CN111723774A - Target identification method for power transmission equipment based on unmanned aerial vehicle inspection - Google Patents
Target identification method for power transmission equipment based on unmanned aerial vehicle inspection Download PDFInfo
- Publication number
- CN111723774A CN111723774A CN202010629415.0A CN202010629415A CN111723774A CN 111723774 A CN111723774 A CN 111723774A CN 202010629415 A CN202010629415 A CN 202010629415A CN 111723774 A CN111723774 A CN 111723774A
- Authority
- CN
- China
- Prior art keywords
- target
- aerial vehicle
- unmanned aerial
- power transmission
- image
- 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
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Multimedia (AREA)
- Probability & Statistics with Applications (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a target identification method of power transmission equipment based on unmanned aerial vehicle inspection, which comprises the steps of firstly collecting image data through an unmanned aerial vehicle, extracting an image of a target to be identified in the image data and establishing a target sample set, then manually marking the image in the target sample set, establishing a defect sample library of an inspection image, then adopting the defect sample library of the inspection image as a training sample, establishing a defect calibration model based on a Faster-RCNN network, then adopting the unmanned aerial vehicle to collect real-time inspection image data, inputting the real-time inspection image data into the defect calibration model, and finally outputting a target identification result and carrying out structured storage on the target identification result; the unmanned aerial vehicle inspection system can carry out unified intelligent management on massive unmanned aerial vehicle inspection images, improves the automatic image classification capability of a pole tower level and a hardware fitting level, improves the accuracy of analysis, and greatly reduces the workload of manual analysis processing.
Description
Technical Field
The invention relates to the technical field of unmanned aerial vehicle routing inspection and power transmission equipment maintenance, in particular to a power transmission equipment target identification method based on unmanned aerial vehicle routing inspection.
Background
The power transmission line inspection is to regularly inspect the power transmission line and pole tower equipment, and timely record and eliminate defects found in the inspection process. The power transmission line inspection can find hidden dangers and faults in the operation process of equipment in advance, and timely process the defects, so that the power grid faults are effectively prevented, and the threats to the personal safety of residents, property safety and industrial generation are reduced.
At present, the power transmission line inspection usually adopts a manual inspection mode, so that a large amount of manpower is consumed, the efficiency is low, and the accuracy is not high. Along with unmanned aerial vehicle's application, unmanned aerial vehicle patrols and examines the proportion of operation in the transmission line daily work of patrolling and examining and increase day by day, and the circuit coverage also can promote fast, but the machine patrols the work of image recognition processing and data management but obvious lags behind, and the later stage work of patrolling and examining has brought a large amount of obstacles.
Disclosure of Invention
The invention aims to provide a target identification method of power transmission equipment based on unmanned aerial vehicle inspection, which can perform unified intelligent management on massive unmanned aerial vehicle inspection images, improve the automatic image classification capability of a pole tower level and a hardware level, improve the accuracy of analysis and greatly reduce the workload of manual analysis processing.
The technical scheme adopted by the invention is as follows:
a power transmission equipment target identification method based on unmanned aerial vehicle routing inspection comprises the following steps:
A. the method comprises the steps of shooting image data of an area where power transmission equipment is located by adopting unmanned aerial vehicle routing inspection, extracting an image of a target to be identified in the image data and establishing a target sample set;
classifying, recombining, positioning and randomly renaming image data shot by the unmanned aerial vehicle in a polling way, and converting randomly-named and disordered images into structured data;
the target to be recognized comprises N targets, and the image of the target to be recognized i (i is less than or equal to N) is collected to form a sample subset AiSample subset a1 to sample subset aNForming a target sample set;
B. manually marking the images in the target sample set, and creating a defect sample library of the inspection images;
C. adopting a defect sample library of the inspection image as a training sample, and creating a defect calibration model based on a Faster-RCNN network;
detecting a target to be identified in a training sample by adopting a training model based on a Faster-RCNN network, training by utilizing a neural network algorithm, and then establishing a defect calibration model by utilizing a YOLOV3 algorithm;
D. collecting real-time polling image data by using an unmanned aerial vehicle, and inputting the real-time polling image data into a defect calibration model;
E. and outputting a target identification result by the defect calibration model and performing structured storage on the target identification result.
Further, the specific process of manually labeling the images in the target sample set in the step B is as follows: firstly, labeling the content of the region in the image, and labeling the hidden danger and the defect found in the region.
Furthermore, the manual marking process can calibrate the defects in the image in a frame selection mode, the image can be zoomed and dragged in the calibration process, the modification, deletion and addition of the calibration frame are supported, and meanwhile, the defect information is structurally stored and displayed; the defect information includes the type and urgency of the defect.
Further, the specific process of detecting the target to be recognized in the training sample by the training model in the step C is as follows:
c1, determining a candidate frame of the training sample by using the candidate frame selection network RPN; candidate in-frame images include three cases: the system comprises power transmission equipment, non-power transmission equipment, power transmission equipment and non-power transmission equipment;
c2, feature mapping; extracting features of the image in the candidate frame by using a convolutional neural network;
c3, classifying the extracted features by adopting a classifier, and judging whether the images in the candidate frame contain the power transmission equipment or not;
c4, refining the area, and enabling the edge of the candidate frame determined as the equipment to tightly wrap the power transmission equipment, so that the area of the candidate frame is enough to accommodate the contained power transmission equipment and the area is minimum.
Further, the classifier in the step c3 adopts a multi-classifier softmax.
The invention has the following beneficial effects:
(1) target identification is carried out based on unmanned aerial vehicle routing inspection data analysis, discrete and disordered image data can be subjected to uniform and ordered standardized management, image renaming is automatically carried out according to an unmanned aerial vehicle routing inspection task, the routing inspection image is guaranteed to have use value and historical traceability, unmanned aerial vehicle routing inspection standardized operation is realized, and the unmanned aerial vehicle routing inspection standardized operation level is improved;
(2) through intelligent operations such as the analysis, cluster, discernment, filing of patrolling and examining the data to unmanned aerial vehicle, the artifical work load that reducible interior image was handled is at least 30%, and the work load of greatly reduced personnel analysis processes, and reduces artifical discernment and handles wrong report rate and missing report rate, improves the accuracy and the validity of analysis, improves unmanned aerial vehicle and patrols and examines work efficiency, benefit, promotes the coverage of patrolling and examining.
Detailed Description
The invention discloses a power transmission equipment target identification method based on unmanned aerial vehicle routing inspection, which comprises the following steps:
A. the method comprises the steps of shooting image data of an area where power transmission equipment is located by adopting unmanned aerial vehicle routing inspection, extracting an image of a target to be identified in the image data and establishing a target sample set;
classifying, recombining, positioning and randomly renaming image data shot by the unmanned aerial vehicle in a polling way, and converting randomly-named and disordered images into structured data;
the target to be recognized comprises N targets, the image of the target to be recognized i (i is less than or equal to N) is collected to form a sample subset Ai, and the sample subset A1To the sample subset ANForming a target sample set;
B. manually marking the images in the target sample set, and creating a defect sample library of the inspection images;
C. adopting a defect sample library of the inspection image as a training sample, and creating a defect calibration model based on a Faster-RCNN network;
detecting a target to be identified in a training sample by adopting a training model based on a Faster-RCNN network, training by utilizing a neural network algorithm, and then establishing a defect calibration model by utilizing a YOLOV3 algorithm;
D. collecting real-time polling image data by using an unmanned aerial vehicle, and inputting the real-time polling image data into a defect calibration model;
E. and outputting a target identification result by the defect calibration model and performing structured storage on the target identification result.
For a better understanding of the present invention, the following embodiments are provided to further explain the technical solutions of the present invention.
A power transmission equipment target identification method based on unmanned aerial vehicle routing inspection comprises the following steps:
A. and adopting an unmanned aerial vehicle to inspect and shoot image data of the area where the power transmission equipment is located, extracting an image of a target to be identified in the image data and establishing a target sample set.
The image data shot by the unmanned aerial vehicle inspection is classified, recombined, positioned and randomly renamed, and randomly named and disordered images are converted into structured data.
The target to be recognized comprises N targets, and the image of the target to be recognized i (i is less than or equal to N) is collected to form a sample subset AiSample subset A1To the sample subset ANA set of target samples is constructed.
The target to be identified in the power transmission line inspection comprises multiple categories such as a tower bird nest, a vibration damper, pins and the like, the image data of each category forms a corresponding sample subset, and the sample subset engineering of all the categories of the target to be identified forms a target sample set. When the identification model is established, a corresponding database is established for each classified target to be identified.
B. And manually marking the images in the target sample set, and creating a defect sample library of the inspection images.
The specific process of manually labeling the images in the target sample set comprises the following steps: firstly, labeling the content of the region in the image, and labeling the hidden danger and the defect found in the region.
The defects in the image can be calibrated in a frame selection mode in the manual marking process, the image can be zoomed and dragged in the calibration process, modification, deletion and addition of the calibration frame are supported, and meanwhile, the defect information is structurally stored and displayed; the defect information includes the type and urgency of the defect.
In order to improve the accuracy and comprehensiveness of the defect calibration model training sample, the following principles need to be followed in the manual labeling process:
(1) marking the whole and the components of all targets to be identified appearing in the image, wherein the marks cannot be missed; for example: all nut areas appearing in the pin area need to be marked, and label omission cannot be realized.
(2) Marking whether the target to be recognized can be seen clearly or not as long as the target to be recognized exists in the image; for example, the position of the installation pin in the image needs to be marked when the pin can be clearly displayed, and the pin cannot be completely displayed due to occlusion or the like.
(3) For the area with a large background and a complex background, due to the existence of excessive interference items, the target to be recognized cannot be clearly distinguished, so that the marking is not needed.
(4) And (4) the complete framing of the target to be identified in the framing process is required, and the interference item cannot be framed in the framing process.
(5) The area of the calibration frame should be as small as possible on the premise that the target to be identified can be completely wrapped.
C. Adopting a defect sample library of the inspection image as a training sample, and creating a defect calibration model based on a Faster-RCNN network; detecting the target to be recognized in the training sample by adopting a training model based on a Faster-RCNN network, training by utilizing a neural network algorithm, and then establishing a defect calibration model by utilizing a YOLOV3 algorithm.
Compared with the fast-RCNN, the SSD only has a main neural network (base network), the network of the RPN is removed, feature maps of various scales are extracted from the middle, small convolution layers are added on the side faces to predict the position and the type of an object, and all the positions and the types are integrated to output a final result. SSD operates Faster because it eliminates a network, when some special target detections are to be verified compared to fast-RCNN.
The specific process of detecting the target to be recognized in the training sample by the training model is as follows:
c1, determining a candidate frame of the training sample by using the candidate frame selection network RPN; candidate in-frame images include three cases: the power transmission equipment comprises power transmission equipment, non-power transmission equipment, power transmission equipment and non-power transmission equipment.
Because the position and the scale of the equipment in the image are unknown, if a smooth window traversal method is adopted, multiple scales need to be adopted to translate in different step lengths, a huge number of windows to be judged are generated, and further, the calculation amount is complicated. While the RPN method or Selective Search method may produce only a few most likely candidate boxes so that the entire recognition process can be completed quickly. In contrast, the RPN method is faster to compute than the Selective Search method, and the candidate box Search is more accurate.
c2, feature mapping; the convolutional neural network is used for extracting the features of the images in the candidate frames, is the most basic classification network for deep learning, and is excellent in classification performance and feature extraction accuracy.
c3, classifying the extracted features by adopting a multi-classifier softmax, and judging whether the images in the candidate frame contain the power transmission equipment.
Softmax is a multi-classifier, and is different from a mode that two classifiers such as logistic regression need to be judged one by one (namely, whether the equipment A is judged, whether the equipment B is judged, and the like), Softmax can finish classification once, and both the calculation performance and the classification precision are greatly improved.
c4, refining the area, so that the edges of the candidate frames determined as the equipment tightly wrap the power transmission equipment, and the area of the candidate frames is enough to accommodate the contained power transmission equipment and is the minimum; since the original candidate region may only contain most of the device or other background parts besides the device, the accuracy of the candidate region can be improved through region refinement.
D. And acquiring real-time inspection image data by adopting an unmanned aerial vehicle, and inputting the real-time inspection image data into a defect calibration model.
E. And outputting a target identification result by the defect calibration model and performing structured storage on the target identification result.
At present, the main targets for power transmission line inspection include power transmission lines, pole tower equipment and the like, and the following are partial experimental data for unmanned aerial vehicle inspection of power supply companies in Shandong electric power sunshine cities of China network:
identifying a tower bird nest:
and (3) training by using a deep learning image recognition positioning algorithm, namely fast-RCNN, using a sample set containing not less than 2000 images of the tower containing the bird nest, and directly positioning the rectangular frame of the bird nest as a recognition result by using a trained recognition model after training.
Identifying the falling of the vibration damper part:
and respectively training the whole of the shockproof hammer, the left hammer body of the shockproof hammer and the recognition model of the right hammer body of the shockproof hammer by using a deep learning image recognition and positioning algorithm Faster-RCNN and using a sample set containing not less than 2000 images of the tower containing the shockproof hammer, and judging that the shockproof hammer partially falls off if the recognized left hammer body of the shockproof hammer or the recognized right hammer body of the shockproof hammer is not in the whole area of the shockproof hammer in the image recognition process.
Identifying the displacement of the vibration damper:
and (3) training the whole shockproof hammer by using a tower image containing the shockproof hammer and a sample set of not less than 2000 pieces by using a deep learning image recognition and positioning algorithm, recognizing a rectangular frame of the whole shockproof hammer in the image, and judging that at least one shockproof hammer in a pair of shockproof hammers with the horizontal distance of the rectangular frame in the same height range smaller than the width of the rectangular frame has displacement.
And (3) identifying the missing of the pin of the nut:
the method comprises the steps of using a deep learning image recognition positioning algorithm, namely fast-RCNN, using a sample set which contains not less than 2000 pieces of tower images of link fittings such as a hanging ring hanging plate, training a pin recognition model, a link fitting recognition model and a nut recognition model. The connecting hardware is firstly identified, then the nut is identified, and then the pin is identified, if only the nut but not the pin is missing.
Identifying missing of cross arm nuts:
and (3) training a cross arm nut and bolt overall recognition model by using a deep learning image recognition positioning algorithm, namely fast-RCNN, using a tower image containing a cross arm nut to be not less than 2000 samples, recognizing the whole nut and bolt, recognizing the nut in the whole, and considering that the nut is lost if the number of the nuts is less than two.
Through the experiment, the method provided by the invention is used for realizing the result of reducing the manual workload of the interior image processing by at least 30% in the intelligent operations of analyzing, clustering, identifying, filing and the like of the unmanned aerial vehicle routing inspection data of the power supply company in Shandong electric power sunshine city in China, and has an excellent effect.
The unmanned aerial vehicle inspection system disclosed by the invention carries out unified intelligent management on massive unmanned aerial vehicle inspection images, namely, an artificial intelligence technology is applied, the automatic image classification capability of a pole tower level and a hardware level is improved, the defect objects in the images are subjected to structural conversion, the structural analysis of defect data is realized, the workload of manual analysis and processing is greatly reduced, the analysis accuracy is improved, the change of an unmanned aerial vehicle inspection mode and a management mode is promoted, and the full coverage of unmanned aerial vehicle inspection operation is further promoted.
Claims (5)
1. A transmission equipment target identification method based on unmanned aerial vehicle routing inspection is characterized in that: the method comprises the following steps:
A. the method comprises the steps of shooting image data of an area where power transmission equipment is located by adopting unmanned aerial vehicle routing inspection, extracting an image of a target to be identified in the image data and establishing a target sample set;
classifying, recombining, positioning and randomly renaming image data shot by the unmanned aerial vehicle in a polling way, and converting randomly-named and disordered images into structured data;
the target to be recognized comprises N targets, and the image of the target to be recognized i (i is less than or equal to N) is collected to form a sample subset AiSample subset A1To the sample subset ANForming a target sample set;
B. manually marking the images in the target sample set, and creating a defect sample library of the inspection images;
C. adopting a defect sample library of the inspection image as a training sample, and creating a defect calibration model based on a Faster-RCNN network;
detecting a target to be identified in a training sample by adopting a training model based on a Faster-RCNN network, training by utilizing a neural network algorithm, and then establishing a defect calibration model by utilizing a YOLOV3 algorithm;
D. collecting real-time polling image data by using an unmanned aerial vehicle, and inputting the real-time polling image data into a defect calibration model;
E. and outputting a target identification result by the defect calibration model and performing structured storage on the target identification result.
2. The unmanned aerial vehicle inspection-based power transmission equipment target identification method according to claim 1, characterized in that: the specific process of manually labeling the images in the target sample set in the step B is as follows: firstly, labeling the content of the region in the image, and labeling the hidden danger and the defect found in the region.
3. The unmanned aerial vehicle inspection-based power transmission equipment target identification method according to claim 2, characterized in that: the manual marking process can calibrate the defects in the image in a frame selection mode, the image can be zoomed and dragged in the calibration process, modification, deletion and addition of the calibration frame are supported, and meanwhile, defect information is structurally stored and displayed; the defect information includes the type and urgency of the defect.
4. The unmanned aerial vehicle inspection-based power transmission equipment target identification method according to claim 1, characterized in that: the specific process of detecting the target to be recognized in the training sample by the training model in the step C is as follows:
c1, determining a candidate frame of the training sample by using the candidate frame selection network RPN; candidate in-frame images include three cases: the first method comprises the following steps: comprising a power transmission apparatus; and the second method comprises the following steps: comprising a non-power transmission device; and the third is that: the system comprises power transmission equipment and non-power transmission equipment;
c2, feature mapping; extracting features of the image in the candidate frame by using a convolutional neural network;
c3, classifying the extracted features by adopting a classifier, and judging whether the images in the candidate frame contain the power transmission equipment or not;
c4, refining the area, and enabling the edge of the candidate frame determined as the equipment to tightly wrap the power transmission equipment, so that the area of the candidate frame is enough to accommodate the contained power transmission equipment and the area is minimum.
5. The unmanned aerial vehicle inspection-based power transmission equipment target identification method according to claim 4, wherein: the classifier in the step c3 adopts a multi-classifier softmax.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010629415.0A CN111723774A (en) | 2020-07-03 | 2020-07-03 | Target identification method for power transmission equipment based on unmanned aerial vehicle inspection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010629415.0A CN111723774A (en) | 2020-07-03 | 2020-07-03 | Target identification method for power transmission equipment based on unmanned aerial vehicle inspection |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111723774A true CN111723774A (en) | 2020-09-29 |
Family
ID=72571325
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010629415.0A Pending CN111723774A (en) | 2020-07-03 | 2020-07-03 | Target identification method for power transmission equipment based on unmanned aerial vehicle inspection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111723774A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112229845A (en) * | 2020-10-12 | 2021-01-15 | 国网河南省电力公司濮阳供电公司 | Unmanned aerial vehicle high-precision winding tower intelligent inspection method based on visual navigation technology |
CN112288711A (en) * | 2020-10-28 | 2021-01-29 | 浙江华云清洁能源有限公司 | Unmanned aerial vehicle inspection image defect image identification method, device, equipment and medium |
CN112464881A (en) * | 2020-12-11 | 2021-03-09 | 国网辽宁省电力有限公司抚顺供电公司 | Unmanned aerial vehicle optical remote sensing autonomous inspection method for power transmission line |
CN112597322A (en) * | 2020-11-16 | 2021-04-02 | 中国电力科学研究院有限公司 | Method and system for extracting target image from power transmission line inspection image |
CN112669264A (en) * | 2020-12-17 | 2021-04-16 | 国网山西省电力公司运城供电公司 | Artificial intelligence defect identification method and system for unmanned aerial vehicle routing inspection of distribution network line |
CN112699817A (en) * | 2021-01-04 | 2021-04-23 | 浙江工业大学 | Unmanned aerial vehicle image pipeline extraction method based on deep convolutional neural network |
CN113192222A (en) * | 2021-04-14 | 2021-07-30 | 山东理工大学 | Equipment-level alarm strategy method for power line visual alarm |
CN113312991A (en) * | 2021-05-14 | 2021-08-27 | 华能阜新风力发电有限责任公司 | Front-end intelligent recognition system based on unmanned aerial vehicle |
CN115656181A (en) * | 2022-10-27 | 2023-01-31 | 重庆红岩建设机械制造有限责任公司 | Automatic inspection and identification method for apparent defects outside main tower of bridge |
-
2020
- 2020-07-03 CN CN202010629415.0A patent/CN111723774A/en active Pending
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112229845A (en) * | 2020-10-12 | 2021-01-15 | 国网河南省电力公司濮阳供电公司 | Unmanned aerial vehicle high-precision winding tower intelligent inspection method based on visual navigation technology |
CN112288711A (en) * | 2020-10-28 | 2021-01-29 | 浙江华云清洁能源有限公司 | Unmanned aerial vehicle inspection image defect image identification method, device, equipment and medium |
CN112597322A (en) * | 2020-11-16 | 2021-04-02 | 中国电力科学研究院有限公司 | Method and system for extracting target image from power transmission line inspection image |
CN112464881A (en) * | 2020-12-11 | 2021-03-09 | 国网辽宁省电力有限公司抚顺供电公司 | Unmanned aerial vehicle optical remote sensing autonomous inspection method for power transmission line |
CN112669264A (en) * | 2020-12-17 | 2021-04-16 | 国网山西省电力公司运城供电公司 | Artificial intelligence defect identification method and system for unmanned aerial vehicle routing inspection of distribution network line |
CN112699817A (en) * | 2021-01-04 | 2021-04-23 | 浙江工业大学 | Unmanned aerial vehicle image pipeline extraction method based on deep convolutional neural network |
CN112699817B (en) * | 2021-01-04 | 2024-02-13 | 浙江工业大学 | Unmanned aerial vehicle image pipeline extraction method based on deep convolutional neural network |
CN113192222A (en) * | 2021-04-14 | 2021-07-30 | 山东理工大学 | Equipment-level alarm strategy method for power line visual alarm |
CN113192222B (en) * | 2021-04-14 | 2022-11-15 | 山东理工大学 | Equipment-level alarm strategy method for power line visual alarm |
CN113312991A (en) * | 2021-05-14 | 2021-08-27 | 华能阜新风力发电有限责任公司 | Front-end intelligent recognition system based on unmanned aerial vehicle |
CN115656181A (en) * | 2022-10-27 | 2023-01-31 | 重庆红岩建设机械制造有限责任公司 | Automatic inspection and identification method for apparent defects outside main tower of bridge |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111723774A (en) | Target identification method for power transmission equipment based on unmanned aerial vehicle inspection | |
CN106712713B (en) | The monitoring system and monitoring and abnormal localization method of a kind of photovoltaic plant | |
CN112734692A (en) | Transformer equipment defect identification method and device | |
CN111444939A (en) | Small-scale equipment component detection method based on weak supervision cooperative learning in open scene of power field | |
CN111160791A (en) | Abnormal user identification method based on GBDT algorithm and factor fusion | |
CN111881970A (en) | Intelligent outer broken image identification method based on deep learning | |
CN113849542A (en) | System and method for checking regional greenhouse gas emission list based on artificial intelligence | |
CN101738998A (en) | System and method for monitoring industrial process based on local discriminatory analysis | |
CN115269958A (en) | Internet reliability data information acquisition and analysis system | |
CN117114420B (en) | Image recognition-based industrial and trade safety accident risk management and control system and method | |
CN115082432B (en) | Small target bolt defect detection method and device based on fine-grained image classification | |
CN113408630A (en) | Transformer substation indicator lamp state identification method | |
CN117372956A (en) | Method and device for detecting state of substation screen cabinet equipment | |
CN112683916A (en) | Method and device for identifying missing or mounting error of small hardware fittings of current collecting line tower | |
CN117764547A (en) | Photovoltaic string fault diagnosis method and system | |
CN113284103B (en) | Substation equipment defect online detection method based on space transformation fast R-CNN model | |
CN115619763A (en) | Power transmission line small part image target detection method based on multistage reasoning detection | |
CN111783899B (en) | Method for autonomously learning and identifying novel contraband | |
CN111930982A (en) | Intelligent labeling method for power grid images | |
Su et al. | Detection and State Classification of Bolts Based on Faster R-CNN | |
Liangzhi et al. | Research on fault prediction and diagnosis of power equipment based on big data | |
CN116703333B (en) | Power plant technical supervision informatization management system | |
Shen et al. | CAD Fabric Model Defect Detection Based on Improved Yolov5 Based on Self-Attention Mechanism | |
CN118211942B (en) | Semiconductor gaseous molecular pollutant space distribution management system and method | |
CN117708720B (en) | Equipment fault diagnosis system based on knowledge graph |
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 |