CN110705414A - Power transmission line construction machinery hidden danger detection method based on deep learning - Google Patents
Power transmission line construction machinery hidden danger detection method based on deep learning Download PDFInfo
- Publication number
- CN110705414A CN110705414A CN201910903974.3A CN201910903974A CN110705414A CN 110705414 A CN110705414 A CN 110705414A CN 201910903974 A CN201910903974 A CN 201910903974A CN 110705414 A CN110705414 A CN 110705414A
- Authority
- CN
- China
- Prior art keywords
- hidden danger
- construction machinery
- power transmission
- transmission line
- picture
- 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
- 238000010276 construction Methods 0.000 title claims abstract description 57
- 230000005540 biological transmission Effects 0.000 title claims abstract description 46
- 238000001514 detection method Methods 0.000 title claims abstract description 38
- 238000013135 deep learning Methods 0.000 title claims abstract description 15
- 238000012544 monitoring process Methods 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000003062 neural network model Methods 0.000 claims abstract description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000010410 layer Substances 0.000 description 9
- 238000007689 inspection Methods 0.000 description 3
- 238000011176 pooling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of power transmission line detection, in particular to a power transmission line construction machinery hidden danger detection method based on deep learning, which comprises the following steps: a. the camera shoots pictures in the range of the power transmission line and transmits the pictures to the server through the 4G network; b. the method comprises the steps that a picture containing hidden danger construction machinery is obtained and marked at a server side, a neural network model is used for training the marked picture, and a construction machinery monitoring model is obtained for detection; c. deploying a monitoring model of the construction machine into a server to load model parameters, detecting whether hidden danger exists in a picture newly shot and uploaded by a camera or not, alarming if the hidden danger exists, and continuously detecting the picture newly shot and uploaded by the camera if the hidden danger does not exist; the invention uses a two-stage target detection algorithm, performs targeted improvement work and realizes high-precision channel hidden danger target detection. In the image data concentration of the power transmission channel of the national power grid, the accuracy rate of the hidden danger detection is improved to 87%.
Description
Technical Field
The invention relates to the technical field of power transmission line detection, in particular to a power transmission line construction machinery hidden danger detection method based on deep learning.
Background
In the power grid industry, the safety problem of the power transmission line is always of great importance. Whether the transmission line is safe or not has important significance for safe and reliable operation of the power grid, so that the operation state of the transmission line needs to be monitored regularly. However, the technological process is accelerated, so that the mechanical construction gradually extends into the channel of the power transmission line, the safe operation of the power transmission line is seriously threatened, and the power transmission line generates huge potential safety hazards. On the other hand, the current maintainer adopts the manual line patrol mode to monitor the power transmission line, the manual line patrol has a patrol vacuum period, the line corridor change condition cannot be timely mastered, and the power transmission line has large longitudinal and transverse span and complex distribution topography. The line state parameters are various, many links are not easy to be found manually, manpower is greatly consumed, and efficiency is reduced.
In summary, how to provide an efficient and reliable detection system and method for a power transmission channel construction machine to provide technical support for detection of hidden troubles of a power transmission line is a problem to be solved urgently by technical personnel in the field at present.
Disclosure of Invention
In order to solve the deficiencies in the above technical problems, the present invention aims to: the method for detecting the hidden danger of the power transmission line construction machinery based on deep learning is provided, accurate and efficient detection of the hidden danger construction machinery of the power transmission line is achieved, the cost of manual inspection of a power transmission channel is lowered, the efficiency of inspection of the hidden danger is improved, and intelligent management of power grid operation is facilitated.
The technical scheme adopted by the invention for solving the technical problem is as follows:
the method for detecting the hidden danger of the power transmission line construction machinery based on the deep learning comprises the following steps:
a. the camera shoots pictures in the range of the power transmission line and transmits the pictures to the server through the 4G network;
b. the method comprises the steps that a picture containing hidden danger construction machinery is obtained and marked at a server side, a neural network model is used for training the marked picture, and a construction machinery monitoring model is obtained for detection;
c. and deploying the monitoring model of the construction machine into a server to load model parameters, detecting whether hidden danger exists in the picture newly shot and uploaded by the camera or not, alarming if the hidden danger exists, and continuously detecting the picture newly shot and uploaded by the camera if the hidden danger does not exist.
Preferably, the step a of transmitting the picture to the server through the 4G network means that the camera includes an access and control device of the 4G network, the control device automatically transmits the monitoring picture captured by the camera to the server, and the access device of the 4G network can be connected to a wireless network of a telecommunication operator to remotely transmit the picture.
Preferably, the specific method of step b is: marking the power transmission channel construction machinery image to be detected, training a construction machinery monitoring model by using a deep learning framework Tensorflow, wherein the input of the construction machinery monitoring model is a monitoring picture, and the output of the construction machinery monitoring model is the pixel position and the type of the construction machinery contained in the picture.
Preferably, the neural network model in the step b is a more advanced object detection algorithm which is more suitable for construction machinery, the algorithm is improved based on a fast-RCNN two-stage object detection algorithm, a thinner feature layer and a smaller R-CNN sub-network are used, and the thinner feature layer not only improves the improvement precision, but also saves the memory and the calculation amount during training and use.
Preferably, the specific mode of step c is to operate the power transmission line channel hidden danger detection model in the server, wait for the camera to upload the monitoring picture, automatically perform hidden danger detection once the picture is uploaded, and store the analysis result.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention can rapidly realize the inspection of the hidden trouble construction machinery on the transmission line, and the detection speed on the GPU reaches hundreds of milliseconds in an embedded processor platform in the monitoring camera.
2. The invention uses a two-stage target detection algorithm, performs targeted improvement work and realizes high-precision channel hidden danger target detection. In the image data concentration of the power transmission channel of the national power grid, the accuracy rate of the hidden danger detection is improved to 87%.
3. Compared with a common target detection algorithm, the method has the following innovation points in ZHYNet:
(1) the backbone network (backbone) covers objects of various sizes based on the FPN method;
(2) RPN outputs thin feature layers (thin feature maps) to accelerate model reasoning;
(3) the head uses a single-layer RCNN sub-network, so that the weight is reduced, and overfitting is avoided;
(4) according to the characteristics of the data set, a special data enhancement method and a detailed parameter analysis and comparison test are realized.
Drawings
FIG. 1 is a flow chart of the present invention.
Fig. 2, global context module structure diagram.
Detailed Description
The embodiments of the present invention are further described below in conjunction with:
example 1
As shown in fig. 1-2, the method for detecting hidden danger of construction machinery of power transmission line based on deep learning of the invention comprises the following steps:
a. the camera shoots pictures in the range of the power transmission line and transmits the pictures to the server through the 4G network; specifically, the step of transmitting the picture to the server through the 4G network means that the camera includes an access and control device of the 4G network, the control device automatically transmits the monitoring picture captured by the camera to the server, and the access device of the 4G network can be connected to a wireless network of a telecommunication operator to remotely transmit the picture.
b. The method comprises the steps that a picture containing hidden danger construction machinery is obtained and marked at a server side, a neural network model is used for training the marked picture, and a construction machinery monitoring model is obtained for detection; specifically, a power transmission channel construction machine image to be detected is labeled, a construction machine monitoring model is trained by using a deep learning framework Tensorflow, the input of the construction machine monitoring model is a monitoring picture, and the output of the construction machine monitoring model is the pixel position and the type of the construction machine contained in the picture. The neural network model is an advanced object detection algorithm more suitable for construction machinery, the algorithm is improved based on a fast-RCNN two-stage object detection algorithm, a thinner feature layer and a smaller R-CNN sub-network are used, and the thinner feature layer not only improves the precision, but also saves the memory and the calculated amount during training and use. An improved two-stage object detection framework model, referred to herein as ZHYNet. The overall structure of the ZHYNet is shown in fig. 1, a larger receptive field is efficiently brought to a detection network by using separable large convolution, a global context information module shown in fig. 2 is introduced, the receptive field can be increased by increasing the value of k, and the global context information is widely applied to a classification task. In addition, the feature layer for the RoI Pooling is thinned, namely, the feature map channels used by the RoI Pooling are compressed to be small, on one hand, the number of the channels of the feature map is reduced, and on the other hand, the number of the channels of the R-CNN part input feature map is reduced. This model also uses a lightweight R-CNN, unlike Faster-RCNN, which uses two powerful fully connected layers to regress and classify each candidate box. In ZHYNet, RCNN uses only one fully connected layer. By combining the "RoI-pooled thin feature layers", the second stage of the two-stage object detector really achieves low time consumption, thereby achieving overall network acceleration.
c. And deploying the monitoring model of the construction machine into a server to load model parameters, detecting whether hidden danger exists in the picture newly shot and uploaded by the camera or not, alarming if the hidden danger exists, and continuously detecting the picture newly shot and uploaded by the camera if the hidden danger does not exist. The specific method is that a power transmission line channel hidden danger detection model is operated in a server, monitoring pictures are uploaded by a camera, once the pictures are uploaded, hidden danger detection is automatically carried out, and analysis results are stored.
Example 2
On the basis of embodiment 1, in the detection system of the construction machinery of a certain province power transmission line, 44000 field pictures are captured together within a certain time period in the process of capturing the power transmission line region by a camera, wherein 18596 pictures containing hidden danger construction machinery and 25404 pictures not containing construction machinery are included, and by using the detection method of the construction machinery, the identification false alarm rate of the hidden danger of the construction machinery is finally obtained and is 5.2%, the identification false alarm rate is 7.2% and the accuracy is 88.1%.
a. Training 18596 marked data sets containing hidden danger construction machinery by using the neural network model provided by the invention, and repeatedly performing data iterative training to finally obtain a construction machinery detection model;
b. acquiring 2000 uploaded pictures newly captured by a camera from a server as a verification set, loading a model in the server, and testing the performance of the verification set by using the trained model;
c. and detecting whether the power transmission line has hidden danger construction machinery, alarming if the power transmission line has hidden danger construction machinery, continuously detecting the newly shot and uploaded picture of the monitoring camera if the power transmission line does not have hidden danger construction machinery, and acquiring the recognition missing report rate, the recognition false report rate and the accuracy rate.
In the embodiment 1, the calculated recognition missing report rate is 5.2%, the recognition false report rate is 7.2%, and the recognition accuracy rate is 88.1%. The calculated recognition missing report rate, recognition false report rate and accuracy all meet the technical requirements.
Example 3
On the basis of embodiment 1, in a detection system of construction machinery of a power transmission line in a certain province, 44000 field pictures are commonly captured within a certain time period in the process of capturing a power transmission line area by a camera.
The steps similar to those of the embodiment 1 are adopted, all samples in the training set are used for repeated iterative training, the calculation and recognition missing report rate is 5.9%, the recognition false report rate is 6.8%, and the recognition accuracy rate is 85.5% in the embodiment 2, and the technical requirements are met.
Claims (5)
1. A method for detecting hidden danger of construction machinery of a power transmission line based on deep learning is characterized by comprising the following steps:
a. the camera shoots pictures in the range of the power transmission line and transmits the pictures to the server through the 4G network;
b. the method comprises the steps that a picture containing hidden danger construction machinery is obtained and marked at a server side, a neural network model is used for training the marked picture, and a construction machinery monitoring model is obtained for detection;
c. and deploying the monitoring model of the construction machine into a server to load model parameters, detecting whether hidden danger exists in the picture newly shot and uploaded by the camera or not, alarming if the hidden danger exists, and continuously detecting the picture newly shot and uploaded by the camera if the hidden danger does not exist.
2. The method for detecting the hidden danger of the construction machinery of the power transmission line based on the deep learning of the claim 1 is characterized in that the step a of transmitting the pictures into the server through the 4G network means that the camera comprises an access device and a control device of the 4G network, the control device automatically transmits the monitoring pictures captured by the camera into the server, and the access device of the 4G network can be connected to a wireless network of a telecommunication operator to remotely transmit the pictures.
3. The method for detecting hidden danger of construction machinery of power transmission line based on deep learning of claim 1, wherein the specific method in step b is as follows: marking the power transmission channel construction machinery image to be detected, training a construction machinery monitoring model by using a deep learning framework Tensorflow, wherein the input of the construction machinery monitoring model is a monitoring picture, and the output of the construction machinery monitoring model is the pixel position and the type of the construction machinery contained in the picture.
4. The method for detecting the hidden danger of the power transmission line construction machinery based on the deep learning of claim 1, wherein the neural network model in the step b is a more advanced object detection algorithm more suitable for the construction machinery, the algorithm is improved based on a fast-RCNN two-stage object detection algorithm, a thinner feature layer and a smaller R-CNN sub-network are used, the thinner feature layer not only improves the improvement precision, but also saves the memory and the calculation amount during training and use.
5. The method for detecting the hidden danger of the construction machinery of the power transmission line based on the deep learning of the claim 1 is characterized in that the specific mode of the step c is to operate a detection model of the hidden danger of the power transmission line channel in a server, wait for a camera to upload a monitoring picture, automatically detect the hidden danger once the picture is uploaded, and store an analysis result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910903974.3A CN110705414A (en) | 2019-09-24 | 2019-09-24 | Power transmission line construction machinery hidden danger detection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910903974.3A CN110705414A (en) | 2019-09-24 | 2019-09-24 | Power transmission line construction machinery hidden danger detection method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110705414A true CN110705414A (en) | 2020-01-17 |
Family
ID=69195616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910903974.3A Pending CN110705414A (en) | 2019-09-24 | 2019-09-24 | Power transmission line construction machinery hidden danger detection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110705414A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111611984A (en) * | 2020-06-30 | 2020-09-01 | 山西振中电力股份有限公司 | Intelligent detection device and method for three-span-point construction large machinery of power transmission line |
| CN112613453A (en) * | 2020-12-29 | 2021-04-06 | 国网山东省电力公司建设公司 | Method and system for checking violation of regulations on construction site of electric power infrastructure |
| CN112906454A (en) * | 2020-12-22 | 2021-06-04 | 安徽康能电气有限公司 | Power transmission channel AI visual monitoring method and device |
| CN115797357A (en) * | 2023-02-10 | 2023-03-14 | 智洋创新科技股份有限公司 | Transmission channel hidden danger detection method based on improved YOLOv7 |
| CN116597390A (en) * | 2023-07-18 | 2023-08-15 | 南方电网数字电网研究院有限公司 | Method and device for detecting construction hidden danger around power transmission line and computer equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5712796A (en) * | 1994-07-14 | 1998-01-27 | Hitachi Cable, Ltd. | Method for evaluating the faulted sections and states in a power transmission line |
| US20070124093A1 (en) * | 2005-11-05 | 2007-05-31 | Myongji University Industry And Academia Cooperation Foundation | Method for locating line-to-ground fault point of underground power cable system |
| CN108307146A (en) * | 2017-12-12 | 2018-07-20 | 张宝泽 | A kind of ultra-high-tension power transmission line Security Vulnerability Detecting System and method |
| CN108734225A (en) * | 2018-06-08 | 2018-11-02 | 浙江大学 | A kind of transmission line construction subject image detection method based on deep learning |
| CN108921826A (en) * | 2018-06-13 | 2018-11-30 | 山东信通电子股份有限公司 | The transmission line of electricity that super-pixel segmentation is combined with deep learning invades object detecting method |
| CN109977921A (en) * | 2019-04-11 | 2019-07-05 | 广东电网有限责任公司 | A kind of transmission line of electricity perils detecting method |
| CN110008901A (en) * | 2019-04-04 | 2019-07-12 | 天津工业大学 | A kind of insulator breakdown identification and localization method based on Mask R-CNN |
-
2019
- 2019-09-24 CN CN201910903974.3A patent/CN110705414A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5712796A (en) * | 1994-07-14 | 1998-01-27 | Hitachi Cable, Ltd. | Method for evaluating the faulted sections and states in a power transmission line |
| US20070124093A1 (en) * | 2005-11-05 | 2007-05-31 | Myongji University Industry And Academia Cooperation Foundation | Method for locating line-to-ground fault point of underground power cable system |
| CN108307146A (en) * | 2017-12-12 | 2018-07-20 | 张宝泽 | A kind of ultra-high-tension power transmission line Security Vulnerability Detecting System and method |
| CN108734225A (en) * | 2018-06-08 | 2018-11-02 | 浙江大学 | A kind of transmission line construction subject image detection method based on deep learning |
| CN108921826A (en) * | 2018-06-13 | 2018-11-30 | 山东信通电子股份有限公司 | The transmission line of electricity that super-pixel segmentation is combined with deep learning invades object detecting method |
| CN110008901A (en) * | 2019-04-04 | 2019-07-12 | 天津工业大学 | A kind of insulator breakdown identification and localization method based on Mask R-CNN |
| CN109977921A (en) * | 2019-04-11 | 2019-07-05 | 广东电网有限责任公司 | A kind of transmission line of electricity perils detecting method |
Non-Patent Citations (2)
| Title |
|---|
| ZEMING LI等: "Light-Head R-CNN: In Defense of Two-Stage Object Detector", 《COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 吕宁: "基于深度学习的输电线路防外力破坏监测系统研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111611984A (en) * | 2020-06-30 | 2020-09-01 | 山西振中电力股份有限公司 | Intelligent detection device and method for three-span-point construction large machinery of power transmission line |
| CN112906454A (en) * | 2020-12-22 | 2021-06-04 | 安徽康能电气有限公司 | Power transmission channel AI visual monitoring method and device |
| CN112613453A (en) * | 2020-12-29 | 2021-04-06 | 国网山东省电力公司建设公司 | Method and system for checking violation of regulations on construction site of electric power infrastructure |
| CN115797357A (en) * | 2023-02-10 | 2023-03-14 | 智洋创新科技股份有限公司 | Transmission channel hidden danger detection method based on improved YOLOv7 |
| CN116597390A (en) * | 2023-07-18 | 2023-08-15 | 南方电网数字电网研究院有限公司 | Method and device for detecting construction hidden danger around power transmission line and computer equipment |
| CN116597390B (en) * | 2023-07-18 | 2023-12-12 | 南方电网数字电网研究院有限公司 | Method and device for detecting construction hidden danger around power transmission line and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110705414A (en) | Power transmission line construction machinery hidden danger detection method based on deep learning | |
| CN108189043B (en) | Inspection method and inspection robot system applied to high-speed rail machine room | |
| CN108288055A (en) | Block of bow collector of electric locomotive based on depth network and placement test and arc method for measuring | |
| CN112115927B (en) | Intelligent machine room equipment identification method and system based on deep learning | |
| CN110516529A (en) | It is a kind of that detection method and system are fed based on deep learning image procossing | |
| KR102346676B1 (en) | Method for creating damage figure using the deep learning-based damage image classification of facility | |
| CN108711148A (en) | A kind of wheel tyre defect intelligent detecting method based on deep learning | |
| US11769333B2 (en) | Method and system for detecting block falling disaster in a tunnel based on visual and mechanical perception | |
| CN111401437A (en) | Deep learning-based power transmission channel hidden danger early warning grade analysis method | |
| CN113205039A (en) | Power equipment fault image identification and disaster investigation system and method based on multiple DCNNs | |
| CN114565845A (en) | Intelligent inspection system for underground tunnel | |
| CN113487821A (en) | Power equipment foreign matter intrusion identification system and method based on machine vision | |
| CN113988573A (en) | Risk judgment method, system and medium for routing inspection unmanned aerial vehicle based on power system | |
| CN117235443A (en) | Electric power operation safety monitoring method and system based on edge AI | |
| CN116624225A (en) | Tunnel defect picture and text inspection information real-time transmission early warning system | |
| CN109522845A (en) | Distribution transformer based on intelligent robot tests safety work measure of supervision | |
| CN114627388A (en) | Power transmission line foreign matter detection equipment and foreign matter detection method thereof | |
| CN111931748B (en) | Worker fatigue detection method suitable for storage battery production workshop | |
| CN117686033A (en) | Intelligent recognition system for coal mine robot | |
| CN113776783A (en) | Machine room server fault lamp detection method based on inspection robot | |
| CN113408630A (en) | Transformer substation indicator lamp state identification method | |
| CN117114420A (en) | Image recognition-based industrial and trade safety accident risk management and control system and method | |
| CN109902647A (en) | The portable online Bird's Nest intelligent identification Method of one kind and system | |
| CN114355083A (en) | Inspection robot fault identification method and system based on artificial intelligence algorithm | |
| CN115201213A (en) | Passenger ropeway steel wire rope detection device and 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200117 |
|
| RJ01 | Rejection of invention patent application after publication |