CN109784336A - A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection - Google Patents
A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection Download PDFInfo
- Publication number
- CN109784336A CN109784336A CN201910089294.2A CN201910089294A CN109784336A CN 109784336 A CN109784336 A CN 109784336A CN 201910089294 A CN201910089294 A CN 201910089294A CN 109784336 A CN109784336 A CN 109784336A
- Authority
- CN
- China
- Prior art keywords
- infrared image
- power equipment
- fault point
- grid
- target detection
- 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.)
- Withdrawn
Links
Landscapes
- Image Analysis (AREA)
Abstract
This application discloses a kind of infrared image fault point recognition methods based on YOLO algorithm of target detection, comprising: obtains infrared image, the prediction frame in infrared image is arranged；Probability in infrared image there are power equipment is set, and according in infrared image, there are the probability of power equipment, and judgment threshold is arranged；Judge whether the value of the confidence is greater than judgment threshold, if the value of the confidence is greater than judgment threshold, there are power equipments, perform the next step；Conversely, then terminating；To there are the infrared images of power equipment to carry out processing and fault signature identification, the power equipment of failure is obtained；It is compared after the power equipment of failure is handled with infrared image fault database, to identify the fault point of infrared image.A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection provided by the present application combines YOLO algorithm of target detection, realize the real-time detection identification to infrared image failure, the error probability for greatly reducing infrared image fault identification, is more clear live effect.
Description
Technical field
This application involves field of image recognition more particularly to a kind of infrared image failures based on YOLO algorithm of target detection
Point recognition methods.
Background technique
Currently, convolutional network method is to be frequently utilized for the fault point of Infrared image to know one of method for distinguishing, however convolution net
Network method can mistake by the plaque detection in infrared image background be target, reason be convolutional network method in the detection without
Method sees global image.
Other than the method for convolutional network method, also many recognition methods for being used for infrared image fault point, and it is current
Method for distinguishing is known in existing infrared image fault point, and there is to the identification of infrared image fault point is difficult, accuracy rate is not high, identification
Therefore the problems such as inefficiency, needs a kind of new infrared image fault point recognition methods, to make up above-mentioned deficiency, improve and know
Other efficiency and accuracy.
Summary of the invention
This application provides a kind of infrared image fault point recognition methods based on YOLO algorithm of target detection, existing to solve
Having infrared image fault point to know method for distinguishing, there is to the identification of infrared image fault point is difficult, accuracy rate is not high, recognition efficiency
Low problem.
This application provides a kind of infrared image fault point recognition methods based on YOLO algorithm of target detection, the methods
The following steps are included:
Step S1: infrared image is obtained, the prediction frame in the infrared image is set, and each prediction frame wraps
Containing five-dimensional information, the five-dimensional information includes the offset for predicting frame center relative to each grid, the five-dimensional information
It further include the wide high relative to the ratio between infrared image described in whole picture of the prediction frame, the five-dimensional information further includes the value of the confidence；
Step S2: being arranged the probability in the infrared image there are power equipment, electric according to existing in the infrared image
Judgment threshold is arranged in the probability of power equipment；
Step S3: judging whether the value of the confidence is greater than the judgment threshold, if the value of the confidence is greater than the judgement threshold
Then there is power equipment in value, execute step S4；Conversely, then terminating；
Step S4: processing is carried out to the infrared image there are power equipment and fault signature identifies, obtains failure
Power equipment；
Step S5: comparing after the power equipment of failure is handled with infrared image fault database, to identify
The fault point of infrared image.
Selectable, the acquisition infrared image, the prediction frame being arranged in the infrared image includes:
Step S11: infrared image is obtained, the infrared image is divided into the grid of S × S；
Step S12: the frame of setting each grid；
Step S13: the friendship of the frame and actual position of each grid and ratio are calculated, the pre- of each grid is obtained
Survey frame.
Selectable, the value of the confidence is obtained by the class probability and confidence calculations of each grid.
Selectable, there are the probability of power equipment in the setting infrared image, according in the infrared image
There are the probability of power equipment setting judgment thresholds to include:
Step S21: there are the probability of power equipment in setting each grid；
Step S22: according to the probability of power equipment in each grid, judgment threshold is set.
It is selectable, it is described to judge whether the value of the confidence is greater than the judgment threshold and includes:
According in each grid the judgment threshold and the value of the confidence, judge described in each grid
Whether the value of the confidence is greater than the judgment threshold, if the value of the confidence is greater than the judgment threshold, in the grid judged
There are power equipments, perform the next step；Conversely, then terminating.
Selectable, described pair there are the infrared images of power equipment to carry out processing and fault signature identification, obtains
The power equipment of failure includes:
Step S41: the power equipment in the grid is extracted in the grid there are power equipment；
Step S42: the power equipment extracted is pre-processed, the power equipment that obtains that treated；
Step S43: to treated, power equipment carries out fault signature identification, obtains the power equipment of failure.
It is selectable, it is described the power equipment of failure is handled after compared with infrared image fault database, thus
The fault point for identifying infrared image includes:
Step S51: cutting is carried out to the fault point of the power equipment of the failure, then the fault point is zoomed in and out
And feature extraction；
Step S52: the fault point for extracting feature is compared with infrared image fault database, to identify infrared
The fault point of image.
Selectable, the pretreatment includes gray processing processing being carried out to the power equipment that extracts, at binaryzation
Reason and denoising.
It is selectable, it is described to compare the fault point for extracting feature with infrared image fault database, to identify
The fault point for haveing infrared image further includes later being trained to infrared image failure.
From the above technical scheme, this application provides a kind of infrared image failures based on YOLO algorithm of target detection
Point recognition methods, the described method comprises the following steps: obtaining infrared image, the prediction frame in the infrared image is arranged；If
The probability in the infrared image there are power equipment is set, there are the settings of the probability of power equipment to sentence according in the infrared image
Disconnected threshold value；Judge whether the value of the confidence is greater than the judgment threshold, if the value of the confidence is greater than the judgment threshold, exists
Power equipment performs the next step；Conversely, then terminating；Processing is carried out to the infrared image there are power equipment and failure is special
Sign identification, obtains the power equipment of failure；It is compared after the power equipment of failure is handled with infrared image fault database,
To identify the fault point of infrared image.A kind of infrared image failure based on YOLO algorithm of target detection provided by the present application
Point recognition methods combines YOLO algorithm of target detection, realizes the real-time detection identification to infrared image failure, substantially reduces
The error probability of infrared image fault identification, is more clear live effect.
Detailed description of the invention
In order to illustrate more clearly of the technical solution of the application, letter will be made to attached drawing needed in the embodiment below
Singly introduce, it should be apparent that, for those of ordinary skills, without any creative labor,
It is also possible to obtain other drawings based on these drawings.
Fig. 1 is the one of a kind of infrared image fault point recognition methods based on YOLO algorithm of target detection provided by the present application
The flow chart of embodiment；
Fig. 2 is a kind of the another of infrared image fault point recognition methods based on YOLO algorithm of target detection provided by the present application
The flow chart of one embodiment；
Fig. 3 is a kind of inspection of the infrared image fault point recognition methods based on YOLO algorithm of target detection provided by the present application
Survey schematic diagram.
Specific embodiment
Below with reference to the attached drawing in the application, the technical scheme in the embodiment of the application is clearly and completely described,
Obviously, described embodiment is only a part of the embodiment of the application, instead of all the embodiments.Based in the application
Embodiment, every other embodiment obtained by those of ordinary skill in the art without making creative efforts,
It shall fall within the protection scope of the present invention.
Many details are explained in the following description in order to fully understand the application, but the application can be with
It is different from the other modes that describe again using other to implement, those skilled in the art can be without prejudice to the application intension
In the case of do similar popularization, therefore the application is not limited by the specific embodiments disclosed below.
Have benefited from convolutional neural networks (Convolution Neural Network:CNN) and candidate region (Region
Proposal) algorithm target detection achieves huge breakthrough, and wherein YOLO is a kind of completely new object detection method, target
Determine and target identification is combined into one, really realizes the detection of end-to-end (End To End), YOLO object detection method
Also reach higher accuracy rate while realizing quick detection.
YOLO object detection method has used the thought of regression forecasting, by object detection task regard as target area prediction and
The regression problem of class prediction.This method is positioned candidate frame extraction, feature extraction, target classification, target using single nerve
It unites.Neural network directly extracts candidate region from image, directly predicted by entire image feature article boundary and
Class probability realizes target detection end to end.
YOLO object detection method predicted using single convolutional neural networks multiple bounding boxes (frame) and
Class probability.YOLO object detection method has relative to conventional method and has advantage as follows:
1, speed is fast.The pre- flow gauge of YOLO object detection method is simple, and speed is quickly.Base edition is on Titan X GPU
It can achieve 45 frames/s or more；Quick version can achieve 150 frames/s or more.Therefore, reality may be implemented in YOLO object detection method
When detect.
2, YOLO object detection method is predicted using full figure information.With sliding window method and region
Proposal-based method is different, and YOLO object detection method can use full figure information during training and prediction.
Plaque detection in background is target by Fast R-CNN (convolutional network method) detection method meeting mistake, and reason is Fast
R-CNN can not see global image in the detection.Relative to Fast R-CNN, YOLO object detection method background forecast error rate
It is at half.
It is a kind of infrared image fault point identification side based on YOLO algorithm of target detection provided by the present application referring to Fig. 1
The flow chart of one embodiment of method, this application provides a kind of infrared image fault point identification side based on YOLO algorithm of target detection
Method the described method comprises the following steps:
Step S101: infrared image is obtained, the prediction frame in the infrared image is set, and each prediction frame is equal
Comprising five-dimensional information, i.e., (x, y, w, h, object_conf), (x, y) is prediction frame center relative to the inclined of cell boundaries
It moves, (w, h) is the width height of frame relative to the ratio between entire image, and object_conf is the value of the confidence；
The each grid divided provides the conditional probability P_r (class | object) in the presence of infrared power equipment, from
And obtain the power equipment probability in whole picture in each small grid；
The value of the confidence, which is multiplied by the class probability of each grid with confidence level, to be calculated, and the value of the confidence can lead to
Following formula is crossed to be calculated:
In test, each frame is multiplied to obtain particular category the value of the confidence with confidence level by class probability, and setting is closed
Suitable judgment threshold, if the value of the confidence is higher than judgment threshold and assert that there are certain class targets in prediction frame:
Step S102: being arranged the probability in the infrared image there are power equipment, exists according in the infrared image
Judgment threshold is arranged in the probability of power equipment.
Step S103: judging whether the value of the confidence is greater than the judgment threshold, if the value of the confidence is greater than the judgement
Then there is power equipment in threshold value, execute step S104；Conversely, then terminating.
Step S104: processing is carried out to the infrared image there are power equipment and fault signature identifies, obtains failure
Power equipment.
Step S105: comparing after the power equipment of failure is handled with infrared image fault database, to identify
The fault point of infrared image out.
It referring to fig. 2, is a kind of infrared image fault point identification side based on YOLO algorithm of target detection provided by the present application
The flow chart of another embodiment of method, the described method comprises the following steps:
Step S201: infrared image is obtained, the infrared image is divided into the grid of S × S.
Step S202: the frame of setting each grid.
Step S203: the friendship of the frame and actual position of each grid and ratio are calculated, each grid is obtained
Predict frame.
Since many grids do not have power equipment presence, so to having the existing prediction frame of power equipment and no electricity
The existing prediction frame of power equipment is provided with different scale factors and is balanced；Calculate the frame and actual position of each grid
Friendship and ratio, obtain best prediction frame as standard, others think that power equipment is not present.
Each prediction frame and actual position are calculated, obtains best prediction frame as standard, others think nothing
Target exists.The design object of loss function (Loss Function) is exactly to allow coordinate and confidence level and classification
(classification) this three aspects reach good balance.According to it is described it is each prediction frame five-dimensional information (x,
Y, w, h, object_conf), the form of loss function such as following formula, testing principle is as shown in Figure 3:
First readjusting image and frame is 448 × 448 (as shown in the first width figures in Fig. 3), then by image
It is divided into 7 × 7 grid (as shown in the second width figure in Fig. 3), and grid cell is detected according to picture centre；So
Grid is trained afterwards, the probability and frame coordinate (as shown in third width figure in Fig. 3) of the type in predicted grid；1st to the 20th dimension
Probability and it is last 4 dimension frame coordinate x, y, w, h (as shown in the 4th width figure in Fig. 3).
Step S204: there are the probability of power equipment in setting each grid.
Step S205: according to the probability of power equipment in each grid, judgment threshold is set.
Step S206: according in each grid the judgment threshold and the value of the confidence, judge each net
Whether the value of the confidence in lattice is greater than the judgment threshold, if the value of the confidence is greater than the judgment threshold, is judged
There are power equipments in the grid, execute 207；Conversely, then terminating.
Step S207: the power equipment in the grid is extracted in the grid there are power equipment.
Step S208: the power equipment extracted is pre-processed, the power equipment that obtains that treated；
Selectable, the pretreatment includes gray processing processing being carried out to the power equipment that extracts, at binaryzation
Reason and denoising.
Step S209: to treated, power equipment carries out fault signature identification, obtains the power equipment of failure.
The feature except power equipment is misidentified in order to prevent, so only identifying the fault signature on power equipment.
Step S210: cutting is carried out to the fault point of the power equipment of the failure, is then contracted to the fault point
It puts and feature extraction.
Step S211: the fault point for extracting feature is compared with infrared image fault database, to identify red
The fault point of outer image.
Step S212: being trained infrared image failure, can be improved the accuracy rate of identification.
From the above technical scheme, this application provides a kind of infrared image failures based on YOLO algorithm of target detection
Point recognition methods, the described method comprises the following steps: obtaining infrared image, the prediction frame in the infrared image is arranged；If
The probability in the infrared image there are power equipment is set, there are the settings of the probability of power equipment to sentence according in the infrared image
Disconnected threshold value；Judge whether the value of the confidence is greater than the judgment threshold, if the value of the confidence is greater than the judgment threshold, exists
Power equipment performs the next step；Conversely, then terminating；Processing is carried out to the infrared image there are power equipment and failure is special
Sign identification, obtains the power equipment of failure；It is compared after the power equipment of failure is handled with infrared image fault database,
To identify the fault point of infrared image.A kind of infrared image failure based on YOLO algorithm of target detection provided by the present application
Point recognition methods combines YOLO algorithm of target detection, realizes the real-time detection identification to infrared image failure, substantially reduces
The error probability of infrared image fault identification, is more clear live effect.
The above is only the specific embodiments of the application, it is noted that those skilled in the art are come
It says, under the premise of not departing from the application principle, several improvements and modifications can also be made, these improvements and modifications also should be regarded as
The protection scope of the application.
Claims (9)
1. a kind of infrared image fault point recognition methods based on YOLO algorithm of target detection, which is characterized in that the method packet
Include following steps:
Step S1: infrared image is obtained, the prediction frame in the infrared image is set, and each prediction frame includes five
Information is tieed up, the five-dimensional information includes the offset for predicting frame center relative to each grid, and the five-dimensional information also wraps
The width height of the prediction frame is included relative to the ratio between infrared image described in whole picture, the five-dimensional information further includes the value of the confidence；
Step S2: being arranged the probability in the infrared image there are power equipment, according to there are electric power to set in the infrared image
Judgment threshold is arranged in standby probability；
Step S3: judging whether the value of the confidence is greater than the judgment threshold, if the value of the confidence is greater than the judgment threshold,
There are power equipments, execute step S4；Conversely, then terminating；
Step S4: processing is carried out to the infrared image there are power equipment and fault signature identifies, obtains the electric power of failure
Equipment；
Step S5: comparing after the power equipment of failure is handled with infrared image fault database, to identify infrared
The fault point of image.
2. the infrared image fault point recognition methods according to claim 1 based on YOLO algorithm of target detection, feature
It is, the acquisition infrared image, the prediction frame being arranged in the infrared image includes:
Step S11: infrared image is obtained, the infrared image is divided into the grid of S × S；
Step S12: the frame of setting each grid；
Step S13: the friendship of the frame and actual position of each grid and ratio are calculated, the prediction side of each grid is obtained
Frame.
3. as claimed in claim 2 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In the value of the confidence is obtained by the class probability and confidence calculations of each grid.
4. as claimed in claim 2 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In there are the probability of power equipment in the setting infrared image, according to there are power equipments in the infrared image
Judgment threshold is arranged in probability
Step S21: there are the probability of power equipment in setting each grid；
Step S22: according to the probability of power equipment in each grid, judgment threshold is set.
5. as claimed in claim 4 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In described to judge whether the value of the confidence is greater than the judgment threshold and includes:
According in each grid the judgment threshold and the value of the confidence, judge the confidence in each grid
Whether value is greater than the judgment threshold, if the value of the confidence is greater than the judgment threshold, exists in the grid judged
Power equipment performs the next step；Conversely, then terminating.
6. as claimed in claim 5 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In described pair there are the infrared images of power equipment to carry out processing and fault signature identification, obtains the power equipment of failure
Include:
Step S41: the power equipment in the grid is extracted in the grid there are power equipment；
Step S42: the power equipment extracted is pre-processed, the power equipment that obtains that treated；
Step S43: to treated, power equipment carries out fault signature identification, obtains the power equipment of failure.
7. as claimed in claim 6 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In, it is described the power equipment of failure is handled after compared with infrared image fault database, to identify infrared image
Fault point include:
Step S51: cutting is carried out to the fault point of the power equipment of the failure, then the fault point is zoomed in and out and special
Sign is extracted；
Step S52: the fault point for extracting feature is compared with infrared image fault database, to identify infrared image
Fault point.
8. as claimed in claim 6 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In the pretreatment includes carrying out gray processing processing, binary conversion treatment and denoising to the power equipment extracted.
9. as claimed in claim 7 based on the infrared image fault point recognition methods of YOLO algorithm of target detection, feature exists
In, it is described to compare the fault point for extracting feature with infrared image fault database, to identify the event of infrared image
It further include being trained to infrared image failure after barrier point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910089294.2A CN109784336A (en) | 2019-01-30 | 2019-01-30 | A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910089294.2A CN109784336A (en) | 2019-01-30 | 2019-01-30 | A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109784336A true CN109784336A (en) | 2019-05-21 |
Family
ID=66503771
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910089294.2A Withdrawn CN109784336A (en) | 2019-01-30 | 2019-01-30 | A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109784336A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110992307A (en) * | 2019-11-04 | 2020-04-10 | 华北电力大学（保定） | Insulator positioning and identifying method and device based on YOLO |
| CN112380952A (en) * | 2020-11-10 | 2021-02-19 | 广西大学 | Power equipment infrared image real-time detection and identification method based on artificial intelligence |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104331521A (en) * | 2014-11-27 | 2015-02-04 | 国家电网公司 | Automatic power transformation equipment abnormity identification method based on image processing |
| CN105868722A (en) * | 2016-04-05 | 2016-08-17 | 国家电网公司 | Identification method and system of abnormal power equipment images |
| CN107563412A (en) * | 2017-08-09 | 2018-01-09 | 浙江大学 | A kind of infrared image power equipment real-time detection method based on deep learning |
| CN107742093A (en) * | 2017-09-01 | 2018-02-27 | 国网山东省电力公司电力科学研究院 | A kind of infrared image power equipment component real-time detection method, server and system |
| US20180211117A1 (en) * | 2016-12-20 | 2018-07-26 | Jayant Ratti | On-demand artificial intelligence and roadway stewardship system |
| CN108711152A (en) * | 2018-05-23 | 2018-10-26 | 深圳市阡丘越科技有限公司 | A kind of image analysis method, device and user terminal based on AI technologies |
| CN109145713A (en) * | 2018-07-02 | 2019-01-04 | 南京师范大学 | A kind of Small object semantic segmentation method of combining target detection |
-
2019
- 2019-01-30 CN CN201910089294.2A patent/CN109784336A/en not_active Withdrawn
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104331521A (en) * | 2014-11-27 | 2015-02-04 | 国家电网公司 | Automatic power transformation equipment abnormity identification method based on image processing |
| CN105868722A (en) * | 2016-04-05 | 2016-08-17 | 国家电网公司 | Identification method and system of abnormal power equipment images |
| US20180211117A1 (en) * | 2016-12-20 | 2018-07-26 | Jayant Ratti | On-demand artificial intelligence and roadway stewardship system |
| CN107563412A (en) * | 2017-08-09 | 2018-01-09 | 浙江大学 | A kind of infrared image power equipment real-time detection method based on deep learning |
| CN107742093A (en) * | 2017-09-01 | 2018-02-27 | 国网山东省电力公司电力科学研究院 | A kind of infrared image power equipment component real-time detection method, server and system |
| CN108711152A (en) * | 2018-05-23 | 2018-10-26 | 深圳市阡丘越科技有限公司 | A kind of image analysis method, device and user terminal based on AI technologies |
| CN109145713A (en) * | 2018-07-02 | 2019-01-04 | 南京师范大学 | A kind of Small object semantic segmentation method of combining target detection |
Non-Patent Citations (4)
| Title |
|---|
| 吴晓烽等: "用于盲人视觉辅助的多目标快速识别并同步测距方法", 《福州大学学报(自然科学版)》 * |
| 沈茂东等: "基于RPN和FCN的电力设备锈迹检测", 《计算机与现代化》 * |
| 蔡成涛等: "基于改进YOLO算法的全景多目标实时检测", 《计算机工程与设计》 * |
| 魏钢等: "输变电设备红外故障诊断技术与试验研究", 《电气技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110992307A (en) * | 2019-11-04 | 2020-04-10 | 华北电力大学（保定） | Insulator positioning and identifying method and device based on YOLO |
| CN112380952A (en) * | 2020-11-10 | 2021-02-19 | 广西大学 | Power equipment infrared image real-time detection and identification method based on artificial intelligence |
| CN112380952B (en) * | 2020-11-10 | 2022-10-11 | 广西大学 | Power equipment infrared image real-time detection and identification method based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111353413A (en) | Low-missing-report-rate defect identification method for power transmission equipment | |
| CN104331710B (en) | On off state identifying system | |
| CN108108736A (en) | A kind of solar energy photovoltaic panel spot identification method | |
| CN110991448A (en) | Text detection method and device for nameplate image of power equipment | |
| CN107481237B (en) | A kind of infrared array image hot spot detection method based on multiframe temperature characterisitic | |
| CN105678760A (en) | Method for recognizing insulator image on the basis of Canny edge detection algorithm | |
| CN107179479B (en) | Visible light image-based transmission conductor strand scattering defect detection method | |
| CN110189344A (en) | A kind of line segment extraction method in power knife switch image | |
| CN109389019B (en) | Face image selection method and device and computer equipment | |
| CN109784336A (en) | A kind of infrared image fault point recognition methods based on YOLO algorithm of target detection | |
| CN111800430A (en) | Attack group identification method, device, equipment and medium | |
| CN112036450B (en) | High-voltage cable partial discharge mode identification method and system based on transfer learning | |
| CN113837991A (en) | Image processing method, device, equipment and storage medium | |
| CN111539924A (en) | Defect detection method, device and equipment for suspension clamp and storage medium | |
| CN116228670A (en) | Yolov 5-based photovoltaic defect detection method, device, computer equipment and medium | |
| CN113536944A (en) | Distribution line inspection data identification and analysis method based on image identification | |
| CN111310899B (en) | Power defect identification method based on symbiotic relation and small sample learning | |
| Dang et al. | Wireless multimedia sensor network for plant disease detections | |
| CN110853010B (en) | High-speed railway cable detection method based on FWA and SM | |
| CN111626104B (en) | Cable hidden trouble point detection method and device based on unmanned aerial vehicle infrared thermal image | |
| CN114140793A (en) | Matching method and device for terminal block and terminal block wiring | |
| CN110929800B (en) | Business body abnormal electricity utilization detection method based on sax algorithm | |
| CN114061476A (en) | Deflection detection method for insulator of power transmission line | |
| Zhang et al. | Snow loss modeling for solar modules using image processing and deep learning | |
| JP4723771B2 (en) | Lightning determination system and lightning determination method |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190521 |
|
| WW01 | Invention patent application withdrawn after publication |