transmission line are solved, the transportability is high, the method can be applied to an embedded development platform, and the application prospect is wide.
Drawings
Fig. 1 is a schematic working flow diagram of a method for detecting obstacles in a power transmission line based on an YN-Net convolutional neural network according to an embodiment of the present invention;
fig. 2 is a diagram of a transmission line inspection robot in a real object provided by the embodiment of the invention;
fig. 3 is a data set image acquisition field diagram of the power transmission line inspection robot provided by the embodiment of the invention;
fig. 4 is an image of a power transmission line obstacle cut based on a sensitive area according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating a Reduce-Group convolution structure according to an embodiment of the present invention;
FIG. 6 is a block diagram of a YN-Net convolutional neural network according to an embodiment of the present invention;
fig. 7 is a diagram illustrating a classification result of obstacles on the transmission line according to an embodiment of the present invention;
FIG. 8 is a graph of a set of actual test results provided by an example of the present invention.
Detailed Description
The present invention is further illustrated by the following figures and specific examples, which are to be understood as illustrative only and not as limiting the scope of the invention, which is to be given the full breadth of the appended claims and any and all equivalent modifications thereof which may occur to those skilled in the art upon reading the present specification.
The invention provides a method for detecting obstacles of a power transmission line based on an YN-Net convolutional neural network, which comprises the following steps as shown in figure 1:
s1: acquiring a barrier image of the power transmission line based on a ZED binocular camera carried by the power transmission line inspection robot;
s2: preprocessing the acquired images of the obstacles of the power transmission line and making the images into a data set;
s3: building an YN-Net convolutional neural network, and fully training the YN-Net convolutional neural network by using the data set of the step S2 to obtain a trained YN-Net convolutional neural network model;
s4: and classifying the acquired real-time images of the transmission line obstacles by using the trained YN-Net convolutional neural network model, and judging whether the obstacles exist in front of the transmission line inspection robot or not according to the probability value of the existence and nonexistence of the obstacles in the input images.
The physical diagram of the power transmission line inspection robot in step S1 of this embodiment is specifically shown in fig. 2.
As shown in fig. 3, the data set image acquisition site of the power transmission line inspection robot includes 6 types of obstacles in the power transmission line in step S1: the inspection robot comprises a vibration damper, a wire clamp, a bridge fitting, a plastic bag, a kite and a balloon, the transmission line inspection robot is enabled to walk on an acquisition line in a reciprocating mode on sunny days, cloudy days and rainy days, a ZED binocular camera carried by the transmission line inspection robot acquires images of 6 obstacles respectively, visible light images acquired by a lens on the left side in the ZED binocular camera are obtained and stored until stored images of the obstacles of each transmission line are larger than 2000, images without the obstacles are larger than 4000, and an image set is obtained through integration.
In step S2, the collected images of the obstacles in the power transmission line are preprocessed by image screening and image augmentation. The pretreatment operation process comprises the following steps:
a1: screening the acquired image set, and deleting the images with fuzzy shooting;
a2: image augmentation is performed on the acquired image set, and the operation comprises the following steps: horizontally turning, vertically turning, adding Gaussian noise, and adjusting the saturation, contrast and brightness of the image;
a3: the method includes the steps of dividing a sensitive area of an acquired image set, dividing the sensitive area by taking the middle point of the image as the center, enabling the longitudinal length of the sensitive area to be the same as that of the image, enabling the transverse length to be one third of that of the image, cutting the image in the image set according to the sensitive area, only keeping the image in the sensitive area, and enabling the image to be a power transmission line obstacle image cut according to the sensitive area as shown in fig. 4.
The data set acquiring method in step S2 of this embodiment is as follows: and according to whether an obstacle exists or not, performing labeling operation on the preprocessed image of the obstacle of the power transmission line, wherein the labeling operation is to label the image according to whether the obstacle exists or not in the image, if the obstacle exists, the image is labeled as 1, if the obstacle does not exist, the image is labeled as 0, and a labeled data set is obtained and named as an Obs-R data set.
The method for building the YN-Net convolutional neural network in step S3 in this embodiment is as follows:
b1: designing a novel convolution structure unit named as Reduce-Group convolution, wherein the Reduce-Group convolution firstly uses 1 × 1 standard convolution to Reduce the dimension of an input feature map, then uses 2 different convolution channels to extract the features of the feature map after dimension reduction, uses channel 1 to perform 3 × 3 grouped convolution feature extraction once on the feature map after dimension reduction, uses channel 2 to perform 3 × 3 grouped convolution feature extraction twice on the feature map after dimension reduction, and finally superposes the results of the two channels and uses 1 × 1 grouped convolution to perform feature extraction to obtain a final output feature map; the structure of the Reduce-Group convolution obtained in this embodiment is shown in fig. 5.
B2: and (3) building an YN-Net convolution neural network based on Reduce-Group convolution:
firstly, the size of an input image is adjusted to 256 × 384 × 3 (pixels), a layer of 3 × 3 standard convolution layer is used for carrying out feature extraction on the adjusted input image to obtain a preliminary feature map, then, five layers of 3 × 3 Reduce-Group convolution layers are used for carrying out further feature extraction on the preliminary feature map to obtain a depth feature map, feature fusion is carried out on the depth feature map by two layers of 1 × 1 standard convolution layers to obtain a final feature map of an YN-Net convolutional neural network, the final feature map is processed by using global mean pooling to obtain an output vector out with the length of 2, and the composition of the out is shown in formula 1, wherein a and b respectively represent probability values of the existence of obstacles and the nonexistence of obstacles in the current input image.
out=[a,b] (1)
The structure of the YN-Net convolutional neural network acquired in this embodiment is shown in fig. 6.
The method for training the YN-Net convolutional neural network model in step S3 in this embodiment is as follows:
and dividing the Obs-R data set into a training set and a testing set according to the ratio of 4:1, and training the built YN-Net convolutional neural network model, wherein the loss function uses a cross entropy loss function, and the network optimizer adopts a adam optimizer to carry out 300 times of iterative training on the network, so that the trained YN-Net convolutional neural network model is obtained.
In this embodiment, as shown in fig. 7, the classification result in step S4 is to determine whether there is an obstacle in front of the power transmission line inspection robot according to the probability value between the existence of the obstacle and the absence of the obstacle in the input image. The actual experimental effect of the method is shown in fig. 8, and it can be seen that the detection accuracy of the method is very high.
Claims (7)
1. A method for detecting obstacles of a power transmission line based on an YN-Net convolutional neural network is characterized by comprising the following steps:
s1: acquiring a barrier image of the power transmission line based on a ZED binocular camera carried by the power transmission line inspection robot;
s2: preprocessing the acquired images of the obstacles of the power transmission line and making the images into a data set;
s3: building an YN-Net convolutional neural network, and fully training the YN-Net convolutional neural network by using the data set of the step S2 to obtain a trained YN-Net convolutional neural network model;
s4: and classifying the acquired real-time images of the transmission line obstacles by using the trained YN-Net convolutional neural network model, and judging whether the obstacles exist in front of the transmission line inspection robot or not according to the probability value of the existence and nonexistence of the obstacles in the input images.
2. The method for detecting the obstacle of the power transmission line based on the YN-Net convolutional neural network as claimed in claim 1, wherein the acquired image of the obstacle of the power transmission line is preprocessed by image screening and image augmentation in the step S2.
3. The method for detecting the obstacle of the power transmission line based on the YN-Net convolutional neural network as claimed in claim 1 or 2, wherein the data set acquisition method in the step S2 is as follows: and according to whether the obstacles exist, performing labeling operation on the preprocessed images of the obstacles of the power transmission line to obtain a labeled data set.
4. The method for detecting the obstacle of the power transmission line based on the YN-Net convolutional neural network as claimed in claim 2, wherein the preprocessing operation in the step S2 is as follows:
a1: screening the acquired image set, and deleting the images with fuzzy shooting;
a2: image augmentation is performed on the acquired image set, and the operation comprises the following steps: horizontally turning, vertically turning, adding Gaussian noise, and adjusting the saturation, contrast and brightness of the image;
a3: the method comprises the steps of dividing a sensitive area of an acquired image set, dividing the sensitive area by taking the middle point of the image as the center, wherein the longitudinal length of the sensitive area is the same as that of the image, the transverse length of the sensitive area is one third of that of the image, cutting the image in the image set according to the sensitive area, and only keeping the image in the sensitive area.
5. The method for detecting the obstacle of the power transmission line based on the YN-Net convolutional neural network as claimed in claim 1, wherein the method for constructing the YN-Net convolutional neural network in the step S3 is as follows:
b1: designing a novel convolution structure unit named as Reduce-Group convolution, wherein the Reduce-Group convolution firstly uses standard convolution to Reduce the dimension of an input feature map, then uses 2 different convolution channels to extract the features of the feature map after dimension reduction, uses channel 1 to extract the features of the feature map after dimension reduction in a grouping convolution mode, uses channel 2 to extract the features of the feature map after dimension reduction in a grouping convolution mode twice, and finally superposes the results of the two channels and uses the grouping convolution to extract the features to obtain a final output feature map;
b2: and constructing an YN-Net convolution neural network based on the Reduce-Group convolution.
6. The method for detecting the obstacle of the power transmission line based on the YN-Net convolutional neural network as claimed in claim 5, wherein the method for building the YN-Net convolutional neural network based on the Reduce-Group convolution in the step B2 is as follows:
firstly, carrying out size adjustment on an input image, carrying out feature extraction on the adjusted input image by using a layer of standard convolutional layer to obtain a preliminary feature map, then carrying out further feature extraction on the preliminary feature map by using five layers of Reduce-Group convolutional layers to obtain a depth feature map, carrying out feature fusion on the depth feature map by using two layers of standard convolutional layers to obtain a final feature map of an YN-Net convolutional neural network, processing the final feature map by using global mean pooling to obtain an output vector out, wherein the composition of the out is shown in formula 1, and a and b respectively represent probability values of obstacles and obstacles which do not exist in the current input image.
out=[a,b] (1)
7. The method for detecting the obstacle of the power transmission line based on the YN-Net convolutional neural network of claim 1, wherein the training method of the YN-Net convolutional neural network model in the step S3 is as follows:
and dividing the data set into a training set and a testing set according to a proportion, training the built YN-Net convolutional neural network model, wherein the loss function uses a cross entropy loss function, and the network optimizer adopts a adam optimizer to carry out iterative training on the network, so that the trained YN-Net convolutional neural network model is obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110504527.8A CN113392701B (en) | 2021-05-10 | 2021-05-10 | YN-Net convolutional neural network-based transmission line obstacle detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110504527.8A CN113392701B (en) | 2021-05-10 | 2021-05-10 | YN-Net convolutional neural network-based transmission line obstacle detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113392701A true CN113392701A (en) | 2021-09-14 |
CN113392701B CN113392701B (en) | 2024-04-26 |
Family
ID=77616841
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110504527.8A Active CN113392701B (en) | 2021-05-10 | 2021-05-10 | YN-Net convolutional neural network-based transmission line obstacle detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113392701B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101277118B1 (en) * | 2013-01-22 | 2013-06-20 | (주) 주암전기통신 | A apparatus for overcoming obstacles in power transmission line |
CN108734143A (en) * | 2018-05-28 | 2018-11-02 | 江苏迪伦智能科技有限公司 | A kind of transmission line of electricity online test method based on binocular vision of crusing robot |
CN110298330A (en) * | 2019-07-05 | 2019-10-01 | 东北大学 | A kind of detection of transmission line polling robot monocular and localization method |
WO2020156028A1 (en) * | 2019-01-28 | 2020-08-06 | 南京航空航天大学 | Outdoor non-fixed scene weather identification method based on deep learning |
-
2021
- 2021-05-10 CN CN202110504527.8A patent/CN113392701B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101277118B1 (en) * | 2013-01-22 | 2013-06-20 | (주) 주암전기통신 | A apparatus for overcoming obstacles in power transmission line |
CN108734143A (en) * | 2018-05-28 | 2018-11-02 | 江苏迪伦智能科技有限公司 | A kind of transmission line of electricity online test method based on binocular vision of crusing robot |
WO2020156028A1 (en) * | 2019-01-28 | 2020-08-06 | 南京航空航天大学 | Outdoor non-fixed scene weather identification method based on deep learning |
CN110298330A (en) * | 2019-07-05 | 2019-10-01 | 东北大学 | A kind of detection of transmission line polling robot monocular and localization method |
Also Published As
Publication number | Publication date |
---|---|
CN113392701B (en) | 2024-04-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109255334B (en) | Remote sensing image ground feature classification method based on deep learning semantic segmentation network | |
CN106290388B (en) | A kind of insulator breakdown automatic testing method | |
US20220309772A1 (en) | Human activity recognition fusion method and system for ecological conservation redline | |
CN109635748B (en) | Method for extracting road characteristics in high-resolution image | |
CN110866494B (en) | Urban group extraction method and system based on optical remote sensing image | |
CN109670405A (en) | A kind of complex background pedestrian detection method based on deep learning | |
CN111582029A (en) | Traffic sign identification method based on dense connection and attention mechanism | |
CN112418049B (en) | Water body change detection method based on high-resolution remote sensing image | |
CN104732524B (en) | The random weight network dividing method of blood leucocyte micro-image | |
CN108734691A (en) | A kind of transmission line of electricity defect image recognition methods | |
CN111832398A (en) | Unmanned aerial vehicle image distribution line pole tower ground wire broken strand image detection method | |
Chen et al. | Surface defect detection of electric power equipment in substation based on improved YOLOv4 algorithm | |
CN115439483B (en) | High-quality welding seam and welding seam defect identification system, method and storage medium | |
CN115661032A (en) | Intelligent pavement disease detection method suitable for complex background | |
CN115984698A (en) | Litchi fruit growing period identification method based on improved YOLOv5 | |
CN111047585A (en) | Pavement crack detection method | |
CN112785610B (en) | Lane line semantic segmentation method integrating low-level features | |
CN112734732B (en) | Railway tunnel cable leakage clamp detection method based on improved SSD algorithm | |
CN113392701A (en) | YN-Net convolution neural network-based power transmission line obstacle detection method | |
CN115830302B (en) | Multi-scale feature extraction fusion power distribution network equipment positioning identification method | |
CN112102236A (en) | Polycrystalline subfissure detection method based on two deep stages | |
Lian et al. | Insulator defect detection algorithm based on improved YOLOv5 | |
CN116129327A (en) | Infrared vehicle detection method based on improved YOLOv7 algorithm | |
CN116503750A (en) | Large-range remote sensing image rural block type residential area extraction method and system integrating target detection and visual attention mechanisms | |
CN115984603A (en) | Fine classification method and system for urban green land based on GF-2 and open map data |
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 |