CN114092888A - Electric power operation laboratory personnel risk detection system that electrocutes - Google Patents
Electric power operation laboratory personnel risk detection system that electrocutes Download PDFInfo
- Publication number
- CN114092888A CN114092888A CN202111470160.9A CN202111470160A CN114092888A CN 114092888 A CN114092888 A CN 114092888A CN 202111470160 A CN202111470160 A CN 202111470160A CN 114092888 A CN114092888 A CN 114092888A
- Authority
- CN
- China
- Prior art keywords
- layer
- image
- convolution
- laboratory
- category
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
- G06N3/04—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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an electric shock risk detection system for laboratory personnel, which comprises an image sensing subsystem, an image recognition subsystem and a judgment module, wherein the image sensing subsystem is used for acquiring an electric shock risk point image of an electric power operation laboratory; the image identification subsystem is used for identifying and outputting the category of the electric shock risk point image; and the judgment module is used for judging whether the laboratory personnel of the electric power operation laboratory have electric shock risks or not according to the identification result. The invention aims to provide an electric power operation laboratory worker electric shock risk detection system, which takes image live-action monitoring of an electric power operation site as a breakthrough, realizes electric power operation site supervision through image recognition, and solves the problem that the electric power operation laboratory in the prior power transmission technology lacks a worker electric shock risk detection system integrating equipment state detection and worker operation supervision.
Description
Technical Field
The invention relates to the technical field of electric power operation accident prevention, in particular to an electric shock risk detection system for laboratory personnel.
Background
With the development of economy and the continuous progress of society in China, the power demand is increasing day by day, but the resource distribution in China is extremely unbalanced: 70% of hydraulic resources are distributed in the middle and southwest areas; 76% of coal power resources are distributed in northwest regions; wind power generation and solar power generation resources are also mainly distributed in northwest regions; but 70% of the electricity demand is present in the eastern coastal and mid-sea areas. Therefore, the research and development of the power transmission technology have important significance for China.
The research on the transmission technology in China is many, as early as 5 months in 2009, national engineering laboratories of extra-high voltage engineering technology have formally started, and each university also builds a power operation laboratory of its own to research the transmission technology. Compared with weak current, strong current has greater danger, and accidents are easily caused by improper operation or equipment aging of a power operation laboratory, so that serious casualties are caused. In the power operation laboratory, especially the system equipment for operation supervision of the dead personnel and safety control of the power operation. Therefore, in a laboratory for electric power operation, a personnel electric shock risk detection system integrating equipment state detection and personnel operation supervision is required, and currently, the market is temporarily lack of such a safety system.
Disclosure of Invention
The invention aims to provide an electric power operation laboratory worker electric shock risk detection system, which takes image live-action monitoring of an electric power operation site as a breakthrough, realizes electric power operation site supervision through image sensing and image recognition, and solves the problem that the electric power operation laboratory in the current power transmission technology lacks a worker electric shock risk detection system integrating equipment state detection and worker operation supervision.
The invention is realized by the following technical scheme:
an electric power operation laboratory worker electric shock risk detection system for electric shock risk detection system for electric shock risk laboratory:
the image sensing subsystem is used for acquiring an electric shock risk point image of the electric power operation laboratory;
the electric shock risk point image comprises a display screen image, a wall corner grounding wire image, a laboratory bench grounding wire image, an insulating pad area image and an insulating rod area image;
the image identification subsystem is used for identifying and outputting the category of the electric shock risk point image;
the categories include category 1 and category 2;
the type 1 corresponds to a display screen displaying a first preset picture, a wall corner grounding wire, a laboratory bench grounding wire, an experimenter standing on an insulating pad for experimental operation and a discharge rod hanging on an insulating rod;
the category 2 corresponds to a display screen displaying a second preset picture, a wall corner grounding wire is not grounded, a laboratory bench grounding wire is not grounded, an experimenter does not stand on an insulating pad for experimental operation, and a discharging rod is not hung on an insulating rod;
and the judgment module is used for judging whether the laboratory personnel of the electric power operation laboratory have electric shock risks or not according to the identification result.
Preferably, the image recognition subsystem comprises:
the experiment state image identification module is used for identifying whether the power operation experiment starts or ends according to the display screen image, considering that the power operation experiment starts when the display screen displays a first preset image, and outputting a category 1; when the display screen displays a second preset image, the power operation experiment is considered to be finished, and the category 2 is output;
the corner grounding image recognition module is used for recognizing whether the corner grounding wire is grounded according to the corner grounding wire image and outputting a category 1 when the corner grounding wire is grounded; otherwise, outputting the category 2;
the experiment table grounding image identification module is used for identifying whether the experiment table grounding wire is grounded according to the experiment table grounding wire image; outputting a category 1 when the ground wire of the experiment table is grounded; otherwise, outputting the category 2;
the insulating pad image identification module is used for identifying whether an experimenter stands on the insulating pad to perform experimental operation according to the image of the area where the insulating pad is located; outputting class 1 when the experimenter stands on the insulating pad for experimental operation; otherwise, outputting the category 2;
the insulating rod image identification module is used for identifying whether the discharge rod is hung on the insulating rod according to the image of the area where the insulating rod is located; outputting the category 1 when the discharge rod is not hung on the insulating rod; otherwise, category 2 is output.
Preferably, when the experimental state image recognition module, the corner grounding image recognition module, the laboratory table grounding image recognition module and the insulation pad image recognition module all output the category 1, the judgment module judges that the experimenter in the power operation laboratory has no electric shock risk;
when the experimental state image recognition module and the insulating rod image recognition module output the category 2, the judgment module judges that the experimenters in the electric power operation laboratory have no electric shock risk.
Preferably, the experimental state image recognition module, the corner grounding image recognition module, the experimental table grounding image recognition module, the insulating pad image recognition module and the insulating rod image recognition module all adopt an image recognition depth neural network model for image recognition.
Preferably, the image recognition deep neural network model includes: the first rolling layer, the first maximum pooling layer, the second rolling layer, the second maximum pooling layer, the third rolling layer, the third maximum pooling layer, the fourth rolling layer, the fourth maximum pooling layer, the first batch normalizing layer, the second batch normalizing layer, the third batch normalizing layer, the feature combining layer, the fifth rolling layer, the fifth maximum pooling layer, the first full-connection layer, the second full-connection layer, the third full-connection layer and the loss layer;
the output end of the first convolution layer is connected with the input end of the first maximum pooling layer, and the output end of the first maximum pooling layer is connected with the input end of the second convolution layer and the input end of the first batch normalization layer; the output end of the second convolution layer is connected with the input end of the second maximum pooling layer, and the output end of the second maximum pooling layer is connected with the input end of the third convolution layer and the input end of the second batch normalization layer; the output end of the third convolutional layer is connected with the input end of the third largest pooling layer; the output end of the third largest pooling layer is connected with the input end of the fourth convolutional layer; the output end of the fourth convolutional layer is connected with the input end of the fourth largest pooling layer; the output end of the fourth largest pooling layer is connected with the input end of the third batch normalization layer;
the characteristic combination layer is used for performing matrix transformation on output characteristic graphs of the first batch normalization layer, the second batch normalization layer and the third batch normalization layer and performing weighting calculation to obtain a characteristic vector, a first input end of the characteristic combination layer is connected with an output end of the first batch normalization layer, a second input end of the characteristic combination layer is connected with an output end of the second batch normalization layer, a third input end of the characteristic combination layer is connected with an output end of the third batch normalization layer, and an output end of the characteristic combination layer is connected with an input end of the fifth convolution layer;
the fifth convolution layer, the fifth maximum pooling layer, the first full-connection layer, the second full-connection layer, the third full-connection layer and the loss layer are sequentially connected in series, and the output end of the loss layer is used as the output end of the image recognition deep neural network model.
Preferably, the convolution kernel size of the first convolution layer is 7 × 7, the feature map dimension is 224 × 224, the convolution kernel depth is 16, and the convolution operation step size is 1;
the feature map dimension of the first maximum pooling layer is 112 × 112, and the depth is 16;
the convolution kernel size of the second convolution layer is 5 multiplied by 5, the feature map size is 112 multiplied by 113, the convolution kernel depth is 32, and the convolution operation step length is 1;
the feature map dimension of the second maximum pooling layer is 56 × 56, and the depth is 32;
the convolution kernel size of the third convolution layer is 3 × 3, the feature map size is 56 × 56, the convolution kernel depth is 64, and the convolution operation step size is 1;
the feature map dimension of the third maximum pooling layer is 28 × 28, and the depth is 64;
the convolution kernel size of the fourth convolution layer is 3 × 3, the feature map size is 28 × 28, the convolution kernel depth is 128, and the convolution operation step size is 1;
the feature map dimension of the fourth maximum pooling layer is 14 × 14, and the depth is 128;
the convolution kernel size of the fifth convolution layer is 3 × 3, the feature map size is 14 × 14, the convolution kernel depth is 256, and the convolution operation step size is 1;
the feature map dimension of the fifth maximum pooling layer is 7 × 7, and the depth is 256.
Preferably, the activation function of the first, second, third, fourth and/or fifth convolutional layer is:
wherein x is the convolution operation result of the first to fifth convolution layers, ReLU is the activation function value, and α is the proportionality coefficient.
Preferably, the loss function of the loss layer is:
Loss＝-y·ln(s)-(1-y)·ln(1-s)
y is a category label corresponding to an image training sample adopted when the deep neural network model is recognized by training an image, wherein the category label comprises a category 1 and a category 2; s is the output value of the loss layer; ln (·) is a natural logarithmic function; loss is the Loss function value.
Preferably, the power operation laboratory further comprises a plurality of temperature sensing units which are installed in a distributed manner in the power operation laboratory, the temperature sensing units are used for monitoring the equipment temperature of the power operation laboratory, and when the equipment temperature of the power operation laboratory is greater than a threshold temperature, the judgment module gives an early warning.
Preferably, the insulation mat further comprises a detection module, wherein the detection module is arranged at the insulation mat and used for judging whether a worker stands on the insulation mat;
and when the detection result of the detection module is inconsistent with the identification result of the insulating pad image identification module, the judgment module carries out early warning.
Compared with the prior art, the invention has the following advantages and beneficial effects:
the image live-action monitoring of the power operation site is taken as a breakthrough, the monitoring of the power operation site is realized through image sensing and image recognition, and the problem that a power operation laboratory in the current power transmission technology lacks a personnel electric shock risk detection system integrating equipment state detection and personnel operation monitoring is solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
fig. 1 is a structural diagram of a personnel electric shock risk detection system according to an embodiment of the present invention;
FIG. 2 is a diagram of an image recognition deep neural network model architecture;
FIG. 3 is a circuit diagram of an insulation pad detection circuit module according to an embodiment of the present invention;
FIG. 4 is a diagram of a resistance based weight sensor according to an embodiment of the present invention;
wherein the reference numerals are:
1. a weight sensor substrate; 2. a protective film; 3. a metal foil strain sensitive grid; 4. a first conductive line; 5. a second conductive line.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention.
Examples
An electric shock risk detection system for laboratory personnel, as shown in fig. 1, comprises:
the image sensing subsystem is used for acquiring an electric shock risk point image of the electric power operation laboratory; the electric shock risk point images comprise display screen images, wall corner grounding wire images, experiment table grounding wire images, insulating pad area images and insulating rod area images;
specifically, the image sensing subsystem in this embodiment includes a first image sensing module, a second image sensing module, a third image sensing module, a fourth image sensing module, and a fifth image sensing module. The first image sensing module is used for shooting a display screen image of the power operation laboratory; the second image sensing module is used for shooting a grounding state image of a wall corner grounding wire of the power operation laboratory; the third image sensing module is used for shooting a grounding state image of a laboratory table grounding wire of the electric power operation laboratory; the fourth image sensing module is used for shooting an image of an area where an insulating pad of the power operation laboratory is located; the fifth image sensing module is used for shooting an image of an area where an insulating rod of the power operation laboratory is located.
The image identification subsystem is used for identifying and outputting the category of the electric shock risk point image;
the categories include category 1 and category 2;
the type 1 is corresponding to a display screen displaying a first preset picture, a wall corner grounding wire, a laboratory bench grounding wire, an experimenter standing on an insulating pad for experimental operation and a discharging rod hanging on an insulating rod; the first preset picture in the embodiment is set as an experiment start word;
the category 2 corresponds to a display screen displaying a second preset picture, a wall corner grounding wire is not grounded, a laboratory bench grounding wire is not grounded, an experimenter does not stand on an insulating pad for experimental operation, and a discharge rod is not hung on an insulating rod; the second preset frame in this embodiment is set to the word "experiment end".
Specifically, the image recognition subsystem in the present embodiment includes:
the experiment state image identification module is used for identifying whether the electric power operation experiment starts or ends according to the display screen image and outputting a category 1 when the display screen displays an experiment starting character; outputting a category 2 when the display screen displays a character of 'experiment end';
the corner grounding image recognition module is used for recognizing whether the corner grounding wire is grounded according to the corner grounding wire image, and outputting a category 1 when the corner grounding wire is grounded; otherwise, outputting the category 2;
the experiment table grounding image identification module is used for identifying whether the experiment table grounding wire is grounded according to the experiment table grounding wire image; outputting a category 1 when the ground wire of the experiment table is grounded; otherwise, outputting the category 2;
the insulating pad image identification module is used for identifying whether an experimenter stands on the insulating pad to perform experimental operation according to the image of the area where the insulating pad is located; outputting category 1 when an experimenter stands on the insulating pad to perform experimental operation; otherwise, outputting the category 2;
the insulating rod image identification module is used for identifying whether the discharge rod is hung on the insulating rod according to the image of the area where the insulating rod is located; outputting the category 1 when the discharge rod is not hung on the insulating rod; otherwise, category 2 is output.
And the risk judgment module is used for judging the electric shock risk of the experimenters in the electric power operation laboratory according to the identification result.
Specifically, the method comprises the following steps:
if the wall corner grounding wire is not grounded, the experiment table grounding wire is not grounded or the experiment personnel do not stand on the insulating pad for experiment operation in the experiment process, the experiment personnel have electric shock risks; during an experiment, the output of the corner grounding image recognition module, the experiment table grounding image recognition module and the insulation pad image recognition module is obtained in real time to recognize the states of a corner grounding wire, an experiment table grounding wire and an experimenter, and then whether the experimenter has an electric shock risk is judged; and whether the test device is in the experimental state is judged through a first preset picture displayed by the display screen. Therefore, during an experiment, if the experimental state image recognition module, the corner grounding image recognition module, the laboratory table grounding image recognition module, the insulating pad image recognition module and the insulating rod image recognition module all output the category 1, it is described that the connection of equipment and the operation of experimenters are all in a standard range in the experimental process, no contact risk exists, and the risk judgment module judges that the experimenters in the electric power operation laboratory have no electric shock risk;
after the experiment is finished, if the discharging rod is not hung on the insulating rod, the electric shock risk also exists, so after the experiment is finished, whether the electric shock risk exists in the laboratory personnel in the electric power operation laboratory or not needs to be judged, and whether the electric shock risk exists in the experiment finished state or not is judged through a second preset picture displayed by the display screen. Therefore, after the experiment is finished, if the experimental state image recognition module and the insulating rod image recognition module both output the category 2, it is shown that after the experiment is finished, the discharge rod is hung on the insulating rod, no contact risk exists, and the risk judgment module judges that the experimenter in the electric power operation laboratory has no electric shock risk;
otherwise, the risk judgment module judges that the experimenter of the electric power operation laboratory has electric shock risk.
In electric power experiment operation process, because in the experimentation because of experimenter carelessness or to electrocute the problem attach inadequately, often appear because of the corner earth connection not ground connection, the laboratory bench earth connection is not ground connection, the staff does not stand the operation on the insulating pad and the staff hangs the discharge rod on the insulator spindle and leads to the phenomenon that the experimenter electrocuted after the experiment, based on this, this application is guarantee experimenter's personal safety, an electric power operation laboratory experimenter electrocute risk detection system is provided, through establishing independent image sensing and the recognition technology who intakes and separately discerns, come to discern latent electrocute risk, avoid the emergence of electrocute accident, and simultaneously, each module independent work in this embodiment, can effectively reduce the whole malfunctioning probability of system.
Further, the image recognition subsystem in this embodiment performs image recognition by using an image recognition deep neural network model, and the image recognition deep neural network model includes, as shown in fig. 2, a first convolution layer, a first maximum pooling layer, a second convolution layer, a second maximum pooling layer, a third convolution layer, a third maximum pooling layer, a fourth convolution layer, a fourth maximum pooling layer, a first batch normalization layer, a second batch normalization layer, a third batch normalization layer, a feature combination layer, a fifth convolution layer, a fifth maximum pooling layer, a first full-link layer, a second full-link layer, a third full-link layer, and a loss layer.
The input end of the first convolution layer is used as the input end of the image recognition deep neural network model, and the output end of the first convolution layer is connected with the input end of the first maximum pooling layer. The output end of the first maximum pooling layer is respectively connected with the input end of the second convolution layer and the input end of the first batch normalization layer; the output end of the second convolution layer is connected with the input end of the second maximum pooling layer; the output end of the second maximum pooling layer is respectively connected with the input end of the third convolution layer and the input end of the second batch normalization layer; the output end of the third convolution layer is connected with the input end of the third maximum pooling layer; the output end of the third maximum pooling layer is connected with the input end of the fourth convolution layer; the output end of the fourth convolution layer is connected with the input end of the fourth maximum pooling layer; and the output end of the fourth maximum pooling layer is connected with the input end of the third batch normalization layer.
The characteristic combination layer is used for performing matrix transformation on output characteristic graphs of the first batch normalization layer, the second batch normalization layer and the third batch normalization layer, and performing weighted calculation to obtain characteristic vectors, wherein a first input end of the characteristic combination layer is connected with an output end of the first batch normalization layer, a second input end of the characteristic combination layer is connected with an output end of the second batch normalization layer, a third input end of the characteristic combination layer is connected with an output end of the third batch normalization layer, and an output end of the characteristic combination layer is connected with an input end of the fifth convolution layer.
In this embodiment, the feature combination layer converts the output feature map of the first batch normalization layer, the output feature map of the second batch normalization layer, and the output feature map of the third batch normalization layer into corresponding intermediate feature vectors, respectively, and then performs the following steps according to a ratio of 1: 2: and 7, adding the three intermediate feature vectors, and calculating to obtain the feature vectors.
The fifth convolution layer, the fifth maximum pooling layer, the first full-connection layer, the second full-connection layer, the third full-connection layer and the loss layer are sequentially connected in series. And the output end of the loss layer is used as the output end of the image recognition deep neural network model.
Through the design, the image recognition deep neural network model with high applicability is realized, and the image recognition deep neural network model can be used for image recognition of different scenes after image samples with different contents and different labels are input for training. Different from the existing neural network model structure, the invention designs the feature combination layer while sequentially stacking the convolution layer and the maximum pooling layer to extract features step by step, and fuses feature information of convolution operation and maximum pooling operation of each layer. The theoretical basis is that the more layers of the convolutional neural network are, the more abstract the extracted features are, the easier the analysis processing is, but the image part information is easy to lose, and the accuracy of image identification and classification is reduced. Therefore, the accuracy of image recognition can be improved by adopting a characteristic combination mode.
Further, the convolution kernel size of the first convolution layer in this embodiment is 7 × 7, the feature map dimension is 224 × 224, the convolution kernel depth is 16, and the convolution operation step size is 1. The first largest pooling layer has a feature map dimension of 112 x 112 with a depth of 16. The convolution kernel size of the second convolution layer is 5 × 5, the feature map size is 112 × 112, the convolution kernel depth is 32, and the convolution step size is 1. The second largest pooling layer had a feature map dimension of 56 x 56 with a depth of 32. The convolution kernel size of the third convolution layer is 3 × 3, the feature map size is 56 × 56, the convolution kernel depth is 64, and the convolution operation step size is 1. The third largest pooling layer has a feature map dimension of 28 x 28 and a depth of 64. The convolution kernel size of the fourth convolution layer is 3 × 3, the feature map size is 28 × 28, the convolution kernel depth is 128, and the convolution operation step size is 1. The feature map dimension of the fourth largest pooling layer is 14 × 14, with a depth of 128. The convolution kernel size of the fifth convolution layer is 3 × 3, the feature map size is 14 × 14, the convolution kernel depth is 256, and the convolution operation step size is 1. The feature map dimension of the fifth largest pooling layer is 7 × 7, with a depth of 256. The activation functions of the first, second, third, fourth, and fifth convolutional layers are:
wherein x is the convolution operation result of the first to fifth convolution layers, ReLU is the activation function value, and α is the proportionality coefficient. In this example, α is 0.001 and the loss function of the loss layer is:
Loss＝-y·ln(s)-(1-y)·ln(1-s)
y is a category label corresponding to an image training sample adopted when the deep neural network model is recognized by training an image, wherein the category label comprises a category 1 and a category 2; s is the output value of the loss layer; ln (·) is a natural logarithmic function; loss is the Loss function value.
The design of each layer parameter of the image recognition deep neural network model can fully utilize the multilayer structure of the image recognition deep neural network model designed by the invention, so that the convolution operation and the maximum pooling operation have good effects and do not occupy larger computing resources; in order to prevent the gradient from disappearing, save the calculated amount, relieve the overfitting phenomenon of the neural network training effectively, the activation function of every convolution layer does not adopt the ordinary linear rectification function that the conventional technical means provides, but has designed the activation function of the sectional variable slope, the activation function can keep the negative axis value, make the gradient in the course of training of the neural network propagate normally, avoid the neural cell of the neural network to shut down the state; the loss function design of the invention combines the requirements of the invention on the image recognition deep neural network model, and each image recognition module is used for the Boolean type logical judgment of the electric shock risk, so that the output value s of the loss layer represents the probability judgment of whether the input image has the electric shock risk of personnel or not by the image recognition deep neural network model, therefore, the binary loss function meets the requirements of the invention, and compared with the conventional loss function design based on mean square error and other modes, the training progress can be improved, and the calculation cost is saved.
It should be noted that the method of identifying whether the power operation experiment is started or ended, whether the wall corner grounding wire is grounded, whether the experiment table grounding wire is grounded, whether the experimenter stands on the insulating pad to perform the experiment operation, and whether the discharging rod is hung on the insulating rod through the image identification subsystem is not limited to the above method, and the method of image comparison may be adopted to perform the judgment. Specifically, the method comprises the following steps:
carrying out feature extraction on the picture acquired by any image sensing subsystem through an SIFT feature extraction algorithm to obtain all feature points and the number F of the feature points1i；
Performing feature extraction on the experimenter operation picture or the equipment wiring picture meeting the specification through an SIFT feature extraction algorithm to obtain all feature points and the number F of the feature points2i；
Matching the picture acquired by the image sensing subsystem with the experimental personnel operation picture or the equipment wiring picture which meets the specification to obtain the number M of matched characteristic pointsi；
Calculating the similarity R between the picture acquired by the image sensing subsystem and the experimental personnel operation picture or equipment wiring picture meeting the specificationiThe similarity calculation formula is as follows:
at a degree of similarity RiWhen the value is larger than the threshold value, the operation specification or the equipment wiring specification of an experimenter is considered, namely: degree of similarity RiWhen the voltage is larger than the threshold value, the electric power operation experiment can be started, the wall corner grounding wire is grounded, the experiment table grounding wire is grounded, an experimenter stands on the insulating pad to perform experiment operation or the discharging rod is hung on the insulating rod.
Furthermore, considering that the power equipment is different from the weak current equipment, when the power equipment has a potential fault hazard, the power equipment usually generates heat along with an abnormality, so that a plurality of temperature sensing units are further arranged in the embodiment, the temperature sensing units are installed on the power equipment in the power operation laboratory in a distributed manner, the power operation laboratory is subjected to omnibearing temperature distributed detection, and when the temperature of any one power equipment in the power operation laboratory is detected to be greater than a threshold temperature, the judgment module gives an early warning to predict the electric shock risk in advance.
Further, in order to prevent the insulating pad image recognition module from having false detection or missing detection, namely: in the experiment process, an experimenter does not stand on the insulating pad to perform experiment operation, but the insulating pad image recognition module outputs the condition of the category 1, and the detecting module is further arranged at the insulating pad to judge whether the experimenter stands on the insulating pad; and when the detection data of the detection module is inconsistent with the image identification module of the insulating pad, the judgment module carries out early warning.
Specifically, when the detection module detects that an experimenter stands on the insulating pad, the detection module transmits a detection result to the judgment module, the judgment module judges whether the experimenter is performing an electric power experiment in an electric power operation laboratory according to the output of the experimental state image identification module, if the experimenter is performing the electric power experiment, the identification result of the insulating pad image identification module is obtained, if the detection result of the detection module is inconsistent with the insulating pad image identification module, the condition that the insulating pad image identification module is subjected to false detection or missing detection is indicated, the electric shock risk detection system has a fault, and at the moment, the judgment module performs early warning and error reporting; and if the judgment module judges that no laboratory personnel are performing the electric power experiment in the electric power operation laboratory according to the output of the experiment state image identification module, no early warning is performed.
Preferably, in the specific implementation, in order to reduce the number of times of early warning or avoid the early warning situation caused by false detection or missed detection of the insulating pad image identification module due to accidental faults, when the detection result of the detection module is inconsistent with the identification result of the insulating pad image identification module, the fourth image sensing module restarts and performs self-checking, and if the restart identification result is still in conflict with the detection result of the detection module, the judgment module performs system error reporting.
Specifically, as shown in fig. 3, the detection module in this embodiment includes: a resistor R1, a resistor R2, a resistance type weight sensor R3, a resistor R4, a resistor R5, a resistor R6, a resistor R7, a resistor R8, a resistor R9, a resistor R10, a resistor R11, a resistor R12, a resistor R13, a resistor R14, a resistor R15, a resistor R16, a resistor R17, a resistor R18, a resistor R19, a resistor R20, a capacitor C1, a capacitor C2, a capacitor C3 and a capacitor C4, a capacitor C5, a PNP transistor Q1, an NPN transistor Q2, an NPN transistor Q3, a PNP transistor Q4, a PNP transistor Q5, an NPN transistor Q6, an NPN transistor Q7, a PNP transistor Q8, a PNP transistor Q9, an NPN transistor Q10, an NPN transistor Q11, an NPN transistor Q12, an NPN transistor Q13, a PNP transistor Q14, an NPN transistor Q15, an NPN transistor Q16, an NPN transistor Q17, a PNP transistor Q18, an NPN transistor Q19, a PNP transistor Q20, an NPN transistor Q21, an NPN transistor Q22, and a PNP transistor Q23;
one end of the resistor R1 is respectively connected with one end of the resistor R2, an emitter of the PNP type triode Q1, an emitter of the PNP type triode Q4, an emitter of the PNP type triode Q5, a collector of the NPN type triode Q10, a collector of the NPN type triode Q13, an emitter of the PNP type triode Q14 and a collector of the NPN type triode Q22, and is used as a power supply end VCC of the insulating pad detection circuit module;
the other end of the resistor R1 is respectively connected with one end of the resistor R5 and a first electrical interface of the resistance weight sensor R3;
the other end of the resistor R2 is respectively connected with one end of the resistor R4 and one end of the resistor R7;
the other end of the resistor R4 is connected with the second electrical interface of the resistance-type weight sensor R3 and is grounded;
the other end of the resistor R5 is respectively connected with one end of a resistor R6, one end of a capacitor C1, one end of a capacitor C2 and the base electrode of an NPN type triode Q6;
the other end of the capacitor C1 is grounded;
the other end of the resistor R6 is respectively connected with the other end of the capacitor C2, one end of the resistor R19, one end of the resistor R20, the emitter of the NPN type triode Q19 and the emitter of the PNP type triode Q20, and is used as an output end Vout of the insulating pad detection circuit module;
the other end of the resistor R7 is respectively connected with one end of the capacitor C3, one end of the resistor R8 and the base electrode of the NPN type triode Q7;
the other end of the capacitor C3 is grounded; the other end of the resistor R8 is grounded;
the base electrode of the PNP type triode Q1 is respectively connected with the collector electrode of the PNP type triode Q1, one end of the resistor R9 and the base electrode of the PNP type triode Q14;
the other end of the resistor R9 is respectively connected with the collector of an NPN type triode Q2, the base of an NPN type triode Q2 and the base of an NPN type triode Q3;
the emitter of the NPN type triode Q2 is grounded;
the base electrode of the PNP type triode Q4 is respectively connected with the base electrode of the PNP type triode Q5, the collector electrode of the PNP type triode Q5, the collector electrode of the NPN type triode Q6 and the collector electrode of the NPN type triode Q7;
the collector of the PNP type triode Q4 is respectively connected with the collector of the NPN type triode Q3, the base of the PNP type triode Q8 and the base of the PNP type triode Q9;
an emitter of the NPN type triode Q3 is connected with one end of the resistor R10;
the other end of the resistor R10 is grounded;
an emitting electrode of the NPN type triode Q6 is connected with an emitting electrode of the PNP type triode Q8;
the collector of the PNP triode Q8 is respectively connected with the base of the NPN triode Q10 and the collector of the NPN triode Q11;
an emitter of the NPN type triode Q10 is respectively connected with a base of the NPN type triode Q11, one end of the resistor R12 and a base of the NPN type triode Q12;
the other end of the resistor R12 is grounded;
an emitter of the NPN type triode Q11 is connected with one end of the resistor R11;
the other end of the resistor R11 is grounded;
an emitting electrode of the NPN type triode Q7 is connected with an emitting electrode of the PNP type triode Q9;
a collector of the PNP triode Q9 is respectively connected with a collector of the NPN triode Q12, a base of the NPN triode Q13, one end of the resistor R16 and one end of the capacitor C5;
an emitter of the NPN type triode Q12 is connected with one end of the resistor R13;
the other end of the resistor R13 is grounded;
an emitter of the NPN type triode Q13 is respectively connected with one end of a resistor R14 and a base of the NPN type triode Q15;
the other end of the resistor R14 is grounded;
the other end of the resistor R16 is connected with one end of the capacitor C4;
the other end of the capacitor C4 is respectively connected with a collector of a PNP type triode Q14, a base of an NPN type triode Q16, a collector of an NPN type triode Q16, a collector of an NPN type triode Q17, a collector of an NPN type triode Q19, a base of an NPN type triode Q22, a base of the PNP type triode Q18 and a collector of the NPN type triode Q15;
an emitter of the NPN type triode Q15 is connected with one end of the resistor R15;
the other end of the resistor R15 is grounded;
an emitter of the NPN type triode Q16 is respectively connected with a base of the NPN type triode Q17 and one end of the resistor R17;
the other end of the resistor R17 is respectively connected with an emitting electrode of an NPN type triode Q17, the other end of the capacitor C5, an emitting electrode of a PNP type triode Q18 and a base electrode of the PNP type triode Q23;
the collector of the PNP type triode Q18 is grounded;
the base electrode of the NPN type triode Q19 is respectively connected with the emitter electrode of the NPN type triode Q22 and the other end of the resistor R19;
the base electrode of the PNP type triode Q20 is respectively connected with the other end of the resistor R20 and the emitter electrode of the PNP type triode Q23;
the collector of the PNP type triode Q20 is respectively connected with the collector of the NPN type triode Q21, the base of the NPN type triode Q21 and one end of the resistor R18;
the emitter of the NPN type triode Q21 is grounded;
the other end of the resistor R18 is grounded;
the collector of the PNP transistor Q23 is grounded.
In the present embodiment, a bridge resistance value measuring circuit of the resistance weight sensor R3 is configured by the resistor R1, the resistor R2, and the resistor R4; the input end of the cascade amplifying circuit is high in equivalent impedance, the output end of the cascade amplifying circuit is high in loading capacity, and the power supply rejection ratio is high; frequency compensation is carried out through the resistor R16, the capacitor C4 and the capacitor C5, and the frequency stability of the circuit is guaranteed; the resistor R5 and the resistor R6 form deep negative feedback, and proportional amplification controlled by the resistance values of the resistor R5 and the resistor R6 is achieved. The resistance value of the resistance-type weight sensor R3 is converted into a recognizable voltage signal, and the resistance value of the resistance-type weight sensor R3 directly reflects the weight borne by the sensor, so that whether a worker stands on an insulating pad to perform electric power experiment operation can be recognized by placing the resistance-type weight sensor R3 on the insulating pad.
As shown in fig. 4, the resistive weight sensor R3 includes a weight sensor substrate 1, a protective film 2, a metal foil strain sensitive grid 3, a first lead 4 and a second lead 5;
the metal foil strain sensitive grid 3 is a snakelike wound metal film resistor, one end of the metal foil strain sensitive grid is connected with a first lead 4, the other end of the metal foil strain sensitive grid is connected with a second lead 5, the metal foil strain sensitive grid is arranged on the weight sensor substrate 1, and the surface of the metal foil strain sensitive grid is pasted with a protective film 2; the first wire 4 serves as a first electrical interface for the resistive weight sensor R3; the second wire 5 serves as a second electrical interface for the resistive weight sensor R3.
The structural design of the resistance type weight sensor R3 is convenient for being placed on the insulating pad, and the performance of the insulating pad is not influenced.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. The utility model provides an electric power operation laboratory is electrocuted and is operated, includes:
the image sensing subsystem is used for acquiring an electric shock risk point image of the electric power operation laboratory;
the electric shock risk point image comprises a display screen image, a wall corner grounding wire image, a laboratory bench grounding wire image, an insulating pad area image and an insulating rod area image;
the image identification subsystem is used for identifying and outputting the category of the electric shock risk point image;
the categories include category 1 and category 2;
the type 1 corresponds to a display screen displaying a first preset picture, a wall corner grounding wire, a laboratory bench grounding wire, an experimenter standing on an insulating pad for experimental operation and a discharge rod hanging on an insulating rod;
the category 2 corresponds to a display screen displaying a second preset picture, a wall corner grounding wire is not grounded, a laboratory bench grounding wire is not grounded, an experimenter does not stand on an insulating pad for experimental operation, and a discharging rod is not hung on an insulating rod;
and the judgment module is used for judging whether the laboratory personnel of the electric power operation laboratory have electric shock risks or not according to the identification result.
2. The electric power operations laboratory worker electric shock risk detection system of claim 1, wherein the image recognition subsystem comprises:
the experiment state image recognition module is used for recognizing whether the power operation experiment starts or ends according to the display screen image, judging that the power operation experiment starts when the display screen displays a first preset image, and outputting a category 1; when a second preset image is displayed on the display screen, judging that the power operation experiment is finished, and outputting a category 2;
the corner grounding image recognition module is used for recognizing whether the corner grounding wire is grounded according to the corner grounding wire image and outputting a category 1 when the corner grounding wire is grounded; otherwise, outputting the category 2;
the experiment table grounding image identification module is used for identifying whether the experiment table grounding wire is grounded according to the experiment table grounding wire image; outputting a category 1 when the ground wire of the experiment table is grounded; otherwise, outputting the category 2;
the insulating pad image identification module is used for identifying whether an experimenter stands on the insulating pad to perform experimental operation according to the image of the area where the insulating pad is located; outputting class 1 when the experimenter stands on the insulating pad for experimental operation; otherwise, outputting the category 2;
the insulating rod image identification module is used for identifying whether the discharge rod is hung on the insulating rod according to the image of the area where the insulating rod is located; outputting the category 1 when the discharge rod is not hung on the insulating rod; otherwise, category 2 is output.
3. The electric power operation laboratory worker electric shock risk detection system according to claim 2, wherein when the experiment state image recognition module, the corner grounding image recognition module, the laboratory table grounding image recognition module and the insulating mat image recognition module all output category 1, the judgment module judges that the electric power operation laboratory worker has no electric shock risk;
when the experimental state image recognition module and the insulating rod image recognition module output the category 2, the judgment module judges that the experimenters in the electric power operation laboratory have no electric shock risk.
4. The electric power operation laboratory worker electric shock risk detection system according to any one of claims 1 to 3, wherein the experiment state image recognition module, the corner grounding image recognition module, the laboratory table grounding image recognition module, the insulation pad image recognition module and the insulation rod image recognition module are all used for image recognition by adopting an image recognition deep neural network model.
5. The electric power operations laboratory worker electric shock risk detection system of claim 4 wherein the image recognition deep neural network model comprises: the first rolling layer, the first maximum pooling layer, the second rolling layer, the second maximum pooling layer, the third rolling layer, the third maximum pooling layer, the fourth rolling layer, the fourth maximum pooling layer, the first batch normalizing layer, the second batch normalizing layer, the third batch normalizing layer, the feature combining layer, the fifth rolling layer, the fifth maximum pooling layer, the first full-connection layer, the second full-connection layer, the third full-connection layer and the loss layer;
the output end of the first convolution layer is connected with the input end of the first maximum pooling layer, and the output end of the first maximum pooling layer is connected with the input end of the second convolution layer and the input end of the first batch normalization layer; the output end of the second convolution layer is connected with the input end of the second maximum pooling layer, and the output end of the second maximum pooling layer is connected with the input end of the third convolution layer and the input end of the second batch normalization layer; the output end of the third convolutional layer is connected with the input end of the third largest pooling layer; the output end of the third largest pooling layer is connected with the input end of the fourth convolutional layer; the output end of the fourth convolutional layer is connected with the input end of the fourth largest pooling layer; the output end of the fourth largest pooling layer is connected with the input end of the third batch normalization layer;
the characteristic combination layer is used for performing matrix transformation on output characteristic graphs of the first batch normalization layer, the second batch normalization layer and the third batch normalization layer and performing weighting calculation to obtain a characteristic vector, a first input end of the characteristic combination layer is connected with an output end of the first batch normalization layer, a second input end of the characteristic combination layer is connected with an output end of the second batch normalization layer, a third input end of the characteristic combination layer is connected with an output end of the third batch normalization layer, and an output end of the characteristic combination layer is connected with an input end of the fifth convolution layer;
the fifth convolution layer, the fifth maximum pooling layer, the first full-connection layer, the second full-connection layer, the third full-connection layer and the loss layer are sequentially connected in series, and the output end of the loss layer is used as the output end of the image recognition deep neural network model.
6. The electric power operation laboratory worker electric shock risk detection system of claim 5, wherein the convolution kernel size of the first convolution layer is 7 x 7, the feature map dimension is 224 x 224, the convolution kernel depth is 16, and the convolution operation step size is 1;
the feature map dimension of the first maximum pooling layer is 112 × 112, and the depth is 16;
the convolution kernel size of the second convolution layer is 5 multiplied by 5, the feature map size is 112 multiplied by 112, the convolution kernel depth is 32, and the convolution operation step length is 1;
the feature map dimension of the second maximum pooling layer is 56 × 56, and the depth is 32;
the convolution kernel size of the third convolution layer is 3 × 3, the feature map size is 56 × 56, the convolution kernel depth is 64, and the convolution operation step size is 1;
the feature map dimension of the third maximum pooling layer is 28 × 28, and the depth is 64;
the convolution kernel size of the fourth convolution layer is 3 × 3, the feature map size is 28 × 28, the convolution kernel depth is 128, and the convolution operation step size is 1;
the feature map dimension of the fourth maximum pooling layer is 14 × 14, and the depth is 128;
the convolution kernel size of the fifth convolution layer is 3 × 3, the feature map size is 14 × 14, the convolution kernel depth is 256, and the convolution operation step size is 1;
the feature map dimension of the fifth maximum pooling layer is 7 × 7, and the depth is 256.
7. An electric power operations laboratory personnel electric shock risk detection system according to claim 5 wherein the activation function of said first convolutional layer, said second convolutional layer, said third convolutional layer, said fourth convolutional layer and/or said fifth convolutional layer is:
wherein x is the convolution operation result of the first to fifth convolution layers, ReLU is the activation function value, and α is the proportionality coefficient.
8. The electric power operation laboratory worker electric shock risk detection system of claim 5, wherein the loss function of the loss layer is:
Loss＝-y·ln(s)-(1-y)·ln(1-s)
y is a category label corresponding to an image training sample adopted when the deep neural network model is recognized by training an image, wherein the category label comprises a category 1 and a category 2; s is the output value of the loss layer; ln (·) is a natural logarithmic function; loss is the Loss function value.
9. The electric power operation laboratory worker electric shock risk detection system according to claim 1, further comprising a plurality of temperature sensing units installed in the electric power operation laboratory in a distributed manner, wherein the temperature sensing units are used for monitoring the equipment temperature of the electric power operation laboratory, and when the equipment temperature of the electric power operation laboratory is greater than a threshold temperature, the judgment module gives an early warning.
10. The electric power operation laboratory worker electric shock risk detection system according to claim 1, further comprising a detection module, wherein the detection module is arranged at the insulating pad and used for judging whether a worker stands on the insulating pad;
and when the detection result of the detection module is inconsistent with the identification result of the insulating pad image identification module, the judgment module carries out early warning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111470160.9A CN114092888B (en) | 2021-12-03 | 2021-12-03 | Electric shock risk detection system for laboratory personnel in electric power operation laboratory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111470160.9A CN114092888B (en) | 2021-12-03 | 2021-12-03 | Electric shock risk detection system for laboratory personnel in electric power operation laboratory |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114092888A true CN114092888A (en) | 2022-02-25 |
| CN114092888B CN114092888B (en) | 2023-05-02 |
Family
ID=80306488
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111470160.9A Active CN114092888B (en) | 2021-12-03 | 2021-12-03 | Electric shock risk detection system for laboratory personnel in electric power operation laboratory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114092888B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114926969A (en) * | 2022-05-10 | 2022-08-19 | 国网湖北省电力有限公司营销服务中心(计量中心) | Intelligent laboratory safety control system and method |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105809355A (en) * | 2016-03-14 | 2016-07-27 | 国网福建省电力有限公司 | Personnel information identification method based on working field remote monitoring system |
| CN105825329A (en) * | 2016-03-14 | 2016-08-03 | 国网福建省电力有限公司 | Safety and quality inspection method based on remote working site monitoring system |
| CN106376987A (en) * | 2016-08-25 | 2017-02-08 | 国网山东省电力公司博兴县供电公司 | Power system electrician work arm portion safety protector |
| CN109118081A (en) * | 2018-08-08 | 2019-01-01 | 成都保源酷码科技有限公司 | A kind of operation safety supervisory systems and method based on image procossing mode |
| JP2019053134A (en) * | 2017-09-13 | 2019-04-04 | 一般財団法人九州電気保安協会 | Electric facility inspection device |
| CN111507308A (en) * | 2020-05-07 | 2020-08-07 | 广东电网有限责任公司 | Transformer substation safety monitoring system and method based on video identification technology |
| CN111983396A (en) * | 2020-07-16 | 2020-11-24 | 中广核核电运营有限公司 | Electric high-voltage test area safety protection system |
| CN112347916A (en) * | 2020-11-05 | 2021-02-09 | 安徽继远软件有限公司 | Power field operation safety monitoring method and device based on video image analysis |
| CN112530115A (en) * | 2020-11-17 | 2021-03-19 | 云南电网有限责任公司 | Electric power operation personnel protection against electric shock scene intelligence supervises integrated equipment |
| US20210182615A1 (en) * | 2019-12-16 | 2021-06-17 | Yongchuan Power Supply Branch, State Grid Chongqing Electric Power Company | Alexnet-based insulator self-explosion recognition method |
| CN112990586A (en) * | 2021-03-22 | 2021-06-18 | 海南电网有限责任公司澄迈供电局 | Intelligent video monitoring method and system for distribution network operation |
-
2021
- 2021-12-03 CN CN202111470160.9A patent/CN114092888B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105809355A (en) * | 2016-03-14 | 2016-07-27 | 国网福建省电力有限公司 | Personnel information identification method based on working field remote monitoring system |
| CN105825329A (en) * | 2016-03-14 | 2016-08-03 | 国网福建省电力有限公司 | Safety and quality inspection method based on remote working site monitoring system |
| CN106376987A (en) * | 2016-08-25 | 2017-02-08 | 国网山东省电力公司博兴县供电公司 | Power system electrician work arm portion safety protector |
| JP2019053134A (en) * | 2017-09-13 | 2019-04-04 | 一般財団法人九州電気保安協会 | Electric facility inspection device |
| CN109118081A (en) * | 2018-08-08 | 2019-01-01 | 成都保源酷码科技有限公司 | A kind of operation safety supervisory systems and method based on image procossing mode |
| US20210182615A1 (en) * | 2019-12-16 | 2021-06-17 | Yongchuan Power Supply Branch, State Grid Chongqing Electric Power Company | Alexnet-based insulator self-explosion recognition method |
| CN111507308A (en) * | 2020-05-07 | 2020-08-07 | 广东电网有限责任公司 | Transformer substation safety monitoring system and method based on video identification technology |
| CN111983396A (en) * | 2020-07-16 | 2020-11-24 | 中广核核电运营有限公司 | Electric high-voltage test area safety protection system |
| CN112347916A (en) * | 2020-11-05 | 2021-02-09 | 安徽继远软件有限公司 | Power field operation safety monitoring method and device based on video image analysis |
| CN112530115A (en) * | 2020-11-17 | 2021-03-19 | 云南电网有限责任公司 | Electric power operation personnel protection against electric shock scene intelligence supervises integrated equipment |
| CN112990586A (en) * | 2021-03-22 | 2021-06-18 | 海南电网有限责任公司澄迈供电局 | Intelligent video monitoring method and system for distribution network operation |
Non-Patent Citations (6)
| Title |
|---|
| RAO B等: "A Method of Extracting and Recognizing Discharge Current of Bird Hazard Hidden Danger Based on Wavelet Transform" * |
| ZHAO Y B等: "Using Communication Networks in Control Systems: The Theoretical and Practical Challenges" * |
| 刘家军等: "电力检修作业挂接地线可视化监测装置" * |
| 尹远: "电力设备运行环境监控图像异常检测算法研究" * |
| 常政威等: "基于机器学习和图像识别的电力作业现场安全监督方法" * |
| 颜廷良: "基于机器学习和图像识别的电力作业现场安全监督" * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114926969A (en) * | 2022-05-10 | 2022-08-19 | 国网湖北省电力有限公司营销服务中心(计量中心) | Intelligent laboratory safety control system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114092888B (en) | 2023-05-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Afrasiabi et al. | Integration of accelerated deep neural network into power transformer differential protection | |
| CN109766952B (en) | Photovoltaic array fault detection method based on partial least square method and extreme learning machine | |
| CN107907791A (en) | A kind of automatic checkout system and detection method of distribution network failure indicator | |
| CN114092888A (en) | Electric power operation laboratory personnel risk detection system that electrocutes | |
| WO2023280115A1 (en) | Vehicle electric leakage detection method and apparatus | |
| CN111323734A (en) | General detection system and detection method for conductivity of external hanging device of airplane weapon | |
| CN108872803A (en) | A kind of detection method of the transformer insulation state based on dielectric return voltage | |
| CN114255562A (en) | Wisdom fire control early warning system based on thing networking | |
| CN117173100A (en) | Polymer lithium ion battery production control system and method thereof | |
| CN115077722B (en) | Partial discharge and temperature comprehensive monitoring system and method applied to high-voltage cabinet | |
| CN217156655U (en) | Non-contact electrostatic detector | |
| CN111352365A (en) | Dustproof ventilation type electric power and electrical equipment cabinet and control method | |
| CN102299511A (en) | Novel intelligent network monitoring type power supply lightning protection device system | |
| CN113269941A (en) | Electrical fire alarm device based on multi-information fusion judgment and control method | |
| CN202350927U (en) | Intelligently universal digital display device | |
| CN211123037U (en) | Lightning protection ground resistance detecting system | |
| CN108845220B (en) | Battery system ground fault detection device and method | |
| Zhijian et al. | Fault location of transformer winding deformation using frequency response analysis | |
| CN114387391A (en) | Safety monitoring method and device for transformer substation equipment, computer equipment and medium | |
| CN207798304U (en) | A kind of strain-type rectangular sensor | |
| CN206074691U (en) | The test device of becket DC impedance in fingerprint recognition module | |
| CN105258729A (en) | Temperature and humidity detection system for switch cabinet based on single chip microcomputer | |
| CN111159650A (en) | Artificial intelligence electric line aging degree detection method and system | |
| CN111289853A (en) | Channel-space attention mechanism-based insulator detection system and algorithm | |
| CN219799745U (en) | Parameter display system |
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 |