CN113591645A - Power equipment infrared image identification method based on regional convolutional neural network - Google Patents
Power equipment infrared image identification method based on regional convolutional neural network Download PDFInfo
- Publication number
- CN113591645A CN113591645A CN202110826482.6A CN202110826482A CN113591645A CN 113591645 A CN113591645 A CN 113591645A CN 202110826482 A CN202110826482 A CN 202110826482A CN 113591645 A CN113591645 A CN 113591645A
- Authority
- CN
- China
- Prior art keywords
- power equipment
- infrared image
- rcnn
- image
- fast
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000013527 convolutional neural network Methods 0.000 title claims description 13
- 238000001514 detection method Methods 0.000 claims abstract description 28
- 238000010586 diagram Methods 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 8
- 230000007613 environmental effect Effects 0.000 claims abstract description 5
- 238000000605 extraction Methods 0.000 claims abstract description 5
- 238000012545 processing Methods 0.000 claims description 9
- 230000002238 attenuated effect Effects 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 abstract description 3
- 230000007547 defect Effects 0.000 abstract description 2
- 238000013528 artificial neural network Methods 0.000 abstract 2
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000013135 deep learning Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- OLBCVFGFOZPWHH-UHFFFAOYSA-N propofol Chemical compound CC(C)C1=CC=CC(C(C)C)=C1O OLBCVFGFOZPWHH-UHFFFAOYSA-N 0.000 description 1
- 229960004134 propofol Drugs 0.000 description 1
- 238000010809 targeting technique Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- 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
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Economics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Human Resources & Organizations (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Bioinformatics & Computational Biology (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a power equipment infrared image identification method based on a regional convolution neural network, which comprises the steps of firstly obtaining infrared images of various power equipment under different environmental conditions, constructing a power equipment infrared image sample library and providing a sample set for training the neural network; in the identification process, firstly, inputting an infrared image with a power equipment label into a ZF network for feature extraction to obtain a feature map of the power equipment image; generating a suggestion window by the characteristic diagram of the obtained electric power equipment image through the fast-RCNN, and simultaneously predicting relevant anchor point parameters through the fast-RCNN by utilizing the characteristic diagram to obtain an object frame finally output by the electric power equipment in the infrared image; and (4) classifying and positioning the infrared images of the electric equipment by utilizing Softmax Loss and Smooth L1 Loss. The method can output an identification result which is intuitive, safe, reliable and accurate in analysis, and overcomes the defects of low detection efficiency, false detection and missing detection in the traditional manual inspection process.
Description
Technical Field
The invention relates to the field of computer image processing and deep learning, in particular to an infrared image identification method for power equipment based on a regional convolutional neural network.
Background
The safe operation of the power equipment in the power system is the basis for ensuring the reliable work of the power grid. Currently, diagnostics for electrical devices have been widely accomplished using infrared targeting techniques. The thermal infrared imager-based electrical equipment fault detection method has the characteristics of visual diagnosis, safety, reliability, accurate analysis and the like. Because the equipment diagnosis process adopting the thermal imager also needs operation and maintenance personnel to carry out on-site detection, the problems of low detection efficiency, false detection and missed detection are inevitable in the manual inspection process. Therefore, how to realize the automatic detection with the assistance of a computer becomes a problem to be solved urgently by practitioners of the same industry.
Disclosure of Invention
In view of the above problems, the invention provides an infrared image identification method for power equipment based on a regional convolutional neural network, which aims to solve the problems of low detection efficiency, false detection and missing detection inevitably existing in the current manual routing inspection process.
The embodiment of the invention provides a power equipment infrared image identification method based on a regional convolutional neural network, which comprises the following steps:
s100, acquiring infrared images of various electric equipment under different environmental conditions, and constructing an electric equipment infrared image sample library; each infrared image is provided with a label of the power equipment;
s200, inputting the infrared image with the electric equipment label into a ZF network for feature extraction to obtain a feature diagram of the electric equipment image; the ZF network consists of 5 convolutional layers and 3 full-connection layers;
s300, generating a suggestion window through the characteristic diagram of the obtained electric power equipment image through the fast-RCNN, and simultaneously predicting relevant anchor point parameters through the characteristic diagram through the fast-RCNN to obtain an object frame finally output by the electric power equipment in the infrared image;
s400, processing a suggested target area provided by the Faster-RCNN network by using a RoI Pooling layer, obtaining feature maps of the suggested target area, and mapping and outputting candidate areas corresponding to the object frame output in the step S300 into a feature map with uniform size;
and S500, mapping and outputting the candidate areas into a feature map with a uniform size, and classifying and positioning the infrared images of the electric power equipment by utilizing Softmax Loss and Smooth L1 Loss.
In an embodiment, in step S300, predicting anchor point related parameters through fast-RCNN using the feature map to obtain an object frame finally output by the power device in the infrared image, including:
predicting relevant parameters of anchor points through fast-RCNN by utilizing the characteristic diagram, and adding predicted offset to the input candidate frame in the training process;
and performing NMS operation once according to the prediction score to obtain the final output object frame of the target.
In one embodiment, performing an NMS operation once based on the predicted score to obtain a final output object box of the target includes:
in the NMS algorithm, according to the intersection ratio of any anchor frame and the anchor frame M with the highest confidence coefficient, when the intersection ratio is greater than a high point threshold value, the confidence coefficient is attenuated more quickly;
when the intersection ratio is smaller than the low-point threshold, the detection result has the trend of non-identical targets, and the confidence coefficient of the detection result is ensured to be slowly attenuated;
the confidence level decays linearly when the cross-over ratio is between 0.2 and 0.8.
The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:
according to the electric power equipment infrared image identification method based on the regional convolutional neural network, which is provided by the embodiment of the invention, based on the improved Faster RCNN, faults of electric power equipment are directly and accurately detected through deep learning, an identification result which is visual, safe, reliable and accurate in analysis is given, and the defects of low detection efficiency, false detection and missing detection in the traditional manual inspection process are overcome. The method skillfully concatenates the convolutional neural network according to the characteristic extraction and classification characteristics of deep learning, and the detection of the power equipment fault is simple and effective.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
fig. 1 is a flowchart of an infrared image recognition method for an electrical device based on a regional convolutional neural network according to an embodiment of the present invention;
fig. 2 is a structural diagram of a regional convolutional neural network according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
Referring to fig. 1, an infrared image identification method for a power device based on a regional convolutional neural network according to an embodiment of the present invention includes:
s100, acquiring infrared images of various electric equipment under different environmental conditions, and constructing an electric equipment infrared image sample library; each infrared image is provided with a label of the power equipment;
s200, inputting the infrared image with the electric equipment label into a ZF network for feature extraction to obtain a feature diagram of the electric equipment image; the ZF network consists of 5 convolutional layers and 3 full-connection layers;
s300, generating a suggestion window through the characteristic diagram of the obtained electric power equipment image through the fast-RCNN, and simultaneously predicting relevant anchor point parameters through the characteristic diagram through the fast-RCNN to obtain an object frame finally output by the electric power equipment in the infrared image;
s400, processing a suggested target area provided by the Faster-RCNN network by using a RoI Pooling layer, obtaining feature maps of the suggested target area, and mapping and outputting candidate areas corresponding to the object frame output in the step S300 into a feature map with uniform size;
and S500, mapping and outputting the candidate areas into a feature map with a uniform size, and classifying and positioning the infrared images of the electric power equipment by utilizing Softmax Loss and Smooth L1 Loss.
As shown in FIG. 2, a block diagram of a regional convolutional neural network is modified on the algorithm framework of fast-RCNN.
In the step S100, an infrared image sample library of the power equipment is manufactured, images in the sample library are labeled on a target (power equipment) through visual image calibration tool software Labeli, and the image labels are used for creating a data set, so that subsequent deep learning training is facilitated. And converting the XML tag file into an XML tag file conforming to the PASCAL VOC standard format, which means that the data dataset comprises training pictures and test pictures, wherein the pictures are named in a fixed format, such as a 'year _ number.jpg' format. The size of the pixels of the pictures is different, for example, the size of the horizontal graph is about 500 × 375, and the size of the vertical graph is about 375 × 500, and the deviation is basically not over 100. As image data for training and test validation. After the picture frame is labeled and the category is selected, the XML tag file is stored and generated, and then a sample library comprising infrared images of various power equipment under different environmental conditions is formed. The method and process for acquiring the infrared image are not limited.
In step S200, a ZF network is used to extract features, where ZFNet is modified in some details based on AlexNet, and the network structure has no major breakthrough, but is modified in some ways based on convolution kernel and stride. The use of lightweight ZF networks has two advantages:
(1) a multi-layer deconvolution network is used for visualizing the evolution of the characteristics in the training process and discovering potential problems;
(2) and simultaneously, the input information of the part is more important for the classification task according to the influence of the local part of the occlusion image on the classification result.
In step S300, generating an advice window from the obtained characteristic diagram of the electrical device through a fast-RCNN regional advice network, and predicting anchor point related parameters through the regional advice network by using the characteristic diagram; in the training process, adding the predicted offset to the input Region Proposal (candidate frame), then performing NMS operation again according to the predicted score to obtain the final output object frame of the target, and in the process, according to the set threshold, positioning and outputting the object frame with the score larger than the threshold; and calibrating the rectangular frame of the target object to be recognized.
Since there may be multiple positioning frames corresponding to one object, non-maximum suppression is generally used to remove the overlapping frames. In a classical NMS (non-maximum suppression) algorithm, a penalty factor is selected to reduce the confidence of an anchor frame according to the intersection ratio of any anchor frame to the anchor frame M with the highest anchor frame position confidence, and a preset threshold. The penalty factor function adopts a piecewise function form, and the change of the penalty factor is realized according to the dynamic change of the cross-over ratio. Wherein: describing the probability that the prediction bounding box contains a certain class of objects, the calculation of confidence is done by sofmax. The method provides a fraction resetting function of a preset segment.
When the intersection ratio is greater than the high point threshold, the same target has the repeated detection trend, so that the confidence coefficient of the target is quickly attenuated, and the confidence coefficient is the maximum prediction probability with the identification target; when the intersection ratio is smaller than the low-point threshold, the detection result has the trend of non-identical targets, so that the confidence coefficient of the detection result is ensured to be slowly attenuated; when the cross-over ratio is between 0.2 and 0.8, the detection result has no tendency, so the confidence level is linearly attenuated. In the step, the detection network is trained by adopting the positioning confidence coefficient of the anchor frame, so that the positioning precision is improved.
In step S400, the RoI posing layer is responsible for processing the proposed target Region provided by the fast-RCNN RPN network, and obtaining feature maps of the proposed Region, and the role of sending the feature maps to the subsequent network RoI posing layer is to map and output the Region propofol (candidate boxes) with different sizes into a feature map with uniform size.
In step S500, classification and positioning of the infrared image of the power device are completed by using Softmax Loss and Smooth L1Loss (detection frame regression).
According to the power equipment infrared image identification method based on the regional convolutional neural network, firstly, a light ZF network is adopted to extract the characteristics of an infrared image target, then a proper anchor window proportion is selected according to the image characteristics to improve the detection precision of the target image, and finally a segmented penalty factor is adopted in an NMS algorithm, so that the power equipment detection problems under some special occasions, such as the conditions of overlapping shielding and the like between equipment, are solved.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (3)
1. An electric power equipment infrared image identification method based on a regional convolutional neural network is characterized by comprising the following steps:
s100, acquiring infrared images of various electric equipment under different environmental conditions, and constructing an electric equipment infrared image sample library; each infrared image is provided with a label of the power equipment;
s200, inputting the infrared image with the electric equipment label into a ZF network for feature extraction to obtain a feature diagram of the electric equipment image; the ZF network consists of 5 convolutional layers and 3 full-connection layers;
s300, generating a suggestion window through the characteristic diagram of the obtained electric power equipment image through the fast-RCNN, and simultaneously predicting relevant anchor point parameters through the characteristic diagram through the fast-RCNN to obtain an object frame finally output by the electric power equipment in the infrared image;
s400, processing a suggested target area provided by the Faster-RCNN network by using a RoI Pooling layer, obtaining feature maps of the suggested target area, and mapping and outputting candidate areas corresponding to the object frame output in the step S300 into a feature map with uniform size;
and S500, mapping and outputting the candidate areas into a feature map with a uniform size, and classifying and positioning the infrared images of the electric power equipment by utilizing Softmax Loss and Smooth L1 Loss.
2. The method according to claim 1, wherein in step S300, the anchor point related parameters are predicted through fast-RCNN using the feature map, and an object frame finally output by the power device in the infrared image is obtained, including:
predicting relevant parameters of anchor points through fast-RCNN by utilizing the characteristic diagram, and adding predicted offset to the input candidate frame in the training process;
and performing NMS operation once according to the prediction score to obtain the final output object frame of the target.
3. The method of claim 2, wherein performing an NMS operation on the prediction score to obtain a final output object box of the target comprises:
in the NMS algorithm, deleting the repeated target frame according to the intersection ratio of any anchor frame and the anchor frame M with the highest confidence coefficient and a preset threshold;
when the intersection ratio is larger than a preset highest IOU _ max threshold value, the confidence coefficient of the intersection ratio is attenuated quickly;
when the intersection ratio is smaller than a preset lowest IOU _ min threshold value, the detection result has the trend of a non-identical target, and the confidence coefficient of the detection result is ensured to be slowly attenuated;
when IOU _ max is 0.8 and IOU _ min is 0.2, the change in confidence selects a linear decay when the value of the cross-over ratio is between the minimum and maximum values.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110826482.6A CN113591645B (en) | 2021-07-21 | 2021-07-21 | Regional convolutional neural network-based infrared image recognition method for power equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110826482.6A CN113591645B (en) | 2021-07-21 | 2021-07-21 | Regional convolutional neural network-based infrared image recognition method for power equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113591645A true CN113591645A (en) | 2021-11-02 |
| CN113591645B CN113591645B (en) | 2024-09-13 |
Family
ID=78248869
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110826482.6A Active CN113591645B (en) | 2021-07-21 | 2021-07-21 | Regional convolutional neural network-based infrared image recognition method for power equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113591645B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114199381A (en) * | 2021-12-13 | 2022-03-18 | 国网陕西省电力公司延安供电公司 | Electrical equipment fault detection method for improving infrared detection model |
| CN114611666A (en) * | 2022-03-08 | 2022-06-10 | 安谋科技（中国）有限公司 | NMS function quantization method, electronic device and medium |
| CN115082318A (en) * | 2022-07-13 | 2022-09-20 | 东北电力大学 | Electrical equipment infrared image super-resolution reconstruction method |
| CN117197097A (en) * | 2023-09-13 | 2023-12-08 | 西南科技大学 | Power equipment component detection method based on infrared image |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA1080827A (en) * | 1976-11-12 | 1980-07-01 | Societe Nationale Industrielle Aerospatiale | Method of steering a spacecraft and of regulating its onboard supply voltage |
| CN107480730A (en) * | 2017-09-05 | 2017-12-15 | 广州供电局有限公司 | Power equipment identification model construction method and system, the recognition methods of power equipment |
| CN107607207A (en) * | 2017-08-22 | 2018-01-19 | 中国科学院深圳先进技术研究院 | A kind of power equipment thermal fault detection method, system and electronic equipment |
| CN110210463A (en) * | 2019-07-03 | 2019-09-06 | 中国人民解放军海军航空大学 | Radar target image detecting method based on Precise ROI-Faster R-CNN |
| CN110334661A (en) * | 2019-07-09 | 2019-10-15 | 国网江苏省电力有限公司扬州供电分公司 | Infrared power transmission and transformation abnormal heating point target detecting method based on deep learning |
| WO2019232836A1 (en) * | 2018-06-04 | 2019-12-12 | 江南大学 | Multi-scale sensing pedestrian detection method based on improved full convolutional network |
| CN111209958A (en) * | 2020-01-03 | 2020-05-29 | 南方电网科学研究院有限责任公司 | Transformer substation equipment detection method and device based on deep learning |
| CN111539422A (en) * | 2020-04-13 | 2020-08-14 | 南京理工大学 | Flight target cooperative identification method based on fast RCNN |
| CN111669223A (en) * | 2020-05-08 | 2020-09-15 | 东北电力大学 | Data transmission system based on LoRa and visible light communication |
| CN112288773A (en) * | 2020-10-19 | 2021-01-29 | 慧视江山科技(北京)有限公司 | Multi-scale human body tracking method and device based on Soft-NMS |
| US20210082144A1 (en) * | 2019-09-12 | 2021-03-18 | Nec Laboratories America, Inc | Keypoint based pose-tracking using entailment |
| CN112581430A (en) * | 2020-12-03 | 2021-03-30 | 厦门大学 | Deep learning-based aeroengine nondestructive testing method, device, equipment and storage medium |
-
2021
- 2021-07-21 CN CN202110826482.6A patent/CN113591645B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA1080827A (en) * | 1976-11-12 | 1980-07-01 | Societe Nationale Industrielle Aerospatiale | Method of steering a spacecraft and of regulating its onboard supply voltage |
| CN107607207A (en) * | 2017-08-22 | 2018-01-19 | 中国科学院深圳先进技术研究院 | A kind of power equipment thermal fault detection method, system and electronic equipment |
| CN107480730A (en) * | 2017-09-05 | 2017-12-15 | 广州供电局有限公司 | Power equipment identification model construction method and system, the recognition methods of power equipment |
| WO2019232836A1 (en) * | 2018-06-04 | 2019-12-12 | 江南大学 | Multi-scale sensing pedestrian detection method based on improved full convolutional network |
| CN110210463A (en) * | 2019-07-03 | 2019-09-06 | 中国人民解放军海军航空大学 | Radar target image detecting method based on Precise ROI-Faster R-CNN |
| CN110334661A (en) * | 2019-07-09 | 2019-10-15 | 国网江苏省电力有限公司扬州供电分公司 | Infrared power transmission and transformation abnormal heating point target detecting method based on deep learning |
| US20210082144A1 (en) * | 2019-09-12 | 2021-03-18 | Nec Laboratories America, Inc | Keypoint based pose-tracking using entailment |
| CN111209958A (en) * | 2020-01-03 | 2020-05-29 | 南方电网科学研究院有限责任公司 | Transformer substation equipment detection method and device based on deep learning |
| CN111539422A (en) * | 2020-04-13 | 2020-08-14 | 南京理工大学 | Flight target cooperative identification method based on fast RCNN |
| CN111669223A (en) * | 2020-05-08 | 2020-09-15 | 东北电力大学 | Data transmission system based on LoRa and visible light communication |
| CN112288773A (en) * | 2020-10-19 | 2021-01-29 | 慧视江山科技(北京)有限公司 | Multi-scale human body tracking method and device based on Soft-NMS |
| CN112581430A (en) * | 2020-12-03 | 2021-03-30 | 厦门大学 | Deep learning-based aeroengine nondestructive testing method, device, equipment and storage medium |
Non-Patent Citations (3)
| Title |
|---|
| SONGTAO LIU等: "Adaptive NMS: Refining Pedestrian Detection in a Crowd", Retrieved from the Internet <URL:http://arxiv.org/abs/1904.03629> * |
| 刘云鹏等: "基于深度学习的输变电设备异常发热点红外图片目标检测方法", 《南方电网技术》, vol. 13, no. 02, 20 February 2019 (2019-02-20), pages 27 - 33 * |
| 胡鹏等: "基于YOLO算法和RK3399平台的行人检测实时性研究", 《数字技术与应用》, vol. 38, no. 03, 25 March 2020 (2020-03-25), pages 78 - 81 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114199381A (en) * | 2021-12-13 | 2022-03-18 | 国网陕西省电力公司延安供电公司 | Electrical equipment fault detection method for improving infrared detection model |
| CN114611666A (en) * | 2022-03-08 | 2022-06-10 | 安谋科技（中国）有限公司 | NMS function quantization method, electronic device and medium |
| CN114611666B (en) * | 2022-03-08 | 2024-05-31 | 安谋科技（中国）有限公司 | Quantification method of NMS function, electronic equipment and medium |
| CN115082318A (en) * | 2022-07-13 | 2022-09-20 | 东北电力大学 | Electrical equipment infrared image super-resolution reconstruction method |
| CN117197097A (en) * | 2023-09-13 | 2023-12-08 | 西南科技大学 | Power equipment component detection method based on infrared image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113591645B (en) | 2024-09-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113591645A (en) | Power equipment infrared image identification method based on regional convolutional neural network | |
| US12039712B2 (en) | Underground infrastructure feature identification using multi-sensor inspection data analysis | |
| CN113139543B (en) | Training method of target object detection model, target object detection method and equipment | |
| CN113439227B (en) | Capturing and storing enlarged images | |
| CN113052295B (en) | Training method of neural network, object detection method, device and equipment | |
| US20220343642A1 (en) | High-precision identification method and system for substations | |
| US20240037724A1 (en) | Plant detection and display system | |
| CN115909059A (en) | Natural resource sample library establishing method and device | |
| CN113222913A (en) | Circuit board defect detection positioning method and device and storage medium | |
| CN117689943A (en) | Training method of screw detection model, screw detection method and device | |
| CN112561885A (en) | YOLOv 4-tiny-based gate valve opening detection method | |
| CN115861922B (en) | Sparse smoke detection method and device, computer equipment and storage medium | |
| KR20210037199A (en) | Apparatus for dividing, tagging an image and for detecting defect of facilities using the same | |
| CN112287938B (en) | Text segmentation method, system, device and medium | |
| Kee et al. | Cracks identification using mask region-based denoised deformable convolutional network | |
| KR102621884B1 (en) | Deep learning-based image analysis method for selecting defective products and a system therefor | |
| CN117036670B (en) | Training method, device, equipment, medium and program product of quality detection model | |
| CN118505704B (en) | General model modeling detection method for detecting defects of panel production line | |
| WO2023112302A1 (en) | Training data creation assistance device and training data creation assistance method | |
| KR102721266B1 (en) | Deep learning-based sewerage defect detection method and apparatus | |
| CN113762248B (en) | Target landing detection method and device, electronic equipment and storage medium | |
| CN113901859A (en) | Pipe gallery video intelligent attribute detection method and system based on embedded front-end processing | |
| CN118736465A (en) | Defect screw detection method and device and electronic equipment | |
| CN117371553A (en) | Machine learning model iteration method and system of intelligent operation and maintenance system | |
| CN118505699A (en) | Gap detection method, device, equipment and storage medium based on visual detection |
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 |