CN109544476A - A kind of power equipment Infrared Image Denoising method based on deep learning - Google Patents
A kind of power equipment Infrared Image Denoising method based on deep learning Download PDFInfo
- Publication number
- CN109544476A CN109544476A CN201811392020.2A CN201811392020A CN109544476A CN 109544476 A CN109544476 A CN 109544476A CN 201811392020 A CN201811392020 A CN 201811392020A CN 109544476 A CN109544476 A CN 109544476A
- Authority
- CN
- China
- Prior art keywords
- infrared image
- power equipment
- model
- image
- deep learning
- 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
- 238000013135 deep learning Methods 0.000 title claims abstract description 26
- 238000000034 method Methods 0.000 title claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 6
- 230000004913 activation Effects 0.000 claims description 20
- 238000012549 training Methods 0.000 claims description 18
- 238000012360 testing method Methods 0.000 claims description 13
- 238000009826 distribution Methods 0.000 claims description 9
- 230000007480 spreading Effects 0.000 claims description 8
- 238000003892 spreading Methods 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 7
- 230000033228 biological regulation Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 9
- 238000013527 convolutional neural network Methods 0.000 abstract description 6
- 238000007796 conventional method Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 3
- 230000000644 propagated effect Effects 0.000 description 3
- 238000013136 deep learning model Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000001931 thermography Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The power equipment Infrared Image Denoising method based on deep learning that the invention discloses a kind of.This method carries out power equipment Infrared Image Denoising using deep learning, the processing of power equipment Infrared Image Denoising is carried out using the denoising model that depth convolutional neural networks are established, recursive call and expansion convolution are applied on proposed depth convolutional neural networks simultaneously, better denoising effect has been reached under identical parameter amount.The invention feature is strong innovation, practical, can be realized that power equipment Infrared Image Denoising effect is good, fireballing demand.
Description
Technical field
The present invention relates to power equipment monitoring technical fields, more specifically refer to that a kind of electric power based on deep learning is set
Standby Infrared Image Denoising method.
Background technique
With the rapid development of our country's economy, country proposes to build strong reliable smart grid.Electrical equipment is located for a long time
Under operation, vulnerable to the influence of external environment, often there are all kinds of failures, therefore, to different types of electrical equipment into
Row effective monitoring becomes the hot spot studied at present.Infrared thermal imaging technique is one in electric system on-line real time monitoring
Powerful tool prevents potential wind by thermal imagery graph discovery power equipment because of the position of temperature anomaly caused by failure or hidden danger
Danger, to promote operation of power networks stability and reliability.But due to the interference such as electronic device in external environment and equipment because
The influence of element, the interference that infrared image can inevitably by various noises during acquisition, transimission and storage, such as Gauss
Noise, impulsive noise etc., this not only results in the decline of picture quality, but also can be to the operation such as infrared image analysis and identification
It has an impact, to reduce the reliability of detection system, therefore it is very crucial for carrying out denoising to the infrared image of acquisition
's.
Deep learning has powerful nonlinear fitting ability, and in various image processing tasks, and deep learning is calculated
Method is better than the performance of traditional algorithm.And be only the time required to deep learning algorithm is in test propagated forward when
Between, it is completely superior to traditional algorithm based on iteration, meets the needs of real-time.
The research of power equipment Ways of Removing Noises of Infrared Image guarantees operation of power networks for promoting Power System Faults Detection effect
Stabilization is of great significance, and has broad application prospects and profoundly social value.In this context, it is proposed that based on deep
Spend the method for the power equipment Infrared Image Denoising of study.
Summary of the invention
A kind of power equipment Infrared Image Denoising method based on deep learning provided by the invention, its object is to solve
The above-mentioned problems in the prior art.
The technical solution adopted by the invention is as follows:
A kind of power equipment Infrared Image Denoising method based on deep learning, comprising the following steps: the following steps are included:
A, the training sample pair of power equipment infrared image is constructed；
B, the power equipment Infrared Image Denoising model based on deep learning is built；
C, parameter regulation and optimization are carried out to the model, the training model simultaneously tests it；
Specifically, above-mentioned steps A the specific steps are under:
Step 1: carrying out normalizing using the infrared image of infrared thermoviewer acquisition power equipment, and to collected infrared image
Change processing obtains the clean infrared image of plural number；
Step 2: after by rotation, overturning or translation, by each clean infrared image random cropping at several lesser
Image block obtains enough clean image patterns；
Step 3: being obtained to every clean image pattern plus white Gaussian noise using clean image pattern as the label of model
Corresponding noisy image sample, obtains a plurality of training samples pair, and using the 70% of total sample pair as training set, 15% as verifying
Collection, 15% is used as test set.
Specifically, under the concrete operations of above-mentioned steps B are:
The model is made of three parts, and first part is characterized extraction module, by a convolutional layer and a nonlinear activation
Layer is constituted；Second part be recursive call module, include three expansion convolutional layers, each expansion convolutional layer later immediately one criticize
Normalize layer and a nonlinear activation layer；Part III is image reconstruction module, is made of a convolutional layer；Using residual error
Thought, the learning model of the task.
More specifically, convolutional layer convolution kernel size is 3 × 3 in features described above extraction module, the weight of convolution kernel initialization
Distribution meets Gaussian Profile, inputs as noisy image, the characteristic pattern in 32 channels of output, the activation primitive that nonlinear activation layer uses
For ReLU.
More specifically, the input for expanding convolutional layer is all the characteristic pattern in 32 channels, and exports in above-mentioned recursive call module
It is all the characteristic pattern in 32 channels, convolution kernel size is all 3 × 3, and the weight distribution of convolution kernel initialization meets Gaussian Profile, first
The spreading rate of a expansion convolutional layer is 2, and the spreading rate of second expansion convolutional layer is 3, and third expands the spreading rate of convolutional layer
It is 4, the activation primitive that nonlinear activation layer uses is ReLU.
More specifically, convolutional layer convolution kernel size is 3 × 3 in above-mentioned image reconstruction module, the weight of convolution kernel initialization
Distribution meets Gaussian Profile, inputs as the characteristic pattern in 32 channels, and output is negative noise pattern；It is above-mentioned specific using the thought of residual error
It is: in a model plus a global bridge joint, model is made to be changed into the negative noise of study by learning clean image, grandfather tape is made an uproar figure
As passing through global bridging, in addition negative noise pattern, just obtains clean image.
Specifically, the concrete operations of above-mentioned steps C are as follows:
Step a: construction loss function, loss function are, wherein I0What is indicated is that the band of mode input is made an uproar
Image, IGWhat is indicated is the clean image of model output；
Step b: carrying out parameter regulation to each convolutional layer, select suitable optimizer training pattern, saves trained model power
Weight；
Step c: being loaded into trained Model Weight, is tested using test the set pair analysis model.
Specifically, in above-mentioned steps a, using the error backpropagation algorithm training model, using adam optimizer, always
Iteration optimization 100,000 times altogether saves the weight of model obtained by iteration optimization.
By the above-mentioned description of this invention it is found that being compared with existing technology, the present invention has the advantages that
Innovative point of the invention includes: that (1) using deep learning realizes power equipment Infrared Image Denoising Algorithm.It is of the invention first
Secondary that deep learning is applied in power equipment Infrared Image Denoising, trained deep learning model only needs one in test phase
Secondary propagated forward expends the time well below the conventional method based on iterative solution, meets Power System Intelligent, real-time
Demand.(2) the model application residual error thought is by Infrared Image Denoising question simplification, directly learn noise rather than band is made an uproar figure
Picture.(3) recursive call and expansion convolution are applied in power equipment Infrared Image Denoising problem simultaneously for the first time, in lower ginseng
Better performance is obtained under quantity.
The present invention carries out the processing of power equipment Infrared Image Denoising using the denoising model that depth convolutional neural networks are established,
Preferably denoising effect can be achieved compared to conventional method, while recursive call and expansion convolution are applied to proposed depth
On convolutional neural networks, better denoising effect has been reached under identical parameter amount.The invention feature is strong innovation, practical
Property is strong, can be realized that power equipment Infrared Image Denoising effect is good, fireballing demand.
Detailed description of the invention
Fig. 1 is the power equipment Infrared Image Denoising model structure based on depth.
Fig. 2 is recursive call model schematic.
Fig. 3 is receptive field contrast table.
Specific embodiment
Illustrate a specific embodiment of the invention with reference to the accompanying drawings.In order to fully understand the present invention, it is described below and is permitted
More details, but to those skilled in the art, the present invention can also be realized without these details.
A kind of power equipment Infrared Image Denoising method based on deep learning, comprising the following steps: the following steps are included:
A, the training sample pair of power equipment infrared image is constructed, it is specifically as follows.
Step 1: being carried out using the infrared image of infrared thermoviewer acquisition power equipment, and to collected infrared image
Normalized obtains the clean infrared image of plural number.
Step 2: by rotation, overturning or translation after, by each clean infrared image random cropping at it is several compared with
Small image block obtains enough clean image patterns.
Step 3: adding white Gaussian noise to every clean image pattern using clean image pattern as the label of model
Corresponding noisy image sample is obtained, a plurality of training samples pair are obtained, using the 70% of total sample pair as training set, 15% conduct
Verifying collection, 15% is used as test set.
B, the power equipment Infrared Image Denoising model based on deep learning is built.
Referring to Fig.1, which is made of three parts, and first part is characterized extraction module, by a convolutional layer and one
Nonlinear activation layer is constituted；Second part is recursive call module, includes three expansion convolutional layers, after each expansion convolutional layer
Immediately one batch of normalization layer and a nonlinear activation layer；Part III is image reconstruction module, is made of a convolutional layer；
Using the thought of residual error, the learning model of the task.
Referring to Fig.1, more specifically, characteristic extracting module is made of a convolutional layer and a nonlinear activation layer, convolution
Layer convolution kernel size is 3 × 3, and the weight distribution of convolution kernel initialization meets Gaussian Profile, inputs as noisy image, and output 32 is logical
The characteristic pattern in road, the activation primitive that nonlinear activation layer uses are ReLU.
Referring to FIG. 1, FIG. 2 and FIG. 3, more specifically, recursive call module is made of three expansion convolutional layers, convolutional layer it is defeated
Enter the characteristic pattern all for 32 channels, and export the characteristic pattern all for 32 channels, convolution kernel size is all 3 × 3, convolution kernel initialization
Weight distribution meet Gaussian Profile, the spreading rate of first expansion convolutional layer is 2, and the spreading rate of second expansion convolutional layer is
3, the spreading rate that third expands convolutional layer is 4.Using the model of expansion convolution and using only four layers before the model of common convolution
Receptive field compare as shown in the table in Fig. 3, after each expansion convolutional layer immediately one criticize normalize layer and one it is non-thread
Property active coating, the activation primitive that nonlinear activation layer uses be ReLU, recursive call is exactly by third nonlinear activation layer
The importation of the module is sent into output, reuses the parameter of the module, is obtained in the case where not increasing parameter amount more preferable
Nonlinear fitting ability.
Referring to FIG. 1, FIG. 2 and FIG. 3, more specifically, image reconstruction module is made of a convolutional layer, convolutional layer convolution kernel is big
Small is 3 × 3, and the weight distribution of convolution kernel initialization meets Gaussian Profile, is inputted as the characteristic pattern in 32 channels, and output is negative noise
Figure.It is specifically using the thought of residual error: in a model plus a global bridge joint, is changed into model by learning clean image
Learn negative noise, original noisy image passes through global bridging, in addition negative noise pattern, just obtains clean image.
C, parameter regulation and optimization are carried out to the model, the training model simultaneously tests it；
The training and test of model can be divided into three steps:
Step a: construction loss function, loss function are, wherein I0What is indicated is the band of mode input
It makes an uproar image, IGWhat is indicated is the clean image of model output；
Step b: carrying out parameter regulation to each convolutional layer, select suitable optimizer training pattern, saves trained model power
Weight；
Step c: being loaded into trained Model Weight, is tested using test the set pair analysis model.
Innovative point of the invention includes: that (1) using deep learning realizes power equipment Infrared Image Denoising Algorithm.This hair
Bright that deep learning is applied in power equipment Infrared Image Denoising for the first time, trained deep learning model is in test phase
A propagated forward is needed, expends the time well below the conventional method based on iterative solution, meets Power System Intelligent, in real time
The demand of property.(2) Infrared Image Denoising question simplification is directly learnt noise rather than band by the model application residual error thought
It makes an uproar image.(3) recursive call and expansion convolution are applied in power equipment Infrared Image Denoising problem simultaneously for the first time, lower
Parameter amount under obtain better performance.
The present invention carries out the processing of power equipment Infrared Image Denoising using the denoising model that depth convolutional neural networks are established,
Preferably denoising effect can be achieved compared to conventional method, while recursive call and expansion convolution are applied to proposed depth
On convolutional neural networks, better denoising effect has been reached under identical parameter amount.The invention feature is strong innovation, practical
Property is strong, can be realized that power equipment Infrared Image Denoising effect is good, fireballing demand.
The above is only a specific embodiment of the present invention, but the design concept of the present invention is not limited to this, all to utilize this
Design makes a non-material change to the present invention, and should all belong to behavior that violates the scope of protection of the present invention.
Claims (8)
1. a kind of power equipment Infrared Image Denoising method based on deep learning, which comprises the following steps: including
Following steps:
A, the training sample pair of power equipment infrared image is constructed；
B, the power equipment Infrared Image Denoising model based on deep learning is built；
C, parameter regulation and optimization are carried out to the model, the training model simultaneously tests it.
2. a kind of power equipment Infrared Image Denoising method based on deep learning as described in claim 1, which is characterized in that
The step A the specific steps are under:
Step 1: carrying out normalizing using the infrared image of infrared thermoviewer acquisition power equipment, and to collected infrared image
Change processing obtains the clean infrared image of plural number；
Step 2: after by rotation, overturning or translation, by each clean infrared image random cropping at several lesser
Image block obtains enough clean image patterns；
Step 3: being obtained to every clean image pattern plus white Gaussian noise using clean image pattern as the label of model
Corresponding noisy image sample, obtains a plurality of training samples pair, and using the 70% of total sample pair as training set, 15% as verifying
Collection, 15% is used as test set.
3. a kind of power equipment Infrared Image Denoising method based on deep learning as described in claim 1, which is characterized in that
Under the concrete operations of the step B are:
The model is made of three parts, and first part is characterized extraction module, by a convolutional layer and a nonlinear activation
Layer is constituted；Second part be recursive call module, include three expansion convolutional layers, each expansion convolutional layer later immediately one criticize
Normalize layer and a nonlinear activation layer；Part III is image reconstruction module, is made of a convolutional layer；Using residual error
Thought, the learning model of the task.
4. a kind of power equipment Infrared Image Denoising method based on deep learning as claimed in claim 3, it is characterised in that:
In the characteristic extracting module, convolutional layer convolution kernel size is 3 × 3, and the weight distribution of convolution kernel initialization meets Gaussian Profile,
Input is noisy image, exports the characteristic pattern in 32 channels, and the activation primitive that nonlinear activation layer uses is ReLU.
5. a kind of power equipment Infrared Image Denoising method based on deep learning as claimed in claim 3, it is characterised in that:
In the recursive call module, the input for expanding convolutional layer is all the characteristic pattern in 32 channels, and exports the feature all for 32 channels
Figure, convolution kernel size is all 3 × 3, and the weight distribution of convolution kernel initialization meets Gaussian Profile, the expansion of first expansion convolutional layer
The rate of opening is 2, and the spreading rate of second expansion convolutional layer is 3, and it is 4 that third, which expands the spreading rate of convolutional layer, nonlinear activation layer
The activation primitive used is ReLU.
6. a kind of power equipment Infrared Image Denoising method based on deep learning as claimed in claim 3, it is characterised in that:
Described image is rebuild in module, and convolutional layer convolution kernel size is 3 × 3, and the weight distribution of convolution kernel initialization meets Gaussian Profile,
Input is the characteristic pattern in 32 channels, exports the noise pattern that is negative；It is described to be specifically using the thought of residual error: to add one in a model
Global bridge joint makes model be changed into the negative noise of study by learning clean image, and original noisy image passes through global bridging, in addition
Negative noise pattern just obtains clean image.
7. a kind of power equipment Infrared Image Denoising method based on deep learning as described in claim 1, which is characterized in that
The concrete operations of the step C are as follows:
Step a: construction loss function, loss function are, wherein I0What is indicated is the band of mode input
It makes an uproar image, IGWhat is indicated is the clean image of model output；
Step b: carrying out parameter regulation to each convolutional layer, select suitable optimizer training pattern, saves trained model power
Weight；
Step c: being loaded into trained Model Weight, is tested using test the set pair analysis model.
8. a kind of power equipment Infrared Image Denoising method based on deep learning as claimed in claim 7, it is characterised in that:
In the step a, using the error backpropagation algorithm training model, using adam optimizer, iteration optimization 100,000 in total
It is secondary, save the weight of model obtained by iteration optimization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811392020.2A CN109544476A (en) | 2018-11-21 | 2018-11-21 | A kind of power equipment Infrared Image Denoising method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811392020.2A CN109544476A (en) | 2018-11-21 | 2018-11-21 | A kind of power equipment Infrared Image Denoising method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109544476A true CN109544476A (en) | 2019-03-29 |
Family
ID=65849075
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811392020.2A Pending CN109544476A (en) | 2018-11-21 | 2018-11-21 | A kind of power equipment Infrared Image Denoising method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109544476A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110596774A (en) * | 2019-09-09 | 2019-12-20 | 中国电子科技集团公司第十一研究所 | Method and device for infrared detection of submarine |
| CN110782406A (en) * | 2019-10-15 | 2020-02-11 | 深圳大学 | Image denoising method and device based on information distillation network |
| CN111881927A (en) * | 2019-05-02 | 2020-11-03 | 三星电子株式会社 | Electronic device and image processing method thereof |
| CN113687326A (en) * | 2021-07-13 | 2021-11-23 | 广州杰赛科技股份有限公司 | Vehicle-mounted radar echo noise reduction method, device, equipment and medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106204468A (en) * | 2016-06-27 | 2016-12-07 | 深圳市未来媒体技术研究院 | A kind of image de-noising method based on ReLU convolutional neural networks |
| CN106408522A (en) * | 2016-06-27 | 2017-02-15 | 深圳市未来媒体技术研究院 | Image de-noising method based on convolution pair neural network |
| US20180240219A1 (en) * | 2017-02-22 | 2018-08-23 | Siemens Healthcare Gmbh | Denoising medical images by learning sparse image representations with a deep unfolding approach |
-
2018
- 2018-11-21 CN CN201811392020.2A patent/CN109544476A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106204468A (en) * | 2016-06-27 | 2016-12-07 | 深圳市未来媒体技术研究院 | A kind of image de-noising method based on ReLU convolutional neural networks |
| CN106408522A (en) * | 2016-06-27 | 2017-02-15 | 深圳市未来媒体技术研究院 | Image de-noising method based on convolution pair neural network |
| US20180240219A1 (en) * | 2017-02-22 | 2018-08-23 | Siemens Healthcare Gmbh | Denoising medical images by learning sparse image representations with a deep unfolding approach |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111881927A (en) * | 2019-05-02 | 2020-11-03 | 三星电子株式会社 | Electronic device and image processing method thereof |
| US11257189B2 (en) | 2019-05-02 | 2022-02-22 | Samsung Electronics Co., Ltd. | Electronic apparatus and image processing method thereof |
| US11861809B2 (en) | 2019-05-02 | 2024-01-02 | Samsung Electronics Co., Ltd. | Electronic apparatus and image processing method thereof |
| CN110596774A (en) * | 2019-09-09 | 2019-12-20 | 中国电子科技集团公司第十一研究所 | Method and device for infrared detection of submarine |
| CN110782406A (en) * | 2019-10-15 | 2020-02-11 | 深圳大学 | Image denoising method and device based on information distillation network |
| CN110782406B (en) * | 2019-10-15 | 2022-10-11 | 深圳大学 | Image denoising method and device based on information distillation network |
| CN113687326A (en) * | 2021-07-13 | 2021-11-23 | 广州杰赛科技股份有限公司 | Vehicle-mounted radar echo noise reduction method, device, equipment and medium |
| CN113687326B (en) * | 2021-07-13 | 2024-01-05 | 广州杰赛科技股份有限公司 | Vehicle-mounted radar echo noise reduction method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109544476A (en) | A kind of power equipment Infrared Image Denoising method based on deep learning | |
| CN109583344B (en) | A kind of power equipment Infrared Image Processing Method based on deep learning | |
| CN112733725B (en) | Hyperspectral image change detection method based on multistage cyclic convolution self-coding network | |
| CN109842373B (en) | Photovoltaic array fault diagnosis method and device based on space-time distribution characteristics | |
| CN109873610B (en) | Photovoltaic array fault diagnosis method based on IV characteristic and depth residual error network | |
| Li et al. | An unmanned inspection system for multiple defects detection in photovoltaic plants | |
| US11251749B2 (en) | Methods and systems for fault detection, diagnosis and localization in solar panel network | |
| Li et al. | Intelligent fault pattern recognition of aerial photovoltaic module images based on deep learning technique | |
| CN111695620B (en) | Method and system for detecting and correcting time sequence abnormal data of power system | |
| CN112200178B (en) | Transformer substation insulator infrared image detection method based on artificial intelligence | |
| CN110728195B (en) | Power quality disturbance detection method based on YOLO algorithm | |
| CN111680788A (en) | Equipment fault diagnosis method based on deep learning | |
| CN111008641B (en) | Power transmission line tower external force damage detection method based on convolutional neural network | |
| CN113283580A (en) | Automatic fault detection method for solar cell panel | |
| CN111191406A (en) | Method for determining an electrical model of a string of photovoltaic modules, diagnostic method and device related thereto | |
| Ding et al. | Fault detection of photovoltaic array based on Grubbs criterion and local outlier factor | |
| Xu et al. | Deep learning in photovoltaic power generation forecasting: Cnn-lstm hybrid neural network exploration and research | |
| CN105374029A (en) | Segmenting method and system of transformer substation equipment infrared image interest areas | |
| Tao et al. | Electric insulator detection of UAV images based on depth learning | |
| CN116754213A (en) | Electric gate valve fault diagnosis method, device and equipment based on strong noise background | |
| CN116452574A (en) | Gap detection method, system and storage medium based on improved YOLOv7 | |
| CN112861670B (en) | Transmission line hardware detection method and system | |
| Karakose et al. | A Novel Reconfiguration Method Using Image Processing Based Moving Shadow Detection, Optimization, and Analysis for PV Arrays. | |
| CN113067805B (en) | Internet of things weak electromagnetic interference attack detection method and system based on edge calculation | |
| CN112001956A (en) | CNN-based schlieren strong laser far-field focal spot measurement image denoising method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190329 |