CN106980827A - A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images - Google Patents
A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images Download PDFInfo
- Publication number
- CN106980827A CN106980827A CN201710156551.0A CN201710156551A CN106980827A CN 106980827 A CN106980827 A CN 106980827A CN 201710156551 A CN201710156551 A CN 201710156551A CN 106980827 A CN106980827 A CN 106980827A
- Authority
- CN
- China
- Prior art keywords
- image
- bird
- nest
- network
- aerial images
- 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
- G06V20/13—Satellite images
-
- 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
Abstract
The present invention relates to a kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images, including following steps：Make data set.Train neutral net：Using convolutional neural networks AlexNet and the VGG16 network of two kinds of different structures, it is respectively trained and obtains two graders based on depth convolutional neural networks, the two graders, which can be distinguished in image, whether there is Bird's Nest presence, referred to as network 1 and network 2.Test identification model：Aerial Images to be tested are classified with network 1 first, the image existed with the presence of the image of Bird's Nest and without Bird's Nest is identified out, is individually placed in set 1 and set 2.The all images gathered in 2 are classified using network 2 again, will wherein identify that the image with the presence of Bird's Nest is placed into set 1.Finally, remaining all images are Aerial Images that the system is identified, in the absence of Bird's Nest in set 2, and the image gathered in 1 is the Aerial Images that there may be Bird's Nest.
Description
Technical field
The invention belongs to image classification field, it is related to a kind of identification transmission line of electricity of utilization depth convolutional neural networks and takes photo by plane
The method of Bird's Nest in image.
Background technology
With the fast development of Chinese national economy, increasingly vigorous, the corresponding electric power work of the demand to electric power energy
Cheng Jianshe dynamics is also constantly being strengthened[1].Transmission line of electricity plays very important effect in power system, is directly connected to
The electrical problem of huge numbers of families, major power outage will bring immeasurable loss to national economy.Therefore, the peace of transmission line of electricity
Full property is one of the problem of power department is paid high attention to[2]。
In recent years, with the continuous improvement of China's natural environment, and correlation protects the continuous complete of the laws and regulations of animal
Kind, the procreation of birds is gradually accelerated, and birds activity is increasingly frequent, and overhead transmission line bird pest failure increases year by year, serious to threaten
The safe operation of power network.Bird pest is to jeopardize one of major failure of transmission line safety operation, bird pest, damage to crops caused by thunder, external force destruction
It is the big major obstacle of overhead transmission line three, the ratio shared by bird pest is 32% according to statistics, and wherein Bird's Nest or bird does nest and cause line
What road was tripped constitutes about 80% or so.How Bird's Nest harm is effectively taken precautions against, to ensuring that transmission line safety operation has important meaning
Justice.
At present, the identification to Bird's Nest in transmission line of electricity mainly has two kinds of means, and one is manual identified, and two be using at image
The method identification of reason.The former is the traditional monitoring mode of transmission line of electricity, and mainly circuit is checked by manpower, finds Bird's Nest
Afterwards, staff is got rid of the danger.This surveillance and control measure consumes substantial amounts of manpower and materials, it is difficult to ensure the ageing of monitoring result
And accuracy, it is not appropriate for large-scale promotion.China territory is vast, with a varied topography, and Plain is few, hills and mountain area are more, meteorological
Complicated condition, daily inspection after being built up for heavy constructions such as extra-high voltage and interregional grids safeguards, existing conventionally test and
Detection methods can not meet its requirement rapidly and efficiently.With automatic control technology, GPS navigation technology, air remote sensing mapping
The development of technology and wireless communication technology etc., the use of unmanned plane is extended to many civil areas, such as earth from military field
Physical detecting, disaster surveillance, crops monitoring.Electric power inspection task can be efficiently completed using unmanned plane Aerial photography[1]。
Transmission line of electricity monitoring technology without Aerial surveillance and based on remote image identifying system is emerging at present
Transmission line of electricity monitoring method.Manpower is substituted using patrol UAV and carries out power transmission line inspection operation；And based on Aerial Images
Recognition methods be by unmanned plane and helicopter collection come image transmitting to backstage monitoring center, utilize the image of monitoring center
Handle identifying system and automatic identification is carried out to the state of Bird's Nest presence or absence on electric power line pole tower.With traditional manual inspection side
Method is compared, and these monitoring methods are simpler, accurate, real-time, economy.
Bird's Nest identification based on Aerial Images depends on the appropriate selection and extraction to characteristics of image.Convolutional neural networks
The parameter that (Convolutional Neural Network, CNN) is provided in a kind of learning model end to end, model passes through
Gradient descent method is trained, and the convolutional neural networks after training can learn the feature into image, and complete
Extraction and classification to characteristics of image.Compared with traditional image-recognizing method, CNN is with very outstanding feature extraction and feature
Learning ability[3]。
[1] Liu Guosong, application [J] the Northeast Electric Power University after strong unmanned planes in power system that goes into business is learned
Report, 2012, (01):53-56.
[2] transmission line of electricity identifications of the Tong Weiguo based on Aerial Images and condition detection method research [D] North China electric power are big
Learn, 2011.
[3] application [J] of Li Xiaolong, Zhang Zhaoxiang, Wang Yunhong, Liu Qing outstanding person's deep learnings in scene classification of taking photo by plane is calculated
Machine science is with exploring, 2014, (03):305-312.
The content of the invention
It is an object of the invention to propose a kind of method that Bird's Nest in transmission line of electricity Aerial Images is recognized based on Aerial Images,
The method is compared with traditional images recognition methods, with more preferable feature extraction and feature learning ability, so as to quick high
Realize that power transmission line Bird's Nest is recognized to effect.Technical scheme is as follows：
A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images, including following steps：
The first step, makes data set：The Aerial Images of transmission line of electricity are collected, and use image enhancement technique, making meets
The data set of call format, matching network structure.
Second step, trains neutral net：Using convolutional neural networks AlexNet and the VGG16 network of two kinds of different structures,
The data set obtained using step one, is respectively trained and obtains two graders based on depth convolutional neural networks, the two points
Class device, which can be distinguished in image, whether there is Bird's Nest presence, referred to as network 1 and network 2, by two cascades, composition Bird's Nest identification
System.
3rd step, identification：Aerial Images to be tested are classified with network 1 first, Bird's Nest is identified out and deposits
Image and without Bird's Nest exist image, be individually placed to set 1 and set 2 in.The all images gathered in 2 are used again
Network 2 is classified, and will wherein identify that the image with the presence of Bird's Nest is placed into set 1.Finally, it is remaining complete in set 2
Portion's image is Aerial Images that the system is identified, in the absence of Bird's Nest, and the image gathered in 1 is the boat that there may be Bird's Nest
Clap image.
The present invention is designed a kind of transmission line of electricity Bird's Nest based on Aerial Images and known using the algorithm of depth convolutional neural networks
Other system.The system is using unmanned plane and the picture library of taking photo by plane of helicopter line walking acquisition as research object, to the image collected
Carry out a series of processing and identification, including the making of data set, the training of neutral net and the test of identification model.The system
To ensure that safe operation of power system provides a kind of newly efficient and accurate means, greatly reduce conventional method manual identified
Workload.
Brief description of the drawings
Fig. 1 as network 1 neural network structure figure
Fig. 2 as network 2 neural network structure figure
The recognition effect contrast of Fig. 3 networks 1 and network 2 before and after fine setting
The final cascade network identification models of Fig. 4 and the recognition effect of the network 1 after fine setting are contrasted
Embodiment
To make technical scheme clearer, is further elaborated to the present invention below in conjunction with the accompanying drawings.The present invention
Implement according to the following steps：
The first step, prepares data set.
(1) preparing pictures data and label data.
The transmission line of electricity that will be obtained using unmanned plane line walking and helicopter style of shooting taken photo by plane picture, including shaft tower is taken photo by plane
Whether picture is collected, classified in the transmission line of electricity in foundation picture with the presence of Bird's Nest, and according to certain quantity ratio
Example is divided into training set, checking collection and test set.In order to match the structure of different neutral nets, size normalizing is carried out to all pictures
Change, two picture set that size is 256*256 and 224*224 are obtained, for network 1 and network 2 to be respectively trained.According to people
For the good picture of classifying packing, the label file for meeting call format is made.
(2) image enhaucament is carried out.
In view of transmission line of electricity Aerial Images data volume it is small the characteristics of, in order to improve system robustness and identification essence
Degree, we carry out image enhaucament to training data.The present invention is increased using five kinds of natural image Enhancement Methods to Aerial Images
By force, wherein S (o) is image after enhancing, and S (i) is original image：
A) picture noise is added.Gaussian noise, salt-pepper noise, three kinds of common image noises of poisson noise are chosen, and are changed
Become different signal to noise ratio generation data.N (θ) is noise in formula, and θ is noise parameter.
S (o)=S (i)+N (θ)
B) image blurringization.Choose the common filtering such as mean filter, gaussian filtering, motion blur, contrast enhancing filtering
Device, sets different parameters and obtains in the image-type after device after filtering that F () is wave filter, φ is filter parameter.
S (o)=F (S (i) | φ)
C) image intensities are changed.Different brightness ratios 20%, 50%, 80% are chosen, original image is subjected to brightness value
Conversion.F () is that brightness changes function in formula.
S (o)=f (S (i))
D) picture quality is adjusted.By jpeg images according to jpeg coding criterions reduce quality, obtain 75%, 90% two kind not
Image under homogenous quantities.M () is jpeg coding quality Tuning functions in formula.
S (o)=M (S (i))
The several frequently seen strategy more than, becomes by using such strategy to the obtained training image of the first step
Change, realize the image enhaucament to micro- Aerial Images training dataset.Enhanced data volume is 13 times of original data volume.
Second step, trains neutral net.
(1) training AlexNet convolutional neural networks are used as network 1.AlexNet uses 5 layers of convolutional layer, 3 layers of full articulamentum
Convolutional neural networks structure, using nonlinear activation function ReLu, and add Dropout technologies and prevent over-fitting, in image
Classification is obtained a wide range of applications in field, and network structure is as shown in Figure 1.
Initial training parameter is set first during concrete operations, including the learning rate of gradient decline is set and whether used
(finetune) is finely tuned on the model of pre-training, then to the calculating average file of all training images, by used in training
Image subtracts average file, then is input in neutral net, and the ginseng of primary network is updated after forward conduction, reverse conduction
Number weight, after successive ignition, you can the neutral net trained is used as network 1.
(2) training VGG16 convolutional neural networks are used as network 2.Relative to AlexNet, VGG16 has deeper network knot
Structure, is made up of 13 layers of convolutional layer, 3 layers of full articulamentum, and has used in every layer of convolutional layer with more the identification, size to be
3*3 convolution kernel, VGG16 generally has preferably classification accuracy in image classification field, and network structure is as shown in Figure 2.
We equally set various training parameters, it is considered to network weight initialization mode.By identical with first neutral net of training
Step method after the i.e. neutral net that is trained be referred to as network 2.
(3) network 1 is cascaded with network 2.Using the output of network 1 as the input of network 2, final knowledge is combined as
Other model.Requirement due to two network structures to inputting size is different, and the dimension of picture into network 1 is 256*256, is passed through
Picture after network 1 will re-start size change over, be transformed to the size 224*224 of matching network 2.
3rd step, tests identification model.
(1) Aerial Images to be tested are classified with network 1 first, the probability exported according to network is identified
Go out the image existed with the presence of the image of Bird's Nest and without Bird's Nest, be individually placed in set 1 and set 2.
(2) all images gathered in 2 are classified using network 2 again, will wherein identifies the figure with the presence of Bird's Nest
As placing into set 1.Finally, remaining all images are figure that the system is identified, in the absence of Bird's Nest of taking photo by plane in set 2
Picture, and the image gathered in 1 is the Aerial Images that there may be Bird's Nest.
4th step, is analyzed experimental data and is handled, and evaluates the recognition accuracy of the system.
(1) it is ALL by all picture numbers of system testing, system identification goes out in the picture set 1 with the presence of Bird's Nest
Amount of images be designated as F, including identifying Bird's Nest without Bird's Nest originally and there is Bird's Nest to recognize correctly originally
Picture number, is designated as FP and FN, then F=FP+FN respectively；Image system identification gone out in the picture set 2 that no Bird's Nest is present
Quantity is designated as T, including without Bird's Nest being originally that identification is correct and the original picture for having Bird's Nest without identifying Bird's Nest
Quantity, is designated as TP and TN, then T=TP+TN respectively.The system defines two kinds of distinguishing indexes according to actual identification needs：
In view of the requirement of the actual identification mission of system, we select, and loss is minimum, work the maximum model of slip
It is used as the identifying system of the present invention.
(2) according to the identification situation of network 1 and network 2, and whether network is respectively obtained using fine setting (finetune)
The loss and workload slip of two networks, four kinds of situations.As shown in Figure 3.Two cascades are constituted to the knowledge of the present invention
Recognition result after other model with network after fine setting 1 as shown in figure 4, and contrasted.It can find that two nets will be used alone
Network recognizes Bird's Nest defect, and loss is consistent before and after fine setting, and network 1 is more slightly higher than the workload slip of network 2.And by two
What network was cascaded the model 100% that constitutes afterwards have identified the sample that there is Bird's Nest defect in test set, and reduce
The manual identified workload of half, with good effect.
Claims (2)
1. a kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images, including following steps：
The first step, makes data set：The Aerial Images of transmission line of electricity are collected, and use image enhancement technique, making meets form
It is required that, the data set of matching network structure.
Second step, trains neutral net：Using convolutional neural networks AlexNet and the VGG16 network of two kinds of different structures, use
The data set that step one is obtained, is respectively trained and obtains two graders based on depth convolutional neural networks, the two graders
It can distinguish in image and whether there is Bird's Nest presence, referred to as network 1 and network 2, by two cascades, composition Bird's Nest identification system
System.
3rd step, identification：Aerial Images to be tested are classified with network 1 first, are identified out with the presence of Bird's Nest
Image and the image existed without Bird's Nest, are individually placed in set 1 and set 2.The all images gathered in 2 are used into network again
2 are classified, and will wherein identify that the image with the presence of Bird's Nest is placed into set 1.Finally, remaining all figures in set 2
As Aerial Images being identified for the system, in the absence of Bird's Nest, and the image gathered in 1 is the figure of taking photo by plane that there may be Bird's Nest
Picture.
2. according to the method described in claim 1, it is characterised in that the image after original image and various image enhaucaments is used for
Data set is made, the method for carrying out image enhaucament to original image is as follows：
A) picture noise is added；
B) to image blurringization 2)；
C) image intensities are changed；Different brightness ratios 20%, 50%, 80% are chosen, original image is subjected to brightness value change
Change；
D) picture quality is adjusted.Quality is reduced according to coding criterion, the image under 75%, 90% two kind of different quality is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710156551.0A CN106980827A (en) | 2017-03-16 | 2017-03-16 | A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710156551.0A CN106980827A (en) | 2017-03-16 | 2017-03-16 | A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106980827A true CN106980827A (en) | 2017-07-25 |
Family
ID=59339315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710156551.0A Pending CN106980827A (en) | 2017-03-16 | 2017-03-16 | A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106980827A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107784634A (en) * | 2017-09-06 | 2018-03-09 | 广东工业大学 | A kind of power transmission line shaft tower Bird's Nest recognition methods based on template matches |
| CN108090518A (en) * | 2017-12-29 | 2018-05-29 | 美的集团股份有限公司 | A kind of cereal recognition methods, device and computer storage media |
| CN108090517A (en) * | 2017-12-29 | 2018-05-29 | 美的集团股份有限公司 | A kind of cereal recognition methods, device and computer storage media |
| CN108510451A (en) * | 2018-02-09 | 2018-09-07 | 杭州雄迈集成电路技术有限公司 | A method of the reconstruction car plate based on the double-deck convolutional neural networks |
| CN109086716A (en) * | 2018-08-01 | 2018-12-25 | 北京嘀嘀无限科技发展有限公司 | A kind of method and device of seatbelt wearing detection |
| CN109190545A (en) * | 2018-08-27 | 2019-01-11 | 国网天津市电力公司 | Bird's Nest automatic recognition system and its methods and applications in transmission line of electricity |
| CN109272043A (en) * | 2018-09-21 | 2019-01-25 | 北京京东金融科技控股有限公司 | Training data generation method, system and electronic equipment for optical character identification |
| CN109472229A (en) * | 2018-10-30 | 2019-03-15 | 福州大学 | Shaft tower Bird's Nest detection method based on deep learning |
| CN109902647A (en) * | 2019-03-08 | 2019-06-18 | 成都国铁电气设备有限公司 | The portable online Bird's Nest intelligent identification Method of one kind and system |
| CN110070530A (en) * | 2019-04-19 | 2019-07-30 | 山东大学 | A kind of powerline ice-covering detection method based on deep neural network |
| CN112183191A (en) * | 2020-08-18 | 2021-01-05 | 天津大学 | Method for detecting bird nest in power transmission line based on countermeasure training |
| CN112307851A (en) * | 2019-08-02 | 2021-02-02 | 上海交通大学烟台信息技术研究院 | Method and system for identifying bird nest on electric power iron tower |
| WO2021263010A3 (en) * | 2020-06-26 | 2022-03-10 | X Development Llc | Electrical power grid modeling |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530620A (en) * | 2013-10-30 | 2014-01-22 | 华北电力大学 | Method for identifying bird nest on electric transmission line tower |
-
2017
- 2017-03-16 CN CN201710156551.0A patent/CN106980827A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530620A (en) * | 2013-10-30 | 2014-01-22 | 华北电力大学 | Method for identifying bird nest on electric transmission line tower |
Non-Patent Citations (6)
| Title |
|---|
| CONNOR BOWLEY等: "Detecting Wildlife in Uncontrolled Outdoor Video using Convolutional Neural Networks", 《2016 IEEE 12TH INTERNATIONAL CONFERENCE ON E-SCIENCE》 * |
| QIAOSONG WANG等: "Fast, Deep Detection and Tracking of Birds & Nests", 《INTERNATIONAL SYMPOSIUM ON VISUAL COMPUTING》 * |
| 付敏: "基于卷积神经网络的人群密度估计", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 刘宁: "航拍图像中绝缘子定位与状态检测研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 张帅等: "基于分层卷积深度学习系统的植物叶片识别研究", 《北京林业大学学报》 * |
| 王桢: "基于视频图像处理技术的输电线路杆塔鸟巢识别", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107784634A (en) * | 2017-09-06 | 2018-03-09 | 广东工业大学 | A kind of power transmission line shaft tower Bird's Nest recognition methods based on template matches |
| CN108090518A (en) * | 2017-12-29 | 2018-05-29 | 美的集团股份有限公司 | A kind of cereal recognition methods, device and computer storage media |
| CN108090517A (en) * | 2017-12-29 | 2018-05-29 | 美的集团股份有限公司 | A kind of cereal recognition methods, device and computer storage media |
| CN108510451B (en) * | 2018-02-09 | 2021-02-12 | 杭州雄迈集成电路技术股份有限公司 | Method for reconstructing license plate based on double-layer convolutional neural network |
| CN108510451A (en) * | 2018-02-09 | 2018-09-07 | 杭州雄迈集成电路技术有限公司 | A method of the reconstruction car plate based on the double-deck convolutional neural networks |
| CN109086716A (en) * | 2018-08-01 | 2018-12-25 | 北京嘀嘀无限科技发展有限公司 | A kind of method and device of seatbelt wearing detection |
| CN109190545A (en) * | 2018-08-27 | 2019-01-11 | 国网天津市电力公司 | Bird's Nest automatic recognition system and its methods and applications in transmission line of electricity |
| CN109272043A (en) * | 2018-09-21 | 2019-01-25 | 北京京东金融科技控股有限公司 | Training data generation method, system and electronic equipment for optical character identification |
| CN109472229A (en) * | 2018-10-30 | 2019-03-15 | 福州大学 | Shaft tower Bird's Nest detection method based on deep learning |
| CN109902647A (en) * | 2019-03-08 | 2019-06-18 | 成都国铁电气设备有限公司 | The portable online Bird's Nest intelligent identification Method of one kind and system |
| CN110070530A (en) * | 2019-04-19 | 2019-07-30 | 山东大学 | A kind of powerline ice-covering detection method based on deep neural network |
| CN112307851A (en) * | 2019-08-02 | 2021-02-02 | 上海交通大学烟台信息技术研究院 | Method and system for identifying bird nest on electric power iron tower |
| WO2021263010A3 (en) * | 2020-06-26 | 2022-03-10 | X Development Llc | Electrical power grid modeling |
| US11580728B2 (en) | 2020-06-26 | 2023-02-14 | X Development Llc | Electrical power grid modeling |
| EP4220228A1 (en) * | 2020-06-26 | 2023-08-02 | X Development LLC | Electrical power grid modeling |
| US11875561B2 (en) | 2020-06-26 | 2024-01-16 | X Development Llc | Electrical power grid modeling |
| US11954907B2 (en) | 2020-06-26 | 2024-04-09 | X Development Llc | Electrical power grid modeling |
| CN112183191A (en) * | 2020-08-18 | 2021-01-05 | 天津大学 | Method for detecting bird nest in power transmission line based on countermeasure training |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106980827A (en) | A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images | |
| CN106971152A (en) | A kind of method of Bird's Nest in detection transmission line of electricity based on Aerial Images | |
| CN109117858B (en) | Method and device for monitoring icing of wind driven generator blade | |
| CN110261394A (en) | Online fan blade damages real-time diagnosis system and method | |
| Xu et al. | Wind turbine blade surface inspection based on deep learning and UAV-taken images | |
| CN108596886A (en) | Aerial Images insulator based on deep learning falls piece fault rapid detecting method | |
| CN110070530A (en) | A kind of powerline ice-covering detection method based on deep neural network | |
| CN108416294A (en) | A kind of fan blade fault intelligent identification method based on deep learning | |
| CN111368690A (en) | Deep learning-based video image ship detection method and system under influence of sea waves | |
| CN112367273B (en) | Flow classification method and device of deep neural network model based on knowledge distillation | |
| CN110084094A (en) | A kind of unmanned plane target identification classification method based on deep learning | |
| CN113205039B (en) | Power equipment fault image recognition disaster investigation system and method based on multiple DCNN networks | |
| CN106056069A (en) | Unmanned aerial vehicle image analysis-based forest land resource asset evaluation method and evaluation system | |
| Zhang et al. | Automatic recognition of oil industry facilities based on deep learning | |
| CN115249331B (en) | Mine ecological safety identification method based on convolutional neural network model | |
| CN112229845A (en) | Unmanned aerial vehicle high-precision winding tower intelligent inspection method based on visual navigation technology | |
| Hao et al. | Detection of bird nests on power line patrol using single shot detector | |
| CN104280253B (en) | A kind of fault diagnosis method and system based on immune detectors | |
| Yang et al. | Lightweight Bird's Nest Location Recognition Method Based on YOLOv4-tiny | |
| Zhou et al. | A method of CNN traffic classification based on sppnet | |
| CN110705693A (en) | Unmanned aerial vehicle abnormal behavior recognition module and recognition method thereof | |
| Wan et al. | Research on the method of grass mouse hole target detection based on deep learning | |
| CN116205905A (en) | Power distribution network construction safety and quality image detection method and system based on mobile terminal | |
| CN114359578A (en) | Application method and system of pest and disease damage identification intelligent terminal | |
| CN115311601A (en) | Fire detection analysis method based on video analysis 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: 20170725 |