CN112070730A - Anti-vibration hammer falling detection method based on power transmission line inspection image - Google Patents
Anti-vibration hammer falling detection method based on power transmission line inspection image Download PDFInfo
- Publication number
- CN112070730A CN112070730A CN202010875888.9A CN202010875888A CN112070730A CN 112070730 A CN112070730 A CN 112070730A CN 202010875888 A CN202010875888 A CN 202010875888A CN 112070730 A CN112070730 A CN 112070730A
- Authority
- CN
- China
- Prior art keywords
- shockproof hammer
- vibration damper
- image
- fixing clamp
- detection model
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 80
- 238000007689 inspection Methods 0.000 title claims abstract description 24
- 230000005540 biological transmission Effects 0.000 title claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000002372 labelling Methods 0.000 claims abstract description 7
- 238000012360 testing method Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 20
- 230000004927 fusion Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000003707 image sharpening Methods 0.000 description 2
- 208000025274 Lightning injury Diseases 0.000 description 1
- 239000006096 absorbing agent Substances 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004643 material aging Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000035939 shock Effects 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30204—Marker
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)
- Image Processing (AREA)
Abstract
The invention relates to a stockbridge damper falling detection method based on a power transmission line inspection image, which comprises the following steps of S1: preprocessing the power inspection image data; step S2, constructing a shockproof hammer target detection data set; s3, constructing a vibration damper target detection model based on DSSD, and S4, training the vibration damper target detection model; step S5, intercepting and labeling the shockproof hammer target after the shockproof hammer target detection data set passes through the shockproof hammer target detection model test, and constructing a shockproof hammer fixing clamp detection data set; step S6, constructing a shockproof hammer fixing clamp detection model and training; and step S7, sequentially passing the shockproof hammer power image to be detected through the trained shockproof hammer target detection model and the shockproof hammer fixing clamp detection model, judging the shockproof hammer handle falling fault, and judging the shockproof hammer missing fault according to the fixing clamp identification result. The invention carries out the drop-out detection of the vibration damper based on the inspection image of the power transmission line, and improves the accuracy and the safety of the detection.
Description
Technical Field
The invention relates to the technical field of power transmission line detection, in particular to a method for detecting falling of a vibration damper based on a power transmission line inspection image.
Background
The transmission line is an important component of the power system and takes on the task of transmitting electric energy at a long distance. The vibration damper in the power line is exposed in the natural environment for a long time, not only bears the self damage of natural tension, material aging and the like, but also is eroded by the external world such as lightning stroke, storm wind, rainwater and the like. If the phenomenon that the vibration damper falls off exists in the power transmission line, the wire can vibrate strongly under the action of strong wind, and the wire can be subjected to fatigue damage after long-term periodic bending, so that the safety potential hazard is serious. The traditional line inspection work of the power transmission line depends on manual inspection, but the manual inspection efficiency is low, the working strength is high, and when special lines, particularly high-voltage lines, the risk of personal safety of crisis detection personnel exists.
Disclosure of Invention
In view of this, the invention aims to provide a method for detecting the falling of the vibration damper based on the power transmission line inspection image, so that the detection accuracy and the detection safety are improved.
In order to achieve the purpose, the invention adopts the following technical scheme:
a method for detecting falling of a vibration damper based on a power transmission line inspection image comprises the following steps:
step S1: preprocessing the power inspection image data to obtain preprocessed image data;
step S2, constructing a shockproof hammer target detection data set based on the preprocessed image data;
s3, constructing a vibration damper target detection model based on DSSD;
step S4, training a shockproof hammer target detection model according to the shockproof hammer target detection data set;
step S5, intercepting and labeling the shockproof hammer target after the shockproof hammer target detection data set passes through the shockproof hammer target detection model test, and constructing a shockproof hammer fixing clamp detection data set;
step S6, constructing a shockproof hammer fixing clamp detection model, and training the shockproof hammer fixing clamp detection model according to the shockproof hammer fixing clamp detection data set;
and step S7, sequentially passing the shockproof hammer power image to be detected through the trained shockproof hammer target detection model and the shockproof hammer fixing clamp detection model, judging the shockproof hammer handle falling fault, and judging the shockproof hammer missing fault according to the fixing clamp identification result.
Further, the step S1 is specifically:
step S11, carrying out sharpening processing on the power patrol inspection image data by utilizing a Laplacian operator to obtain a power image with clear edges;
and step S12, defogging the power image with clear edges, and defogging the power image by using a dark channel prior defogging method to obtain preprocessed image data.
Further, the step S2 is specifically:
step S21, recording target information of the shockproof hammer target in the preprocessed image data by using LabelME, and generating an electric power image annotation file in an xml format;
and step S22, dividing the shockproof hammer image and the annotation file into a shockproof hammer training set and a shockproof hammer testing set according to a preset proportion.
Further, the electric power image labeling file specifically includes coordinate information, category information and difficult sample information of all targets to be detected in the image.
Furthermore, the shockproof hammer target detection model is based on a DSSD network model, a top-down network structure is added into the DSSD to perform high-low layer feature fusion, and multiplication operation is performed on the shallow layer feature graph and the deep layer feature graph on corresponding channels; and adding a residual error unit during prediction, and performing inter-channel addition on the processed convolution kernel used by the original characteristic graph and the characteristic graph of the network main channel in a residual error bypass.
Further, the detection model of the shockproof hammer fixing clamp is specifically as follows:
the method comprises the steps that a shockproof hammer target detection result is obtained after a shockproof hammer image to be detected is subjected to network forward propagation through a trained shockproof hammer target detection model;
connecting all the detected vibration dampers in the image to form a vibration damper central point connecting line, and transversely expanding the line to obtain a vibration damper connecting area;
marking the single shockproof hammer and the fixing clamp in the area, and constructing a training data set of the single shockproof hammer and the fixing clamp;
and training the constructed data set through the model, not loading a pre-training model, and training until the network converges to obtain a trained vibration damper fixing clamp detection model.
Further, the step S7 is specifically:
step S71, after the electric power image of the vibration damper to be detected passes through the trained vibration damper target detection model, extracting the target position of the vibration damper, and establishing a candidate search area according to the coordinate position;
and step S72, removing the vibration damper target ROI from the image in the candidate searching area according to the detected vibration damper target result, judging the vibration damper handle falling fault through a trained vibration damper fixing clamp detection model of the image, and judging the vibration damper missing fault according to the fixing clamp identification result.
Compared with the prior art, the invention has the following beneficial effects:
the invention can effectively and timely find the loss fault of the vibration damper, and improves the accuracy and the safety of detection.
Drawings
FIG. 1 illustrates an image sharpening and defogging process according to an embodiment of the present invention;
FIG. 2 is a top-down structure of a DSSD according to an embodiment of the present invention;
FIG. 3 is a prediction model of a DSSD according to an embodiment of the present invention;
FIG. 4 is a network architecture diagram of a DSSD in accordance with an embodiment of the present invention;
FIG. 5 is a section of the damper attachment area in accordance with an embodiment of the present invention;
FIG. 6 illustrates the area to be inspected after the target of the shakeproof hammer is removed in one embodiment of the present invention.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments.
The invention provides a method for detecting falling of a vibration damper based on a power transmission line inspection image, which comprises the following steps of:
step S1: preprocessing the power inspection image data to obtain preprocessed image data;
in the present embodiment, all the power images in the image library are subjected to image sharpening and image defogging processing. Firstly, sharpening the image by using a Laplacian operator to obtain a power image with a clear edge; secondly, defogging is carried out on the image, defogging is carried out on the power image by utilizing a dark channel prior defogging method, a clear image is obtained, and the preprocessed data are shown in figure 1.
Step S2, constructing a shockproof hammer target detection data set based on the preprocessed image data;
in the present embodiment, LabelME is used to record target information for a damper target in a power image, and a power image annotation file in an xml format is generated. The file comprises coordinate information, belonging category information and difficult sample information of all targets to be detected in the image. Integrating the jar damper image with the annotation file, in a 2: the ratio of 8 is divided into the desired set of stockbridge and test sets.
In this embodiment, the hard sample is specifically a shock absorber that is difficult to distinguish and detect by human eyes is marked as the hard sample in the labeling process.
S3, constructing a vibration damper target detection model based on DSSD;
in this embodiment, the DSSD network model is an improvement over the SSD target detection model, where the SSD predicts large objects using a higher layer network and predicts small objects using a lower layer network. A top-down network structure shown in figure 2 is added in the DSSD to perform high-low layer feature fusion, shallow and deep feature maps are subjected to multiplication operation on corresponding channels, and the problem of poor detection effect due to lack of high-layer semantic features when low-layer network feature information used by the SSD is used for predicting small objects is solved. And adding a residual error unit during prediction, processing a convolution kernel used by the original characteristic diagram in a residual error bypass, and adding the processed convolution kernel and the characteristic diagram of the network main trunk path between the channels, as shown in fig. 3. The overall network structure of the DSSD is shown in fig. 4.
Referring to fig. 2, in the top-down network structure of the DSSD in this embodiment, the deconvolution is used on the right side to replace the upsampling operation of the conventional FPN and TDM, and the Eltw Product fusion is performed on the high-bottom layer information, that is, multiplication is performed on the corresponding channel. Addition is used in FPN and overlap is used in TDM;
referring to fig. 3, in the present embodiment, the DSSD model uses a residual unit in a prediction module. The method is characterized in that (a) is a method used by SSD, a multi-scale feature map in a network is directly extracted for classification and prediction of frame regression; (b) is a network structure of a resnet residual error unit; (c) in order to improve a prediction model only comprising a residual error unit, a convolution kernel used by an original characteristic graph is processed in a residual error bypass and then added with the characteristic graph of a network main road; (d) a prediction model containing only two residual units.
Referring to fig. 5, in the present embodiment, the DSSD takes six layers of the seven-layer feature map for prediction based on the SSD model, inputs the six layers into the deconvolution model, and outputs the modified feature map pyramid, forming an hourglass structure composed of feature maps. And finally, inputting the data to box regression and classification through a prediction module.
Step S4, training the shockproof hammer data set through the model, selecting a model without loading pre-training, and training the network weight from zero to the loss value convergence to obtain a trained shockproof hammer target detection model;
step S5, intercepting and labeling the shockproof hammer target after the shockproof hammer target detection data set passes through the shockproof hammer target detection model test, and constructing a shockproof hammer fixing clamp detection data set
Step S6, constructing a shockproof hammer fixing clamp detection model, and training the shockproof hammer fixing clamp detection model according to the shockproof hammer fixing clamp detection data set;
another object detection model is constructed. And (5) the shockproof hammer target detection model obtained in the step (S4) is used for carrying out network forward propagation on the shockproof hammer image to be detected to obtain a shockproof hammer target detection result, and the shockproof hammer target detection result passes through the shockproof hammer connection area. And training the constructed detection data set of the shockproof hammer fixing clamp through the model, not loading a pre-training model, and training until the network converges to obtain a trained detection model of the shockproof hammer fixing clamp.
Preferably, in this embodiment, the vibration damper connection area is specifically: connecting all the detected vibration dampers in the image to form a vibration damper central point connecting line, and transversely expanding the line, wherein the width of the connecting line is 2-3 times of the target width of the vibration damper, and the length of the connecting line is the length of the vibration damper connected with the head end and the tail end plus twice the length of the vibration damper to form a vibration damper connecting area as shown in figure 5
And step S7, sequentially passing the shockproof hammer power image to be detected through the trained shockproof hammer target detection model and the shockproof hammer fixing clamp detection model, judging the shockproof hammer handle falling fault, and judging the shockproof hammer missing fault according to the fixing clamp identification result.
And (3) enabling the image of the vibration damper to be detected to pass through a vibration damper target detection model, deducting the vibration damper target in the obtained detection result to obtain an area shown in fig. 6, and detecting the single vibration damper handle and the vibration damper fixing clamp to obtain a vibration damper handle and fixing clamp detection result.
And judging whether the shockproof hammer falls or lacks faults in the image or not by obtaining the detection result of the shockproof hammer connection area. And if the detected number of the single bundles is more than 0, judging that the image has the fault that the vibration damper falls off the bundles. If the detected number of the single bundles is equal to 0 and the number of the fixing clips is larger than 0, the image is judged to have the defect of the loss of the vibration damper.
The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.
Claims (7)
1. A method for detecting falling of a vibration damper based on a power transmission line inspection image is characterized by comprising the following steps:
step S1: preprocessing the power inspection image data to obtain preprocessed image data;
step S2, constructing a shockproof hammer target detection data set based on the preprocessed image data;
s3, constructing a vibration damper target detection model based on DSSD;
step S4, training a shockproof hammer target detection model according to the shockproof hammer target detection data set;
step S5, intercepting and labeling the shockproof hammer target after the shockproof hammer target detection data set passes through the shockproof hammer target detection model test, and constructing a shockproof hammer fixing clamp detection data set;
step S6, constructing a shockproof hammer fixing clamp detection model, and training the shockproof hammer fixing clamp detection model according to the shockproof hammer fixing clamp detection data set;
and step S7, sequentially passing the shockproof hammer power image to be detected through the trained shockproof hammer target detection model and the shockproof hammer fixing clamp detection model, judging the shockproof hammer handle falling fault, and judging the shockproof hammer missing fault according to the fixing clamp identification result.
2. The method for detecting the drop of the vibration damper based on the power transmission line inspection image according to claim 1, wherein the step S1 specifically comprises:
step S11, carrying out sharpening processing on the power patrol inspection image data by utilizing a Laplacian operator to obtain a power image with clear edges;
and step S12, defogging the power image with clear edges, and defogging the power image by using a dark channel prior defogging method to obtain preprocessed image data.
3. The method for detecting the drop of the vibration damper based on the power transmission line inspection image according to claim 1, wherein the step S2 specifically comprises:
step S21, recording target information of the shockproof hammer target in the preprocessed image data by using LabelME, and generating an electric power image annotation file in an xml format;
and step S22, dividing the shockproof hammer image and the annotation file into a shockproof hammer training set and a shockproof hammer testing set according to a preset proportion.
4. The method for detecting the falling of the vibration damper based on the power transmission line inspection image according to claim 3, wherein the electric power image labeling file specifically comprises coordinate information, belonging category information and difficult sample information of all targets to be detected in the image.
5. The method for detecting the falling of the vibration damper based on the power transmission line inspection image according to claim 1, wherein the target detection model of the vibration damper is based on a DSSD network model, a top-down network structure is added into the DSSD to perform high-low layer feature fusion, and the shallow-layer feature graph and the deep-layer feature graph are subjected to multiplication operation on corresponding channels; and adding a residual error unit during prediction, and performing inter-channel addition on the processed convolution kernel used by the original characteristic graph and the characteristic graph of the network main channel in a residual error bypass.
6. The method for detecting the falling of the vibration damper based on the power transmission line inspection image according to claim 1, wherein the detection model of the vibration damper fixing clamp specifically comprises:
the method comprises the steps that a shockproof hammer target detection result is obtained after a shockproof hammer image to be detected is subjected to network forward propagation through a trained shockproof hammer target detection model;
connecting all the detected vibration dampers in the image to form a vibration damper central point connecting line, and transversely expanding the line to obtain a vibration damper connecting area;
marking the single shockproof hammer and the fixing clamp in the area, and constructing a training data set of the single shockproof hammer and the fixing clamp;
and training the constructed data set through the model, not loading a pre-training model, and training until the network converges to obtain a trained vibration damper fixing clamp detection model.
7. The method for detecting the drop of the vibration damper based on the power transmission line inspection image according to claim 1, wherein the step S7 specifically comprises:
step S71, after the electric power image of the vibration damper to be detected passes through the trained vibration damper target detection model, extracting the target position of the vibration damper, and establishing a candidate search area according to the coordinate position;
and step S72, removing the vibration damper target ROI from the image in the candidate searching area according to the detected vibration damper target result, judging the vibration damper handle falling fault through a trained vibration damper fixing clamp detection model of the image, and judging the vibration damper missing fault according to the fixing clamp identification result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010875888.9A CN112070730A (en) | 2020-08-27 | 2020-08-27 | Anti-vibration hammer falling detection method based on power transmission line inspection image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010875888.9A CN112070730A (en) | 2020-08-27 | 2020-08-27 | Anti-vibration hammer falling detection method based on power transmission line inspection image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112070730A true CN112070730A (en) | 2020-12-11 |
Family
ID=73659040
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010875888.9A Pending CN112070730A (en) | 2020-08-27 | 2020-08-27 | Anti-vibration hammer falling detection method based on power transmission line inspection image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112070730A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112906654A (en) * | 2021-03-29 | 2021-06-04 | 杭州电力设备制造有限公司 | Anti-vibration hammer detection method based on deep learning algorithm |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108037133A (en) * | 2017-12-27 | 2018-05-15 | 武汉市智勤创亿信息技术股份有限公司 | A kind of power equipments defect intelligent identification Method and its system based on unmanned plane inspection image |
CN108596883A (en) * | 2018-04-12 | 2018-09-28 | 福州大学 | It is a kind of that method for diagnosing faults is slid based on the Aerial Images stockbridge damper of deep learning and distance restraint |
WO2020071701A1 (en) * | 2018-10-05 | 2020-04-09 | 재단법인대구경북과학기술원 | Method and device for detecting object in real time by means of deep learning network model |
-
2020
- 2020-08-27 CN CN202010875888.9A patent/CN112070730A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108037133A (en) * | 2017-12-27 | 2018-05-15 | 武汉市智勤创亿信息技术股份有限公司 | A kind of power equipments defect intelligent identification Method and its system based on unmanned plane inspection image |
CN108596883A (en) * | 2018-04-12 | 2018-09-28 | 福州大学 | It is a kind of that method for diagnosing faults is slid based on the Aerial Images stockbridge damper of deep learning and distance restraint |
WO2020071701A1 (en) * | 2018-10-05 | 2020-04-09 | 재단법인대구경북과학기술원 | Method and device for detecting object in real time by means of deep learning network model |
Non-Patent Citations (1)
Title |
---|
CHENG-YANG FU,ET AL: "DSSD:Deconvolutional Single Shot Detector", 《ARXIV:1701.06659V1》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112906654A (en) * | 2021-03-29 | 2021-06-04 | 杭州电力设备制造有限公司 | Anti-vibration hammer detection method based on deep learning algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109785289B (en) | Transmission line defect detection method and system and electronic equipment | |
CN108009591B (en) | Contact network key component identification method based on deep learning | |
CN110569841B (en) | Contact gateway key component target detection method based on convolutional neural network | |
CN111797890A (en) | Method and system for detecting defects of power transmission line equipment | |
CN109344905A (en) | A kind of transmission facility automatic fault recognition methods based on integrated study | |
Li et al. | Sewer pipe defect detection via deep learning with local and global feature fusion | |
CN108711148B (en) | Tire defect intelligent detection method based on deep learning | |
CN112364754A (en) | Bolt defect detection method and system | |
CN112906654A (en) | Anti-vibration hammer detection method based on deep learning algorithm | |
CN111008641A (en) | Power transmission line tower external force damage detection method based on convolutional neural network | |
CN114627082A (en) | Photovoltaic module EL defect detection method based on improved YOLOv5 network | |
CN111199213B (en) | Method and device for detecting defects of equipment for transformer substation | |
CN114022715A (en) | Method and device for detecting lead sealing defect of high-voltage cable terminal | |
CN116883801A (en) | YOLOv8 target detection method based on attention mechanism and multi-scale feature fusion | |
CN112070730A (en) | Anti-vibration hammer falling detection method based on power transmission line inspection image | |
Manninen et al. | Multi-stage deep learning networks for automated assessment of electricity transmission infrastructure using fly-by images | |
Bahreini et al. | Point cloud semantic segmentation of concrete surface defects using dynamic graph CNN | |
CN111079645A (en) | Insulator self-explosion identification method based on AlexNet network | |
CN110428398A (en) | A kind of high iron catenary bracing wire defect inspection method based on deep learning | |
CN117541534A (en) | Power transmission line inspection method based on unmanned plane and CNN-BiLSTM model | |
Ramlal et al. | Toward automated utility pole condition monitoring: A deep learning approach | |
CN116579998A (en) | Defect detection method, device, equipment, medium and program product of transformer equipment | |
CN111652834A (en) | Steel coil strapping tape identification method based on support vector machine | |
Chen et al. | Attention mechanism-based CNN for surface damage detection of wind turbine blades | |
CN116229278B (en) | Method and system for detecting rust defect of vibration damper of power transmission line |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201211 |
|
RJ01 | Rejection of invention patent application after publication |