CN111008641B - Power transmission line tower external force damage detection method based on convolutional neural network - Google Patents
Power transmission line tower external force damage detection method based on convolutional neural network Download PDFInfo
- Publication number
- CN111008641B CN111008641B CN201911019078.7A CN201911019078A CN111008641B CN 111008641 B CN111008641 B CN 111008641B CN 201911019078 A CN201911019078 A CN 201911019078A CN 111008641 B CN111008641 B CN 111008641B
- Authority
- CN
- China
- Prior art keywords
- tower
- external force
- remote sensing
- model
- neural network
- 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.)
- Active
Links
Images
Classifications
-
- 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
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention relates to a power transmission line tower external force damage detection method based on a convolutional neural network, and belongs to the technical field of satellite remote sensing image processing in operation and maintenance and application of power transmission lines. The detection method comprises the steps of firstly preprocessing an original remote sensing image, and dividing a training set, a verification set and a test set from the original remote sensing image; the convolutional neural network is learned through training samples and optimized by using an Adam algorithm. And after the optimal model is obtained through training, inputting complete image data into the model, finally obtaining a detection result of each base tower in the remote sensing image, and judging whether external force damage exists or not. The detection method can reduce line faults caused by external force damage, timely discover possible damage to the power transmission line tower through satellite remote sensing detection, manage and control as soon as possible, and reduce tripping and outage of the power transmission line caused by external force damage.
Description
Technical Field
The invention belongs to the technical field of satellite remote sensing image processing in operation and maintenance and application of a power transmission line, and particularly relates to a power transmission line tower external force damage detection method based on a convolutional neural network, which is used for observing whether a tower is damaged by external force.
Background
With the development and maturity of satellite remote sensing technology, the picture information in the remote sensing image plays an important role in the fields of geological exploration, environmental monitoring and the like. The classification technology of the remote sensing images is one of the concerned application directions all the time, and the purpose of the classification technology is to accurately judge the surface object type corresponding to each pixel point in the remote sensing images. However, in practical applications, it is very challenging to obtain high classification accuracy due to the limitation of resolution.
In recent years, deep neural networks, particularly convolutional neural networks, have remarkable performances in the field of natural images, and hyperspectral image classification methods based on convolutional neural networks are also proposed continuously. A great deal of research results show that the feature extraction and learning capacity of the convolutional neural network is superior to that of the traditional feature extraction method, and the feature extraction mode is more scientific and accurate.
The transmission line towers are wide in distribution area, severe in environment along the line, various in external force damage distribution points, various in types, and poor in coverage area and timeliness in the existing detection technology. Meanwhile, when technologies in various fields develop towards automation and intellectualization, tower external force damage detection also depends on manual investigation, which wastes time and labor, so that how to overcome the defects of the prior art is a problem to be solved in the field.
Disclosure of Invention
The invention aims to provide a convolution neural network-based external force damage detection method for remote sensing images of a power transmission line tower, aiming at the conditions of high maintenance cost, insufficient timeliness and the like of the power transmission line tower caused by the limitation of the prior art. The damage analysis of the tower is carried out through the satellite remote sensing image, so that the loss caused by not timely discovering and processing the damage of the tower can be reduced while the manual investigation cost is saved.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a method for detecting external force damage of a power transmission line tower based on a convolutional neural network comprises the following steps:
step (1), obtaining original data of a remote sensing image;
step (2), preprocessing the remote sensing images, separating and dividing images of all towers to obtain a training set, a verification set and a test set with labels, wherein the labels are divided into two types, namely the transmission line tower is damaged by external force and the transmission line tower is not damaged by the external force;
step (3), a convolutional neural network model is constructed, each parameter is initialized, and a training set is input into the network for learning;
step (4), after the training set is input into the model, an output value is obtained through forward propagation, and the output value is the probability that the tower is damaged; calculating a cost function to measure the error between the output value and the true value, then feeding back through a back propagation error, updating the parameter of the convolution kernel and continuing training; after the first round of training is finished, taking the verification set as input, operating the model and observing the result, and repeating the operation for multiple times until the convergence of the cost function is observed to obtain a neural network model with the optimal current parameters;
step (5), inputting the tower remote sensing image concentrated in the test into the trained optimal model to obtain a classification result of whether the tower is damaged by external force or not, and evaluating the accuracy of the model;
and (6) when the accuracy of the model reaches the standard, the obtained model can be used for new tower external force damage inspection.
Further, it is preferable that the specific process of step (2) is:
step (2.1), processing the satellite remote sensing image: intercepting the position of the tower, and ensuring that the position of the tower is in the center of each picture;
step (2.2), dividing a training set: 60% of the towers in one line are taken as a training set, 20% of the towers are taken as a verification set, and 20% of the towers are taken as a test set. (but not limited thereto, it may be divided into 30%, and 40%)
And when the training set is divided, a random mode is adopted.
Further, preferably, the last layer in the step (4) adopts logistic regression, and the output value of each neuron corresponds to the probability of whether the tower is damaged or not; the cost function adopts a cross entropy loss function, and the expression is
: whereinyRepresenting a true value, and a is a model output value; the error is then back propagated through the Adam algorithm, updating the parameters of all convolution kernels so that the cost function tends to converge. The labels are divided into two types, namely, the transmission line tower is damaged by external force and the transmission line tower is not damaged by the external force; the label "1" can be used to indicate that the material is damaged by external force, and "0" can be used to indicate that the material is not damaged by external force.
Further, it is preferable that in step (5), when the accuracy is higher than 80%, the model is fixed, and if the accuracy is lower than 80%, 25% of data in the test set is input into the network for retraining. However, the accuracy and the like may be set according to actual conditions.
The parameters initialized in the step (3) of the invention comprise network layer, hidden layer unit number and the like.
Compared with the prior art, the invention has the beneficial effects that:
1. the human cost is reduced, the frequency of manual investigation can be obviously reduced through the satellite remote sensing detection, and the investment of manpower and material resources is reduced.
2. The line faults caused by external damage are reduced, the damage to the power transmission line tower possibly caused is timely found through satellite remote sensing detection, the control is carried out as soon as possible, and the tripping and outage of the power transmission line caused by the external damage are reduced.
3. In 2015 to 2019, external force damage faults occur for 220kV and above power transmission lines of southern power grid companies for 436 times, accounting for 18.7% of the total number of faults, 274 times of outage accounting for 35.12% of the total number of outage, and the external force damage faults tend to increase year by year, and become main reasons of outage of the power transmission lines.
Drawings
FIG. 1 is a flowchart of a method for detecting external force damage of a power transmission line tower based on a convolutional neural network;
FIG. 2 is a convolutional neural network model constructed in step (3) of the present invention.
The specific thinking of the model is as follows: convolution-activation-pooling-full join-classification;
Detailed Description
The present invention will be described in further detail with reference to examples.
It will be appreciated by those skilled in the art that the following examples are illustrative of the invention only and should not be taken as limiting the scope of the invention. The examples do not specify particular techniques or conditions, and are performed according to the techniques or conditions described in the literature in the art or according to the product specifications. The materials or equipment used are not indicated by manufacturers, but are all conventional products available by purchase.
As shown in fig. 1, a method for detecting external force damage of a power transmission line tower based on a convolutional neural network comprises the following steps:
step (1), obtaining original data of a remote sensing image;
step (2), preprocessing the remote sensing image, separating and dividing images of each tower to obtain a training set, a verification set and a test set with labels, wherein the labels are divided into two types, namely, the transmission line tower is damaged by external force and the transmission line tower is not damaged by the external force;
step (3), a convolutional neural network model is constructed, each parameter is initialized, and a training set is input into the network for learning;
step (4), after the training set is input into the model, an output value is obtained through forward propagation, and the output value is the probability that the tower is damaged; calculating a cost function to measure the error between the output value and the true value, then feeding back through back propagation error, updating the parameter of the convolution kernel and continuing training; after the first round of training is finished, taking the verification set as input, operating the model and observing the result, and repeating the operation for multiple times until the convergence of the cost function is observed to obtain a neural network model with the optimal current parameters;
inputting the tower remote sensing graph with the concentrated test into a trained optimal model to obtain a classification result of whether the tower is damaged by external force or not, and evaluating the accuracy of the model;
and (6) when the accuracy of the model reaches the standard, the obtained model can be used for new tower external force damage inspection.
The specific process of the step (2) is as follows:
step (2.1), processing the satellite remote sensing image: intercepting the position of the tower, and ensuring that the position of the tower is in the center of each picture;
step (2.2), dividing a training set: 60% of the towers in one line are taken as a training set, 20% of the towers are taken as a verification set, and 20% of the towers are taken as a test set.
And when the training set, the verification set and the test set are divided, a random mode is adopted.
The last layer in the step (4) adopts logistic regression, and the output value of each neuron corresponds to the probability of whether the tower is damaged or not; the cost function adopts a cross entropy loss function, and the expression is
: whereinyRepresenting a true value, and a is a model output value; the error is then back propagated through the Adam algorithm, updating the parameters of all convolution kernels so that the cost function tends to converge.
And (5) when the accuracy is higher than 80%, fixing the model, and if the accuracy is lower than 80%, inputting 25% of data in the test set into the network for retraining.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (4)
1. A method for detecting external force damage of a power transmission line tower based on a convolutional neural network is characterized by comprising the following steps:
step (1), obtaining original data of a remote sensing image;
step (2), preprocessing the remote sensing image, separating and dividing images of each tower to obtain a training set, a verification set and a test set with labels, wherein the labels are divided into two types, namely, the transmission line tower is damaged by external force and the transmission line tower is not damaged by the external force;
step (3), a convolutional neural network model is constructed, each parameter is initialized, and a training set is input into the network for learning;
step (4), after the training set is input into the model, an output value is obtained through forward propagation, and the output value is the probability that the tower is damaged; calculating a cost function to measure the error between the output value and the true value, then feeding back through back propagation error, updating the parameter of the convolution kernel and continuing training; after the first round of training is finished, taking the verification set as input, operating the model and observing the result, and repeating the operation for multiple times until the convergence of the cost function is observed to obtain a neural network model with the optimal current parameters;
inputting the tower remote sensing graph with the concentrated test into a trained optimal model to obtain a classification result of whether the tower is damaged by external force or not, and evaluating the accuracy of the model;
and (6) when the accuracy of the model reaches the standard, the obtained model can be used for new tower external force damage inspection.
2. The method for detecting the external force damage of the power transmission line tower based on the convolutional neural network as claimed in claim 1, wherein the specific process of the step (2) is as follows:
step (2.1), processing the satellite remote sensing image: intercepting the position of the tower, and ensuring that the position of the tower is in the center of each picture;
step (2.2), dividing a training set: 60% of the towers in one line are taken as a training set, 20% of the towers are taken as a verification set, and 20% of the towers are taken as a test set.
3. The remote sensing image classification method for the electric wire towers based on the convolutional neural network as claimed in claim 1, wherein the last layer in the step (4) adopts logistic regression, and the output value of each neuron corresponds to the probability of whether the tower is damaged or not; the cost function adopts a cross entropy loss function, and the expression is
: whereinyRepresenting a true value, and a is a model output value; the error is then propagated back through the Adam algorithm, updating the parameters of all convolution kernels so that the cost function tends to converge.
4. The convolutional neural network-based electric wire tower remote sensing image two-classification method as claimed in claim 1, wherein in the step (5), when the accuracy is higher than 80%, the model is fixed, and if the accuracy is lower than 80%, 25% of data in the test set is input into the network for retraining.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911019078.7A CN111008641B (en) | 2019-10-24 | 2019-10-24 | Power transmission line tower external force damage detection method based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911019078.7A CN111008641B (en) | 2019-10-24 | 2019-10-24 | Power transmission line tower external force damage detection method based on convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111008641A CN111008641A (en) | 2020-04-14 |
| CN111008641B true CN111008641B (en) | 2023-04-07 |
Family
ID=70110875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911019078.7A Active CN111008641B (en) | 2019-10-24 | 2019-10-24 | Power transmission line tower external force damage detection method based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111008641B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111614969A (en) * | 2020-05-14 | 2020-09-01 | 深圳供电局有限公司 | Unmanned aerial vehicle tour video live broadcast method and system |
| CN111881760B (en) * | 2020-06-30 | 2021-10-08 | 深圳金三立视频科技股份有限公司 | Transmission line external damage prevention identification method and terminal |
| CN112380944B (en) * | 2020-11-06 | 2021-12-21 | 中国电力科学研究院有限公司 | Method and system for evaluating structural state of transmission tower based on satellite remote sensing |
| CN113592078A (en) * | 2021-08-09 | 2021-11-02 | 郑州大学 | Deep learning network training method based on artificial intelligence |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108009525A (en) * | 2017-12-25 | 2018-05-08 | 北京航空航天大学 | A kind of specific objective recognition methods over the ground of the unmanned plane based on convolutional neural networks |
| CN110188720A (en) * | 2019-06-05 | 2019-08-30 | 上海云绅智能科技有限公司 | A kind of object detection method and system based on convolutional neural networks |
-
2019
- 2019-10-24 CN CN201911019078.7A patent/CN111008641B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108009525A (en) * | 2017-12-25 | 2018-05-08 | 北京航空航天大学 | A kind of specific objective recognition methods over the ground of the unmanned plane based on convolutional neural networks |
| CN110188720A (en) * | 2019-06-05 | 2019-08-30 | 上海云绅智能科技有限公司 | A kind of object detection method and system based on convolutional neural networks |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111008641A (en) | 2020-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111008641B (en) | Power transmission line tower external force damage detection method based on convolutional neural network | |
| CN109785289B (en) | Transmission line defect detection method and system and electronic equipment | |
| CN109507535B (en) | Method and device for predicting operation stage and operation life of transformer substation grounding grid | |
| CN111444939B (en) | Small-scale equipment component detection method based on weak supervision cooperative learning in open scene of power field | |
| CN114092832B (en) | High-resolution remote sensing image classification method based on parallel hybrid convolutional network | |
| CN106991666B (en) | A kind of disease geo-radar image recognition methods suitable for more size pictorial informations | |
| CN110619479B (en) | Fault judgment model modeling method and fault judgment method for photovoltaic module | |
| CN111369057A (en) | Air quality prediction optimization method and system based on deep learning | |
| CN113591948A (en) | Defect pattern recognition method and device, electronic equipment and storage medium | |
| CN106951863B (en) | Method for detecting change of infrared image of substation equipment based on random forest | |
| Zhu et al. | Identifying carrot appearance quality by an improved dense CapNet | |
| CN114006369A (en) | Regional wind and light station power joint prediction method and device, electronic equipment and storage medium | |
| CN115859099A (en) | Sample generation method and device, electronic equipment and storage medium | |
| CN105374029A (en) | Segmenting method and system of transformer substation equipment infrared image interest areas | |
| CN112767351B (en) | Substation equipment defect detection method based on sensitive position dependence analysis | |
| CN117252459A (en) | Fruit quality evaluation system based on deep learning | |
| CN116702005A (en) | Neural network-based data anomaly diagnosis method and electronic equipment | |
| He et al. | A novel deep learning model for transmission line icing thickness prediction | |
| CN114519293A (en) | Cable body fault identification method based on hand sample machine learning model | |
| CN114662976A (en) | Power distribution network overhead line state evaluation method based on convolutional neural network | |
| CN113344915B (en) | Method and system for detecting defects of key parts of power transmission line | |
| CN112861670A (en) | Transmission line hardware detection method and system | |
| CN112070730A (en) | Anti-vibration hammer falling detection method based on power transmission line inspection image | |
| CN112579429A (en) | Problem positioning method and device | |
| CN112634216B (en) | Insulator self-explosion detection method based on deep learning model |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |