CN113343392A - Intelligent detection and positioning method for oil conservator - Google Patents
Intelligent detection and positioning method for oil conservator Download PDFInfo
- Publication number
- CN113343392A CN113343392A CN202110760371.XA CN202110760371A CN113343392A CN 113343392 A CN113343392 A CN 113343392A CN 202110760371 A CN202110760371 A CN 202110760371A CN 113343392 A CN113343392 A CN 113343392A
- Authority
- CN
- China
- Prior art keywords
- conservator
- data
- oil
- module
- main body
- 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 43
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 17
- 238000013480 data collection Methods 0.000 claims abstract description 15
- 239000006096 absorbing agent Substances 0.000 claims abstract description 13
- 230000006870 function Effects 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 9
- 238000002372 labelling Methods 0.000 claims description 6
- 238000013136 deep learning model Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims 1
- 238000012360 testing method Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Geometry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Medical Informatics (AREA)
- Mathematical Analysis (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Housings And Mounting Of Transformers (AREA)
Abstract
The invention discloses an intelligent detection and positioning method for a conservator, which comprises a conservator main body and a conservator data collection and processing module, wherein the upper end of the conservator main body is respectively provided with an air release plug of an oil conservator and a moisture absorber pipe joint, the conservator data collection and processing module is internally composed of a conservator data acquisition unit, a conservator data arrangement unit and a conservator data marking unit together, the conservator data acquisition unit, the conservator data arrangement unit and the conservator data marking unit are electrically connected, the conservator data collection and processing module is in signal connection with a network input module, the network input module is in signal connection with a built conservator target detection network module, and the built conservator target detection network module is in signal connection with a conservator detection result module. According to the invention, the detection result after 5 classifications of the conservators is as follows: the mAP is 94%, and the average IoU is 84.3%, so the proposed method improves the accuracy of the conservator detection.
Description
Technical Field
The invention relates to the technical field of the power industry, in particular to an intelligent detection and positioning method for an oil conservator.
Background
At present, a method for detecting electric equipment mainly uses an infrared picture as a research basis to research the problem of abnormal heating of the electric equipment. The infrared picture comprises various pseudo colors, the shooting environment is complex, the interference of the shot equipment is serious, the training data types are few, and the like, so that the existing method has the defects of low detection accuracy, poor model generalization capability and the like.
The precise positioning of the conservator is closely related to the abnormal detection and structural division of the conservator. The invention designs a fast RCNN target detection model based on a regional suggestion network, which is used for improving the detection accuracy of a conservator.
Disclosure of Invention
The invention aims to provide an intelligent detection and positioning method for an oil conservator, which aims to solve the problems of low detection accuracy, poor model generalization capability and the like of the existing method provided in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a method for intelligently detecting and positioning a conservator comprises a conservator main body and a conservator data collecting and processing module, wherein an oil conservator air release plug and a moisture absorber pipe joint are respectively arranged at the upper end of the conservator main body, the oil conservator air release plug is positioned at one side of the moisture absorber pipe joint, manhole cover plates are arranged on the surfaces of two sides of the conservator main body, a pointer type oil level meter is arranged on the surface of each manhole cover plate, a viewing window is arranged on the surface of the front end of the conservator main body, the conservator data collecting and processing module is composed of an conservator data collecting unit, an conservator data arranging unit and an conservator data marking unit together, the conservator data collecting unit, the conservator data arranging unit and the conservator data marking unit are all electrically connected, the conservator data collecting and processing module is in signal connection with a network input module, and the network input module is in signal connection with a conservator target detection network model module, and the conservator target detection network model building module is in signal connection with the conservator detection result module.
Preferably, an oil outlet pipe is installed at the lower end of the oil conservator main body, and an oil outlet control valve is installed inside the oil outlet pipe.
Preferably, the conservator main body is divided into five categories, and the categories are divided according to the arrangement positions of components such as a pointer type oil level gauge, a manhole cover plate, an oil conservator air release plug, a moisture absorber pipe joint and a viewing window in the conservator equipment.
Preferably, for the above five categories of the conservators, 2000 pieces of data are screened and labeled for each category, and 5 × 2000 pieces of data are generated as a data set.
Preferably, the conservator target detection network model based on the fast RCNN is built, a deep learning model trained by adopting a tensrflow framework is constructed, a VGG16 is selected as a main feature extraction network, and pre-training is performed on ImageNet.
Preferably, the pre-trained VGG16 model is subjected to 1000 iterations on the data set in the conservator data collection processing module, the learning rate is set to 1e-4, the Batch size is 16, and the loss function selects the cross entropy loss function commonly used in image classification.
Preferably, before the network training, the temperature data input into the network is normalized, and finally, a target detection model of the conservator device is obtained, where the mapp is 94% and the average IoU is 84.3%.
Compared with the prior art, the invention has the beneficial effects that:
the conservators are divided into 5 types, and the types are divided into 5 types according to different placement positions of components such as a pointer type oil level gauge, a manhole cover plate, an oil conservator air release plug, a moisture absorber pipe joint, a viewing window and the like in the conservator equipment. For the above 5 conservator categories, 2000 pieces of data are screened and labeled for each category, and 5 × 2,000 pieces of data are generated as a data set. And building a conservator target detection network model based on Faster R-CNN. We adopt a Tensorflow framework to train a deep learning model, select VGG16 as a backbone feature extraction network, and pre-train on ImageNet. The pre-trained VGG16 model performed 1,000 iterations on the dataset, the learning rate was set to 1e-4, the Batch size was taken to be 16, and the loss function selected the Cross Entropy cross entropy loss function commonly used in image classification. Before the network training, the temperature data input into the network is standardized. And finally, obtaining a target detection model of the conservator device, wherein mAP is 94%, and average IoU is 84.3%. For 3,000 pieces of test data, the same experimental environment and network structure, the results of the test on a single category of conservator (not classified by 5) were: mAP 86%, average IoU 63.2%; and the detection result after 5 classifications are made to the conservator: mAP 94%, average IoU 84.3%. Therefore, the method improves the accuracy of the detection of the conservator.
Drawings
FIG. 1 is a front view of the main body of the conservator of the present invention;
FIG. 2 is a schematic view of the manhole cover plate structure of the present invention;
FIG. 3 is an overall flow chart of the method for building a conservator target detection network;
fig. 4 is a schematic block diagram of a conservator data collection processing module according to the present invention.
In the figure: 1. a viewing window; 2. a connector of the moisture absorber pipe; 3. a conservator main body; 4. an air release plug of the oil conservator; 5. a manhole cover plate; 6. an oil outlet pipe; 7. an oil outlet control valve; 8. a pointer type oil level meter; 9. the conservator data collection and processing module; 10. a network input module; 11. building a conservator target detection network model module; 12. a conservator detection result module; 13. an oil conservator data acquisition unit; 14. the conservator data arrangement unit; 15. and an oil conservator data labeling unit.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
Referring to fig. 1-4, an embodiment of the present invention is shown: an intelligent detection and positioning method for an oil conservator comprises an oil conservator main body 3 and an oil conservator data collection processing module 9, wherein the upper end of the oil conservator main body 3 is respectively provided with an oil conservator air release plug 4 and a moisture absorber pipe connector 2, the oil conservator air release plug 4 is positioned at one side of the moisture absorber pipe connector 2, the two side surfaces of the oil conservator main body 3 are provided with manhole cover plates 5, the surface of the manhole cover plate 5 is provided with a pointer type oil level meter 8, the front end surface of the oil conservator main body 3 is provided with a viewing window 1, the oil conservator data collection processing module 9 is internally composed of an oil conservator data acquisition unit 13, an oil conservator data sorting unit 14 and an oil conservator data labeling unit 15, the oil conservator data acquisition unit 13, the oil conservator data sorting unit 14 and the oil conservator data labeling unit 15 are all electrically connected, the oil conservator data collection processing module 9 is in signal connection with a network input module 10, the network input module 10 is in signal connection with the conservator target detection building network model module 11, and the conservator target detection building network model module 11 is in signal connection with the conservator detection result module 12.
Further, an oil outlet pipe 6 is installed at the lower end of the conservator main body 3, and an oil outlet control valve 7 is installed inside the oil outlet pipe 6.
Further, the conservator main body 3 is divided into five categories, and the categories are divided according to the different arrangement positions of the pointer type oil level gauge 8, the manhole cover plate 5, the oil conservator air release plug 4, the moisture absorber pipe joint 2, the observation window 1 and other components in the conservator device.
Further, 2000 pieces of data are screened and labeled for each of the five conservator categories, and 5 x 2000 pieces of data are generated to serve as a data set.
Further, a conservator target detection network model based on fast RCNN is built, a deep learning model trained by a Tensorflow frame is adopted, a VGG16 is selected as a main feature extraction network, and pre-training is carried out on ImageNet.
Further, the pre-trained VGG16 model performs 1000 iterations on the data set in the conservator data collection processing module 9, the learning rate is set to 1e-4, the Batch size is 16, and the loss function selects the cross entropy loss function commonly used in image classification.
Further, before network training, the temperature data input into the network is normalized, and finally, a target detection model of the conservator device is obtained, wherein the mAP is 94%, and the average IoU is 84.3%.
The working principle is as follows: when the device is used, the conservators are divided into 5 categories, and the categories are divided into 5 categories according to different placement positions of components such as a pointer type oil level gauge, a manhole cover plate, an air release plug of an oil conservator, a connector of a moisture absorber and a viewing window in the conservator device. For the above 5 conservator categories, 2000 pieces of data are screened and labeled for each category, and 5 × 2,000 pieces of data are generated as a data set. And building a conservator target detection network model based on Faster R-CNN. We adopt a Tensorflow framework to train a deep learning model, select VGG16 as a backbone feature extraction network, and pre-train on ImageNet. The pre-trained VGG16 model performed 1,000 iterations on the dataset, the learning rate was set to 1e-4, the Batch size was taken to be 16, and the loss function selected the Cross Entropy cross entropy loss function commonly used in image classification. Before the network training, the temperature data input into the network is standardized. And finally, obtaining a target detection model of the conservator device, wherein mAP is 94%, and average IoU is 84.3%. For 3,000 pieces of test data, the same experimental environment and network structure, the results of the test on a single category of conservator (not classified by 5) were: mAP 86%, average IoU 63.2%; and the detection result after 5 classifications are made to the conservator: mAP 94%, average IoU 84.3%. Therefore, the method improves the accuracy of the detection of the conservator.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Claims (7)
1. The intelligent detection and positioning method of the conservator comprises an conservator main body (3) and an conservator data collection and processing module (9), and is characterized in that the upper end of the conservator main body (3) is respectively provided with an oil conservator air release plug (4) and a moisture absorber pipe joint (2), the oil conservator air release plug (4) is positioned at one side of the moisture absorber pipe joint (2), the two side surfaces of the conservator main body (3) are provided with a manhole cover plate (5), the surface of the manhole cover plate (5) is provided with a pointer type oil level gauge (8), the front end surface of the conservator main body (3) is provided with a viewing window (1), the inside of the conservator data collection and processing module (9) is jointly composed of an conservator data collection unit (13), an conservator data arrangement unit (14) and an conservator data labeling unit (15), the conservator data collection unit (13), the conservator data arrangement unit (14) and the conservator data labeling unit (15) are all electrically connected, the conservator data collecting and processing module (9) is in signal connection with the network input module (10), the network input module (10) is in signal connection with the conservator target detection network model building module (11), and the conservator target detection network model building module (11) is in signal connection with the conservator detection result module (12).
2. The method for intelligently detecting and positioning the conservator as claimed in claim 1, wherein: an oil outlet pipe (6) is installed at the lower end of the oil conservator main body (3), and an oil outlet control valve (7) is installed inside the oil outlet pipe (6).
3. The method for intelligently detecting and positioning the conservator as claimed in claim 1, wherein: the conservator main body (3) is divided into five categories, and the categories are divided into five categories according to different placing positions of components such as a pointer type oil level meter (8), a manhole cover plate (5), an oil conservator air release plug (4), a moisture absorber pipe joint (2) and a viewing window (1) in conservator equipment.
4. The method for intelligently detecting and positioning the conservator as claimed in claim 3, wherein: and screening 2000 pieces of data for each of the five types of the conservators, and labeling to generate 5 x 2000 pieces of data as a data set.
5. The method for intelligently detecting and positioning the conservator as claimed in claim 4, wherein: the method comprises the steps of establishing a conservator target detection network model based on fast RCNN, adopting a Tensorflow framework to train a deep learning model, selecting a VGG16 as a main feature extraction network, and pre-training on ImageNet.
6. The method for intelligently detecting and positioning the conservator as claimed in claim 5, wherein: the pre-trained VGG16 model performs 1000 iterations on a data set in a conservator data collection processing module (9), the learning rate is set to be 1e-4, the size of Batch size is 16, and a cross entropy loss function commonly used in image classification is selected as the loss function.
7. The method for intelligently detecting and positioning the conservator as claimed in claim 6, wherein: before network training, the temperature data input into the network is standardized, and finally a target detection model of the conservator device is obtained, wherein mAP is 94%, and average IoU is 84.3%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110760371.XA CN113343392A (en) | 2021-07-06 | 2021-07-06 | Intelligent detection and positioning method for oil conservator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110760371.XA CN113343392A (en) | 2021-07-06 | 2021-07-06 | Intelligent detection and positioning method for oil conservator |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113343392A true CN113343392A (en) | 2021-09-03 |
Family
ID=77482596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110760371.XA Pending CN113343392A (en) | 2021-07-06 | 2021-07-06 | Intelligent detection and positioning method for oil conservator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113343392A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109489770A (en) * | 2018-11-29 | 2019-03-19 | 国网上海市电力公司 | Intelligent transformer oil level detection system based on pressure capsule system |
| WO2019109771A1 (en) * | 2017-12-05 | 2019-06-13 | 南京南瑞信息通信科技有限公司 | Power artificial-intelligence visual-analysis system on basis of multi-core heterogeneous parallel computing |
| CN110501055A (en) * | 2018-05-16 | 2019-11-26 | 国网福建省电力有限公司漳州供电公司 | The measurement method of installation for transformer oil level based on supersonic sounding |
| CN113052255A (en) * | 2021-04-07 | 2021-06-29 | 浙江天铂云科光电股份有限公司 | Intelligent detection and positioning method for reactor |
-
2021
- 2021-07-06 CN CN202110760371.XA patent/CN113343392A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019109771A1 (en) * | 2017-12-05 | 2019-06-13 | 南京南瑞信息通信科技有限公司 | Power artificial-intelligence visual-analysis system on basis of multi-core heterogeneous parallel computing |
| CN110501055A (en) * | 2018-05-16 | 2019-11-26 | 国网福建省电力有限公司漳州供电公司 | The measurement method of installation for transformer oil level based on supersonic sounding |
| CN109489770A (en) * | 2018-11-29 | 2019-03-19 | 国网上海市电力公司 | Intelligent transformer oil level detection system based on pressure capsule system |
| CN113052255A (en) * | 2021-04-07 | 2021-06-29 | 浙江天铂云科光电股份有限公司 | Intelligent detection and positioning method for reactor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110598736B (en) | Power equipment infrared image fault positioning, identifying and predicting method | |
| CN112258496A (en) | Underground drainage pipeline disease segmentation method based on full convolution neural network | |
| CN105004972B (en) | Porcelain insulator Condition assessment of insulation method based on day blind ultraviolet imagery characteristics of image | |
| CN102445640B (en) | GIS device intelligent recognition method based on vector machine and artificial fish swarm optimization | |
| CN112183667B (en) | Insulator fault detection method in cooperation with deep learning | |
| CN103439342B (en) | Based on the Infrared Non-destructive Testing method of thermal map temporal aspect | |
| CN106981063A (en) | A kind of grid equipment state monitoring apparatus based on deep learning | |
| CN111127449B (en) | Automatic crack detection method based on encoder-decoder | |
| CN109376792A (en) | Photovoltaic cell open defect classification method based on multichannel residual error neural network | |
| CN108875719B (en) | Air cooler dust accumulation state sensing system and calculation method based on deep learning and infrared image recognition | |
| CN109784326A (en) | A kind of vehicle chassis detection method based on deep learning | |
| CN110070530A (en) | A kind of powerline ice-covering detection method based on deep neural network | |
| CN107092906A (en) | A kind of Chinese traditional medicinal materials recognition device based on deep learning | |
| CN107490769B (en) | A kind of fuel cell water distribution visualization on-line detecting system | |
| CN109345586A (en) | Electrical equipment discharge characteristic extracting method based on ultraviolet imagery technology | |
| CN106709903A (en) | PM2.5 concentration prediction method based on image quality | |
| CN117788950A (en) | Pavement disease detection method based on improved YOLOv8 model | |
| CN205157546U (en) | Air quality test equipment with quality class appraises function | |
| CN109884049B (en) | Harmful substance detection device capable of detecting kitchen fume | |
| CN113343392A (en) | Intelligent detection and positioning method for oil conservator | |
| CN112330675B (en) | Traffic road image atmospheric visibility detection method based on AOD-Net | |
| CN112132088A (en) | Inspection point location missing inspection identification method | |
| CN117352088A (en) | Prediction method of spatial pollutant distribution based on convolutional neural network | |
| CN116704387A (en) | Power line channel inspection system and method based on video structuring | |
| CN110910352A (en) | Solar cell defect detection system and detection method based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210903 |
|
| RJ01 | Rejection of invention patent application after publication |