CN110147757A - Passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks - Google Patents
Passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks Download PDFInfo
- Publication number
- CN110147757A CN110147757A CN201910412356.9A CN201910412356A CN110147757A CN 110147757 A CN110147757 A CN 110147757A CN 201910412356 A CN201910412356 A CN 201910412356A CN 110147757 A CN110147757 A CN 110147757A
- Authority
- CN
- China
- Prior art keywords
- engineering truck
- image
- convolutional neural
- neural networks
- transmission line
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
Abstract
Present disclose provides a kind of passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks.Wherein, the passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks, including obtain transmission line of electricity monitoring image and carry out unified pixel size processing;By treated, transmission line of electricity monitoring image is input to the identification model of the engineering truck based on convolutional neural networks, goes out image category by engineering truck identification Model Distinguish;Wherein, image category is classified with the engineering truck classification for including in image;Image category result is mapped in original transmission line of electricity monitoring image, if image category result is to contain engineering truck, the then transmission line of electricity monitoring image of engineering truck identification model output mark engineering machinery bounding box position and type label, continues the detection of next image;Otherwise, this image is skipped, the detection of next image is continued.
Description
Technical field
The disclosure belongs to field of image processing more particularly to a kind of passway for transmitting electricity engineering truck based on convolutional neural networks
Discrimination method and system.
Background technique
Only there is provided background technical informations relevant to the disclosure for the statement of this part, it is not necessary to so constitute first skill
Art.
As the continuous promotion of urban construction speed and science and technology and expanding economy, the construction of national grid are also constantly pushing away
Also increase into, the erection quantity and range of transmission line of electricity being increasing, the operation of power networks environment on city and its periphery is increasingly multiple
It is miscellaneous, threat is caused to electric power netting safe running, wherein cause transmission line malfunction in route because of foreign body intrusions such as engineering trucks therefore
A possibility that in barrier, is also increasing, this brings huge security risk to transmission line of electricity, also controls to power grid real-time monitoring and analysis
System proposes more acute challenge.
In order to ensure power supply reliability and stability, Guo Wang company has carried out a wide range of upgrading to grid monitoring system and has changed
It makes, is mounted with the first-class equipment of monitoring camera on shaft tower along the important transmission line of electricity, to carrying out real-time online monitoring along transmission of electricity,
Transmission line of electricity context situation is recorded, and is sent images in fortune inspection monitoring room, You Yunjian responsible person is to power transmission line
Operation safety monitoring provides safeguard.
Inventors have found that with the raising of monitoring device coverage rate, occur that related such as monitoring data is big, effective image
The problems such as ratio data is small low with image data validity, artificial monitoring are difficult to fast and effeciently find the hair of foreign body intrusion event
It is raw.
Summary of the invention
To solve the above-mentioned problems, the first aspect of the disclosure provides a kind of passway for transmitting electricity based on convolutional neural networks
Engineering truck discrimination method can effectively identify the engineering truck of invasion transmission line of electricity, guarantee higher accurate rate and robust
Property while, there is faster detection speed, safe and reliable guarantee can be provided for transmission line of electricity.
To achieve the goals above, the disclosure adopts the following technical scheme that
A kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks, comprising:
It obtains transmission line of electricity monitoring image and carries out unified pixel size processing;
By treated, transmission line of electricity monitoring image is input to the identification model of the engineering truck based on convolutional neural networks, by
Engineering truck identification Model Distinguish goes out image category;Wherein, image category is divided with the engineering truck classification for including in image
Class;
Image category result is mapped in original transmission line of electricity monitoring image, if image category result is to contain engineering truck
, then the output of engineering truck identification model marks the transmission line of electricity monitoring image of engineering machinery bounding box position and type label,
Continue the detection of next image;Otherwise, this image is skipped, the detection of next image is continued.
To solve the above-mentioned problems, the second aspect of the disclosure provides a kind of passway for transmitting electricity based on convolutional neural networks
Engineering truck identification system can effectively identify the engineering truck of invasion transmission line of electricity, guarantee higher accurate rate and robust
Property while, there is faster detection speed, safe and reliable guarantee can be provided for transmission line of electricity.
To achieve the goals above, the disclosure adopts the following technical scheme that
A kind of passway for transmitting electricity engineering truck identification system based on convolutional neural networks, comprising:
Image pre-processing module is used to obtain transmission line of electricity monitoring image and carries out unified pixel size processing;
Image category recognizes module, is used for that transmission line of electricity monitoring image to be input to based on convolutional Neural net by treated
The engineering truck of network recognizes model, goes out image category by engineering truck identification Model Distinguish;Wherein, image category in image to wrap
The engineering truck classification contained is classified;
Image category output module is used to for image category result being mapped in original transmission line of electricity monitoring image, if
Image category result is to contain engineering truck, then engineering truck identification model output mark engineering machinery bounding box position and type
The transmission line of electricity monitoring image of label continues the detection of next image;Otherwise, this image is skipped, next image is continued
Detection.
To solve the above-mentioned problems, a kind of computer readable storage medium is provided in terms of the third of the disclosure, it can
The effectively engineering truck of identification invasion transmission line of electricity while guaranteeing higher accurate rate and robustness, has faster detection
Speed can provide safe and reliable guarantee for transmission line of electricity.
To achieve the goals above, the disclosure adopts the following technical scheme that
A kind of computer readable storage medium, is stored thereon with computer program, realization when which is executed by processor
Step in passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks as described above.
To solve the above-mentioned problems, the 4th aspect of the disclosure provides a kind of computer equipment, can effectively identify
The engineering truck for invading transmission line of electricity while guaranteeing higher accurate rate and robustness, has faster detection speed, can
Safe and reliable guarantee is provided for transmission line of electricity.
To achieve the goals above, the disclosure adopts the following technical scheme that
A kind of computer equipment can be run on a memory and on a processor including memory, processor and storage
Computer program, the processor realize the passway for transmitting electricity work described above based on convolutional neural networks when executing described program
Step in journey vehicle recognition method.
The beneficial effect of the disclosure is:
(1) disclosure can shoot image by camera on batch electric power line pole tower, mark out engineering truck in the picture
Position and classification.
(2) disclosure can real-time judge may threaten the engineering truck of transmission line of electricity, reduce the work of manual identified
Amount improves identification accuracy, reduces fortune inspection personnel line walking work load.
Detailed description of the invention
The Figure of description for constituting a part of this disclosure is used to provide further understanding of the disclosure, and the disclosure is shown
Meaning property embodiment and its explanation do not constitute the improper restriction to the disclosure for explaining the disclosure.
Fig. 1 is a kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks that the embodiment of the present disclosure provides
Flow chart.
Fig. 2 be the embodiment of the present disclosure provide to engineering truck identification model be trained flow chart.
Specific embodiment
The disclosure is described further with embodiment with reference to the accompanying drawing.
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the disclosure.Unless another
It indicates, all technical and scientific terms used herein has usual with disclosure person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the disclosure.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Embodiment 1
As shown in Figure 1, a kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks of the present embodiment, packet
It includes:
Step 1: obtaining transmission line of electricity monitoring image and carry out unified pixel size processing.
Specifically, transmission line of electricity monitoring image is obtained, cutting, unified pixel size, such as pixel are zoomed in and out to image
Size is collectively referred to as 256*256.
It should be noted that those skilled in the art can carry out specifically to be arranged according to the actual situation the size of unified pixel,
Magnitude numerical value has no effect on the result entirely recognized.
In this way by the uniformity of input data, the speed and accuracy of engineering truck identification model training are improved, finally
Improve the accuracy of journey vehicle recognition.
Step 2: by treated, transmission line of electricity monitoring image is input to the identification of the engineering truck based on convolutional neural networks
Model goes out image category by engineering truck identification Model Distinguish;Wherein, engineering truck classification of the image category to include in image
Classify.
By treated transmission line of electricity monitoring image the is input to identification of the engineering truck based on convolutional neural networks model it
Before, it further include being trained to engineering truck identification model, as shown in Fig. 2, its process are as follows:
Step 2.1: obtaining the transmission line of electricity monitoring image for containing engineering truck as training set;
Specifically, the image containing engineering truck is extracted from transmission line of electricity monitoring image, as convolutional neural networks
Training set.
As a kind of optional embodiment, also the transmission line of electricity monitoring image in training set by overturning, rotation or is put down
It is moved into row data augmentation.
Data augmentation can rich image training set quantity, preferably extraction characteristics of image, extensive model (prevents model mistake
Fitting).
Data augmentation process is to carry out original image:
1) it overturns, image is flipped an angle at random along horizontally or vertically method;
2) it rotates, by image by Random-Rotation certain angle clockwise or counterclockwise;
3) it translates, image is translated into a fixed step size along horizontally or vertically method.
Herein it is noted that the relative position of the relatively whole picture of engineering truck can occur in image after data augmentation
Transformation, has difference with original image annotation results.
Step 2.2: engineering truck position and type to every transmission line of electricity monitoring image in training set are formed into mark
Bounding box with type label.
Specifically, position is carried out to engineering truck in image using annotation tool and type marks.Wherein, annotation process
It is using LabelImg software, to take engineering truck using boundary circle, and mark its classification, output and image xml of the same name
Mark file.
Herein it should be noted that xml mark file can not be used directly in model training, need to be converted to model energy
The TXT text file enough identified, wherein records engineering truck image file name in TXT file, target marks bounding box coordinates
With category classification label.
It is understood that other existing annotation tools can also be used in annotation tool, those skilled in the art can basis
Actual conditions are specifically chosen.
Step 2.3: will be trained in the convolutional neural networks of the sample set input initialization marked, adjustment convolution mind
Through convolutional layer convolution nuclear parameter in network, engineering truck identification model is obtained.
Specifically, convolutional neural networks model parameter, main weight and biasing and convolution including convolution kernel are initialized
The learning rate of neural network.
Convolution nuclear parameter in convolutional neural networks is initialized, the image marked is inputted in convolutional neural networks, monitoring
Image is divided into tri- chrominance channel RGB extraction pixel data as convolutional neural networks input and obtains output valve by propagated forward, obtains
Bounding box position and type result to convolutional neural networks reality output.
Convolutional network reality output result and annotation results are calculated and obtain bounding box recurrence loss function and Classification Loss
Functional value seeks error gradient using back-propagation method, is minimised as target in convolutional neural networks with loss function
Carry out right value update.
Propagated forward: sample enters neural network model by input layer, obtains model by propagated forward and calculates output
Bounding box and type.
Backpropagation: the recurrence between computation model output boundary frame position result and practical mark bounding box position is lost
Function;Classification Loss function between model output class and practical mark type, according to two kinds of penalty values, using backpropagation
Algorithm continues to optimize the convolution kernel weight and biasing of convolutional neural networks.
The bounding box that backpropagation uses returns loss function Jlocal(θ) are as follows:
In formula (1), θ indicates convolution kernel weight in convolutional neural networks, and m indicates sample size,Indicate that sample passes through
The bounding box of convolutional neural networks propagated forward output is as a result, d indicates sample actual boundary frame result.
The Classification Loss function J that backpropagation usesclass(θ) are as follows:
In formula (2), x(i)Indicate i-th of sample, hθ(x(i)) indicate i-th sample by before convolutional neural networks to biography
Broadcast the classification results of output, y(i)Indicate sample actual classification result.
Whole loss function is J (θ):
J (θ)=λclass·Jclass+λlocal·Jlocal (3)
In formula (3), λclassAccounting of the presentation class loss function in whole loss function;λlocalIndicate bounding box
Return accounting of the loss function in whole loss function.
Convolution kernel weight in convolutional neural networks is updated using gradient descent method:
In formula (4), convolution kernel weight in the updated convolutional neural networks of θ ' expression;α indicates convolutional neural networks
Habit rate.
Specifically, in the convolutional neural networks that the monitoring image input training after will be processed is completed, pass through convolution kernel pair
Image is calculated, and the correlated characteristic of image is extracted;Image is calculated by multilayer convolution kernel, and image low-level features are changed into height
Grade feature divides an image into the classification set by the full linking layer positioning classification of the last layer.
Step 3: image category result is mapped in original transmission line of electricity monitoring image, if image category result be containing
Engineering truck, then the transmission line of electricity of engineering truck identification model output mark engineering machinery bounding box position and type label monitors
Image continues the detection of next image;Otherwise, this image is skipped, the detection of next image is continued.
The present embodiment can shoot image by camera on batch electric power line pole tower, mark out engineering truck in the picture
Position and classification.
The present embodiment can real-time judge may threaten the engineering truck of transmission line of electricity, reduce the work of manual identified
Amount improves identification accuracy, reduces fortune inspection personnel line walking work load.
Embodiment 2
A kind of passway for transmitting electricity engineering truck identification system based on convolutional neural networks of the present embodiment, comprising:
(1) image pre-processing module is used to obtain transmission line of electricity monitoring image and carries out unified pixel size processing;
Specifically, transmission line of electricity monitoring image is obtained, cutting, unified pixel size, such as pixel are zoomed in and out to image
Size is collectively referred to as 256*256.
It should be noted that those skilled in the art can carry out specifically to be arranged according to the actual situation the size of unified pixel,
Magnitude numerical value has no effect on the result entirely recognized.
In this way by the uniformity of input data, the speed and accuracy of engineering truck identification model training are improved, finally
Improve the accuracy of journey vehicle recognition.
(2) image category recognizes module, is used for that transmission line of electricity monitoring image to be input to based on convolution mind by treated
Engineering truck through network recognizes model, goes out image category by engineering truck identification Model Distinguish;Wherein, image category is with image
In include engineering truck classification classify;
Described image classification recognizes module, further includes engineering truck identification model training module, is used for:
(2.1) the transmission line of electricity monitoring image for containing engineering truck is obtained as training set;
Specifically, the image containing engineering truck is extracted from transmission line of electricity monitoring image, as convolutional neural networks
Training set.
As a kind of optional embodiment, also the transmission line of electricity monitoring image in training set by overturning, rotation or is put down
It is moved into row data augmentation.
Data augmentation can rich image training set quantity, preferably extraction characteristics of image, extensive model (prevents model mistake
Fitting).
Data augmentation process is to carry out original image:
1) it overturns, image is flipped an angle at random along horizontally or vertically method;
2) it rotates, by image by Random-Rotation certain angle clockwise or counterclockwise;
3) it translates, image is translated into a fixed step size along horizontally or vertically method.
Herein it is noted that the relative position of the relatively whole picture of engineering truck can occur in image after data augmentation
Transformation, has difference with original image annotation results.
(2.2) the engineering truck position to every transmission line of electricity monitoring image in training set and type form band into mark
There is the bounding box of type label.
Specifically, position is carried out to engineering truck in image using annotation tool and type marks.Wherein, annotation process
It is using LabelImg software, to take engineering truck using boundary circle, and mark its classification, output and image xml of the same name
Mark file.
Herein it should be noted that xml mark file can not be used directly in model training, need to be converted to model energy
The TXT text file enough identified, wherein records engineering truck image file name in TXT file, target marks bounding box coordinates
With category classification label.
It is understood that other existing annotation tools can also be used in annotation tool, those skilled in the art can basis
Actual conditions are specifically chosen.
(2.3) it will be trained in the convolutional neural networks of the sample set input initialization marked, adjust convolutional Neural
Convolutional layer convolution nuclear parameter in network obtains engineering truck identification model.
Specifically, convolutional neural networks model parameter, main weight and biasing and convolution including convolution kernel are initialized
The learning rate of neural network.
Convolution nuclear parameter in convolutional neural networks is initialized, the image marked is inputted in convolutional neural networks, monitoring
Image is divided into tri- chrominance channel RGB extraction pixel data as convolutional neural networks input and obtains output valve by propagated forward, obtains
Bounding box position and type result to convolutional neural networks reality output.
Convolutional network reality output result and annotation results are calculated and obtain bounding box recurrence loss function and Classification Loss
Functional value seeks error gradient using back-propagation method, is minimised as target in convolutional neural networks with loss function
Carry out right value update.
Propagated forward: sample enters neural network model by input layer, obtains model by propagated forward and calculates output
Bounding box and type.
Backpropagation: the recurrence between computation model output boundary frame position result and practical mark bounding box position is lost
Function;Classification Loss function between model output class and practical mark type, according to two kinds of penalty values, using backpropagation
Algorithm continues to optimize the convolution kernel weight and biasing of convolutional neural networks.
The bounding box that backpropagation uses returns loss function Jlocal(θ) are as follows:
In formula (1), θ indicates convolution kernel weight in convolutional neural networks, and m indicates sample size,Indicate that sample passes through
The bounding box of convolutional neural networks propagated forward output is as a result, d indicates sample actual boundary frame result.
The Classification Loss function J that backpropagation usesclass(θ) are as follows:
In formula (2), x(i)Indicate i-th of sample, hθ(x(i)) indicate i-th sample by before convolutional neural networks to biography
Broadcast the classification results of output, y(i)Indicate sample actual classification result.
Whole loss function is J (θ):
J (θ)=λclass·Jclass+λlocal·Jlocal (3)
In formula (3), λclassAccounting of the presentation class loss function in whole loss function;λlocalIndicate bounding box
Return accounting of the loss function in whole loss function.
Convolution kernel weight in convolutional neural networks is updated using gradient descent method:
In formula (4), convolution kernel weight in the updated convolutional neural networks of θ ' expression;α indicates convolutional neural networks
Habit rate.
Specifically, in the convolutional neural networks that the monitoring image input training after will be processed is completed, pass through convolution kernel pair
Image is calculated, and the correlated characteristic of image is extracted;Image is calculated by multilayer convolution kernel, and image low-level features are changed into height
Grade feature divides an image into the classification set by the full linking layer positioning classification of the last layer.
(3) image category output module is used to for image category result being mapped in original transmission line of electricity monitoring image,
If image category result is to contain engineering truck, engineering truck recognizes model output mark engineering machinery bounding box position and kind
The transmission line of electricity monitoring image of class label continues the detection of next image;Otherwise, this image is skipped, next figure is continued
The detection of picture.
The present embodiment can shoot image by camera on batch electric power line pole tower, mark out engineering truck in the picture
Position and classification.
The present embodiment can real-time judge may threaten the engineering truck of transmission line of electricity, reduce the work of manual identified
Amount improves identification accuracy, reduces fortune inspection personnel line walking work load.
Embodiment 3
A kind of computer readable storage medium is present embodiments provided, computer program is stored thereon with, which is located
Manage the step realized in the passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks as shown in Figure 1 when device executes.
The present embodiment can shoot image by camera on batch electric power line pole tower, mark out engineering truck in the picture
Position and classification.
The present embodiment can real-time judge may threaten the engineering truck of transmission line of electricity, reduce the work of manual identified
Amount improves identification accuracy, reduces fortune inspection personnel line walking work load.
Embodiment 4
Present embodiments provide a kind of computer equipment, including memory, processor and storage are on a memory and can be
The computer program run on processor, the processor are realized as shown in Figure 1 based on convolutional Neural when executing described program
Step in the passway for transmitting electricity engineering truck discrimination method of network.
The present embodiment can shoot image by camera on batch electric power line pole tower, mark out engineering truck in the picture
Position and classification.
The present embodiment can real-time judge may threaten the engineering truck of transmission line of electricity, reduce the work of manual identified
Amount improves identification accuracy, reduces fortune inspection personnel line walking work load.
It should be understood by those skilled in the art that, embodiment of the disclosure can provide as method, system or computer program
Product.Therefore, the shape of hardware embodiment, software implementation or embodiment combining software and hardware aspects can be used in the disclosure
Formula.Moreover, the disclosure, which can be used, can use storage in the computer that one or more wherein includes computer usable program code
The form for the computer program product implemented on medium (including but not limited to magnetic disk storage and optical memory etc.).
The disclosure is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present disclosure
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the program can be stored in a computer-readable storage medium
In, the program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, the storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random
AccessMemory, RAM) etc..
The foregoing is merely preferred embodiment of the present disclosure, are not limited to the disclosure, for the skill of this field
For art personnel, the disclosure can have various modifications and variations.It is all within the spirit and principle of the disclosure, it is made any to repair
Change, equivalent replacement, improvement etc., should be included within the protection scope of the disclosure.
Claims (10)
1. a kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks characterized by comprising
It obtains transmission line of electricity monitoring image and carries out unified pixel size processing;
By treated, transmission line of electricity monitoring image is input to the identification model of the engineering truck based on convolutional neural networks, by engineering
Vehicle recognition Model Distinguish goes out image category;Wherein, image category is classified with the engineering truck classification for including in image;
Image category result is mapped in original transmission line of electricity monitoring image, if image category result is to contain engineering truck,
The then transmission line of electricity monitoring image of engineering truck identification model output mark engineering machinery bounding box position and type label, continues
The detection of next image;Otherwise, this image is skipped, the detection of next image is continued.
2. a kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks as described in claim 1, feature
It is, by before treated transmission line of electricity monitoring image the is input to identification model of the engineering truck based on convolutional neural networks,
It further include being trained to engineering truck identification model, process are as follows:
The transmission line of electricity monitoring image for containing engineering truck is obtained as training set;
Engineering truck position and type to every transmission line of electricity monitoring image in training set form into mark and have type label
Bounding box;
It will be trained in the convolutional neural networks of the sample set input initialization marked, adjust convolution in convolutional neural networks
Layer convolution nuclear parameter obtains engineering truck identification model.
3. a kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks as claimed in claim 2, feature
It is, during being trained to engineering truck identification model, further includes: to the transmission line of electricity monitoring image in training set
Data augmentation is carried out by overturning, rotation or translation.
4. a kind of passway for transmitting electricity engineering truck discrimination method based on convolutional neural networks as claimed in claim 2, feature
It is, during being trained in the convolutional neural networks of the sample set input initialization marked, by convolutional network reality
Border exports result and annotation results calculate and obtain bounding box recurrence loss function and Classification Loss functional value, using backpropagation side
Method seeks error gradient, is minimised as target to the carry out right value update in convolutional neural networks with loss function.
5. a kind of passway for transmitting electricity engineering truck identification system based on convolutional neural networks characterized by comprising
Image pre-processing module is used to obtain transmission line of electricity monitoring image and carries out unified pixel size processing;
Image category recognizes module, is used for that transmission line of electricity monitoring image to be input to based on convolutional neural networks by treated
Engineering truck recognizes model, goes out image category by engineering truck identification Model Distinguish;Wherein, image category to include in image
Engineering truck classification is classified;
Image category output module is used to for image category result being mapped in original transmission line of electricity monitoring image, if image
Category result is to contain engineering truck, then engineering truck identification model output mark engineering machinery bounding box position and type label
Transmission line of electricity monitoring image, continue next image detection;Otherwise, this image is skipped, the inspection of next image is continued
It surveys.
6. a kind of passway for transmitting electricity engineering truck identification system based on convolutional neural networks as claimed in claim 5, feature
It is, described image classification recognizes module, and further include engineering truck identification model training module, is used for:
The transmission line of electricity monitoring image for containing engineering truck is obtained as training set;
Engineering truck position and type to every transmission line of electricity monitoring image in training set form into mark and have type label
Bounding box;
It will be trained in the convolutional neural networks of the sample set input initialization marked, adjust convolution in convolutional neural networks
Layer convolution nuclear parameter obtains engineering truck identification model.
7. a kind of passway for transmitting electricity engineering truck identification system based on convolutional neural networks as claimed in claim 6, feature
It is, the engineering truck recognizes model training module, is also used to: to the transmission line of electricity monitoring image in training set by turning over
Turn, rotation or translation carry out data augmentation.
8. a kind of passway for transmitting electricity engineering truck identification system based on convolutional neural networks as claimed in claim 6, feature
It is, in engineering truck identification model training module, by the convolutional Neural net of the sample set input initialization marked
During being trained in network, convolutional network reality output result and annotation results are calculated and obtain bounding box recurrence loss letter
Several and Classification Loss functional value, seeks error gradient using back-propagation method, is minimised as target to convolution with loss function
Carry out right value update in neural network.
9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor
Such as the passway for transmitting electricity engineering truck discrimination method of any of claims 1-4 based on convolutional neural networks is realized when row
In step.
10. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor realizes such as base of any of claims 1-4 when executing described program
Step in the passway for transmitting electricity engineering truck discrimination method of convolutional neural networks.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910412356.9A CN110147757A (en) | 2019-05-17 | 2019-05-17 | Passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910412356.9A CN110147757A (en) | 2019-05-17 | 2019-05-17 | Passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110147757A true CN110147757A (en) | 2019-08-20 |
Family
ID=67595470
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910412356.9A Pending CN110147757A (en) | 2019-05-17 | 2019-05-17 | Passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110147757A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110602449A (en) * | 2019-09-01 | 2019-12-20 | 天津大学 | Intelligent construction safety monitoring system method in large scene based on vision |
CN111309035A (en) * | 2020-05-14 | 2020-06-19 | 浙江远传信息技术股份有限公司 | Multi-robot cooperative movement and dynamic obstacle avoidance method, device, equipment and medium |
CN111652102A (en) * | 2020-05-27 | 2020-09-11 | 国网山东省电力公司东营供电公司 | Power transmission channel target object identification method and system |
CN111798435A (en) * | 2020-07-08 | 2020-10-20 | 国网山东省电力公司东营供电公司 | Image processing method, and method and system for monitoring invasion of engineering vehicle into power transmission line |
CN115273368A (en) * | 2022-07-20 | 2022-11-01 | 云南电网有限责任公司电力科学研究院 | Method, medium, equipment and system for warning invasion of vehicles in power transmission line corridor construction |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106778472A (en) * | 2016-11-17 | 2017-05-31 | 成都通甲优博科技有限责任公司 | The common invader object detection and recognition method in transmission of electricity corridor based on deep learning |
CN107679495A (en) * | 2017-10-09 | 2018-02-09 | 济南大学 | A kind of detection method of transmission line of electricity periphery activity engineering truck |
CN108109385A (en) * | 2018-01-18 | 2018-06-01 | 南京杰迈视讯科技有限公司 | A kind of vehicle identification of power transmission line external force damage prevention and hazardous act judgement system and method |
-
2019
- 2019-05-17 CN CN201910412356.9A patent/CN110147757A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106778472A (en) * | 2016-11-17 | 2017-05-31 | 成都通甲优博科技有限责任公司 | The common invader object detection and recognition method in transmission of electricity corridor based on deep learning |
CN107679495A (en) * | 2017-10-09 | 2018-02-09 | 济南大学 | A kind of detection method of transmission line of electricity periphery activity engineering truck |
CN108109385A (en) * | 2018-01-18 | 2018-06-01 | 南京杰迈视讯科技有限公司 | A kind of vehicle identification of power transmission line external force damage prevention and hazardous act judgement system and method |
Non-Patent Citations (1)
Title |
---|
闫春江 等: "基于深度学习的输电线路工程车辆入侵检测", 《信息技术》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110602449A (en) * | 2019-09-01 | 2019-12-20 | 天津大学 | Intelligent construction safety monitoring system method in large scene based on vision |
CN111309035A (en) * | 2020-05-14 | 2020-06-19 | 浙江远传信息技术股份有限公司 | Multi-robot cooperative movement and dynamic obstacle avoidance method, device, equipment and medium |
CN111652102A (en) * | 2020-05-27 | 2020-09-11 | 国网山东省电力公司东营供电公司 | Power transmission channel target object identification method and system |
CN111798435A (en) * | 2020-07-08 | 2020-10-20 | 国网山东省电力公司东营供电公司 | Image processing method, and method and system for monitoring invasion of engineering vehicle into power transmission line |
CN115273368A (en) * | 2022-07-20 | 2022-11-01 | 云南电网有限责任公司电力科学研究院 | Method, medium, equipment and system for warning invasion of vehicles in power transmission line corridor construction |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110147757A (en) | Passway for transmitting electricity engineering truck discrimination method and system based on convolutional neural networks | |
CN109784333A (en) | Based on an objective detection method and system for cloud bar power channel characteristics | |
CN110084165A (en) | The intelligent recognition and method for early warning of anomalous event under the open scene of power domain based on edge calculations | |
Shen et al. | Exemplar-based human action pose correction and tagging | |
CN108037770A (en) | Unmanned plane power transmission line polling system and method based on artificial intelligence | |
CN110298265A (en) | Specific objective detection method in a kind of elevator based on YOLO neural network | |
CN109635685A (en) | Target object 3D detection method, device, medium and equipment | |
CN107609512A (en) | A kind of video human face method for catching based on neutral net | |
CN106803071A (en) | Object detecting method and device in a kind of image | |
CN107808133A (en) | Oil-gas pipeline safety monitoring method, system and software memory based on unmanned plane line walking | |
CN109214366A (en) | Localized target recognition methods, apparatus and system again | |
CN108229524A (en) | A kind of chimney and condensing tower detection method based on remote sensing images | |
CN109753898A (en) | A kind of safety cap recognition methods and device | |
Yang et al. | Research on subway pedestrian detection algorithms based on SSD model | |
Teerakawanich et al. | Short term prediction of sun coverage using optical flow with GoogLeNet | |
CN113902792A (en) | Building height detection method and system based on improved RetinaNet network and electronic equipment | |
CN108022257A (en) | Suitable for the high speed convolution neutral net method for tracking target and device of hardware | |
Zhang et al. | Few-shot object detection with self-adaptive global similarity and two-way foreground stimulator in remote sensing images | |
Guan et al. | EARL: an elliptical distribution aided adaptive rotation label assignment for oriented object detection in remote sensing images | |
CN116935473A (en) | Real-time detection method and system for wearing safety helmet based on improved YOLO v7 under complex background | |
Xia et al. | Neural detection of foreign objects for transmission lines in power systems | |
CN110009246A (en) | A kind of transmission line of electricity operating status appraisal procedure based on big data Portrait brand technology | |
CN107295214A (en) | Interpolated frame localization method and device | |
CN102999766A (en) | Target tracking method on basis of online study for context relation of space-time structure | |
NO20210472A1 (en) | Bird detection and species determination |
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: 20190820 |
|
RJ01 | Rejection of invention patent application after publication |