CN109902647B - Portable online bird nest intelligent identification method and system - Google Patents
Portable online bird nest intelligent identification method and system Download PDFInfo
- Publication number
- CN109902647B CN109902647B CN201910176773.8A CN201910176773A CN109902647B CN 109902647 B CN109902647 B CN 109902647B CN 201910176773 A CN201910176773 A CN 201910176773A CN 109902647 B CN109902647 B CN 109902647B
- Authority
- CN
- China
- Prior art keywords
- bird nest
- bird
- nest
- intelligent
- identification
- 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.)
- Active
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a portable online bird nest intelligent identification method and a portable online bird nest intelligent identification system, wherein the system adopts an embedded software and hardware design scheme, is small and portable, can carry out high-definition shooting on a bird nest of a railway line contact net support column and realize online intelligent inspection only by temporarily erecting the bird nest on a locomotive driving platform, and can accurately position and alarm the detected bird nest in time, thereby effectively finding the bird nest on the railway line contact net support column.
Description
Technical Field
The invention relates to the field of contact network power supply safety detection, in particular to a portable online bird nest intelligent identification method and system.
Background
Bird's nest is patrolled and examined to be the important work content of contact net power supply safety inspection, birds nest on the pillar along the railway line or on transmission line shaft tower can cause harmful effects to equipment such as contact net, harm the safe operation of train even, to the bird's nest that exists, generally carry out bird's nest to detect through artifical the mode of patrolling and examining or off-line data analysis, but these traditional modes not only are consuming time and wasting power, can not in time get rid of the potential safety hazard, bring very big puzzlement for railway workers, consequently, the urgent need for a bird's nest of online, in time, intelligence patrols and examines method and system.
Disclosure of Invention
Aiming at the technical problems in the background art, the invention provides a portable online intelligent bird nest identification method, which specifically comprises the following steps:
s1: acquiring a real-time picture;
s2: identifying the real-time picture by utilizing a deep learning network model and giving an identification result;
s3: if the identification result indicates that the bird nest exists, early warning is carried out; otherwise, step S1 is executed or the identification method is ended.
Preferably, the step S2 of the intelligent bird nest identification method further includes a second confirmation of the identification result, and the identification result is updated according to the second confirmation result.
Preferably, the secondary confirmation of the identification result, and the specific method for updating the identification result according to the secondary confirmation result, comprises:
s21: carrying out prior characteristic detection based on the bird nest structural features; if the prior characteristic detection result is judged to be a non-bird nest, updating the identification result, and skipping to the step S3; otherwise, directly jumping to the step S22;
s22: and detecting a large target attached to the bird nest by adopting a multi-target detection algorithm based on deep learning, if the large target does not exist, updating the identification result, and skipping to the step S3, otherwise, directly skipping to the step S3.
Preferably, the prior characteristic detection rule of S21 includes one or more of the following rules:
1) Judging whether the length-width ratio of the bird nest is in a set range, if so, determining the bird nest;
2) Judging whether the height of the bird nest from the ground is within a set range, if so, determining that the bird nest is the bird nest;
3) And judging whether the size of the bird nest is within a set range, and if so, determining that the bird nest is the bird nest.
Preferably, the specific implementation method of S22 is as follows:
s221: establishing a training sample library: classifying and marking the images containing the bird nests according to large targets attached to the bird nests; the large target comprises a tree, a signal transmitting tower, a power line pole and a contact net support;
s222: sample training: carrying out deep learning network training by using the marked target data;
s223: and (4) detecting a result: carrying out large target detection aiming at the real-time picture;
if the contact net support is not detected in the image, the image is regarded as a non-contact net support bird nest, and the identification result is updated;
if the contact network strut is detected, whether the bird nest position and the strut position are in a certain range is judged, and if yes, the contact network strut is considered to have a bird nest.
Preferably, the specific implementation method for performing early warning in S3 is as follows:
and (4) determining the accurate position information of the bird nest by combining the GPS positioning information and the one-pole one-gear basic database, and outputting alarm information.
Preferably, the alarm information is transmitted to the mobile display device by wireless.
Preferably, the specific construction method of the deep learning network in S2 is as follows:
1) Constructing the deep learning network model, wherein the deep learning network model comprises a convolution layer, a pooling layer, a full-link layer and a classifier layer;
2) Calculating and extracting local features of a bird nest in the image through the convolutional layer;
3) Sampling the local features of the image through the pooling layer, and reducing the resolution of the image to the maximum extent;
4) Extracting the features layer by layer through the full connection layer;
5) And predicting the category probability and the precise position of the bird nest through the classifier layer, and defining a classifier loss function, wherein the classifier loss function comprises category loss and position loss.
Meanwhile, the invention also provides a portable bird nest inspection system which comprises a data acquisition module, a central processing module and a mobile display device; the data acquisition module, the central processing module and the mobile display equipment are sequentially connected;
the data acquisition module comprises a high-definition image acquisition unit and a positioning module;
the central processing module comprises a data compression storage unit, a core control panel, a power management unit, an acquisition module control unit, an abnormity alarm unit, an intelligent bird nest detection unit and a database; the data compression storage unit, the power management unit, the acquisition module control unit, the abnormity alarm unit, the bird nest intelligent detection unit and the database are all connected with the core control panel;
and the abnormity alarm unit is connected with the mobile display equipment.
The invention has the beneficial effects that: use portable online bird's nest intelligence system of patrolling and examining to replace traditional artifical mode of patrolling and examining to carry out bird's nest defect detection, can discover in real time the bird's nest on the contact net pillar and pinpoint to railway worker in time carries out the hidden danger and clears away, guarantees that railway power supply safety practices thrift manpower and time cost simultaneously, improves work efficiency. The system is based on an embedded software and hardware design scheme, so the system is small and portable, can shoot bird nests of pillars along the railway and can intelligently patrol only by temporarily erecting the system on a locomotive driving platform, and can accurately position the detected bird nests and timely alarm.
Drawings
FIG. 1 is a schematic structural diagram of a portable online intelligent bird nest identification system;
FIG. 2 is a schematic structural diagram of a deep learning network in the portable online intelligent bird nest identification method;
fig. 3 is a flow chart of a portable online bird nest identification method.
Detailed Description
In order to more clearly understand the technical features, objects, and effects of the present invention, embodiments of the present invention will now be described with reference to the accompanying drawings.
As shown in fig. 1, the portable online bird nest inspection system comprises a data acquisition module, a central processing module and a mobile display device. The data acquisition module, the central processing module and the mobile display device are sequentially connected.
The data acquisition module comprises a high-definition image acquisition unit and a positioning module. The high-definition image acquisition unit adopts an industrial camera to acquire images in real time. The high-speed industrial camera with proper resolution (2048 x 1024) and the lens with proper focal length (25 mm) are selected for the camera, high-frame-rate shooting is carried out on contact net support posts on two sides of a railway at proper angles, and in order to meet the requirement of complete coverage shooting, the acquisition frame rate cannot be lower than 40 frames/second.
The central processing module comprises a data compression storage unit, a core control panel, a power management unit, an acquisition module control unit, an abnormity alarm unit, a bird nest intelligent detection unit and a database, the image data acquired by the camera is stored in a limited storage space, in order to realize high-efficiency compression operation capacity, the data compression storage unit adopts a JPEG image compression chip based on FPGA, extremely high compression capacity and power consumption energy efficiency ratio can be obtained, the positioning unit adopts GPS/Beidou dual-mode positioning to realize accurate positioning, the panoramic image acquisition part is matched with the image positioning unit to realize positioning of the geographic position of the image, the data compression storage unit adopts four Didi Zhizun high-speed series-3D version 500G solid state hard disks to form a 2TB storage space for storing the data output by the data compression storage unit and the corresponding position information from the image positioning unit, the core control panel is used for controlling data transmission in the data acquisition module, and mainly performs three operations: 1) Directly transmitting output data of a JPEG compression unit and position information obtained by an image positioning unit to a data storage unit; 2) Synchronously transmitting the original image data obtained by the input end of a JPEG compression engine and the position information obtained by an image positioning unit to a central processing module; 3) And synchronously transmitting the raw image data obtained by the input end of the JPEG compression engine to the mobile display equipment.
The central processing module adopts an NVIDIA TX2 core module, a built-in 6-core CPU and a 256-core GPU are AI single-module supercomputers based on an NVIDIA Pascal framework, powerful computing power can be provided, the appearance is small and exquisite, energy is saved and high efficiency can be realized, a large and complex deep neural network can be cast, the GPU in the Tx2 can be used for carrying out intelligent bird nest detection in real time while collecting, and the central processing module mainly carries out five tasks, which are respectively:
a. analyzing the image by using a deep learning method to obtain the specific position of the bird nest, wherein the intelligent bird nest identification method specifically comprises the following steps:
s1: acquiring a real-time picture; s2: identifying the real-time picture by utilizing a deep learning network model and giving an identification result; s3: if the bird nest exists in the identification result, early warning is carried out; otherwise, step S1 is executed or the identification method is ended. Preferably, step S2 of the intelligent bird nest identification method further includes a secondary confirmation of the identification result, and the identification result is updated according to the secondary confirmation result. Preferably, the secondary confirmation of the identification result, and the specific method for updating the identification result according to the secondary confirmation result, comprises: s21: carrying out prior characteristic detection based on bird nest structural features; if the prior characteristic detection result is judged to be a non-bird nest, updating the identification result, and skipping to the step S3; otherwise, directly jumping to the step S22; s22: and detecting a large target attached to the bird nest by adopting a multi-target detection algorithm based on deep learning, if the large target does not exist, updating the identification result, and skipping to the step S3, otherwise, directly skipping to the step S3. Preferably, the prior characteristic detection rule of S21 includes one or more of the following rules: 1) Judging whether the length-width ratio of the bird nest is in a set range, if so, determining the bird nest; 2) Judging whether the height of the bird nest from the ground is within a set range, if so, determining that the bird nest is the bird nest; 3) And judging whether the size of the bird nest is within a set range, and if so, determining the bird nest. Preferably, the specific implementation method of S22 is as follows: s221: establishing a training sample library: classifying and marking the images containing the bird nests according to large targets attached to the bird nests; the large target comprises a tree, a signal transmitting tower, a power line pole and a contact net support; s222: sample training: carrying out deep learning network training by using the marked target data; s223: and (4) detecting a result: carrying out large target detection aiming at the real-time picture; if the contact net support is not detected in the image, the image is regarded as a non-contact net support bird nest, and the identification result is updated; if the contact network strut is detected, whether the bird nest position and the strut position are in a certain range is judged, and if yes, the contact network strut is considered to have a bird nest. Preferably, the specific implementation method for performing early warning in S3 is as follows: and (4) determining the accurate position information of the bird nest by combining the GPS positioning information and the one-pole one-gear basic database, and outputting alarm information. Preferably, the alarm information is transmitted to the mobile display device by wireless. Preferably, the specific construction method of the deep learning network in S2 is shown in fig. 2, and specifically includes: 1) Constructing the deep learning network model, wherein the deep learning network model comprises a convolution layer, a pooling layer, a full-link layer and a classifier layer; 2) Calculating and extracting local bird nest features in the image through the convolutional layer; 3) Sampling the local features of the image through the pooling layer, and reducing the resolution of the image to the maximum extent; 4) Extracting the features layer by layer through the full connection layer; 5) And predicting the category probability and the accurate position of the bird nest through the classifier layer, and defining a classifier loss function, wherein the classifier loss function comprises category loss and position loss.
b. The method comprises the steps that images of bird nests which are intelligently identified and corresponding positioning information are transmitted to a bird nest statistical database, the bird nest statistical database is used for storing the images containing the bird nests and the positioning information corresponding to the images, and can generate bird nest early warning reports according to the stored information after inspection is finished, and the bird nest early warning reports are displayed to workers through mobile display equipment for the workers to check and overhaul;
c. the image of the bird nest intelligently identified and the corresponding positioning information are transmitted to a mobile display device, and an abnormal alarm is given;
d. controlling the normal operation of the data acquisition module;
e. and controlling the power supply module to normally supply power.
The power supply module adopts a lithium battery to supply power for the operation of the system.
The mobile display device mainly performs four tasks: 1) Displaying the image information acquired by the data acquisition module in real time; 2) Displaying images and corresponding positioning information of the bird nests which are intelligently identified in real time, and sending bird nest alarm information; 3) A high-definition image acquisition unit control interface is arranged on a display interface, a user can adjust the working process and related parameters of the panoramic camera through the interface, and the mobile display equipment transmits related control information to the central processing module to realize the control of the data acquisition module; 4) After the inspection is finished, the bird nest early warning report is displayed for the working personnel to check.
Data acquisition module, central processing module integrated installation are in high strength machine incasement, the machine case erects on the overhead device, and the overhead device adopts sucking disc, tripod cooperation cloud platform to provide nimble changeable erects position and angle selection, and horizontal rotation angle can be supported to its cloud platform system: 0-360 DEG, vertical rotation angle: the angle is-75 to +40 degrees, a good shooting angle is fully ensured when the contact net support rod and the surrounding environment are subjected to image acquisition, and the system is ensured to be in a stable state when data are acquired.
The mobile display device realizes data transmission with the data acquisition module and the central processing module through Wi-Fi.
The embedded portable bird nest inspection system adopts an embedded software and hardware design scheme, is small and portable, only needs to be temporarily erected on a locomotive cab, can carry out high-definition shooting and intelligent inspection on the bird nest of the railway line contact net pillar, accurately positions the detected bird nest and timely alarms, and can effectively find the bird nest on the railway line contact net pillar.
As shown in fig. 3, regarding the bird nest identification algorithm, firstly, a deep learning algorithm is used to perform preliminary bird nest detection on the collected image data; and secondly, secondarily confirming the primarily detected bird nest by combining an interference filtering algorithm, and outputting a final bird nest detection result. The deep convolutional network used by the bird nest detection algorithm based on deep learning is a binary neural network. Namely: all network weights use binarization approximation, a convolutional neural network with binarization weights is significantly smaller than a non-binary standard deep convolutional network, and when the weight values are binary, the convolutional operation is estimated only by addition and subtraction without multiplication, thereby resulting in accelerated operation. The image data collected by the system is subjected to bird nest detection by using the trained deep learning network model, the image data detected to contain the bird nest is marked, and the position information of the bird nest (coordinates (x, y) of the center point of the bird nest in the image and the length and width of a rectangular frame formed by the range of the bird nest) is recorded.
However, in the detection result, a small number of non-bird nest targets similar to the bird nest characteristics exist (false detection), and all the detected bird nests are not safe for railway catenary power supply, so that the bird nests on the true catenary pillars must be confirmed among the detected bird nests, and the irrelevant bird nests on non-railway power supply devices such as power poles and signal transmission towers are filtered out. Therefore, after the bird nest is preliminarily detected based on the deep learning algorithm, a set of interference filtering algorithm is designed.
The filtering algorithm mainly filters false detection targets and irrelevant normal bird nests of the preliminary detection of the deep learning algorithm, and retains the bird nest defects on the contact net support columns which really affect the railway power supply safety.
The filtering algorithm is divided into 2-layer filtering.
a. Filtering by using a priori characteristics based on the bird nest structure. The layer of filtering is mainly used for filtering out some non-bird nest objects.
(1) Setting the integral proportion of the bird nest according to the length-width ratio of the rectangular frame, regarding the target with the length-width ratio exceeding a certain threshold as not being the bird nest, and filtering; (2) Bird nests generally do not appear near ground level, so if the detected bird nest location is within a certain threshold of the lower edge of the image, it is considered an interfering target and filtered; (3) The size of the bird nest is generally within a fixed range, and is not too small or too large. Therefore, the detected bird nest is considered to be an interference target and filtered if the length and the width of the rectangular frame to which the bird nest belongs are respectively smaller than a certain fixed threshold value or larger than a certain fixed threshold value. (4) And (4) performing deduplication operation, wherein only one frame of detection result with the maximum confidence coefficient is reserved in algorithm design when the same bird nest target appears in the continuous n frames.
b. The areas where the bird nests are located in the analysis image are basically located at treetops, signal transmission towers, power line poles and contact net posts. The four types of target objects have obvious characteristics, so that the invention uses a multi-target detection algorithm based on deep learning to simultaneously detect whether the four types of target objects are contained in the image and the specific positions of the target objects.
Establishing a training sample library: and (4) reserving a nest image after a result detected by the nest is subjected to a first filtering algorithm. These images containing the bird's nest are classified and labeled according to the large target (tree, signal transmission tower, power line pole and overhead line system pillar) to which the bird's nest is attached.
Sample training: using the marked 4 types of target data to perform deep learning network training,
and (3) detection results: if the contact net support is not detected in the image, other three types are detected. This nest is considered not to be a defective nest on the strut of the catenary. Filtering; if the contact net pillar is detected, the position of the pillar is continuously corrected with the previously detected position of the bird nest, and if the position of the bird nest is within a certain range of the pillar position, the bird nest is regarded as a contact net pillar bird nest defect. At the moment, the frame of image is reserved, and the accurate position of the bird nest is positioned according to the longitude and latitude at the moment and by combining a rod-file database, so that an alarm is generated.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.
Claims (8)
1. A portable online intelligent bird nest identification method is characterized by specifically comprising the following steps:
s1: acquiring a real-time picture;
s2: identifying the real-time picture by utilizing a deep learning network model and giving an identification result;
s3: if the identification result indicates that the bird nest exists, early warning is carried out; otherwise, executing step S1 or ending the identification method;
step S2 of the intelligent bird nest identification method further comprises secondary confirmation of the identification result, and the identification result is updated according to the secondary confirmation result;
the secondary confirmation of the identification result, according to the secondary confirmation result, the specific method for updating the identification result comprises the following steps:
s21: carrying out prior characteristic detection based on bird nest structural features; if the prior characteristic detection result is judged to be a non-bird nest, updating the identification result, and skipping to the step S3; otherwise, directly jumping to the step S22;
s22: and detecting a large target attached to the bird nest by adopting a multi-target detection algorithm based on deep learning, if the large target does not exist, updating an identification result, and skipping to the step S3, otherwise, directly skipping to the step S3.
2. The portable online intelligent bird nest identification method of claim 1, wherein the prior characteristic detection rule of S21 includes one or more of the following rules:
1) Judging whether the length-width ratio of the bird nest is within a set range, if so, determining the bird nest;
2) Judging whether the height of the bird nest from the ground is within a set range, if so, determining that the bird nest is the bird nest;
3) And judging whether the size of the bird nest is within a set range, and if so, determining the bird nest.
3. The portable online intelligent bird nest identification method according to claim 1 or 2, characterized in that the specific implementation method of S22 is:
s221: establishing a training sample library: classifying and marking the images containing the bird nests according to large targets attached to the bird nests; the large target comprises a tree, a signal transmitting tower, a power line pole and a contact net support;
s222: sample training: carrying out deep learning network training by using the marked target data;
s223: and (4) detecting a result: carrying out large target detection aiming at the real-time picture;
if the contact net support is not detected in the image, the image is regarded as a non-contact net support bird nest, and the identification result is updated;
if the contact net support is detected, whether the position of the bird nest and the position of the support are within a certain range or not is judged, and if yes, the contact net support is considered to have the bird nest.
4. The portable online intelligent bird nest identification method of claim 3, wherein the specific implementation method for performing early warning in S3 is as follows:
and (4) determining the accurate position information of the bird nest by combining the GPS positioning information and the one-pole one-gear basic database, and outputting alarm information.
5. The portable online intelligent bird nest identification method of claim 4, wherein the alarm information is transmitted to a mobile display device by wireless.
6. The portable online intelligent bird nest identification method of claim 5, wherein the deep learning network model is a binary weight network.
7. The embedded portable online bird nest intelligent identification system is applied to the portable online bird nest intelligent identification method according to claim 1, and is characterized in that the embedded portable bird nest inspection system comprises a portable online bird nest inspection system, a central processing module and a mobile display device, wherein the data acquisition module, the central processing module and the mobile display device are connected with the data acquisition module; the data acquisition module, the central processing module and the mobile display equipment are sequentially connected;
the data acquisition module comprises a high-definition image acquisition unit and a positioning module;
the central processing module comprises a data compression storage unit, a core control panel, a power management unit, an acquisition module control unit, an abnormity alarm unit, an intelligent bird nest detection unit and a database; the data compression storage unit, the power management unit, the acquisition module control unit, the abnormity alarm unit, the bird nest intelligent detection unit and the database are all connected with the core control panel;
and the abnormity alarm unit is connected with the mobile display equipment.
8. The portable online intelligent bird nest identification system of claim 7, wherein the central processing module transmits the image of the bird nest identified intelligently and the corresponding positioning information to the bird nest statistical database, and can generate a bird nest early warning report according to the information stored in the bird nest statistical database after the inspection is finished, and the bird nest early warning report is displayed to the staff through the mobile display device for the staff to check and overhaul; the central processing module transmits the image of the bird nest intelligently identified and the corresponding positioning information to the mobile display equipment for carrying out abnormity alarm;
the data acquisition module and the central processing module are integrally installed in the high-strength shell case, the case is erected on the erecting device, and the erecting device adopts a sucker and a tripod to match with a tripod head to provide flexibly variable erecting positions and angle selection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910176773.8A CN109902647B (en) | 2019-03-08 | 2019-03-08 | Portable online bird nest intelligent identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910176773.8A CN109902647B (en) | 2019-03-08 | 2019-03-08 | Portable online bird nest intelligent identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109902647A CN109902647A (en) | 2019-06-18 |
| CN109902647B true CN109902647B (en) | 2023-04-18 |
Family
ID=66946855
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910176773.8A Active CN109902647B (en) | 2019-03-08 | 2019-03-08 | Portable online bird nest intelligent identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109902647B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112715427B (en) * | 2020-12-30 | 2022-12-09 | 青海保绿丰生态农林科技有限公司 | Automatic monitoring system for nesting, laying number, brooding and predation habit rules of eagle |
| CN114913473B (en) * | 2022-03-21 | 2023-08-15 | 中国科学院光电技术研究所 | Lightweight monomer type imaging contact net safety inspection instrument |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103745224A (en) * | 2013-12-24 | 2014-04-23 | 西南交通大学 | Image-based railway contact net bird-nest abnormal condition detection method |
| CN106250835A (en) * | 2016-07-26 | 2016-12-21 | 国网福建省电力有限公司 | Bird's Nest recognition methods on the transmission line of electricity of feature based identification |
| CN106971152A (en) * | 2017-03-16 | 2017-07-21 | 天津大学 | A kind of method of Bird's Nest in detection transmission line of electricity based on Aerial Images |
| CN108537154A (en) * | 2018-03-28 | 2018-09-14 | 天津大学 | Transmission line of electricity Bird's Nest recognition methods based on HOG features and machine learning |
| CN108764020A (en) * | 2018-03-30 | 2018-11-06 | 广东工业大学 | A kind of Bird's Nest recognition methods on high tension electric tower based on unmanned plane image |
| 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 |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101709954B (en) * | 2009-12-11 | 2011-11-16 | 北京航空航天大学 | Overhead wire icing on-line monitoring device based on machine vision |
| CN103530620A (en) * | 2013-10-30 | 2014-01-22 | 华北电力大学 | Method for identifying bird nest on electric transmission line tower |
| CN105208347B (en) * | 2015-10-08 | 2018-08-14 | 成都时代星光科技有限公司 | The aerial unmanned plane of railway line is patrolled and real time image collection transmitting, monitoring device automatically |
| CN105389555A (en) * | 2015-11-06 | 2016-03-09 | 天津航天中为数据系统科技有限公司 | Hidden danger target analysis method and system based on aerial image |
| EP3376213A1 (en) * | 2017-03-15 | 2018-09-19 | Siemens Aktiengesellschaft | Method and assembly for monitoring the state of a system with operating elements |
| CN106980827A (en) * | 2017-03-16 | 2017-07-25 | 天津大学 | A kind of method of Bird's Nest in identification transmission line of electricity based on Aerial Images |
| CN107665336A (en) * | 2017-09-20 | 2018-02-06 | 厦门理工学院 | Multi-target detection method based on Faster RCNN in intelligent refrigerator |
| CN108037133B (en) * | 2017-12-27 | 2021-01-26 | 武汉市智勤创亿信息技术股份有限公司 | Intelligent electric power equipment defect identification method and system based on unmanned aerial vehicle inspection image |
| CN108681733A (en) * | 2018-05-15 | 2018-10-19 | 苏州华眸智能科技有限公司 | A kind of pantograph abnormal state detecting system and detection method based on image procossing |
-
2019
- 2019-03-08 CN CN201910176773.8A patent/CN109902647B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103745224A (en) * | 2013-12-24 | 2014-04-23 | 西南交通大学 | Image-based railway contact net bird-nest abnormal condition detection method |
| CN106250835A (en) * | 2016-07-26 | 2016-12-21 | 国网福建省电力有限公司 | Bird's Nest recognition methods on the transmission line of electricity of feature based identification |
| CN106971152A (en) * | 2017-03-16 | 2017-07-21 | 天津大学 | A kind of method of Bird's Nest in detection transmission line of electricity based on Aerial Images |
| CN108537154A (en) * | 2018-03-28 | 2018-09-14 | 天津大学 | Transmission line of electricity Bird's Nest recognition methods based on HOG features and machine learning |
| CN108764020A (en) * | 2018-03-30 | 2018-11-06 | 广东工业大学 | A kind of Bird's Nest recognition methods on high tension electric tower based on unmanned plane image |
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109902647A (en) | 2019-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112633535B (en) | Intelligent inspection method and system for photovoltaic power station based on unmanned aerial vehicle image | |
| CN108037770B (en) | Unmanned aerial vehicle power transmission line inspection system and method based on artificial intelligence | |
| CN111582234B (en) | Large-scale oil tea tree forest fruit intelligent detection and counting method based on UAV and deep learning | |
| CN112115927B (en) | Intelligent machine room equipment identification method and system based on deep learning | |
| CN108710126A (en) | Automation detection expulsion goal approach and its system | |
| CN104581076A (en) | Mountain fire monitoring and recognizing method and device based on 360-degree panoramic infrared fisheye camera | |
| CN109711348A (en) | Intelligent monitoring method and system based on the long-term real-time architecture against regulations in hollow panel | |
| CN112101088A (en) | Automatic unmanned aerial vehicle power inspection method, device and system | |
| CN112381356A (en) | Completion acceptance method, completion acceptance system, server and storage medium for engineering project | |
| CN114240868A (en) | Unmanned aerial vehicle-based inspection analysis system and method | |
| CN109902647B (en) | Portable online bird nest intelligent identification method and system | |
| CN112863100B (en) | Intelligent construction safety monitoring system and method | |
| CN108688805A (en) | The unmanned plane detected safely applied to restricted clearance | |
| CN111561943A (en) | Robot inspection method and system | |
| CN116258980A (en) | Unmanned aerial vehicle distributed photovoltaic power station inspection method based on vision | |
| CN116846059A (en) | Edge detection system for power grid inspection and monitoring | |
| CN117292318A (en) | Power transmission line safety monitoring and early warning method and system based on computer vision | |
| CN210511074U (en) | Full-intelligent oil-gas pipeline inspection system based on unmanned aerial vehicle remote sensing and deep learning | |
| CN112529836A (en) | High-voltage line defect detection method and device, storage medium and electronic equipment | |
| Zheng et al. | Forest farm fire drone monitoring system based on deep learning and unmanned aerial vehicle imagery | |
| CN114663818B (en) | Airport operation core area monitoring and early warning system and method based on vision self-supervision learning | |
| CN115912183A (en) | High-voltage transmission line ecological measure inspection method and system and readable storage medium | |
| CN116243725A (en) | Substation unmanned aerial vehicle inspection method and system based on visual navigation | |
| CN115187880A (en) | Communication optical cable defect detection method and system based on image recognition and storage medium | |
| CN114898361A (en) | Peach orchard fruit state identification and counting method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |