CN108734117A - Cable machinery external corrosion failure evaluation method based on YOLO - Google Patents
Cable machinery external corrosion failure evaluation method based on YOLO Download PDFInfo
- Publication number
- CN108734117A CN108734117A CN201810437049.1A CN201810437049A CN108734117A CN 108734117 A CN108734117 A CN 108734117A CN 201810437049 A CN201810437049 A CN 201810437049A CN 108734117 A CN108734117 A CN 108734117A
- Authority
- CN
- China
- Prior art keywords
- cable machinery
- image
- cable
- external corrosion
- evaluation method
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- 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
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The cable machinery external corrosion failure evaluation method based on YOLO that the invention discloses a kind of.Existing video monitoring system, often only function for monitoring, does not have the function that automatic identification is carried out to cable machinery external corrosion breakage state also.The present invention uses image-scaling method to adjust picture size first, and convolutional neural networks is recycled to carry out feature extraction, wherein each layer uses batch method for normalizing normative model, finally by RPN neural network forecast bounding boxes.The method recognition accuracy of the present invention is high, and robustness is good；Carrying out cable machinery external corrosion failure evaluation in the crusing robot system in buried cable tunnel using this method has extraordinary effect.
Description
Technical field
The invention belongs to cable machinery external corrosion failure evaluation field, outside especially a kind of cable machinery based on YOLO
Portion's corrosion failure recognition methods.
Background technology
The development boosting of China's urban modernization pace of construction in buried cable tunnel, in spies such as Beijing, Shanghai, Guangzhou
Large size city, underground distribution network have tended to be perfect, some large- and-medium size cities are also in the construction progress for accelerating cable tunnel.With
The gradual universal of Urban Underground power grid, cable fault will enter the high-incidence season, to the depth to cable tunnel inspection data
Excavation proposes active demand.
Buried cable and related facility are the emphasis in power transmission link.The normal operation of equipment can ensure company
For the reliable supply of electric power service of customers with secure, and can occur in emergency circumstances to have preferable power supply Means of Ensuring and
Ability.In order to ensure the normal operation of underground power grid, parameter and cable machinery external corrosion breakage shape to underground power grid are needed
State is monitored.
Existing video monitoring system, often only function for monitoring, does not have also to cable machinery external corrosion breakage shape
State carries out the function of automatic identification.
Invention content
To solve the above problems, the present invention provides a kind of cable machinery external corrosion failure evaluation method based on YOLO,
It carries out automatic identification by crusing robot to cable machinery external corrosion breakage state, realizes cable tunnel intelligent patrol detection,
Ensure underground electric power netting safe running.
The technical solution adopted by the present invention is as follows：Cable machinery external corrosion failure evaluation method based on YOLO, packet
It includes：
1) include the tunnel internal cable machinery sample graph of object by tunnel crusing robot camera shooting, collecting
Picture, the object for including in sample image are slot box, wind turbine, maintenance power box and the distribution box of tunnel internal；
2) all tunnel internal cable machinery sample images are traversed, object is directed into rower to every image encirclement frame
Note processing, obtains training set；
3) image-scaling method is used to adjust picture size：Image scaling is carried out for the sample image in training set, is adjusted
Whole picture size；
4) feature extraction is carried out to the image that step 3) obtains using convolutional neural networks, wherein convolutional neural networks use
20 layers of trained network of convolutional layer before YOLO, and using batch method for normalizing then every layer network is normalized
It is inputted again, carrys out predicted boundary frame finally by RPN networks；
5) by the continuous repetitive exercise of method in step 4) until model training error tend towards stability, the entire net finally obtained
Network model is as tunnel internal cable machinery detection model；
6) cable machinery testing image is acquired in real time, after identical image-scaling method zooms in and out with step 3)
As the input for the tunnel internal cable machinery detection model that step 5) obtains, the output of tunnel internal cable machinery detection model
The as final recognition result of cable machinery testing image.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 1), tunnel internal electricity
Cable equipment sample image refers near crusing robot walking to cable machinery, and camera is towards cable machinery, with cable machinery
For object, between the poor 30 degree of visual angles of horizontal left avertence of horizontal face object and the poor 30 degree of visual angles of horizontal right avertence with
And it is looked up from 70 degree of visual angles of upper vertical view deviation and from down and acquires the image obtained between 70 degree of visual angles of deviation.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 2), encirclement frame one
Rectangle frame, one is divided into 4 classifications, respectively represents the cable machinery of different types of damage：One kind is the cable of slot box damage
Equipment, one kind are the cable machineries of wind turbine damage, and one kind is the cable machinery of maintenance power box corrosion, and also one kind is distribution box
The cable machinery of damage is handled without the image of above-mentioned 4 classifications damage without label, and the training set in step 2) is by above-mentioned
Method obtains.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 3), image scaling side
Method uses bilinear interpolation.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 3), it is known that Q11、Q21、
Q12、Q22, it is P points the point of interpolation, uses bilinear interpolation；Q11Abscissa be x1, ordinate y1；Q12Abscissa be
x1, ordinate y2；Q21Abscissa be x2, ordinate y1；Q22Abscissa be x2, ordinate y2；The abscissa of P points
For x, ordinate y；R1Abscissa be x, ordinate y1；R2Abscissa be x, ordinate y2；
First in the direction of the x axis, to R1And R2Two points carry out linear interpolation, obtain：
Then linear interpolation is carried out in y-axis direction, obtained：
It is as follows to obtain desired result f (x, y)：
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, if one coordinate system of selection makes f
Four known point Q11、Q21、Q12、Q22Coordinate is respectively (0,0), (0,1), (1,0) and (1,1), then interpolation formula is with regard to abbreviation
For：
f(x,y)≈f(0,0)(1-x)(1-y)+f(1,0)x(1-y)+f(0,1)(1-x)y+f(1,1)xy。
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 3), using bilinearity
Interpolation method carries out image scaling to the sample image in training set, and adjustment picture size is 448 × 448.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 4), batch normalizes
Normalized is made to the input data of a certain layer network using following formula：
Batch stochastic gradient descent is used in training process, wherein：
E[x(k)] refer to every a collection of training data neuron x(k)Average value；Refer to every a batch instruction
Practice data neuron x(k)Standard deviation.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 5), model training misses
Shown in poor following formula：
Wherein：Input picture is divided into S × S-grid by YOLO, if the center of target falls into grid cell, grid
Unit is just responsible for detecting the target, and each grid will predict B bounding box；
xi, yi, wi, hiThe coordinate value of model prediction bounding box, x are indicated respectivelyiIndicate the horizontal seat of bounding box top left corner apex
Mark, yiIndicate the ordinate of bounding box top left corner apex, wiIndicate the width value of bounding box, hiIndicate the height value of bounding box, Ci
Indicate the self-confident value of the bounding box of model prediction, pi(c) probability value for indicating the classification of model prediction, added with the change of horizontal line on head
Amount corresponds to the normalized value of relevant variable respectively；
Indicate whether target appears in grid i；Indicate that j-th of bounding box in grid i predicts correct class
Not；
Since error in classification is different with the significance level of error of coordinate, so making at weighting to their own loss function
Reason more payes attention to the coordinate prediction of 8 dimensions, to larger weights are assigned before these loss functions, is denoted as λcoord, 5 are taken when training；
To the Confidence loss function of aimless bounding box, smaller weights are assigned, λ is denoted asnoobj, 0.5 is taken when training；To there is mesh
The Confidence loss function and classification loss function of target bounding box, assign normal weights 1.
As the supplement of above-mentioned cable machinery external corrosion failure evaluation method, in the step 6), tunnel internal electricity
Cable equipment testing image refers near crusing robot walking to cable machinery, and camera is towards cable machinery, with cable machinery
For object, between the poor 30 degree of visual angles of horizontal left avertence of horizontal face object and the poor 30 degree of visual angles of horizontal right avertence with
And it is looked up from 70 degree of visual angles of upper vertical view deviation and from down and acquires the image obtained between 70 degree of visual angles of deviation.
The invention has the advantages that：Recognition accuracy is high, and robustness is good；Using this method in buried cable tunnel
Crusing robot system in carry out cable machinery external corrosion failure evaluation there is extraordinary effect, and to being similarly in dusk
Secretly, the tunnel internal equipment detection under complex background has versatility.
The present invention carries out automatic identification by crusing robot to cable machinery external corrosion breakage state, is to realize cable
The key problem in technology of tunnel intelligent inspection and the important leverage of underground electric power netting safe running.
Description of the drawings
Fig. 1 is the method flow diagram in the embodiment of the present invention；
Fig. 2 is the network design configuration figure in the embodiment of the present invention；
Fig. 3 is the bilinear interpolation principle schematic in the embodiment of the present invention.
Specific implementation mode
With reference to the accompanying drawings of the specification and specific implementation mode invention is further described in detail.
Embodiment
As shown in Figure 1, the present invention provides a kind of cable machinery external corrosion failure evaluation method based on YOLO, including with
Lower step：
1) include by tunnel crusing robot camera shooting, collecting device target object tunnel internal cable machinery
Sample image, the object for including in sample image are slot box, wind turbine, maintenance power box and the distribution box of tunnel internal；
2) all tunnel internal cable machinery sample images are traversed, object is directed into rower to every image encirclement frame
Note processing, obtains training set；
3) use image-scaling method adjustment picture size for 448 × 448：Figure is carried out for the sample image in training set
As scaling, adjustment picture size is 448 × 448；
4) feature extraction is carried out to the image that step 3) obtains using convolutional neural networks, wherein convolutional neural networks use
20 layers of trained network of convolutional layer before YOLO, and using batch method for normalizing then every layer network is normalized
It is inputted again, carrys out predicted boundary frame finally by RPN networks；
5) by the continuous repetitive exercise of method in step 4) until model training error tend towards stability, the entire net finally obtained
Network model is as tunnel internal cable machinery detection model；
6) after acquisition cable machinery testing image is according to identical image-scaling method zooms in and out with step 3) in real time
As the input for the tunnel internal cable machinery detection model that step 5) obtains, the output of tunnel internal cable machinery detection model
The as final recognition result of cable machinery testing image.
In the step 1), tunnel internal cable machinery sample image refers to that crusing robot walking is attached to cable machinery
Closely, camera is towards cable machinery, using cable machinery as object, the poor 30 degree of visual angles of horizontal left avertence of horizontal face object and
70 degree of visual angles of deviation are looked up between the horizontal poor 30 degree of visual angles of right avertence and from upper 70 degree of visual angles of vertical view deviation and from down
Between acquire the image of acquisition.
In the step 2), encirclement frame is a rectangle frame, and one has been divided into 4 classifications, respectively represents different types of
The cable machinery of damage：One kind is the cable machinery of slot box damage, and one kind is the cable machinery of wind turbine damage, and one kind is maintenance electricity
The cable machinery of source case corrosion, also one kind be distribution box damage cable machinery, without above-mentioned 4 classifications damage image not
Processing is marked, the training set in step 2) obtains as stated above.
In the step 3), image-scaling method uses bilinear interpolation, principle as shown in Fig. 3, it is known that Q11、
Q21、Q12、Q22, but it is P points to want the point of interpolation, uses bilinear interpolation；Q11Abscissa be x1, ordinate y1；Q12Cross
Coordinate is x1, ordinate y2；Q21Abscissa be x2, ordinate y1；Q22Abscissa be x2, ordinate y2；P points
Abscissa is x, ordinate y；The abscissa of R1 is x, ordinate y1；The abscissa of R2 is x, ordinate y2；
First in the direction of the x axis, linear interpolation is carried out to two points of R1 and R2, obtained：
Then linear interpolation is carried out in y-axis direction, obtained：
It is as follows thus to obtain desired result f (x, y)：
If selection one coordinate system make f four known point coordinates be respectively (0,0), (0,1), (1,0) and (1,
1), then interpolation formula can abbreviation be：
f(x,y)≈f(0,0)(1-x)(1-y)+f(1,0)x(1-y)+f(0,1)(1-x)y+f(1,1)xy。
Image scaling is carried out to the sample image in training set using bilinear interpolation method, adjustment picture size is 448
×448。
In the step 4), convolutional neural networks literary grace 20 layers of convolutional layer trained network before YOLO, and use
Batch method for normalizing is normalized before every layer network, then by RPN networks come predicted boundary frame.Network design knot
Structure is as shown in Fig. 2, wherein batch normalization makees normalized using following formula to the input data of a certain layer network.
Batch stochastic gradient descent is used in training process, wherein：
E[x(k)] refer to every a collection of training data neuron x(k)Average value；Refer to every a batch instruction
Practice data neuron x(k)Standard deviation.
After the transformation of above formula, the activation x of some neuron forms the normal state point that mean value is 0, variance is 1
Cloth, it is therefore an objective to the linear zone of value toward the nonlinear transformation subsequently to be carried out be pulled, derivative value is increased, enhance backpropagation information
Mobility accelerates convergence speed.
In the step 5), shown in the following formula of model training error：
Wherein：Input picture is divided into S × S-grid by YOLO.If the center of target falls into grid cell, grid
Unit is just responsible for detecting the target, and each grid will predict B bounding box.In the present invention, S=14, B=2 are taken.
xi, yi, wi, hiThe coordinate value of model prediction bounding box is indicated respectively.xiIndicate the horizontal seat of bounding box top left corner apex
Mark, yiIndicate the ordinate of bounding box top left corner apex, wiIndicate the width value of bounding box, hiIndicate the height value of bounding box.Ci
Indicate the self-confident value of the bounding box of model prediction, pi(c) probability value of the classification of model prediction is indicated.Added with the change of horizontal line on head
Amount corresponds to the normalized value of relevant variable respectively.
Indicate whether target appears in grid i；Indicate that j-th of bounding box in grid i predicts correct class
Not.
Pay attention to：Loss function in above formula only when in a grid there are when target, just can be to error in classification
(classification error) is punished；Only when bounding box has predicted correct classification, bounding box can just be sat
Mark error (coordinate error) is punished.
Since error in classification is different with the significance level of error of coordinate, so being weighted to their own loss function
Processing more payes attention to the coordinate prediction of 8 dimensions, to larger weights are assigned before these loss functions, is denoted as λcoord, taken when training
5；To the Confidence loss function of aimless bounding box, smaller weights are assigned, λ is denoted asnoobj, 0.5 is taken when training；To having
The Confidence loss function and classification loss function of the bounding box of target, assign normal weights 1.
In the step 6), tunnel internal cable machinery testing image refers to that crusing robot walking is attached to cable machinery
Closely, camera is towards cable machinery, using cable machinery as object, the poor 30 degree of visual angles of horizontal left avertence of horizontal face object and
70 degree of visual angles of deviation are looked up between the horizontal poor 30 degree of visual angles of right avertence and from upper 70 degree of visual angles of vertical view deviation and from down
Between acquire the image of acquisition.
Application examples
Experiment picture shares 1000, wherein the picture as training set has 800, inside training set picture normally, slot
The cable machinery that box damage, wind turbine damage, maintenance power box corrosion, distribution box damage respectively has 160, remaining 200 pictures is made
Equally include each 40 of this 5 class cable machinery for test set picture.
The model trained using embodiment removes the test set picture of detection tunnel internal cable machinery, obtained result such as table
Shown in 1：
1 cable machinery state-detection result of table
As it can be seen from table 1 the cable of normal, slot box damage, wind turbine damage, maintenance power box corrosion, distribution box damage
Equipment can reach 100%.
For the cable machinery of normal, slot box damage, wind turbine damage, maintenance power box corrosion, distribution box damage, the present invention
Method accurately can detect and orient the cable machinery in image and complete accurate result detection.
It can be seen that the present invention can realize the cable machinery detection and positioning of crusing robot shooting, have higher
Accuracy rate, and it is good with stability, the advantages that strong antijamming capability, versatility is high, there is Shandong to tunnel internal dim environment
Stick can be applied to tunnel internal cruising inspection system.
Above-mentioned specific implementation mode is used for illustrating the present invention, rather than limits the invention, the present invention's
In spirit and scope of the claims, to any modifications and changes that the present invention makes, the protection model of the present invention is both fallen within
It encloses.
Claims (10)
1. the cable machinery external corrosion failure evaluation method based on YOLO, which is characterized in that including：
1) include the tunnel internal cable machinery sample image of object, sample by tunnel crusing robot camera shooting, collecting
The object for including in this image is slot box, wind turbine, maintenance power box and the distribution box of tunnel internal；
2) all tunnel internal cable machinery sample images are traversed, place is marked for object with encirclement frame to every image
Reason obtains training set；
3) image-scaling method is used to adjust picture size：Image scaling, adjustment figure are carried out for the sample image in training set
As size；
4) feature extraction is carried out to the image that step 3) obtains using convolutional neural networks, wherein convolutional neural networks use YOLO
The trained network of preceding 20 layers of convolutional layer, and using batch method for normalizing to every layer network be normalized then again into
Row input, carrys out predicted boundary frame finally by RPN networks；
5) by the continuous repetitive exercise of method in step 4) until model training error tend towards stability, the whole network mould finally obtained
Type is as tunnel internal cable machinery detection model；
6) cable machinery testing image, conduct after being zoomed in and out according to the identical image-scaling method with step 3) are acquired in real time
The input for the tunnel internal cable machinery detection model that step 5) obtains, the output of tunnel internal cable machinery detection model are
The final recognition result of cable machinery testing image.
2. cable machinery external corrosion failure evaluation method according to claim 1, which is characterized in that the step 1)
In, tunnel internal cable machinery sample image refers near crusing robot walking to cable machinery, and camera is set towards cable
It is standby, using cable machinery as object, the poor 30 degree of visual angles of horizontal left avertence of horizontal face object and the poor 30 degree of visual angles of horizontal right avertence
Between and look up to acquire between 70 degree of visual angles of deviation from upper 70 degree of visual angles of vertical view deviation and from down and obtain
The image obtained.
3. cable machinery external corrosion failure evaluation method according to claim 1, which is characterized in that the step 2)
In, encirclement frame is a rectangle frame, and one is divided into 4 classifications, respectively represents the cable machinery of different types of damage：One kind is slot
The cable machinery of box damage, one kind are the cable machineries of wind turbine damage, and one kind is the cable machinery of maintenance power box corrosion, also
One kind is the cable machinery of distribution box damage, is handled without label without the image of above-mentioned 4 classifications damage, in step 2)
Training set obtains as stated above.
4. cable machinery external corrosion failure evaluation method according to claim 1, which is characterized in that the step 3)
In, image-scaling method uses bilinear interpolation.
5. cable machinery external corrosion failure evaluation method according to claim 4, which is characterized in that the step 3)
In, it is known that Q11、Q21、Q12、Q22, it is P points the point of interpolation, uses bilinear interpolation；Q11Abscissa be x1, ordinate y1；
Q12Abscissa be x1, ordinate y2；Q21Abscissa be x2, ordinate y1；Q22Abscissa be x2, ordinate y2；
The abscissa of P points is x, ordinate y；R1Abscissa be x, ordinate y1；R2Abscissa be x, ordinate y2；
First in the direction of the x axis, to R1And R2Two points carry out linear interpolation, obtain：
Then linear interpolation is carried out in y-axis direction, obtained：
It is as follows to obtain desired result f (x, y)：
6. cable machinery external corrosion failure evaluation method according to claim 5, which is characterized in that if selection one
Coordinate system makes four known point Q of f11、Q21、Q12、Q22Coordinate is respectively (0,0), (0,1), (1,0) and (1,1), then
Interpolation formula is with regard to abbreviation：
f(x,y)≈f(0,0)(1-x)(1-y)+f(1,0)x(1-y)+f(0,1)(1-x)y+f(1,1)xy。
7. cable machinery external corrosion failure evaluation method according to claim 4, which is characterized in that the step 3)
In, image scaling is carried out to the sample image in training set using bilinear interpolation method, adjustment picture size is 448 × 448.
8. cable machinery external corrosion failure evaluation method according to claim 1, which is characterized in that the step 4)
In, batch normalization makees normalized using following formula to the input data of a certain layer network：
Batch stochastic gradient descent is used in training process, wherein：
E[x(k)] refer to every a collection of training data neuron x(k)Average value；Refer to every a collection of training data
Neuron x(k)Standard deviation.
9. cable machinery external corrosion failure evaluation method according to claim 1, which is characterized in that the step 5)
In, shown in the following formula of model training error：
Wherein：Input picture is divided into S × S-grid by YOLO, if the center of target falls into grid cell, grid cell
Just it is responsible for detecting the target, each grid will predict B bounding box；
xi, yi, wi, hiThe coordinate value of model prediction bounding box, x are indicated respectivelyiIndicate the abscissa of bounding box top left corner apex, yi
Indicate the ordinate of bounding box top left corner apex, wiIndicate the width value of bounding box, hiIndicate the height value of bounding box, CiIt indicates
The self-confident value of the bounding box of model prediction, pi(c) probability value for indicating the classification of model prediction, added with the variable of horizontal line point on head
The normalized value of relevant variable is not corresponded to；
Indicate whether target appears in grid i；Indicate that j-th of bounding box in grid i predicts correct classification；
Since error in classification is different with the significance level of error of coordinate, so weighting processing is made to their own loss function,
The coordinate prediction for more paying attention to 8 dimensions, to larger weights are assigned before these loss functions, is denoted as λcoord, 5 are taken when training；It is right
The Confidence loss function of aimless bounding box assigns smaller weights, is denoted as λnoobj, 0.5 is taken when training；To there is target
Bounding box Confidence loss function and classification loss function, assign normal weights 1.
10. cable machinery external corrosion failure evaluation method according to claim 1, which is characterized in that the step
6) in, tunnel internal cable machinery testing image refers to crusing robot walking near cable machinery, and camera is towards cable
Equipment, using cable machinery as object, the poor 30 degree of visual angles of horizontal left avertence of horizontal face object and poor 30 degree of horizontal right avertence regard
It looks up between angle and from upper 70 degree of visual angles of vertical view deviation and from down and is acquired between 70 degree of visual angles of deviation
The image of acquisition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810437049.1A CN108734117A (en) | 2018-05-09 | 2018-05-09 | Cable machinery external corrosion failure evaluation method based on YOLO |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810437049.1A CN108734117A (en) | 2018-05-09 | 2018-05-09 | Cable machinery external corrosion failure evaluation method based on YOLO |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108734117A true CN108734117A (en) | 2018-11-02 |
Family
ID=63938164
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810437049.1A Pending CN108734117A (en) | 2018-05-09 | 2018-05-09 | Cable machinery external corrosion failure evaluation method based on YOLO |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108734117A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109522963A (en) * | 2018-11-26 | 2019-03-26 | 北京电子工程总体研究所 | A kind of the feature building object detection method and system of single-unit operation |
| CN110060233A (en) * | 2019-03-20 | 2019-07-26 | 中国农业机械化科学研究院 | A kind of corn ear damage testing method |
| CN110188780A (en) * | 2019-06-03 | 2019-08-30 | 电子科技大学中山学院 | Method and device for constructing deep learning model for positioning multi-target feature points |
| CN110321853A (en) * | 2019-07-05 | 2019-10-11 | 杭州巨骐信息科技股份有限公司 | Distribution cable external force damage prevention system based on video intelligent detection |
| CN110780164A (en) * | 2019-11-04 | 2020-02-11 | 华北电力大学（保定） | Insulator infrared fault positioning diagnosis method and device based on YOLO |
| CN110907749A (en) * | 2019-11-19 | 2020-03-24 | 湖南国奥电力设备有限公司 | Method and device for positioning fault underground cable |
| CN111582334A (en) * | 2020-04-23 | 2020-08-25 | 浙江大学 | High-speed railway catenary image identification method combining YOLOv3 and SENEt |
| CN113160184A (en) * | 2021-04-26 | 2021-07-23 | 贵州电网有限责任公司 | Unmanned aerial vehicle intelligent inspection cable surface defect detection method based on deep learning |
| CN113341820A (en) * | 2021-06-16 | 2021-09-03 | 江苏纬信工程咨询有限公司 | Intelligent construction site safety monitoring device based on Internet of things and monitoring method thereof |
| CN113486746A (en) * | 2021-06-25 | 2021-10-08 | 海南电网有限责任公司三亚供电局 | Power cable external damage prevention method based on biological induction and video monitoring |
| CN114049354A (en) * | 2022-01-12 | 2022-02-15 | 山东仲良格环保技术有限公司 | Rust remover optimized proportioning method and system based on metal corrosion degree |
| WO2022062242A1 (en) * | 2020-09-27 | 2022-03-31 | 广东海洋大学 | Deep learning-based underwater imaging and fishing net damage identification method and system |
| CN114758288A (en) * | 2022-03-15 | 2022-07-15 | 华北电力大学 | Power distribution network engineering safety control detection method and device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809722A (en) * | 2015-04-13 | 2015-07-29 | 国家电网公司 | Electrical device fault diagnosis method based on infrared thermography |
| CN106198551A (en) * | 2016-08-01 | 2016-12-07 | 南方电网科学研究院有限责任公司 | Method and device for detecting defects of power transmission line |
| CN106251059A (en) * | 2016-07-27 | 2016-12-21 | 中国电力科学研究院 | A kind of cable status appraisal procedure based on probabilistic neural network algorithm |
| CN106408562A (en) * | 2016-09-22 | 2017-02-15 | 华南理工大学 | Fundus image retinal vessel segmentation method and system based on deep learning |
| CN106707109A (en) * | 2017-02-21 | 2017-05-24 | 国网山东省电力公司邹城市供电公司 | Underground cable detection system |
| CN107330437A (en) * | 2017-07-03 | 2017-11-07 | 贵州大学 | Feature extracting method based on the real-time detection model of convolutional neural networks target |
| CN107491781A (en) * | 2017-07-21 | 2017-12-19 | 国家电网公司 | A kind of crusing robot visible ray and infrared sensor data fusion method |
-
2018
- 2018-05-09 CN CN201810437049.1A patent/CN108734117A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809722A (en) * | 2015-04-13 | 2015-07-29 | 国家电网公司 | Electrical device fault diagnosis method based on infrared thermography |
| CN106251059A (en) * | 2016-07-27 | 2016-12-21 | 中国电力科学研究院 | A kind of cable status appraisal procedure based on probabilistic neural network algorithm |
| CN106198551A (en) * | 2016-08-01 | 2016-12-07 | 南方电网科学研究院有限责任公司 | Method and device for detecting defects of power transmission line |
| CN106408562A (en) * | 2016-09-22 | 2017-02-15 | 华南理工大学 | Fundus image retinal vessel segmentation method and system based on deep learning |
| CN106707109A (en) * | 2017-02-21 | 2017-05-24 | 国网山东省电力公司邹城市供电公司 | Underground cable detection system |
| CN107330437A (en) * | 2017-07-03 | 2017-11-07 | 贵州大学 | Feature extracting method based on the real-time detection model of convolutional neural networks target |
| CN107491781A (en) * | 2017-07-21 | 2017-12-19 | 国家电网公司 | A kind of crusing robot visible ray and infrared sensor data fusion method |
Non-Patent Citations (6)
| Title |
|---|
| WEI LIU等: "SSD: Single Shot MultiBox Detector", 《EUROPEAN CONFERENCE ON COMPUTER VISION》 * |
| 杨眷玉: "基于卷积神经网络的物体识别研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 汤踊 等: "深度学习在输电线路中部件识别与缺陷检测的研究", 《电子测量技术》 * |
| 王珏: "基于Android平台的多目标实时跟踪技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 石鑫,朱永利: "深度学习神经网络在电力变压器故障诊断中的应用", 《电力建设》 * |
| 苗向鹏: "基于图像处理的接触网绝缘子识别与破损检测", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109522963A (en) * | 2018-11-26 | 2019-03-26 | 北京电子工程总体研究所 | A kind of the feature building object detection method and system of single-unit operation |
| CN110060233A (en) * | 2019-03-20 | 2019-07-26 | 中国农业机械化科学研究院 | A kind of corn ear damage testing method |
| CN110188780A (en) * | 2019-06-03 | 2019-08-30 | 电子科技大学中山学院 | Method and device for constructing deep learning model for positioning multi-target feature points |
| CN110321853B (en) * | 2019-07-05 | 2021-05-11 | 杭州巨骐信息科技股份有限公司 | Distributed cable external-damage-prevention system based on video intelligent detection |
| CN110321853A (en) * | 2019-07-05 | 2019-10-11 | 杭州巨骐信息科技股份有限公司 | Distribution cable external force damage prevention system based on video intelligent detection |
| CN110780164A (en) * | 2019-11-04 | 2020-02-11 | 华北电力大学（保定） | Insulator infrared fault positioning diagnosis method and device based on YOLO |
| CN110907749A (en) * | 2019-11-19 | 2020-03-24 | 湖南国奥电力设备有限公司 | Method and device for positioning fault underground cable |
| CN111582334A (en) * | 2020-04-23 | 2020-08-25 | 浙江大学 | High-speed railway catenary image identification method combining YOLOv3 and SENEt |
| WO2022062242A1 (en) * | 2020-09-27 | 2022-03-31 | 广东海洋大学 | Deep learning-based underwater imaging and fishing net damage identification method and system |
| CN113160184A (en) * | 2021-04-26 | 2021-07-23 | 贵州电网有限责任公司 | Unmanned aerial vehicle intelligent inspection cable surface defect detection method based on deep learning |
| CN113160184B (en) * | 2021-04-26 | 2022-06-07 | 贵州电网有限责任公司 | Unmanned aerial vehicle intelligent inspection cable surface defect detection method based on deep learning |
| CN113341820A (en) * | 2021-06-16 | 2021-09-03 | 江苏纬信工程咨询有限公司 | Intelligent construction site safety monitoring device based on Internet of things and monitoring method thereof |
| CN113486746A (en) * | 2021-06-25 | 2021-10-08 | 海南电网有限责任公司三亚供电局 | Power cable external damage prevention method based on biological induction and video monitoring |
| CN114049354A (en) * | 2022-01-12 | 2022-02-15 | 山东仲良格环保技术有限公司 | Rust remover optimized proportioning method and system based on metal corrosion degree |
| CN114758288A (en) * | 2022-03-15 | 2022-07-15 | 华北电力大学 | Power distribution network engineering safety control detection method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108734117A (en) | Cable machinery external corrosion failure evaluation method based on YOLO | |
| CN112199993B (en) | Method for identifying transformer substation insulator infrared image detection model in any direction based on artificial intelligence | |
| CN108846418A (en) | A kind of positioning of cable machinery temperature anomaly and recognition methods | |
| CN108275524B (en) | A kind of elevator maintenance operation monitoring and guiding device based on the assessment of the first multi-view video series of operations | |
| CN110826514A (en) | Construction site violation intelligent identification method based on deep learning | |
| CN110070530A (en) | A kind of powerline ice-covering detection method based on deep neural network | |
| CN113903081A (en) | Visual identification artificial intelligence alarm method and device for images of hydraulic power plant | |
| CN113052876B (en) | Video relay tracking method and system based on deep learning | |
| CN108229587A (en) | A kind of autonomous scan method of transmission tower based on aircraft floating state | |
| CN114623049B (en) | Wind turbine tower clearance monitoring method and computer program product | |
| CN113569956B (en) | Mountain fire disaster investigation and identification method based on AI algorithm | |
| CN113763484A (en) | Ship target positioning and speed estimation method based on video image analysis technology | |
| CN115019254A (en) | Method, device, terminal and storage medium for detecting foreign matter invasion in power transmission area | |
| CN113076825A (en) | Transformer substation worker climbing safety monitoring method | |
| CN115880231A (en) | Power transmission line hidden danger detection method and system based on deep learning | |
| Fang et al. | A framework of power pylon detection for UAV-based power line inspection | |
| CN115689206A (en) | Intelligent monitoring method for transformer substation infrastructure progress based on deep learning | |
| CN115880613A (en) | Monitoring method of marine flexible pipeline monitoring system based on deep learning | |
| Rong et al. | A joint faster RCNN and stereovision algorithm for vegetation encroachment detection in power line corridors | |
| Ning et al. | Object detection and danger warning of transmission channel based on improved YOLO network | |
| CN114882682A (en) | High-voltage cable state monitoring platform and monitoring method | |
| Wu et al. | Detection method based on improved faster R-CNN for pin defect in transmission lines | |
| Wang et al. | Research on appearance defect detection of power equipment based on improved faster-rcnn | |
| CN110363877A (en) | Ship lock inspection device and method based on unmanned running gear | |
| Ma et al. | Assisting Wind Turbine Hoisting with Yolov7 and Object Tracking Technology |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181102 |