CN106981063A - A kind of grid equipment state monitoring apparatus based on deep learning - Google Patents
A kind of grid equipment state monitoring apparatus based on deep learning Download PDFInfo
- Publication number
- CN106981063A CN106981063A CN201710143974.9A CN201710143974A CN106981063A CN 106981063 A CN106981063 A CN 106981063A CN 201710143974 A CN201710143974 A CN 201710143974A CN 106981063 A CN106981063 A CN 106981063A
- Authority
- CN
- China
- Prior art keywords
- deep learning
- grid equipment
- equipment state
- video
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/183—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a single remote source
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/02—Recognising information on displays, dials, clocks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
A kind of grid equipment state monitoring apparatus based on deep learning, the device includes video acquisition unit, grid equipment detection unit, display unit and storage element.The industrial camera of video acquisition unit by the video information collected by network cable transmission to Video Decoder, while through network cable transmission to storage element carry out data backup.By extracting the frame of video of monitoring video, an image library dedicated for recognizing grid equipment state is built.Electrical device status data transfer in database module to deep learning training module is carried out deep learning modeling by electrical equipment detection unit, the model that deep learning identification module is set up using deep learning training module carries out identification and classification to electrical device status, while being shown and being stored identification and classification result.The present invention effectively alleviates the pressure of manpower monitoring, reaches intellectual monitoring truly.
Description
Technical field
The invention belongs to artificial intelligence power network video monitoring technical field, it is related to a kind of grid equipment based on deep learning
State monitoring apparatus.
Background technology
For a long time, it is necessary to fixed cycle, fixed line, determining method, determining that people, calibration be accurate, timing during substation inspection
Between, the defect or hidden danger of timely discovering device, but substation inspection periodically or non-periodically sets mainly by artificial mode to scene
It is standby to carry out walkaround inspection or grid equipment state is monitored using red place's line temperature sensing meanses, workload greatly, and by environment because
The influence of each side such as element, peopleware, easily causes and makes an inspection tour not in place, checks situation not in place and occurs, makes an inspection tour efficiency and matter
Amount often falls flat.
At present, there is various sensors and monitoring device in transformer station, produce mass data, including captured by industrial camera
Image and video.Traditional monitoring system only realizes remote viewing function, and large nuber of images resource data is only manually searched for, efficiency
It is relatively low, and be far from playing view data resources advantage.Traditional power generation relies primarily on structural data.In recent years,
Video, the unstructured data such as image has exceeded the growth rate of structural data, and these unstructured datas turn into electricity
The major part of net big data.If image/video can be carried out these unstructured datas into structuring, in the image of magnanimity
In find the state of normal operation and failure in real time, realize that Intelligent Recognition, efficient decision-making will have important value and meaning.
The content of the invention
In order to overcome the deficiencies in the prior art, the present invention provides a kind of grid equipment state based on deep learning
Monitoring device.
In order to solve the above technical problems, the technical solution adopted by the present invention is：
A kind of grid equipment state monitoring apparatus based on deep learning, the grid equipment state monitoring apparatus includes video
Collecting unit, grid equipment detection unit, display unit and storage element；Wherein, video acquisition unit include industrial camera and
Video encoder, for obtaining grid equipment state to be identified；Grid equipment detection unit includes image database module, depth
Learning training module, deep learning identification module；Display unit is LED display, by display recognition result and real-time monitoring
Video recording；Storage element is stored for cloud disk, and video, image and recognition result will be monitored in real time and carries out cloud disk storage.
The industrial camera of described video acquisition unit decodes the video information collected by network cable transmission to video
Device, while carrying out data backup through network cable transmission to storage element；By extracting the frame of video of monitoring video, one is built specially
Image data base for recognizing grid equipment state；All images in image data base include the make and break of disconnecting link, cooler
Stop, open, thermometer, instrument board be in the state such as early warning scale area and non-early warning scale area, each image only include a kind of shape
State, altogether 8 kinds of states.The quantity per class picture is usually arranged as 1000.To each width in the database that establishes
Picture all does some pretreatments, and process is as follows：All pictures are divided into training pictures and test chart on a 50-50 basis at random first
Piece collection, then the size of each pictures is normalized, input data is 256 × 256 pixels；Nothing is carried out to each width picture
Sequence sorts, and obtains a label by mark by hand, shows the identification state of the width picture.
Described electrical equipment detection unit is by the electrical device status data transfer in image data base to deep learning
Training module carries out deep learning modeling, and the model that deep learning identification module is set up using deep learning training module is to electric
Equipment state carries out identification and classification, while being shown and being stored identification and classification result.
Described deep learning training module is divided into two parts：Pretreatment and training；Pretreatment has three parts：Unordered row
Sequence, color, the cluster of figure, the adjustment of size and form；Training include convolutional layer, pond layer, full articulamentum, classify layer and
Connected behind activation primitive, the convolutional layer and connect the full articulamentum behind the pond layer, the pond layer, finally
Classification layer on each output node be the grid equipment state probability, institute is determined according to the grid equipment state probability
State the attribute information of grid equipment state to be identified.
Wherein：
Convolutional layer, by convolution algorithm so that original signal feature strengthens and reduces noise；
Pond layer, many features, including maximum pond, average are reduced using image local principle by the method for sampling
Chi Hua, random pool；
Specific convolutional calculation, pondization are calculated to be realized using the convolutional layer and pond layer in existing CaffeNet；
Full articulamentum, each neuron of full articulamentum is connected with next layer of each neuron, as traditional multilayer
Perceptron neural network is the same；
Classification layer, is divided into 8 classes by grid equipment state, classification results and label is matched；
Activation primitive, calculating and adjustment for parameter；
To avoid the occurrence of the phenomenon of over-fitting, fine- is carried out to deep learning training module on image library CaffeNet
tune。
The deep learning identification module, for computer to be obtained to the real-time pictures from industrial camera automatically, automatically
The feature and content of picture are extracted and described, the state change of equipment in frame out is distinguished, the static nature of target is extracted, is judged
Its classification, and further determine that the attribute information of the grid equipment state；On this basis, that is, give a warning, record information
With startup event handling prediction scheme.
The grid equipment state monitoring apparatus of the present invention, utilizes the image data base of a grid equipment state, the data
Storehouse include with it is various under the conditions of grid equipment state image, all images are in terms of background, angle, illumination, yardstick
All there is very big otherness.And to avoid training obtained grid equipment state model mistake in scale too small database
Fitting, so fine-tune is carried out in CaffeNet databases, so that the grid equipment state model after being trained.Obtaining
Get after grid equipment status image to be identified, classified again without user's manual definition state, be directly by the device
It can recognize that and adopt an effective measure, effectively alleviate the pressure of manpower monitoring, reach intellectual monitoring truly, as a result table
The bright grid equipment state monitoring apparatus proposed by the present invention based on deep learning has very high practicality and feasibility.
Brief description of the drawings
Fig. 1 is the structural representation of grid equipment state monitoring apparatus of the present invention.
Fig. 2 is the schematic flow sheet that grid equipment state identification method of the present invention is implemented.
Embodiment
An embodiment of the present invention is described further below in conjunction with the accompanying drawings.
As shown in figure 1, the device includes video acquisition unit, grid equipment detection unit, display unit and storage element.
Wherein, the video acquisition unit includes industrial camera and video encoder, for obtaining grid equipment to be identified
State.
Wherein, the grid equipment detection unit includes database module, deep learning training module, deep learning identification
Module.
Wherein, the display unit is LED display, and display recognition result and real-time monitoring are recorded a video.
Wherein, the storage element is stored for cloud disk, and video, image and recognition result will be monitored in real time and carries out cloud disk storage
Deposit.
As shown in Fig. 2 first, set up one and include disconnecting link make and break, cooler stops, opened, and thermometer, instrument board are in
The grid equipment status image database of the image such as early warning scale area and non-early warning scale area, all images are all from the prison of reality
Measurement equipment, each class amount of images is 1000, and the background of each class image, angle, illumination, have on yardstick it is very big
Otherness.For every piece image, a label is obtained by mark by hand, the identification state of the width image is indicated.
Some pretreatments are all done to every piece image in the database that establishes, it is first that all images are random on a 50-50 basis
Be divided into training image collection and test chart image set, then the size of each image be normalized, input data be 256 ×
256 pixels.
, it is necessary to solve the over-fitting that training pattern parameter is excessively brought before learning to model.Therefore,
CaffeNet images disclose training in advance neural network model on storehouse, and the grid equipment slip condition database for reusing foundation continues to learn
The model is practised until convergence.
Described deep learning training module is divided into two parts：Pretreatment and training.Pretreatment has three parts：Unordered row
Sequence, color, the cluster of figure, is sized and form.Training includes 5 convolutional layers, 5 pond layers, 3 full articulamentums and 1
Individual classification layer.The convolutional layer 1 is used to filter with pond layer 1, and 2 pairs of filter result of convolutional layer 2 and pond layer carry out absolute value school
Just, convolutional layer 3 and pond layer 3 are used to carry out correction result average and normalized square mean, convolutional layer 4 and pond layer 4 for pair
Normalization result carries out sampling window all values and averaged, and convolutional layer 5 is used for maximum to average value progress with pond layer 5
Change.Wherein, activation primitive uses ReLU unsaturation activation primitives, calculating and adjustment for parameter.
By convolutional layer 1, the grid equipment status image to be identified and convolution kernel are subjected to convolutional calculation, convolution kernel is big
Small is 7*7, and each moving step length is 2 pixels during slip, and the characteristic layer number of input is 96, and the number of the parameter of convolution kernel is
96*7*7*3=14112, the result obtained after convolutional layer 1 is 110*110*96=1161600；
By pond layer 1, pond range size is 3*3, mobile for 2 pixels every time, the figure obtained after pond layer 1
As dimension is 55*55*96=290400；
By convolutional layer 2, last layer is obtained into output and convolution kernel progress convolutional calculation, convolution kernel size is 5*5, is slided
When each moving step length be 2 pixels, the characteristic layer number of input is 256, and the number of the parameter of convolution kernel is 256*5*5*96
=614400, the result obtained after convolutional layer 2 is 26*26*256=173056；
By pond layer 2, pond range size is 3*3, mobile for 2 pixels every time；
By convolutional layer 3, last layer is obtained into output and convolution kernel progress convolutional calculation, convolution kernel size is 3*3, is slided
When each moving step length be 1 pixel, the characteristic layer number of input is 384, and the number of the parameter of convolution kernel is 384*3*3*256
=884736；
By pond layer 3, pond range size is 3*3, mobile for 2 pixels every time；
By convolutional layer 4, last layer is obtained into output and convolution kernel progress convolutional calculation, convolution kernel size is 3*3, is slided
When each moving step length be 1 pixel, the characteristic layer number of input is 384, and the number of the parameter of convolution kernel is 384*3*3*384
=1327104；
By pond layer 4, pond range size is 3*3, mobile for 2 pixels every time；
By convolutional layer 5, last layer is obtained into output and convolution kernel progress convolutional calculation, convolution kernel size is 3*3, is slided
When each moving step length be 1 pixel, the characteristic layer number of input is 256, and the number of the parameter of convolution kernel is 256*3*3*384
=884736；
By pond layer 5, pond range size is 3*3, mobile for 2 pixels every time；
By full articulamentum 1, the node number of full articulamentum is 4096, and the number for the convolution nuclear parameter being related to is
4096*4096=16777216；
By full articulamentum 2, the node number of full articulamentum is 4096, and the number for the convolution nuclear parameter being related to is
4096*4096=16777216；
By full articulamentum 3, the node number of full articulamentum is 4096, and the number for the convolution nuclear parameter being related to is
4096*8=32768；
Eventually pass classification layer to be classified, the numerical value of each output node on full articulamentum is converted between 0 to 1
Probable value, the probability of each class electrical device status of correspondence.
Using the grid equipment status image to be identified as input, by convolution operation from input layer to convolutional layer, volume
Each neuron of lamination can be connected with the local receptor field of certain size in input layer, by obtaining described treat after convolution
The feature of grid equipment status image is recognized, the process for changing layer from convolutional layer to pond is properly termed as pond process, it is therefore intended that subtract
The feature quantity of few last layer, the feature obtained after convolutional layer, pond layer and full articulamentum can be divided by classification layer
Each output node on class, classification layer is the grid equipment state probability, is determined according to the grid equipment state probability
The attribute information of the grid equipment state to be identified.Grid equipment state is divided into 8 classes by classification layer, by classification results and label
Matched.Maximum grid equipment state probability can be regard as final result during specific implementation.
The deep learning identification module, for computer to be obtained to the real-time pictures from industrial camera, foundation automatically
The learning rules of deep learning training module, extract the static nature information of target, judge its classification, and further determine that described
The attribute information of grid equipment state.On the basis of final result, recognition result is shown, and recognition result is subjected to cloud disk storage
Deposit, you can give a warning, record information and start the follow-up such as event handling prediction scheme.
The present invention uses Caffe deep learning frameworks, and this is a very clear and efficient deep learning framework,
The framework can be run and its outstanding model and large-scale data, for it is to be solved the problem of with very strong adaptation
Property.Include basic learning rate using the caffe parameters set：0.01, learn momentum：0.9, weight penalty coefficient：0.0005, repeatedly
Generation number：20000.
Claims (5)
1. a kind of grid equipment state monitoring apparatus based on deep learning, it is characterised in that the grid equipment status monitoring is filled
Put including video acquisition unit, grid equipment detection unit, display unit and storage element；Wherein, video acquisition unit includes
Industrial camera and video encoder, for obtaining grid equipment state to be identified；Grid equipment detection unit includes view data
Library module, deep learning training module, deep learning identification module；Display unit is LED display, will display recognition result with
And monitoring video recording in real time；Storage element is stored for cloud disk, and video, image and recognition result will be monitored in real time and carries out cloud disk storage
Deposit；
The industrial camera of described video acquisition unit by the video information collected by network cable transmission to Video Decoder, together
When through network cable transmission to storage element carry out data backup；By extract monitoring video frame of video, build one dedicated for
Recognize the image data base of grid equipment state；All images in image data base include the make and break of disconnecting link, cooler
Stop, open, thermometer, instrument board are in the state such as early warning scale area and non-early warning scale area, each image only includes a kind of state,
8 kinds of states altogether；
Described electrical equipment detection unit trains the electrical device status data transfer in image data base to deep learning
Module carries out deep learning modeling, and the model that deep learning identification module is set up using deep learning training module is to electrical equipment
State carries out identification and classification, while being shown and being stored identification and classification result.
2. a kind of grid equipment state monitoring apparatus based on deep learning according to claim 1, it is characterised in that institute
The deep learning training module stated is divided into two parts：Pretreatment and training；Pretreatment has three parts：Unordered sequence, color, figure
The adjustment of the cluster of shape, size and form；Training includes convolutional layer, pond layer, full articulamentum, classification layer and activation primitive,
Connected behind the convolutional layer and the full articulamentum is connected behind the pond layer, the pond layer, last classification layer
On each output node be the grid equipment state probability, determined according to the grid equipment state probability described to be identified
The attribute information of grid equipment state；
Wherein：
Convolutional layer, by convolution algorithm so that original signal feature strengthens and reduces noise；
Pond layer, many features, including maximum pond, average pond are reduced using image local principle by the method for sampling
Change, random pool；
Specific convolutional calculation, pondization are calculated to be realized using the convolutional layer and pond layer in existing CaffeNet；
Full articulamentum, each neuron of full articulamentum is connected with next layer of each neuron, as traditional Multilayer Perception
Device neutral net is the same；
Classification layer, is divided into 8 classes by grid equipment state, classification results and label is matched；
Activation primitive, calculating and adjustment for parameter；
To avoid the occurrence of the phenomenon of over-fitting, fine-tune is carried out to deep learning training module on image library CaffeNet.
3. a kind of grid equipment state monitoring apparatus based on deep learning according to claim 1 or 2, its feature exists
In the deep learning identification module, for computer to be obtained to the real-time pictures from industrial camera automatically, is automatically extracted simultaneously
The feature and content of picture are described, the state change of equipment in frame out is distinguished, the static nature of target is extracted, judges its class
Not, and the attribute information of the grid equipment state is further determined that；On this basis, that is, give a warning, record information and open
Dynamic event handling prediction scheme.
4. according to a kind of grid equipment state monitoring apparatus based on deep learning of Claims 2 or 3, it is characterised in that described
1000 are set to per the quantity of class picture.
5. a kind of grid equipment state monitoring apparatus based on deep learning according to claim 4, it is characterised in that right
Each width picture in the database established all does some pretreatments, and all pictures are divided into training on a 50-50 basis at random first
Pictures and test pictures collection, then the size of each pictures is normalized, input data is 256 × 256 pixels；To every
One width picture carries out unordered sequence, obtains a label by mark by hand, shows the state of the width picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710143974.9A CN106981063A (en) | 2017-03-14 | 2017-03-14 | A kind of grid equipment state monitoring apparatus based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710143974.9A CN106981063A (en) | 2017-03-14 | 2017-03-14 | A kind of grid equipment state monitoring apparatus based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106981063A true CN106981063A (en) | 2017-07-25 |
Family
ID=59338204
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710143974.9A Pending CN106981063A (en) | 2017-03-14 | 2017-03-14 | A kind of grid equipment state monitoring apparatus based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106981063A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108012121A (en) * | 2017-12-14 | 2018-05-08 | 安徽大学 | A kind of edge calculations and the real-time video monitoring method and system of cloud computing fusion |
| CN108055529A (en) * | 2017-12-25 | 2018-05-18 | 国家电网公司 | Electric power unmanned plane and robot graphics' data normalization artificial intelligence analysis's system |
| CN108189043A (en) * | 2018-01-10 | 2018-06-22 | 北京飞鸿云际科技有限公司 | A kind of method for inspecting and crusing robot system applied to high ferro computer room |
| CN108275524A (en) * | 2018-01-12 | 2018-07-13 | 东北大学 | A kind of elevator maintenance operation monitoring and guiding device based on the assessment of the first multi-view video series of operations |
| CN108387581A (en) * | 2018-02-24 | 2018-08-10 | 温州宝德电气有限公司 | A kind of general-purpose machines visual identity detection device based on deep learning |
| CN108469845A (en) * | 2018-05-15 | 2018-08-31 | 东北大学 | Packaged type solar tracking system based on the Big Dipper and method |
| CN108600701A (en) * | 2018-05-02 | 2018-09-28 | 广州飞宇智能科技有限公司 | A kind of monitoring system and method judging video behavior based on deep learning |
| CN108875719A (en) * | 2018-09-25 | 2018-11-23 | 浙江浙能兴源节能科技有限公司 | Air cooler dust stratification state perception system and calculation method based on deep learning and infrared image identification |
| CN109325936A (en) * | 2018-08-17 | 2019-02-12 | 国网天津市电力公司 | Controller switching equipment defect image identification terminal and method based on neural network deep learning |
| CN109785289A (en) * | 2018-12-18 | 2019-05-21 | 中国科学院深圳先进技术研究院 | A kind of transmission line of electricity defect inspection method, system and electronic equipment |
| CN111047293A (en) * | 2019-12-12 | 2020-04-21 | 云南云电同方科技有限公司 | Method and system for managing graphic data resources |
| CN111553497A (en) * | 2020-04-29 | 2020-08-18 | 成都新潮传媒集团有限公司 | Equipment working state detection method and device of multimedia terminal |
| CN112164153A (en) * | 2020-09-22 | 2021-01-01 | 厦门德威智联科技有限公司 | AI edge calculation fault diagnosis device |
| CN112240796A (en) * | 2019-07-16 | 2021-01-19 | 张力 | Monitoring management system for device state |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105426908A (en) * | 2015-11-09 | 2016-03-23 | 国网冀北电力有限公司信息通信分公司 | Convolutional neural network based substation attribute classification method |
| CN105678344A (en) * | 2016-02-29 | 2016-06-15 | 浙江群力电气有限公司 | Intelligent classification method for power instrument equipment |
| CN106326932A (en) * | 2016-08-25 | 2017-01-11 | 北京每刻风物科技有限公司 | Power line inspection image automatic identification method based on neural network and power line inspection image automatic identification device thereof |
| CN106339722A (en) * | 2016-08-25 | 2017-01-18 | 国网浙江省电力公司杭州供电公司 | Line knife switch state monitoring method and device |
| CN106407369A (en) * | 2016-09-09 | 2017-02-15 | 华南理工大学 | Photo management method and system based on deep learning face recognition |
-
2017
- 2017-03-14 CN CN201710143974.9A patent/CN106981063A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105426908A (en) * | 2015-11-09 | 2016-03-23 | 国网冀北电力有限公司信息通信分公司 | Convolutional neural network based substation attribute classification method |
| CN105678344A (en) * | 2016-02-29 | 2016-06-15 | 浙江群力电气有限公司 | Intelligent classification method for power instrument equipment |
| CN106326932A (en) * | 2016-08-25 | 2017-01-11 | 北京每刻风物科技有限公司 | Power line inspection image automatic identification method based on neural network and power line inspection image automatic identification device thereof |
| CN106339722A (en) * | 2016-08-25 | 2017-01-18 | 国网浙江省电力公司杭州供电公司 | Line knife switch state monitoring method and device |
| CN106407369A (en) * | 2016-09-09 | 2017-02-15 | 华南理工大学 | Photo management method and system based on deep learning face recognition |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108012121A (en) * | 2017-12-14 | 2018-05-08 | 安徽大学 | A kind of edge calculations and the real-time video monitoring method and system of cloud computing fusion |
| CN108055529A (en) * | 2017-12-25 | 2018-05-18 | 国家电网公司 | Electric power unmanned plane and robot graphics' data normalization artificial intelligence analysis's system |
| CN108189043A (en) * | 2018-01-10 | 2018-06-22 | 北京飞鸿云际科技有限公司 | A kind of method for inspecting and crusing robot system applied to high ferro computer room |
| CN108275524A (en) * | 2018-01-12 | 2018-07-13 | 东北大学 | A kind of elevator maintenance operation monitoring and guiding device based on the assessment of the first multi-view video series of operations |
| CN108387581A (en) * | 2018-02-24 | 2018-08-10 | 温州宝德电气有限公司 | A kind of general-purpose machines visual identity detection device based on deep learning |
| CN108600701B (en) * | 2018-05-02 | 2020-11-24 | 广州飞宇智能科技有限公司 | Monitoring system and method for judging video behaviors based on deep learning |
| CN108600701A (en) * | 2018-05-02 | 2018-09-28 | 广州飞宇智能科技有限公司 | A kind of monitoring system and method judging video behavior based on deep learning |
| CN108469845A (en) * | 2018-05-15 | 2018-08-31 | 东北大学 | Packaged type solar tracking system based on the Big Dipper and method |
| CN109325936A (en) * | 2018-08-17 | 2019-02-12 | 国网天津市电力公司 | Controller switching equipment defect image identification terminal and method based on neural network deep learning |
| CN108875719A (en) * | 2018-09-25 | 2018-11-23 | 浙江浙能兴源节能科技有限公司 | Air cooler dust stratification state perception system and calculation method based on deep learning and infrared image identification |
| CN108875719B (en) * | 2018-09-25 | 2023-09-22 | 浙江浙能兴源节能科技有限公司 | Air cooler dust accumulation state sensing system and calculation method based on deep learning and infrared image recognition |
| CN109785289A (en) * | 2018-12-18 | 2019-05-21 | 中国科学院深圳先进技术研究院 | A kind of transmission line of electricity defect inspection method, system and electronic equipment |
| CN109785289B (en) * | 2018-12-18 | 2021-07-20 | 中国科学院深圳先进技术研究院 | Transmission line defect detection method and system and electronic equipment |
| CN112240796A (en) * | 2019-07-16 | 2021-01-19 | 张力 | Monitoring management system for device state |
| CN111047293A (en) * | 2019-12-12 | 2020-04-21 | 云南云电同方科技有限公司 | Method and system for managing graphic data resources |
| CN111047293B (en) * | 2019-12-12 | 2023-11-03 | 云南云电同方科技有限公司 | Method and system for managing graphic data resources |
| CN111553497A (en) * | 2020-04-29 | 2020-08-18 | 成都新潮传媒集团有限公司 | Equipment working state detection method and device of multimedia terminal |
| CN112164153A (en) * | 2020-09-22 | 2021-01-01 | 厦门德威智联科技有限公司 | AI edge calculation fault diagnosis device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106981063A (en) | A kind of grid equipment state monitoring apparatus based on deep learning | |
| CN108171117B (en) | Electric power artificial intelligence visual analysis system based on multicore heterogeneous Computing | |
| CN106991666B (en) | A kind of disease geo-radar image recognition methods suitable for more size pictorial informations | |
| CN110852347B (en) | Fire detection method using improved YOLO v3 | |
| CN108734143A (en) | A kind of transmission line of electricity online test method based on binocular vision of crusing robot | |
| CN111444939B (en) | Small-scale equipment component detection method based on weak supervision cooperative learning in open scene of power field | |
| CN110135319A (en) | A kind of anomaly detection method and its system | |
| CN111325347B (en) | Automatic danger early warning description generation method based on interpretable visual reasoning model | |
| CN106971152A (en) | A kind of method of Bird's Nest in detection transmission line of electricity based on Aerial Images | |
| Bo et al. | Particle pollution estimation from images using convolutional neural network and weather features | |
| CN110070530A (en) | A kind of powerline ice-covering detection method based on deep neural network | |
| CN108921039A (en) | The forest fire detection method of depth convolution model based on more size convolution kernels | |
| CN108830305A (en) | A kind of real-time fire monitoring method of combination DCLRN network and optical flow method | |
| CN107633511A (en) | A kind of blower fan vision detection system based on own coding neutral net | |
| CN109741320A (en) | A kind of wind electricity blade fault detection method based on Aerial Images | |
| CN102445640A (en) | GIS device intelligent recognition method based on vector machine and artificial fish swarm optimization | |
| CN109801265A (en) | A kind of real-time transmission facility foreign matter detection system based on convolutional neural networks | |
| CN111046728A (en) | Straw combustion detection method based on characteristic pyramid network | |
| CN108229524A (en) | A kind of chimney and condensing tower detection method based on remote sensing images | |
| CN111539325A (en) | Forest fire detection method based on deep learning | |
| CN109086803A (en) | A kind of haze visibility detection system and method based on deep learning and the personalized factor | |
| CN113205039A (en) | Power equipment fault image identification and disaster investigation system and method based on multiple DCNNs | |
| Yandouzi et al. | Investigation of combining deep learning object recognition with drones for forest fire detection and monitoring | |
| CN107944373A (en) | A kind of video anomaly detection method based on deep learning | |
| CN102867214B (en) | Counting management method for people within area range |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170725 |
|
| RJ01 | Rejection of invention patent application after publication |