CN112541399A - Transmission line monitoring control method and device - Google Patents
Transmission line monitoring control method and device Download PDFInfo
- Publication number
- CN112541399A CN112541399A CN202011303082.9A CN202011303082A CN112541399A CN 112541399 A CN112541399 A CN 112541399A CN 202011303082 A CN202011303082 A CN 202011303082A CN 112541399 A CN112541399 A CN 112541399A
- Authority
- CN
- China
- Prior art keywords
- hidden danger
- time
- monitoring device
- targets
- transmission line
- 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.)
- Granted
Links
- 230000005540 biological transmission Effects 0.000 title claims abstract description 59
- 238000012544 monitoring process Methods 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000012806 monitoring device Methods 0.000 claims abstract description 143
- 230000008859 change Effects 0.000 claims description 26
- 238000012549 training Methods 0.000 claims description 16
- 230000009467 reduction Effects 0.000 claims description 13
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 239000013598 vector Substances 0.000 claims description 11
- 238000013507 mapping Methods 0.000 claims description 7
- 238000000513 principal component analysis Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 abstract 1
- 230000008569 process Effects 0.000 description 10
- 238000010276 construction Methods 0.000 description 9
- 230000003203 everyday effect Effects 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001960 triggered effect Effects 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
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- 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
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Economics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
- Alarm Systems (AREA)
Abstract
The application discloses a power transmission line monitoring shooting control method and device, which are used for solving the technical problems that the existing power transmission line monitoring shooting device is not timely in shooting adjustment and high in error rate. Averaging the number of hidden danger targets shot by each power transmission line monitoring device shooting the same scene image within a first preset time length; counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset time lengths, constructing a time series hidden danger target number prediction model, and determining a predicted value of the number of the hidden danger targets within the first time period to be detected through the prediction model; in a plurality of monitoring devices for shooting images of the same scene, any monitoring device shoots the distribution condition of the number of the hidden danger targets changing along with the time in the current time period, and the distribution condition of the number of the shot hidden danger targets changing along with the time in the first time period to be measured is determined. By the method, shooting of the monitoring device is adjusted in time, and the error rate is reduced.
Description
Technical Field
The application relates to the technical field of electric power, in particular to a monitoring and shooting control method and device for a power transmission line.
Background
The transmission line is an important component of the power grid, is influenced by artificial and natural conditions, and various potential safety hazards often appear in the transmission line and the environment where the transmission line is located. Therefore, in order to ensure safe operation of the power transmission line, image monitoring of the power transmission line and its surroundings is required.
In the prior art, the photographing of the monitoring device of the power transmission line is set and mainly completed through manual configuration. And for hidden trouble accidents in the scene of the continuously-occurring power transmission line, such as smoke and fire hidden troubles, the monitoring device needs to be manually triggered to take initiative shooting, so that the purpose of continuously tracking the change condition of the hidden trouble is achieved. Moreover, along with the increasing number of the monitoring devices, the monitoring scenes are complex and various, and along with seasonal changes, the types and the occurrence frequency of the hidden dangers are also changed continuously. The mode that the device was shot is taken in prison manual adjustment, not only can't ensure the promptness of shooing, and misoperation rate is also higher simultaneously, can not ensure transmission line safe operation.
Disclosure of Invention
The embodiment of the application provides a monitoring shooting control method and device for a power transmission line, and aims to solve the technical problems that the existing monitoring shooting device for the power transmission line cannot adjust shooting timely and has high error rate.
On one hand, the embodiment of the application provides a monitoring control method for a power transmission line. Determining the number of hidden danger targets shot by each monitoring device in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time; the same scene image is an image of a hidden danger target of the same type in a scene; sequentially counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset time lengths, constructing a time series hidden danger target number prediction model, and determining a predicted value of the number of the hidden danger targets in the first time period to be measured according to the prediction model; counting the distribution condition of the number of the shot hidden danger targets along with the change of time in the current time period of any one monitoring device in a plurality of monitoring devices for shooting the same scene image, so as to determine the distribution condition of the number of the shot hidden danger targets along with the change of time in the first waiting time period of the monitoring device; wherein the first period to be measured is a period next to the current period.
According to the image scene difference, the image classification method and the image classification device classify different types of images. Different time sequence hidden danger target prediction models are constructed by utilizing the number of the hidden danger targets in the classified images, so that the problem that the obtained predicted value of the hidden danger target is inaccurate due to complex and various monitoring scenes is solved. Meanwhile, the number distribution condition of the hidden danger targets in the next time period is predicted according to the number distribution condition of the hidden danger targets changing along with the time in the current time period. The time period to be measured is divided into a plurality of smaller time periods to respectively predict the number of the hidden danger targets, so that the obtained predicted value is more accurate. The embodiment of the application solves the problem that the generated error rate is high due to the fact that the shooting interval is adjusted manually through experience, and meanwhile, the shooting interval can be dynamically adjusted along with the change of seasons and time, the number of hidden danger targets of the power transmission line can be tracked and predicted in real time, and the safe operation of the power transmission line is guaranteed.
In an implementation manner of the present application, determining a distribution of the number of the hidden danger targets shot by the monitoring device in the first time period to be measured, which changes with time, specifically includes: according to the formula
Determine the firstThe distribution condition of the number of the hidden danger targets shot in the time period to be measured along with the change of time; wherein n isiThe number of the potential risk targets in a second time period to be measured is calculated, and the first time period to be measured is composed of a plurality of second time periods to be measured; alpha is a weight; i is the code of the second time period to be measured; m isiRepresenting the number of hidden danger targets in the i period of the current time period by the monitoring device; and p is a predicted value of the target number of the hidden danger in the first time period to be measured.
According to the method and the device, the first time period to be measured is divided into the second time periods to be measured, and the quantity value of the hidden danger target in each second time period to be measured is predicted respectively. Therefore, the shooting interval of the monitoring device can be dynamically adjusted along with the change of time. Under the condition that the frequency of hidden danger changes suddenly, the shooting frequency is ensured to be adjusted in time.
In one implementation manner of the present application, after determining a distribution of the number of hidden danger targets captured in the first time period to be measured, the method further includes: according to the formula
Determining the photographing interval of the monitoring device in a second time period to be measured; where val is a photographing interval of the monitoring apparatus in the second time period to be measured, and round () represents a rounding function.
In an implementation manner of the present application, after determining that the photographing interval of the monitoring apparatus is within the second time period to be measured, the method further includes: sequentially counting a plurality of second time periods to be measured contained in the first time period to be measured, and respectively corresponding to the photographing intervals of the monitoring device; determining a photographing interval sequence [ val ] of the monitoring device in the first time period to be measured according to the photographing intervals of the monitoring device respectively corresponding to the plurality of second time periods to be measured in the first time period to be measured1,val2,val3...vali]。
According to the embodiment of the application, the photographing interval sequence of the first time period to be measured is obtained by carrying out statistics on the predicted photographing frequency sequence of the second time periods to be measured. The photographing interval sequence of the first time period to be measured obtained in the way is more accurate and concrete. In addition, the shooting frequency of the monitoring device is adjusted according to the change of time, and the problem that the hidden danger occurrence frequency is changeable when the frequency is adjusted manually according to experience is solved.
In an implementation manner of the present application, within a continuous first preset duration, sequentially counting the quantity value of the hidden danger target for a plurality of times, and constructing a time series hidden danger target quantity prediction model specifically includes: respectively acquiring the number value of the hidden danger targets shot by each corresponding monitoring device in each first preset time within a plurality of continuous first preset time; arranging the acquired numerical values of the plurality of hidden danger targets according to a time sequence to construct time sequence data; and dividing the time sequence data into a plurality of samples for predicting the number of the hidden danger targets, and training a neural network by using the samples for predicting the number of the hidden danger targets to obtain a time sequence hidden danger target number prediction model.
According to the method and the device, the images are classified, and then time series hidden danger target quantity prediction models are respectively constructed for the images of different types. The problem of inaccurate prediction of the frequency of the hidden danger caused by different types of hidden danger targets with different frequencies due to season change or weather change in the power transmission line is solved.
In an implementation manner of the present application, dividing the time series data into a plurality of samples of prediction hidden danger target numbers specifically includes: intercepting the time series data by using sliding windows with the same size and sliding step length with the same size to obtain a plurality of samples for predicting the number of the hidden danger targets; dividing a group of data contained in a sample into input data and output data; the output data is the data which is acquired last in time sequence in the group of data, and the input data is all data except the data which is acquired last in time sequence in the group of data.
In the embodiment of the present application, a sample is divided into input data and output data, and data passing through an input/output mode is used as the sample. When the data of the hidden danger targets in the current time period is input, the quantity value of the hidden danger targets predicted in the time period to be measured can be obtained. The predicted quantity value can be closely related to the target quantity value of the near-term hidden danger, and the problem that the shooting frequency of the monitoring device is inaccurate due to the fact that the hidden danger is variable along with time change is solved.
In an implementation manner of the present application, before determining that the number of hidden danger targets shot by each monitoring device is averaged in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time, the method further includes: extracting characteristic vectors in different scene images shot by a plurality of power transmission line monitoring devices; performing principal component analysis and feature dimensionality reduction on the feature vector to obtain a dimensionality-reduced image; and taking the image subjected to dimensionality reduction as an input sample, inputting the input sample into a classification network model, and performing training to obtain the image classification network model.
The classification network is trained by taking the image subjected to the dimensionality reduction as a sample, and only the components with obvious image characteristics are selected, so that the calculated amount is reduced. Moreover, the relevance between the original features of the image after dimension reduction is eliminated, the redundancy of data information is reduced, and the classification of the image is facilitated. The obtained classification network model can classify the images more accurately.
In an implementation manner of the present application, before determining that the number of hidden danger targets shot by each monitoring device is averaged in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time, the method further includes: naming an image shot by a monitoring device, wherein the naming comprises a serial number of the monitoring device and shooting time of the image; and mapping the serial numbers corresponding to the monitoring device with the types of the images shot by the monitoring device, and obtaining scene classification information corresponding to the images shot by the monitoring device according to the mapping relation.
According to the embodiment of the application, the photographed images are named, and the monitoring devices corresponding to the images in the same scene can be clearly classified. The number of monitoring devices that take the same type of image is acquired. And then the quantity value of the hidden danger target shot by each monitoring device is calculated quickly. Therefore, the speed of the shooting interval of the monitoring device is increased.
In an implementation manner of the present application, before determining that, in a plurality of power transmission line monitoring devices that capture images of the same scene within a first preset time period, a number value of a hidden danger target captured by each monitoring device is averaged, the method further includes: counting the total number of hidden danger targets in the same scene image shot by the plurality of power transmission line monitoring devices within a first preset time; counting the number of monitoring devices for shooting the same scene image within the first preset time; and determining the quantity value of the hidden danger targets shot by each monitoring device in the scene according to the total number of the hidden danger targets and the number of the monitoring devices.
On the other hand, the embodiment of the present application further provides a control device for monitoring and shooting of a power transmission line, including: at least one processor; and a memory communicatively coupled to the at least one processor; the storage stores instructions which can be executed by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can determine the number of the hidden danger targets shot by each monitoring device in a plurality of power transmission line monitoring devices which shoot images of the same scene within a first preset time length; the same scene image is an image of a hidden danger target of the same type in a scene; sequentially counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset time lengths, constructing a time series hidden danger target number prediction model, and determining a predicted value of the number of the hidden danger targets in the first time period to be measured according to the prediction model; counting the distribution condition of the number of the shot hidden danger targets along with the change of time in the current time period of any one monitoring device in a plurality of monitoring devices for shooting the same scene image, so as to determine the distribution condition of the number of the shot hidden danger targets along with the change of time in the first waiting time period of the monitoring device; wherein the first period to be measured is a period next to the current period.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a flowchart of a monitoring control method for a power transmission line provided in an embodiment of the present application;
fig. 2 is a flowchart of a scene classification network model training process according to an embodiment of the present disclosure;
fig. 3 is a flowchart of a process for constructing a time series hidden danger target quantity prediction model according to an embodiment of the present application;
fig. 4 is a schematic diagram of an internal structure of the power transmission line monitoring and photographing control device provided in the embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In the prior art, when the monitoring device of the power transmission line is adjusted, manual configuration is mainly used for completing the adjustment. The photographing intervals of 60 minutes, 30 minutes, 20 minutes and 15 minutes are used in addition. In the manual configuration process, the failure rate is high, the fixed frequency is configured, and the shooting frequency cannot be dynamically adjusted according to the specific situation of the continuous hidden danger. Meanwhile, along with the continuous increase of the monitoring and shooting devices, the problems of variable seasons and the like, the timeliness of the existing manual configuration mode is low, the frequency setting can only depend on experience, and the accuracy is not high.
In order to solve the above problems, embodiments of the present application provide a method and an apparatus for monitoring and controlling a transmission line. The shot images are classified according to different scenes, so that the shooting frequency of the monitoring device corresponding to the images of different scenes is calculated, and the accuracy of calculating the shooting frequency of the monitoring device is improved. Meanwhile, the time series hidden danger target quantity prediction model is constructed by taking the hidden danger target quantity in the continuous time period as a sample. And predicting the number of the hidden danger targets in the time period to be measured according to the number of the hidden danger targets in the current time period. The calculated shooting frequency can be changed in time according to the change of the number of the hidden dangers, so that not only are complicated labor and high error rate caused by manual configuration avoided, but also the shooting frequency of the monitoring device is matched with the occurrence frequency of the hidden danger targets, and the number of the shot hidden danger targets is more accurate.
The technical solutions proposed in the embodiments of the present application are described in detail below with reference to the accompanying drawings.
Fig. 1 is a flowchart of a monitoring control method for a power transmission line according to an embodiment of the present disclosure. As shown in fig. 1, the monitoring control method for the power transmission line includes the following steps:
s101, classifying the images by the monitoring platform according to different scenes in the images.
In one embodiment of the application, a monitoring platform receives shot images transmitted by a plurality of monitoring devices and names the received images.
Specifically, when naming a received image, the name includes a serial number of a monitoring device that captured the image, and a time of capturing the image. The specific format is as follows: device string number _ year, month, day, hour, minute, second, jpg.
In one embodiment of the application, images shot by a recent part monitoring device on a power transmission line are collected, and 5 different scenes of the images, namely an outer construction broken region, a floater region, a firework region, a three-span region and a natural disaster region, are classified. The images divided into the same scene are images of hidden danger targets of the same type in the scene.
In the present embodiment, it is preferable to divide the image into 5 types, i.e., an external construction damage susceptibility region, a floating material susceptibility region, a smoke and fire susceptibility region, a three-span susceptibility region, and a natural disaster susceptibility region, but the image is not limited to only the above 5 types.
In one embodiment of the application, feature vectors in different scene images shot by a plurality of collected monitoring devices are extracted.
Specifically, a gabor filter is used to extract the gist eigenvector of the picture. Wherein, the gabor function extracts features in different scales and different directions of a frequency domain. The gist feature vector is a scene feature description.
In one embodiment of the present application, a pca principal component analysis is required on the gist feature vectors. And performing characteristic dimensionality reduction to obtain a dimensionality reduced image. And then, taking the main component image subjected to dimensionality reduction as an input sample, and inputting the input sample into a classification network model for training to obtain an image classification network model. In the pca principal component analysis, a group of variables possibly having correlation are converted into a group of linearly uncorrelated variables through orthogonal transformation, and the group of converted variables is called a principal component.
In an embodiment of the application, normalization processing is performed on the image after dimension reduction, the processed image is used as a feature sample, and an Alexnet network is trained to obtain an image classification network model.
According to the method and the device, the dimension of the high-dimensional data is reduced through the feature dimension reduction technology, the number of prediction variables is reduced, the dimension of a feature space is reduced, and the subsequent classification network model design is easier to realize on calculation. Meanwhile, the correlation degree among original features is eliminated, the redundancy of data information is reduced, and the classification of the images is facilitated.
In one embodiment of the application, an image shot by a monitoring device is input to an image classification network model, and the image is classified according to the scene. And according to the name on the image, mapping the serial number corresponding to the monitoring device with the types of the plurality of images shot by the monitoring device, and according to the mapping relation, obtaining the scene classification information corresponding to the images shot by the monitoring device. That is, according to the mapping relationship, which monitoring devices shoot each type of scene images are obtained, so that the number of the monitoring devices shooting the same type is obtained.
S102, the monitoring platform determines a predicted value of the number of the hidden danger targets in the first time period to be measured.
In one embodiment of the application, a classification scene is selected, an image belonging to the scene is transmitted to an analysis server for hidden danger target identification, and the number of hidden danger targets in the image is obtained.
In an embodiment of the application, a first preset time duration is counted, for example, the total number of hidden danger targets in images of the same scene shot by a plurality of power transmission line monitoring devices in one day is counted. And counting a first preset time, such as the number of monitoring devices for shooting the same scene image in one day. And determining the number of the hidden danger targets shot by each monitoring device in a plurality of power transmission line monitoring devices shooting the same scene image within a first preset time according to the counted total number of the hidden danger targets and the number of the monitoring devices.
For example, if the images in the scene of the construction-prone region are selected for statistics. The total number of hidden danger targets in all images of the scene in one day is counted, for example, 100 hidden danger targets are counted. Then, the number of all monitoring devices for shooting the scene image is counted, for example, 20 monitoring devices shoot the image of the scene of the construction outer broken and easy-to-send area in one day. At this time, the monitoring device for shooting images in the scene of the construction outer broken and easy-to-send area within one day can be calculated, and the average number of the hidden danger targets shot by each monitoring device is 5.
In one embodiment of the present application, the number of hidden danger targets is counted several times in sequence within several consecutive days.
For example, in one month continuously, the monitoring devices that shoot the scene images of the construction-prone region are counted every day, and the number of the hidden danger targets shot by each monitoring device is averaged. And arranging the acquired number of the plurality of hidden danger targets according to a time sequence to construct time sequence data.
It should be noted that, in the embodiment of the present application, it is preferable to count the number of hidden danger targets for one month, but the present application is not limited to only selecting one month.
In one embodiment of the present application, the time series data is divided into a number of samples that predict the target number of potential hazards.
Specifically, the time series data is intercepted by using sliding windows with the same size and sliding step lengths with the same size, and a plurality of samples for predicting the number of the hidden danger targets are obtained.
In one embodiment of the present application, a set of data contained within a sample is divided into input data and output data. The output data is the data which is acquired last in time sequence in the group of data, and the input data is all data except the data which is acquired last in time sequence in the group of data.
Specifically, in the embodiment of the present application, one sample includes 4 data values, data of three time steps is used as an input, and data of one time step is used as an output value. Constituting samples of the input/output pattern.
For example, if the constructed time-series data is [1,2,3,4,5,6,7,8,9,10 ]]The input/output pattern is divided into samples of input/output patterns,
four data in each sample, the first three data being input data and the last data being output data.
It should be noted that, in the embodiment of the present application, it is preferable that one sample includes 4 data values, but it is not limited to include only 4 data values.
In an embodiment of the application, a neural network is trained by using a sample of an input/output mode for predicting the number of the hidden danger targets, so as to obtain a time series hidden danger target number prediction model in the scene.
It should be noted that a time series hidden danger target quantity prediction model constructed in the embodiment of the present application corresponds to an image in a classified scene. According to the method and the device, the images are divided into 5 types according to different scenes, so that time series hidden danger target quantity prediction models under 5 different scenes need to be constructed.
In an embodiment of the application, a first time period to be measured, such as a predicted value of the target number of the hidden danger in one day, is determined according to a constructed time series target number prediction model.
For example, when the number of the hidden danger targets which can be shot in the next day by the monitoring device for shooting the scene of the broken and prone area outside the construction is determined. In each monitoring device for shooting the scene image on the first day, the number of the hidden danger targets which can be shot by each monitoring device is averaged, and the number of the hidden danger targets is input into a time series hidden danger target number prediction model, so that the predicted value of the number of the hidden danger targets in the scene on the second day can be obtained.
S103, the monitoring platform counts the distribution situation of the number of the hidden danger targets shot by any one monitoring device in a first waiting time period, such as a day, along with the change of time.
In an embodiment of the application, the distribution situation of the number of the hidden danger targets shot by any one monitoring device in the current time period, such as a day, along with the change of time is counted. Wherein, the first time period to be measured is the next time period of the current time period.
Specifically, the current time period is divided into 24 hours, the number of the hidden danger targets shot by the monitoring device in each hour is counted, and the counted numerical values are arranged in sequence according to the time sequence.
In one embodiment of the present application, the formula is based on
And determining the distribution condition of the number of the shot hidden danger targets changing along with the time in the next day by the monitoring device. That is, the number of the hidden trouble targets that can be photographed by the monitoring apparatus every hour on the next day.
In an embodiment of the present application, in the above formula for predicting the magnitude of the hidden danger target, niRepresenting the calculated number of the hidden danger targets in the second prediction time period, wherein the first time period to be measured is composed of a plurality of second time periods to be measuredAnd (5) time section composition. α is a weight. i (i ═ 1, 2.. 24) is the code for the second time segment to be measured. m isiRepresenting the number of the hidden danger targets in the i period of the current period, the monitoring device arranges the counted number of the 24 hidden danger targets in a time sequence to obtain a data sequence [ m1, m2, m3 and m4.... m24 ]]. And p is a predicted value of the target number of the hidden danger in the first time period to be measured.
S104, the monitoring platform determines a photographing interval sequence of the selected monitoring device in the first time period to be measured.
In one embodiment of the present application, the formula is based on
The interval between photographs of the selected monitoring device during a first period of time to be measured, such as every hour on the second day, is determined. Where val is the frequency of photographing by the monitoring apparatus every hour on the next day, and round () represents a rounding function.
For example, if the number n1 of the potential targets captured by the monitoring device in the first hour corresponding to the second day is 0 in the first time period to be measured, the monitoring device takes a photograph at a time interval of 1 hour in the first hour, that is, at the time when the first hour starts.
If the number n2 of the hidden danger targets shot by the monitoring device in the second hour corresponding to the second day to be measured is calculated to be 3, the shooting interval of the monitoring device in the first hour is 20 min.
In one embodiment of the present application, 24 hours included in the next day are counted in sequence, and each hour corresponds to a photographing interval of the monitoring device. After the images are arranged in sequence, the photographing interval sequence [ val ] of the monitoring device in every hour on the next day can be determined1,val2,val3...va24]。
S105, the monitoring platform uploads the photographing interval sequence to the corresponding monitoring device.
In one embodiment of the present application, when the shooting interval of the selected monitoring device within two days is determined, the shooting interval is uploaded to the monitoring device in time. The monitoring device shoots the power transmission line according to the shooting interval.
According to the embodiment of the application, the shooting interval sequence is uploaded in time, and the shooting frequency of the monitoring device is controlled, so that the appropriate shooting interval can be adjusted in time according to the change of time, the change of seasons and the change of the occurrence frequency of hidden dangers. The accuracy of shooting the number of the hidden danger targets is ensured, and the safe operation of the power transmission line is guaranteed.
Fig. 2 is a flowchart of a scene classification network model training process provided in the embodiment of the present application.
As shown in fig. 2, in the training process of the scene classification network model according to the embodiment of the present application, the following steps are required:
s201, the monitoring platform receives images shot by the plurality of monitoring devices and stores the received images.
S202, the monitoring platform classifies the received images according to different scenes in the images.
In one embodiment of the application, the image is divided into five types, namely a construction outer broken easy-to-send area, a floater easy-to-send area, a firework easy-to-send area, a three-span easy-to-send area and a natural disaster easy-to-send area according to the type of a hidden danger target.
S203, the server extracts the feature vectors in the image.
In one embodiment of the present application, the gist feature vectors in the same scene image are extracted using gabor filtering.
And S204, the server performs principal component analysis and feature dimension reduction on the extracted feature vector.
In one embodiment of the present application, pca principal component analysis and feature dimensionality reduction is performed on the extracted gist feature vectors. The number of variables is reduced, the dimensionality of a feature space is reduced, and the subsequent classification network model design is more easily realized on calculation. Meanwhile, the redundancy of data information is reduced, and the classification of the images is facilitated.
S205, the server performs normalization processing on the feature data after dimension reduction to obtain a feature sample. Inputting the network into an Alexnet network for training to obtain a classification network model.
It should be noted that the above is made by taking the example of training the neural network through the server to obtain the classification network model. Those skilled in the art will appreciate that the neural network may be trained by the monitoring platform to obtain a classification network model. The training of the neural network by the monitoring platform is the same as the training of the neural network by a server outside the monitoring platform, and is not repeated herein.
Fig. 3 is a flowchart of a process for constructing a time series hidden danger target quantity prediction model according to an embodiment of the present application.
As shown in fig. 3, in the process of constructing the time series hidden danger target quantity prediction model according to the embodiment of the present application, the following steps are required:
s301, the monitoring platform obtains the total number of monitoring devices for shooting the scene images of the same type.
In one embodiment of the application, the total number of monitoring devices for shooting the same scene image in one month is continuously counted by taking days as a unit.
For example, the number of monitoring devices for capturing images of scenes in the construction-prone region is counted every day. Assuming that 20 monitoring devices exist on the first day and 15 monitoring devices exist on the second day, the number change condition of the monitoring devices in one month is counted in sequence.
S302, the monitoring platform sends the collected images within one month to an image analysis server, and the analysis server carries out hidden danger target identification detection on the received images. And identifying hidden danger targets and the number of the hidden danger targets in the image.
For example, the images of the scene of the construction outbreak area which are counted every day in one month are sent to an image analysis server, and hidden danger targets such as an excavator, a crane and the like appearing in the images are identified. Therefore, the total number of the hidden danger targets in the construction external damage and easy occurrence area images shot every day is determined.
And S303, the monitoring platform determines the number of the hidden danger targets shot by each monitoring device every day.
In an embodiment of the application, the number of the hidden danger targets which can be shot by taking each monitoring shooting as a unit is determined by acquiring the number of the monitoring shooting devices which shoot the same scene image every day and the total number of the hidden danger targets shot every day.
For example, the monitoring device that shoots the scene image of the construction outer broken and easy-to-send area on the first day has 20 devices, and if the number of the hidden danger targets is shot totally 100, the number value of the determined hidden danger targets is 5. The other way round is. The monitoring devices for shooting the field of images the next day have 15 devices, and if 60 hidden danger targets are shot in total, the number value of the determined hidden danger targets is 4.
S304, the monitoring platform arranges the number of the shot hidden danger targets in one month according to a time sequence to obtain time sequence data.
S305, the monitoring platform intercepts the time series data to obtain a training sample.
For example, four data are a set of time series data that are truncated. The first three data in each set of data are used as input data and the last data is used as output data, thereby constituting a sample of the input/output pattern.
S306, the server trains the neural network through the obtained samples of the input/output modes to obtain a classification network model.
It should be noted that the above description is made by taking an example of obtaining a time series hidden danger target quantity prediction model through training a neural network by a server. As can be understood by those skilled in the art, the neural network can be trained through the monitoring platform to obtain a time series hidden danger target quantity prediction model. The training of the neural network by the monitoring platform is the same as the training of the neural network by a server outside the monitoring platform, and is not repeated herein.
Fig. 4 is a schematic diagram of an internal structure of a power transmission line monitoring and photographing control device according to an embodiment of the present application.
The embodiment of the application provides a transmission line monitoring control device, includes:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
determining the number of hidden danger targets shot by each monitoring device in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time; and the same scene image is an image of a hidden danger target of the same type in the scene.
Sequentially counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset time lengths, constructing a time series hidden danger target number prediction model, and determining a predicted value of the number of the hidden danger targets in the first time period to be measured according to the prediction model.
Counting the distribution condition of the number of the shot hidden danger targets along with the change of time in the current time period of any one monitoring device in a plurality of monitoring devices for shooting the same scene image, so as to determine the distribution condition of the number of the shot hidden danger targets along with the change of time in the first waiting time period of the monitoring device; wherein the first period to be measured is a period next to the current period.
The embodiments in the present application are described in a progressive manner, and the same and similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, as for the apparatus embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (10)
1. A monitoring and shooting control method for a power transmission line is characterized by comprising the following steps:
determining the number of hidden danger targets shot by each monitoring device in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time; the same scene image is an image of a hidden danger target of the same type in a scene;
sequentially counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset time lengths, constructing a time series hidden danger target number prediction model, and determining a predicted value of the number of the hidden danger targets in the first time period to be measured according to the prediction model;
counting the distribution condition of the number of the shot hidden danger targets along with the change of time in the current time period of any one monitoring device in a plurality of monitoring devices for shooting the same scene image, so as to determine the distribution condition of the number of the shot hidden danger targets along with the change of time in the first waiting time period of the monitoring device; wherein the first period to be measured is a period next to the current period.
2. The power transmission line monitoring control method according to claim 1, wherein the determining of the distribution of the number of the hidden danger targets shot by the monitoring device in the first waiting time period, which changes with time, specifically comprises:
according to the formula
Determining the distribution condition of the number of the hidden danger targets shot in the first time period to be measured along with the change of time;
wherein n isiThe number of the potential risk targets in a second time period to be measured is calculated, and the first time period to be measured is composed of a plurality of second time periods to be measured; alpha is a weight; i is the code of the second time period to be measured; m isiRepresenting the number of hidden danger targets in the i period of the current time period by the monitoring device; and p is a predicted value of the target number of the hidden danger in the first time period to be measured.
3. The monitoring control method for the power transmission line according to claim 2, wherein after determining the distribution of the number of the hidden danger targets shot in the first time period to be measured, the method further comprises:
according to the formula
Determining the photographing interval of the monitoring device in the second time period to be measured;
where val is a photographing interval of the monitoring device in the second time period to be measured, and round () represents a rounding function.
4. The monitoring control method for the power transmission line according to claim 3, wherein the determination is made that the monitoring device is in the second time period to be measured and after the photographing interval, the method further comprises:
sequentially counting a plurality of second time periods to be measured contained in the first time period to be measured, and respectively corresponding to the photographing intervals of the monitoring device;
determining the first waiting time of the monitoring device according to the photographing intervals of the monitoring device respectively corresponding to the second waiting time periodsInterval sequence of intervals of photographing within a period [ val ]1,val2,val3...vali]。
5. The monitoring control method for the power transmission line according to claim 1, wherein the step of sequentially counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset durations to construct a time series hidden danger target number prediction model specifically comprises the steps of:
respectively acquiring the number of the hidden danger targets shot by each monitoring device in each first preset time within a plurality of continuous first preset time;
arranging the number of the acquired hidden danger targets according to a time sequence to construct time sequence data;
and dividing the time sequence data into a plurality of samples for predicting the number of the hidden danger targets, and training a neural network by using the samples for predicting the number of the hidden danger targets to obtain a time sequence hidden danger target number prediction model.
6. The power transmission line monitoring control method according to claim 5, wherein the dividing the time series data into a plurality of samples of predicted hidden danger target quantities specifically comprises:
intercepting the time sequence data by using sliding windows with the same size and sliding step length with the same size to obtain a plurality of samples for predicting the number of the hidden danger targets;
dividing a set of data contained in the sample into input data and output data; the output data is the data which is obtained last in time sequence in the group of data, and the input data is all data except the data which is obtained last in time sequence in the group of data.
7. The power transmission line monitoring control method according to claim 1, wherein before determining that the number of the hidden danger targets shot by each monitoring device is averaged, in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time period, the method further comprises:
extracting characteristic vectors in different scene images shot by a plurality of power transmission line monitoring devices;
performing principal component analysis and feature dimensionality reduction on the feature vector to obtain a dimensionality-reduced image;
and taking the image subjected to dimensionality reduction as an input sample, inputting the input sample into a classification network model, and performing training to obtain the image classification network model.
8. The power transmission line monitoring control method according to claim 1, wherein before determining that the number of the hidden danger targets shot by each monitoring device is averaged, in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time period, the method further comprises:
naming an image shot by a monitoring device, wherein the naming comprises a serial number of the monitoring device and shooting time of the image;
and mapping the serial numbers corresponding to the monitoring device with the types of the images shot by the monitoring device, and obtaining scene classification information corresponding to the images shot by the monitoring device according to the mapping relation.
9. The power transmission line monitoring control method according to claim 1, wherein before determining that the number of the hidden danger targets captured by each monitoring device is averaged, the method further comprises the steps of, in a plurality of power transmission line monitoring devices capturing images of the same scene within a first preset time period:
counting the total number of hidden danger targets in the same scene image shot by the plurality of power transmission line monitoring devices within a first preset time;
counting the number of monitoring devices for shooting the same scene image within the first preset time;
and determining the number of the hidden danger targets shot by each monitoring device in the scene according to the total number of the hidden danger targets and the number of the monitoring devices.
10. The utility model provides a transmission line monitoring controlling means, its characterized in that includes:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to:
determining the number of hidden danger targets shot by each monitoring device in a plurality of power transmission line monitoring devices shooting images of the same scene within a first preset time; the same scene image is an image of a hidden danger target of the same type in a scene;
sequentially counting the number of the hidden danger targets for a plurality of times within a plurality of continuous first preset time lengths, constructing a time series hidden danger target number prediction model, and determining a predicted value of the number of the hidden danger targets in the first time period to be measured according to the prediction model;
counting the distribution condition of the number of the shot hidden danger targets along with the change of time in the current time period of any one monitoring device in a plurality of monitoring devices for shooting the same scene image, so as to determine the distribution condition of the number of the shot hidden danger targets along with the change of time in the first waiting time period of the monitoring device; wherein the first period to be measured is a period next to the current period.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011303082.9A CN112541399B (en) | 2020-11-19 | 2020-11-19 | Transmission line monitoring control method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011303082.9A CN112541399B (en) | 2020-11-19 | 2020-11-19 | Transmission line monitoring control method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112541399A true CN112541399A (en) | 2021-03-23 |
| CN112541399B CN112541399B (en) | 2022-11-15 |
Family
ID=75014359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011303082.9A Active CN112541399B (en) | 2020-11-19 | 2020-11-19 | Transmission line monitoring control method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112541399B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113114935A (en) * | 2021-04-07 | 2021-07-13 | 中煤科工集团重庆研究院有限公司 | Vibration identification method based on video image |
| CN113840122A (en) * | 2021-11-25 | 2021-12-24 | 南方电网数字电网研究院有限公司 | Monitoring shooting control method and device, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106651855A (en) * | 2016-12-30 | 2017-05-10 | 国网山东省电力公司济宁供电公司 | Image monitoring and shooting method for automatic identification and alarming of hidden troubles of power transmission line channel |
| CN107657260A (en) * | 2017-10-09 | 2018-02-02 | 济南大学 | The detection method of ground hidden danger in a kind of transmission line of electricity surrounding enviroment |
| RU2683411C1 (en) * | 2018-03-01 | 2019-03-28 | Общество с ограниченной ответственностью "Лаборатория будущего" | Method of digital control of the process of monitoring, maintenance and local repair of opl and system for its implementation |
| CN109905289A (en) * | 2019-01-23 | 2019-06-18 | 中国人民解放军军事科学院系统工程研究院 | A kind of network service credible attribute dynamic prediction method |
| CN110888788A (en) * | 2019-10-16 | 2020-03-17 | 平安科技（深圳）有限公司 | Anomaly detection method and device, computer equipment and storage medium |
| CN111537841A (en) * | 2020-06-30 | 2020-08-14 | 上海交通大学 | Optimization method and system suitable for ground fault type identification |
-
2020
- 2020-11-19 CN CN202011303082.9A patent/CN112541399B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106651855A (en) * | 2016-12-30 | 2017-05-10 | 国网山东省电力公司济宁供电公司 | Image monitoring and shooting method for automatic identification and alarming of hidden troubles of power transmission line channel |
| CN107657260A (en) * | 2017-10-09 | 2018-02-02 | 济南大学 | The detection method of ground hidden danger in a kind of transmission line of electricity surrounding enviroment |
| RU2683411C1 (en) * | 2018-03-01 | 2019-03-28 | Общество с ограниченной ответственностью "Лаборатория будущего" | Method of digital control of the process of monitoring, maintenance and local repair of opl and system for its implementation |
| CN109905289A (en) * | 2019-01-23 | 2019-06-18 | 中国人民解放军军事科学院系统工程研究院 | A kind of network service credible attribute dynamic prediction method |
| CN110888788A (en) * | 2019-10-16 | 2020-03-17 | 平安科技（深圳）有限公司 | Anomaly detection method and device, computer equipment and storage medium |
| CN111537841A (en) * | 2020-06-30 | 2020-08-14 | 上海交通大学 | Optimization method and system suitable for ground fault type identification |
Non-Patent Citations (1)
| Title |
|---|
| 陈雷 等: "一种具有录波功能智能监拍装置的设计", 《集成电路应用》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113114935A (en) * | 2021-04-07 | 2021-07-13 | 中煤科工集团重庆研究院有限公司 | Vibration identification method based on video image |
| CN113840122A (en) * | 2021-11-25 | 2021-12-24 | 南方电网数字电网研究院有限公司 | Monitoring shooting control method and device, electronic equipment and storage medium |
| CN113840122B (en) * | 2021-11-25 | 2022-03-08 | 南方电网数字电网研究院有限公司 | Monitoring shooting control method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112541399B (en) | 2022-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110555390B (en) | Pedestrian re-identification method, device and medium based on semi-supervised training mode | |
| CN108965826B (en) | Monitoring method, monitoring device, processing equipment and storage medium | |
| CN115600932B (en) | Cultural relic storage environment anomaly evaluation method based on big data | |
| CN108921203B (en) | Detection and identification method for pointer type water meter | |
| US9141184B2 (en) | Person detection system | |
| CN112541399B (en) | Transmission line monitoring control method and device | |
| JP2021533474A (en) | Node classification method, model training method, and its equipment, equipment and computer program | |
| CN108960314B (en) | Training method and device based on difficult samples and electronic equipment | |
| CN115249331B (en) | Mine ecological safety identification method based on convolutional neural network model | |
| CN116744357A (en) | Base station fault prediction method, device, equipment and medium | |
| CN113962282A (en) | Improved YOLOv5L + Deepsort-based real-time detection system and method for ship engine room fire | |
| CN111291627A (en) | Face recognition method and device and computer equipment | |
| CN113935395A (en) | Training of object recognition neural networks | |
| CN109063589A (en) | Instrument and equipment on-line monitoring method neural network based and system | |
| CN108345857A (en) | A kind of region crowd density prediction technique and device based on deep learning | |
| CN114155595A (en) | Behavior detection monitoring method, intelligent camera and intelligent monitoring system | |
| CN117093822B (en) | Industrial brain data analysis platform based on industrial knowledge graph | |
| CN112183380B (en) | Passenger flow volume analysis method and system based on face recognition and electronic equipment | |
| CN112001300B (en) | Building monitoring method and device based on cross entropy according to position and electronic equipment | |
| CN113469587A (en) | Method and device for evaluating space-time influence of climate change on regional landslide | |
| CN110909706A (en) | Method and device for judging person during daytime and night, electronic equipment and storage medium | |
| CN110765817A (en) | Method, device and equipment for selecting crowd counting model and storage medium thereof | |
| CN114821428A (en) | Intelligent fire identification method and device, computer equipment and storage medium | |
| CN110969209B (en) | Stranger identification method and device, electronic equipment and storage medium | |
| JP6883699B1 (en) | Flow estimation system, information processing equipment and programs |
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 |