CN109801284A - A kind of high iron catenary insulator breakdown detection method based on deep learning - Google Patents
A kind of high iron catenary insulator breakdown detection method based on deep learning Download PDFInfo
- Publication number
- CN109801284A CN109801284A CN201910072859.6A CN201910072859A CN109801284A CN 109801284 A CN109801284 A CN 109801284A CN 201910072859 A CN201910072859 A CN 201910072859A CN 109801284 A CN109801284 A CN 109801284A
- Authority
- CN
- China
- Prior art keywords
- insulator
- model
- training
- image
- high iron
- 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
Landscapes
- Insulators (AREA)
- Image Analysis (AREA)
Abstract
A kind of high iron catenary insulator breakdown detection method based on deep learning disclosed by the invention, deep neural network model is trained using the image largely containing normal insulation and faulty insulator, insulator position detection model and insulator breakdown disaggregated model are obtained, by the on-line checking that the position detection model and failure modes model encapsulation are realized to insulator at the function that program can be called；The present invention it is relatively traditional manually see that drawing method is more efficient, while the progress insulator breakdown detection of relatively traditional extractions manual features method it is more robust with it is efficient；The present invention is trained deep neural network as positive negative sample using a large amount of normal insulation and faulty insulator subgraph, improve the accuracy of insulator breakdown detection, the cost of overhaul is saved simultaneously, the early warning period is shortened, the intellectually and automatically for improving high iron catenary maintenance engineering is horizontal.
Description
Technical field
The invention belongs to high iron catenary fault detection technique field, more particularly, to a kind of based on deep learning
High iron catenary insulator breakdown detection method.
Background technique
China possesses maximum High-speed Railway Network in the world, ends to the end of the year in 2017, Chinese high-speed rail operating mileage has accounted for
Nearly 70 the percent of the whole world.High-speed railway touching net is the power transmission line powered to electric locomotive set up along high iron wire overhead
Road.Contact net will directly affect the safe operation of locomotive once powering off, and generate serious consequence.So, it is ensured that the safety of contact net
Operation is a vital job in electrified high-speed rail.
Insulator is the object of emphasis maintenance as the critical component in contact net.It is promulgated according to Chinese Railway parent company
" overhead contact line state-detection monitoring device (4C) provisional technical conditions " it is found that insulator breakdown mainly has breakage, it is dirty
Burning with electric discharge causes enamel to fall off three kinds of situations.High iron catenary insulator is got over by way of manually seeing that figure carries out malfunction elimination
Do not meet the status that China's high-speed rail route mileage increases substantially more, and with the gradually aging of high-speed rail route, components therefore
Barrier rate also can gradually increase, it is therefore desirable to propose that the mode of more intelligent and high-efficiency carries out the investigation diagnosis of high iron catenary failure.
The Intelligentized method for the detection of high iron catenary insulator breakdown is regarded based on traditional computer mostly at this stage
Feel technology is detected and is identified by the feature for manually setting and extracting, and in actual operation, the image of acquisition due to by
Influence insulator geometric shape in the picture to camera site is ever-changing, and often interference occurs and block, and imaging is carried on the back
Scape is complicated, while insulator its external manifestation affected by environment of different defect types is rich and varied, therefore is easy to appear insulation
Sub- missing inspection and erroneous judgement, so that the accuracy of insulator breakdown detection and efficiency reduce.
Summary of the invention
In view of the drawbacks of the prior art, the purpose of the present invention is to provide a kind of high iron catenary based on deep learning is exhausted
Edge fault detection method, it is intended to solve the prior art and be easy to appear insulator missing inspection and erroneous judgement, insulator breakdown is caused to detect
Accuracy and efficiency reduce the problem of.
To achieve the above object, one aspect of the present invention provides a kind of high iron catenary insulator event based on deep learning
Hinder detection model training method, comprising the following steps:
(1) acquisition contact net supports the image with suspension arrangement；
(2) insulator position and classification are labeled in the acquired images, are obtained for training position detection mould
First training dataset of type；
(3) first training dataset is inputted in the YOLOv2 deep neural network model put up and is iterated instruction
Practice, the position detection model for detecting insulator position in the picture is obtained after repetitive exercise；
(4) it screens in the acquired images and cuts out normal and failure insulation subgraph, and to the insulator
Image carries out pretreatment and data extending, obtains the second training dataset for training Fault Model；
(5) after the insulator picture size concentrated to second training data carries out normalization processing, input is put up
Resnet-50 deep neural network model in be iterated training, obtain after repetitive exercise for judging insulator therefore
Hinder the failure modes model of classification.
Further, insulator classification described in the step (2) includes: that insulator is normal, insulator is damaged, insulator
Dirty and insulator enamel falls off.
Further, pretreatment and data extending, packet are carried out described in the step (4) and to the insulation subgraph
It includes:
Gaussian filtering process, removal imaging noise are carried out to the insulation subgraph；
Mode using greyscale transformation, gamma transformation, HSV disturbance, rotation or mirror image expands treated image
It fills.
Another aspect of the present invention provides a kind of high iron catenary insulator breakdown detection method based on deep learning,
It is characterized in that, comprising:
(1) image to be detected is inputted in trained position detection model, obtains position, size and the confidence of insulator
Spend information；
(2) confidence level is chosen to be cut and carried out at picture size normalization more than or equal to the insulation subgraph of setting value
After reason, inputs in trained failure modes model, obtain the fault category of insulator；
(3) position and the fault category of insulator are exported in a manner of visual in image to be detected.
Preferably, the confidence level setting value is 0.3.
Preferably, apply the position detection model and failure modes model to acquisition on high iron catenary inspection car
Image is measured in real time, and is updated training dataset, especially training data using the mass data of storage and concentrated relatively
Few defect image data set, to enhance the Generalization Capability of model, improves position to be trained tuning to original model
The precision of detection and the accuracy of failure modes.
Further, the method can also be applied to the insulator breakdown inspection of general fast railway contact line and high voltage transmission line
It surveys.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, can obtain down and show
Beneficial effect:
(1) present invention extracts the characteristics of image of insulator by deep neural network, manually sees figure compared to traditional utilization
The mode that insulator is detected with the image processing algorithm based on manual features, it is more robust with it is efficient；
(2) present invention uses the image for largely containing normal insulation and faulty insulator as sample to depth nerve net
Network is trained, and improves the precision of position detection and the accuracy of failure modes；
(3) isolator detecting method of the invention shortens the time of high iron catenary insulator breakdown investigation, saves
The cost of overhaul, the intellectually and automatically for improving high iron catenary maintenance engineering are horizontal.
Detailed description of the invention
Fig. 1 is the high iron catenary insulator breakdown detection method flow chart of the invention based on deep learning；
Fig. 2 is normal insulation subgraph example；
Fig. 3 is faulty insulator example images；
Fig. 4 is the enhancing of insulator training dataset and expansion schematic diagram；
Fig. 5 is defects of insulator classification schematic diagram；
Fig. 6 is an isolator detecting and failure modes result example.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
With reference to Fig. 1, a kind of high iron catenary insulator breakdown detection method based on deep learning provided by the invention, packet
Include following steps:
(1) acquisition contact net supports the image data with suspension arrangement；
Specifically, the contact net 4C for the industrial camera shooting installed at the top of high iron catenary inspection car can be directly used
The history image data of device can also be supported to fill with suspension using unmanned plane or other ground crusing robots acquisition contact net
The image data set.
(2) insulator position and classification are labeled in the acquired images, are obtained for training position detection mould
First training dataset of type；
Specifically, 10000 width are chosen in the acquired images and include the image of insulator, and are marked in these images
The position of insulator and classification out；When practical operation, rectangle frame can be used and choose insulator in image, the width of rectangle frame
Degree, the location information of height and apex coordinate as insulator；The classification of the insulator includes: normal insulation, breakage
Insulator, dirty insulator, enamel fall off four kinds of insulator, the position of insulator and classification information and correspondence in every image
The data informations such as width, height, the storage location of image constitute first training dataset.
(3) first training dataset is inputted in the YOLOv2 deep neural network model put up and is iterated instruction
Practice, the position detection model for detecting insulator position in the picture is obtained after repetitive exercise；
Specifically, YOLOv2 (You Only Look Once) deep neural network model is every for detecting in the picture
The size of one insulator position and enclosure rectangle frame, includes 22 layers of convolutional layer altogether, and 5 layers of maximum pond layer are arranged defeated when training
The width for entering image is 416 pixels, is highly 416 pixels, and the port number of image is 3, is iterated trained picture number every time
Amount is 64, momentum 0.9, learning rate 0.001, maximum number of iterations 80200, and network output is the insulator in image
Position, size and the confidence level of several and each insulator are needed since testing result is there are crossing redundancy using non-very big
The method that value inhibits merges testing result, so that the insulator output result for detecting each uniquely corresponds to.
(4) it screens in the acquired images and cuts out normal and failure insulation subgraph, and to the insulator
Image carries out pretreatment and data extending, obtains the second training dataset for training Fault Model；
Specifically, " high-speed railway touching net operating maintenance the rule " (TG/GD124- promulgated with Chinese Railway parent company
It 2015) is standard, tissue this field profession service work personnel screen the normal image of insulator in a manner of manually identifying
5000, the image 1000 of insulator breakage is opened, the image 1000 that insulator is dirty is opened and insulator burn falls off with enamel
Image 1000 open, and insulator is cut from image, it is damaged, dirty, burn wherein normally insulation subgraph is positive sample
Wound is negative sample with the faulty insulator that enamel falls off, and Fig. 2, Fig. 3 respectively illustrate normal insulation and faulty insulator not
With the imaging under shooting angle, insulator form can show variform due to shooting angle difference in practical screening, scheme
In insulator it is for reference only, not as judgment criteria；
Gaussian filtering process is carried out after noise is imaged in removal to the insulation subgraph cut out and uses greyscale transformation, gamma
8000 images are extended to 80000 by transformation, HSV (Hue, Saturation, Value) disturbance, rotation or the mode of mirror image,
It is as shown in Figure 4 that insulator example images are obtained using above-mentioned mapping mode.
(5) after the insulator picture size concentrated to second training data carries out normalization processing, input is put up
Resnet-50 deep neural network model in be iterated training, obtain after repetitive exercise for judging insulator therefore
Hinder the failure modes model of classification；
Specifically, Resnet-50 (Deep Residual Networks-50) deep neural network model is used for insulation
Son carries out failure modes, altogether includes 49 layers of convolutional layer, 1 layer of full articulamentum；The height of setting input picture and width are equal when training
For 224 pixels, image channel number is 3, and initial learning rate is 0.001, fixed learning rate is 0.1, and maximum number of iterations is
75000；Network output is multi-class fiducial probability vector, finds the largest component of the vector and makes its corresponding subscript position
It is encoded with 0-1 hot key, obtains the classification results of insulator, 0001 class represents that insulator is normal, 0010 class represents insulator
Damaged, 0100 class represents that insulator is dirty, 1000 classes represent insulator enamel and fall off, and wherein each classification results is one corresponding
Fiducial probability, such as the above 4 class fiducial probabilities are respectively 0.95,0.14,0.07,0.06, and it is corresponding to choose fiducial probability maximum
Class judges that the class obtained, output result are insulator normal category as network model.
(6) the good position detection model of application training and failure modes model detect testing image, specifically include:
Image to be detected is inputted in trained position detection model, position, size and the confidence level of insulator are obtained
Information；
The insulation subgraph that confidence level is chosen more than or equal to setting value is cut and carries out picture size normalization processing
Afterwards, it inputs in trained failure modes model, obtains the fault category of insulator；
Position and the fault category of insulator are exported in a manner of visual in image to be detected.
Specifically, according to YOLOv2 network export and fused each insulator position, size, confidence information,
It chooses insulation sub-information of the confidence level more than or equal to 0.3 to cut out by insulator in original image, and will cut out as shown in Figure 5
The insulator picture size come carries out normalization processing, obtains 224 pixels × 224 pixel size images as Resnet-50's
Input, can be obtained the fault category of insulator.
When practical operation, position can be examined by deep neural network module (DNN Modual) in Opencv vision library
Surveying model and failure modes model encapsulation is class, and by function call, such method realizes detection, and with graphic user interface
Mode by result presentation to user, as shown in fig. 6, the rectangle frame that different colours or line style can be used chooses the insulation in image
Son, different colours or line style respectively correspond a kind of fault category.
On high iron catenary inspection car using the position detection model and failure modes model to the image of acquisition into
Row real-time detection, and update training dataset, especially training data using the mass data of storage and concentrate relatively small number of lack
It falls into image data set and, to enhance the Generalization Capability of model, improves position detection to be trained tuning to original model
The accuracy of precision and failure modes.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should all include
Within protection scope of the present invention.
Claims (5)
1. a kind of high iron catenary insulator breakdown detection model training method based on deep learning, which is characterized in that including
Following steps:
(1) acquisition contact net supports the image with suspension arrangement；
(2) insulator position and classification are labeled in the acquired images, are obtained for training position detection model
First training dataset；
(3) first training dataset is inputted in the YOLOv2 deep neural network model put up and is iterated training,
The position detection model for detecting insulator position in the picture is obtained after repetitive exercise；
(4) it screens in the acquired images and cuts out normal and failure insulation subgraph, and to the insulation subgraph
Pretreatment and data extending are carried out, the second training dataset for training Fault Model is obtained；
(5) it after the insulator picture size concentrated to second training data carries out normalization processing, inputs and puts up
It is iterated training in Resnet-50 deep neural network model, obtains after repetitive exercise for judging insulator breakdown
The failure modes model of classification.
2. a kind of high iron catenary insulator breakdown detection model training side based on deep learning according to claim 1
Method, which is characterized in that insulator classification described in the step (2) includes: that insulator is normal, insulator is damaged, insulator is dirty
Dirty and insulator enamel falls off.
3. a kind of high iron catenary insulator breakdown detection model instruction based on deep learning according to claim 1 or 2
Practice method, which is characterized in that carry out pretreatment and data extending, packet described in the step (4) and to the insulation subgraph
It includes:
Gaussian filtering process, removal imaging noise are carried out to the insulation subgraph；
Mode using greyscale transformation, gamma transformation, HSV disturbance, rotation or mirror image expands treated image.
4. a kind of high iron catenary insulator breakdown detection method based on deep learning characterized by comprising
Image to be detected is inputted in trained position detection model, position, size and the confidence information of insulator are obtained；
It is defeated after the insulation subgraph that selection confidence level is more than or equal to setting value is cut and carries out picture size normalization processing
Enter in trained failure modes model, obtains the fault category of insulator；
Position and the fault category of insulator are exported in a manner of visual in image to be detected.
5. a kind of insulator breakdown detection method based on deep learning according to claim 4, which is characterized in that described
Confidence level setting value is 0.3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910072859.6A CN109801284A (en) | 2019-01-25 | 2019-01-25 | A kind of high iron catenary insulator breakdown detection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910072859.6A CN109801284A (en) | 2019-01-25 | 2019-01-25 | A kind of high iron catenary insulator breakdown detection method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109801284A true CN109801284A (en) | 2019-05-24 |
Family
ID=66560340
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910072859.6A Pending CN109801284A (en) | 2019-01-25 | 2019-01-25 | A kind of high iron catenary insulator breakdown detection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109801284A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110133443A (en) * | 2019-05-31 | 2019-08-16 | 中国科学院自动化研究所 | Based on the transmission line part detection method of parallel vision, system, device |
| CN110176001A (en) * | 2019-06-03 | 2019-08-27 | 浙江大学 | A kind of high iron catenary insulator breakage accurate positioning method based on Grad-CAM algorithm |
| CN110264448A (en) * | 2019-06-06 | 2019-09-20 | 中南大学 | A kind of insulator breakdown detection method based on machine vision |
| CN110276747A (en) * | 2019-06-06 | 2019-09-24 | 中南大学 | A kind of insulator breakdown detection and failure ranking method based on image analysis |
| CN110288571A (en) * | 2019-06-06 | 2019-09-27 | 中南大学 | A kind of high iron catenary insulator method for detecting abnormality based on image procossing |
| CN110378420A (en) * | 2019-07-19 | 2019-10-25 | Oppo广东移动通信有限公司 | A kind of image detecting method, device and computer readable storage medium |
| CN111091543A (en) * | 2019-12-12 | 2020-05-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring loss fault target detection method |
| CN111598842A (en) * | 2020-04-24 | 2020-08-28 | 云南电网有限责任公司电力科学研究院 | Method and system for generating model of insulator defect sample and storage medium |
| CN111784656A (en) * | 2020-06-28 | 2020-10-16 | 京东数字科技控股有限公司 | Railway contact network fault detection method and device, electronic equipment and storage medium |
| CN111899216A (en) * | 2020-06-16 | 2020-11-06 | 西安交通大学 | Abnormity detection method for insulator fastener of high-speed rail contact network |
| CN112270270A (en) * | 2020-10-30 | 2021-01-26 | 湘潭大学 | Contact network fault type identification method based on ABC optimization neural network |
| CN112634254A (en) * | 2020-12-29 | 2021-04-09 | 北京市商汤科技开发有限公司 | Insulator defect detection method and related device |
| CN112906620A (en) * | 2021-03-09 | 2021-06-04 | 唐山职业技术学院 | Unmanned aerial vehicle-assisted insulator fault detection method and device and electronic equipment |
| CN113469950A (en) * | 2021-06-08 | 2021-10-01 | 海南电网有限责任公司电力科学研究院 | Method for diagnosing abnormal heating defect of composite insulator based on deep learning |
| CN113570563A (en) * | 2021-07-19 | 2021-10-29 | 江苏新绿能科技有限公司 | Method for evaluating health state of insulated terminal of high-speed railway contact net |
| CN114392964A (en) * | 2021-12-24 | 2022-04-26 | 西南交通大学 | Insulator rinse-system that dirty degree of intellectuality was judged |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH03259706A (en) * | 1990-03-09 | 1991-11-19 | Mitsubishi Electric Corp | Appearance inspecting device |
| CN106408025A (en) * | 2016-09-20 | 2017-02-15 | 西安工程大学 | Classification and recognition method of aerial image insulators based on image processing |
| CN106919929A (en) * | 2017-03-09 | 2017-07-04 | 国家电网公司 | Insulator automatic identifying method in a kind of infrared image based on template matches |
| CN107742130A (en) * | 2017-10-25 | 2018-02-27 | 西南交通大学 | High iron catenary based on deep learning supports device fastener failure diagnostic method |
| CN108280855A (en) * | 2018-01-13 | 2018-07-13 | 福州大学 | A kind of insulator breakdown detection method based on Fast R-CNN |
| CN108596886A (en) * | 2018-04-17 | 2018-09-28 | 福州大学 | Aerial Images insulator based on deep learning falls piece fault rapid detecting method |
| CN109166094A (en) * | 2018-07-11 | 2019-01-08 | 华南理工大学 | A kind of insulator breakdown positioning identifying method based on deep learning |
-
2019
- 2019-01-25 CN CN201910072859.6A patent/CN109801284A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH03259706A (en) * | 1990-03-09 | 1991-11-19 | Mitsubishi Electric Corp | Appearance inspecting device |
| CN106408025A (en) * | 2016-09-20 | 2017-02-15 | 西安工程大学 | Classification and recognition method of aerial image insulators based on image processing |
| CN106919929A (en) * | 2017-03-09 | 2017-07-04 | 国家电网公司 | Insulator automatic identifying method in a kind of infrared image based on template matches |
| CN107742130A (en) * | 2017-10-25 | 2018-02-27 | 西南交通大学 | High iron catenary based on deep learning supports device fastener failure diagnostic method |
| CN108280855A (en) * | 2018-01-13 | 2018-07-13 | 福州大学 | A kind of insulator breakdown detection method based on Fast R-CNN |
| CN108596886A (en) * | 2018-04-17 | 2018-09-28 | 福州大学 | Aerial Images insulator based on deep learning falls piece fault rapid detecting method |
| CN109166094A (en) * | 2018-07-11 | 2019-01-08 | 华南理工大学 | A kind of insulator breakdown positioning identifying method based on deep learning |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110133443A (en) * | 2019-05-31 | 2019-08-16 | 中国科学院自动化研究所 | Based on the transmission line part detection method of parallel vision, system, device |
| CN110176001A (en) * | 2019-06-03 | 2019-08-27 | 浙江大学 | A kind of high iron catenary insulator breakage accurate positioning method based on Grad-CAM algorithm |
| CN110176001B (en) * | 2019-06-03 | 2021-09-03 | 浙江大学 | Grad-CAM algorithm-based high-speed rail contact net insulator damage accurate positioning method |
| CN110276747B (en) * | 2019-06-06 | 2021-03-23 | 中南大学 | Insulator fault detection and fault rating method based on image analysis |
| CN110264448A (en) * | 2019-06-06 | 2019-09-20 | 中南大学 | A kind of insulator breakdown detection method based on machine vision |
| CN110276747A (en) * | 2019-06-06 | 2019-09-24 | 中南大学 | A kind of insulator breakdown detection and failure ranking method based on image analysis |
| CN110288571A (en) * | 2019-06-06 | 2019-09-27 | 中南大学 | A kind of high iron catenary insulator method for detecting abnormality based on image procossing |
| CN110264448B (en) * | 2019-06-06 | 2021-04-23 | 中南大学 | Insulator fault detection method based on machine vision |
| CN110288571B (en) * | 2019-06-06 | 2021-03-23 | 中南大学 | High-speed rail contact net insulator abnormity detection method based on image processing |
| CN110378420A (en) * | 2019-07-19 | 2019-10-25 | Oppo广东移动通信有限公司 | A kind of image detecting method, device and computer readable storage medium |
| CN111091543B (en) * | 2019-12-12 | 2020-10-09 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring loss fault target detection method |
| CN111091543A (en) * | 2019-12-12 | 2020-05-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring loss fault target detection method |
| CN111598842A (en) * | 2020-04-24 | 2020-08-28 | 云南电网有限责任公司电力科学研究院 | Method and system for generating model of insulator defect sample and storage medium |
| CN111899216A (en) * | 2020-06-16 | 2020-11-06 | 西安交通大学 | Abnormity detection method for insulator fastener of high-speed rail contact network |
| CN111784656A (en) * | 2020-06-28 | 2020-10-16 | 京东数字科技控股有限公司 | Railway contact network fault detection method and device, electronic equipment and storage medium |
| CN112270270A (en) * | 2020-10-30 | 2021-01-26 | 湘潭大学 | Contact network fault type identification method based on ABC optimization neural network |
| CN112634254A (en) * | 2020-12-29 | 2021-04-09 | 北京市商汤科技开发有限公司 | Insulator defect detection method and related device |
| CN112906620A (en) * | 2021-03-09 | 2021-06-04 | 唐山职业技术学院 | Unmanned aerial vehicle-assisted insulator fault detection method and device and electronic equipment |
| CN113469950A (en) * | 2021-06-08 | 2021-10-01 | 海南电网有限责任公司电力科学研究院 | Method for diagnosing abnormal heating defect of composite insulator based on deep learning |
| CN113570563A (en) * | 2021-07-19 | 2021-10-29 | 江苏新绿能科技有限公司 | Method for evaluating health state of insulated terminal of high-speed railway contact net |
| CN113570563B (en) * | 2021-07-19 | 2024-03-01 | 江苏新绿能科技有限公司 | High-speed railway contact net insulated terminal health state assessment method |
| CN114392964A (en) * | 2021-12-24 | 2022-04-26 | 西南交通大学 | Insulator rinse-system that dirty degree of intellectuality was judged |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109801284A (en) | A kind of high iron catenary insulator breakdown detection method based on deep learning | |
| CN109376768B (en) | Aerial image tower signboard fault diagnosis method based on deep learning | |
| Yuan et al. | A deep convolutional neural network for detection of rail surface defect | |
| Feng et al. | Electrical insulator defects detection method based on yolov5 | |
| CN109409404A (en) | A kind of high iron catenary radix saposhnikoviae bracing wire fault detection method based on deep learning | |
| CN108009591A (en) | A kind of contact network key component identification method based on deep learning | |
| CN107742130A (en) | High iron catenary based on deep learning supports device fastener failure diagnostic method | |
| CN108664997A (en) | High iron catenary equipotential line defective mode detection method based on cascade Faster R-CNN | |
| CN110458798B (en) | Vibration damper defect visual detection method, system and medium based on key point detection | |
| CN111079627A (en) | Railway wagon brake beam body breaking fault image identification method | |
| CN111080600A (en) | Fault identification method for split pin on spring supporting plate of railway wagon | |
| CN111862029A (en) | Fault detection method for bolt part of vertical shock absorber of railway motor train unit | |
| CN105354589A (en) | Method and system for intelligently identifying insulator crack in catenary image | |
| CN111080613B (en) | Image recognition method for damage fault of wagon bathtub | |
| CN109308448A (en) | A method of it prevents from becoming distribution maloperation using image processing techniques | |
| CN109389146A (en) | Equipment state feedback method and system based on neural network image sorting algorithm | |
| CN112560816A (en) | Equipment indicator lamp identification method and system based on YOLOv4 | |
| CN112418155A (en) | Method for detecting position and type of workpiece on subway car side inspection image | |
| CN109785288A (en) | Transmission facility defect inspection method and system based on deep learning | |
| CN116168351B (en) | Inspection method and device for power equipment | |
| CN111859805A (en) | Method for detecting topological relation of power drawing based on artificial intelligence | |
| CN116883801A (en) | YOLOv8 target detection method based on attention mechanism and multi-scale feature fusion | |
| CN115546664A (en) | Cascaded network-based insulator self-explosion detection method and system | |
| CN115239646A (en) | Defect detection method and device for power transmission line, electronic equipment and storage medium | |
| CN106355187A (en) | Application of visual information to electrical equipment monitoring |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhong Sheng Inventor after: Le Mingyang Inventor after: Wang Jianhui Inventor after: Yang Bo Inventor after: Yan Luxin Inventor before: Zhong Sheng Inventor before: Le Mingyang Inventor before: Yang Bo Inventor before: Yan Luxin |
|
| CB03 | Change of inventor or designer information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190524 |
|
| RJ01 | Rejection of invention patent application after publication |