CN116245871A - Power transmission line abnormal target detection method based on improved YOLOX algorithm - Google Patents
Power transmission line abnormal target detection method based on improved YOLOX algorithm Download PDFInfo
- Publication number
- CN116245871A CN116245871A CN202310329483.9A CN202310329483A CN116245871A CN 116245871 A CN116245871 A CN 116245871A CN 202310329483 A CN202310329483 A CN 202310329483A CN 116245871 A CN116245871 A CN 116245871A
- Authority
- CN
- China
- Prior art keywords
- abnormal
- feature
- transmission line
- algorithm
- yolox
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
- G06V10/225—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on a marking or identifier characterising the area
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- G06T2207/10004—Still image; Photographic 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]
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention relates to a transmission line abnormal target detection method based on an improved YOLOX algorithm, which comprises the following steps: 1) Inputting an abnormal image of a transmission line to be identified, preprocessing the abnormal image, adjusting the abnormal image to be a standard input image, and training based on an improved YOLOX algorithm; 2) Splitting the high resolution image into a plurality of low resolution images using a Focus model; 3) Extracting the characteristics of the input image by a characteristic selection module to serve as abnormal characteristics; 4) Continuously extracting abnormal features and dividing the abnormal features into three effective feature layers P3, P4 and P5; 5) Inputting the feature images of the effective feature layers into a PAFPN network for feature extraction, efficiently fusing the feature images of different feature layers, and outputting the feature images of different feature layers; 6) And detecting and positioning abnormal targets on the effective feature graphs of different feature layers. Compared with the prior art, the method has the advantages of high detection precision of the abnormal target, reduction of algorithm complexity and the like.
Description
Technical Field
The invention relates to a transmission line abnormal target detection method, in particular to a transmission line abnormal target detection method based on an improved YOLOX algorithm.
Background
With the development of scientific technology, the laying range of the power transmission line gradually covers all regions of the country, and the structural design forms of power equipment and related power components are also more and more complex. However, the electricity consumption required by daily life production of people is gradually increased, so that the probability of operation faults of the power supply and distribution lines is increased. If the safe and stable operation of the power system cannot be ensured, the daily life of people can be influenced, and serious social loss can be caused. Therefore, how to ensure the safe and reliable operation of the power supply and distribution line becomes a great challenge in the power transmission link.
By detecting targets in the power transmission line image, such as loose cables, damaged equipment and the like, the fault condition on the power transmission line can be timely found, and the occurrence of power interruption or safety accidents caused by the fault condition can be avoided. Meanwhile, the detection of the abnormal target can also improve the overhaul efficiency of the power transmission line, reduce the maintenance cost and improve the reliability and safety of the power system.
However, the traditional detection of the transmission line is realized by manual work, which has low efficiency and consumes manpower and material resources. Therefore, the research on the transmission line abnormal target detection algorithm based on deep learning has very important practical significance and application value. However, the following problems exist with using unmanned aerial vehicles for transmission line detection:
(1) During the acquisition of the power line image, the drone will traverse areas with different topographical features, such as: villages, farmlands, rivers, hills, glaciers, forests, roads, etc., which cause the acquired images to have constantly changing background information.
(2) Factors such as illumination and weather conditions (such as sunny days, overcast and rainy days and the like) cannot be determined in advance, so that contact network images acquired by equipment are seriously interfered by fog, haze and different illumination, and the detection difficulty is increased.
(3) Unmanned aerial vehicle needs the manual operation of electric wire netting staff, in order to avoid accidents such as collision, can't gather accurate fault location. Generally, the unmanned aerial vehicle has a wide cruise field of view, and thus the resolution of the photographed image is high. And most of abnormal targets on the power transmission line are hidden and easily ignored, and detection is easy to deviate.
(4) In the transmission line detection field, the physical size of an abnormal target shot by an unmanned aerial vehicle is smaller. Also, it is difficult for the neural network to extract sufficient target feature information due to complex background, occlusion, edge blurring, and lack of samples. At the same time, small target detection is also a complex problem and a popular research topic in computer vision, which is a challenging task.
(5) The existing detection algorithm based on deep learning is mainly aimed at detecting general targets, and when the detection algorithm is applied to a transmission line abnormal detection scene, the performance of the algorithm is obviously reduced, and the detection performance cannot meet the requirements of industrial application.
Therefore, how to realize a small target detection method capable of detecting an abnormality of a power transmission line in real time is a technical problem to be solved.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a transmission line abnormal target detection method based on an improved YOLOX algorithm.
The aim of the invention can be achieved by the following technical scheme:
a transmission line anomaly target detection method based on an improved YOLOX algorithm, the method comprising the steps of:
1) Inputting an abnormal image of a transmission line to be identified, preprocessing the abnormal image, adjusting the abnormal image into a standard input image, and training the preprocessed standard input image based on an improved YOLOX algorithm;
2) Splitting the high resolution image into a plurality of low resolution images using a Focus model;
3) Extracting features of the input image as abnormal features by a feature selection module;
4) Continuously extracting abnormal features and dividing the abnormal features into three effective feature layers P3, P4 and P5;
5) Inputting the feature images of the effective feature layers into a PAFPN network based on a new dual-channel attention mechanism to perform feature extraction, efficiently fusing the feature images of different feature layers, and outputting the feature images of different feature layers;
6) And detecting and positioning an abnormal target according to the effective feature graphs of different feature layers entering the detector.
Further, the abnormal images of the transmission line to be identified are obtained through carrying photographic equipment on the unmanned aerial vehicle, and are transmission line images of different positions, different weather and different illumination.
Further, the pretreatment specifically comprises: and carrying out normalization processing on the images, cutting all the images to 1368 multiplied by 912 resolution, marking an abnormal target by using labeme software, obtaining an abnormal data set of the power transmission line, and dividing the data set into a training set and a testing set.
Further, the abnormal data set of the power transmission line comprises a high-voltage tower bird nest data set and a self-explosion insulator data set.
Further, the 3) specifically includes:
301 Dividing the input feature map into C feature sub-maps, and obtaining N-dimensional vectors of each feature sub-map through a full connection layer;
302 Constraining the N-dimensional vector into [0,1] using a Softmax activation function, and then combining with the corresponding feature sub-map as a weighting weight;
303 Combining the C feature sub-maps into a new feature map containing all basic features of the abnormal target of the transmission line as abnormal features.
Further, the P3, P4 and P5 are respectively used for detecting a small target, a medium target and a large target; the continuous extraction of the abnormal features uses a FReLU activation function to optimize, and is expressed as follows:
Y＝max(x,T(x
where T (x) is a simple and efficient spatial context feature extractor and x represents the outlier feature.
Further, the 5) specifically includes: the input is divided into two branches along the channel direction, attention patterns are generated respectively, and the attention patterns are combined after being processed, so that new feature patterns of different feature layers are obtained.
Further, the feature map of the new different feature layer includes cross-channel information, direction information and position information of the abnormal target.
An electronic device comprising a memory having a computer program stored thereon and a processor that when executing the program implements a method as described above.
A computer readable storage medium having stored thereon a computer program which when executed by a processor implements a method as described above.
Compared with the prior art, the invention has the following beneficial effects:
1. according to the invention, through an improved YOLOX algorithm, a feature selection module is added in a backbone network, important feature information in an effective feature map is extracted by distributing weights, and practical features are reserved for subsequent feature map extraction, so that the detection precision of an abnormal target in a power transmission line is further improved. And introducing SimAM into the residual edge for processing by optimizing the residual structure in the feature extraction network, and improving the target feature representation and the capability of the network for detecting the abnormal small target of the transmission line by optimizing the neuron on the residual edge.
2. The present invention uses FReLU to enhance the nonlinear representation of abnormally small objects in a transmission line by improving the activation function of the YOLOX network. The FReLU function is used in the residual error module of the feature layer P3, so that the capability of detecting a small target is improved, excessive network overhead is prevented, and the algorithm complexity is reduced.
3. The invention provides a novel attention mechanism characteristic enhancement network (AM-FPNet) based on a dual-channel attention idea, and the attention mechanism is embedded into a shallow characteristic enhancement diagram of an FPN module, namely P3, so as to improve characteristic information of an abnormally small target in a power transmission line.
Drawings
FIG. 1 is a block diagram of the YOLOX algorithm prior to modification of the present invention;
FIG. 2 is a schematic diagram of a feature selection module according to the present invention;
fig. 3 is a schematic diagram of a residual network after SimAM optimization according to the present invention;
FIG. 4 is a schematic diagram of an improved feature extraction network AM-FPNet of the present invention;
fig. 5 is a schematic diagram of a feature enhanced network incorporating an attention mechanism of the present invention.
Detailed Description
The invention will now be described in detail with reference to the drawings and specific examples. The present embodiment is implemented on the premise of the technical scheme of the present invention, and a detailed implementation manner and a specific operation process are given, but the protection scope of the present invention is not limited to the following examples.
Examples
As shown in fig. 1, the structure diagram of the YOLOX algorithm before improvement is mainly aimed at detecting general targets, and when the YOLOX algorithm is applied to a transmission line abnormality detection scene, the performance of the algorithm is obviously reduced, and the detection performance cannot meet the requirements of industrial application.
As shown in fig. 2 to 5, a transmission line abnormal target detection method based on an improved YOLOX algorithm is characterized in that the method comprises the following steps:
1) And inputting an abnormal image of the transmission line to be identified, preprocessing the abnormal image, adjusting the abnormal image to be a standard input image, and training the preprocessed standard input image based on an improved YOLOX algorithm.
The method is characterized in that the abnormal image of the transmission line to be identified is obtained through carrying photographic equipment on the unmanned aerial vehicle, an operator controls the unmanned aerial vehicle to fly on the ground, and the transmission line images of different positions, different weather and different illumination are collected. The pretreatment is specifically as follows: and carrying out normalization processing on the images, cutting all the images to 1368 multiplied by 912 resolution, marking an abnormal target by using labeme software (data marking software), obtaining an abnormal data set of the power transmission line, including a bird nest data set of a high-voltage tower and a self-explosion insulator data set, and dividing the data set into a training set and a testing set. N images are randomly selected as input images at a time during the training process.
2) The Focus model is used to split the high resolution image into a plurality of low resolution images.
3) Features of the input image are extracted as abnormal features by a feature selection module.
301 Dividing the input feature map into C feature sub-maps, and obtaining N-dimensional vectors of each feature sub-map through a full connection layer;
302 Constraining the N-dimensional vector into [0,1] using a Softmax activation function, and then combining with the corresponding feature sub-map as a weighting weight;
303 Combining the C feature sub-maps into a new feature map containing all basic features of the abnormal target of the transmission line as abnormal features.
4) The abnormal features are continuously extracted and divided into three effective feature layers P3, P4 and P5.
Along with the increase of the network layer number, the sizes of the extracted abnormal feature graphs are different, and the P3, the P4 and the P5 are respectively used for detecting a small target, a medium target and a large target; since the targets on the transmission line are mostly small targets, the position of the sppbottcleck structure is adjusted from the P5 layer to the P3 layer. Compared with the original backbone network, the adjusted network can reduce part of network parameters and improve the perception field of view of small targets. In addition, in order to improve the detection performance of the network on the small target, the continuous extraction of the abnormal features uses a FReLU activation function to optimize, which is expressed as:
Y＝max(x,T(x
where T (x) is a simple and efficient spatial context feature extractor and x represents the outlier feature.
5) The novel dual-channel attention mechanism is based on the fact that feature images of effective feature layers are input into a PAFPN network to conduct feature extraction, feature images of different feature layers are fused efficiently, and feature images of different feature layers are output.
Dividing the input into two branches along the channel direction, respectively generating attention patterns by utilizing the relation among the channels, and capturing the direction sensing information and the position sensing information of the target in a pooling mode. In the channel attention sub-graph, a convolution kernel is used to learn the correspondence between channels and multiply the input by the value after calculation by the Sigmoid function. Furthermore, to better learn detailed target information, different pooling kernels are used simultaneously to capture remote dependencies along both spatial directions. Then, sub-graphs of different sizes are combined and converted to have the same number of channels. Finally, a 1×1 convolution operation is performed on the obtained product, and the obtained product is multiplied by an input to be used as an attention weight. Merging after obtaining the two attention subgraphs to obtain new feature graphs of different feature layers; the feature map of the new different feature layers comprises cross-channel information, direction information and position information of the abnormal target. The new dual-channel attention mechanism is embedded into the P3 shallow feature enhancement map to improve the feature information of the abnormally small targets in the transmission line.
6) And detecting and positioning an abnormal target according to the effective feature graphs of different feature layers entering the detector.
Particularly, all parameters of the invention are deep network parameters except for the super parameters indicating the value, and the Adam optimizer method is used for autonomous learning optimization.
An electronic device comprising a memory having a computer program stored thereon and a processor that when executing the program implements a method as described above.
A computer readable storage medium having stored thereon a computer program which when executed by a processor implements a method as described above.
The foregoing describes in detail preferred embodiments of the present invention. It should be understood that numerous modifications and variations can be made in accordance with the concepts of the invention by one of ordinary skill in the art without undue burden. Therefore, all technical solutions which can be obtained by logic analysis, reasoning or limited experiments based on the prior art by the person skilled in the art according to the inventive concept shall be within the scope of protection defined by the claims.
Claims (10)
1. The transmission line abnormal target detection method based on the improved YOLOX algorithm is characterized by comprising the following steps of:
1) Inputting an abnormal image of a transmission line to be identified, preprocessing the abnormal image, adjusting the abnormal image into a standard input image, and training the preprocessed standard input image based on an improved YOLOX algorithm;
2) Splitting the high resolution image into a plurality of low resolution images using a Focus model;
3) Extracting features of the input image as abnormal features by a feature selection module;
4) Continuously extracting abnormal features and dividing the abnormal features into three effective feature layers P3, P4 and P5;
5) Inputting the feature images of the effective feature layers into a PAFPN network based on a new dual-channel attention mechanism to perform feature extraction, efficiently fusing the feature images of different feature layers, and outputting the feature images of different feature layers;
6) And detecting and positioning an abnormal target according to the effective feature graphs of different feature layers entering the detector.
2. The method for detecting the abnormal target of the power transmission line based on the improved YOLOX algorithm according to claim 1, wherein the abnormal image of the power transmission line to be identified is obtained by carrying a photographic device on an unmanned aerial vehicle and is a power transmission line image of different positions, different weather and different illumination.
3. The transmission line abnormal target detection method based on the improved YOLOX algorithm of claim 1, wherein the preprocessing specifically comprises the following steps: and carrying out normalization processing on the images, cutting all the images to 1368 multiplied by 912 resolution, marking an abnormal target by using labeme software, obtaining an abnormal data set of the power transmission line, and dividing the data set into a training set and a testing set.
4. A transmission line anomaly target detection method based on a modified YOLOX algorithm as claimed in claim 3, wherein the transmission line anomaly data set comprises a high voltage tower bird nest data set and a self-explosion insulator data set.
5. The transmission line abnormal target detection method based on the improved YOLOX algorithm of claim 1, wherein the 3) specifically comprises:
301 Dividing the input feature map into C feature sub-maps, and obtaining N-dimensional vectors of each feature sub-map through a full connection layer;
302 Constraining the N-dimensional vector into [0,1] using a Softmax activation function, and then combining with the corresponding feature sub-map as a weighting weight;
303 Combining the C feature sub-maps into a new feature map containing all basic features of the abnormal target of the transmission line as abnormal features.
6. The transmission line abnormal target detection method based on the improved YOLOX algorithm according to claim 1, wherein P3, P4 and P5 are used for detecting a small target, a medium target and a large target respectively; the continuous extraction of the abnormal features uses a FReLU activation function to optimize, and is expressed as follows:
Y＝max(x,T(x
where T (x) is a simple and efficient spatial context feature extractor and x represents the outlier feature.
7. The transmission line abnormal target detection method based on the improved YOLOX algorithm of claim 1, wherein the 5) specifically comprises: the input is divided into two branches along the channel direction, attention patterns are generated respectively, and the attention patterns are combined after being processed, so that new feature patterns of different feature layers are obtained.
8. The method for detecting abnormal targets of power transmission lines based on the improved YOLOX algorithm of claim 7, wherein the feature map of the new different feature layers includes cross-channel information, direction information and position information of the abnormal targets.
9. An electronic device comprising a memory and a processor, the memory having stored thereon a computer program, characterized in that the processor, when executing the program, implements the method according to any of claims 1-8.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310329483.9A CN116245871A (en) | 2023-03-30 | 2023-03-30 | Power transmission line abnormal target detection method based on improved YOLOX algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310329483.9A CN116245871A (en) | 2023-03-30 | 2023-03-30 | Power transmission line abnormal target detection method based on improved YOLOX algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116245871A true CN116245871A (en) | 2023-06-09 |
Family
ID=86624370
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310329483.9A Pending CN116245871A (en) | 2023-03-30 | 2023-03-30 | Power transmission line abnormal target detection method based on improved YOLOX algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116245871A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116665016A (en) * | 2023-06-26 | 2023-08-29 | 中国科学院长春光学精密机械与物理研究所 | Single-frame infrared dim target detection method based on improved YOLOv5 |
| CN117036361A (en) * | 2023-10-10 | 2023-11-10 | 云南大学 | Power grid transmission line smoke detection method, system, electronic equipment and medium |
-
2023
- 2023-03-30 CN CN202310329483.9A patent/CN116245871A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116665016A (en) * | 2023-06-26 | 2023-08-29 | 中国科学院长春光学精密机械与物理研究所 | Single-frame infrared dim target detection method based on improved YOLOv5 |
| CN116665016B (en) * | 2023-06-26 | 2024-02-23 | 中国科学院长春光学精密机械与物理研究所 | Single-frame infrared dim target detection method based on improved YOLOv5 |
| CN117036361A (en) * | 2023-10-10 | 2023-11-10 | 云南大学 | Power grid transmission line smoke detection method, system, electronic equipment and medium |
| CN117036361B (en) * | 2023-10-10 | 2024-02-20 | 云南大学 | Power grid transmission line smoke detection method, system, electronic equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111428748B (en) | HOG feature and SVM-based infrared image insulator identification detection method | |
| Liu et al. | High precision detection algorithm based on improved RetinaNet for defect recognition of transmission lines | |
| CN116245871A (en) | Power transmission line abnormal target detection method based on improved YOLOX algorithm | |
| Liu et al. | Insulator detection in aerial images based on faster regions with convolutional neural network | |
| CN111582092B (en) | Pedestrian abnormal behavior detection method based on human skeleton | |
| Song et al. | Analysis on the impact of data augmentation on target recognition for UAV-based transmission line inspection | |
| Liao et al. | Fusion of infrared-visible images in UE-IoT for fault point detection based on GAN | |
| Wang et al. | Study on the method of transmission line foreign body detection based on deep learning | |
| Li et al. | Transmission line detection in aerial images: An instance segmentation approach based on multitask neural networks | |
| CN112419333A (en) | Remote sensing image self-adaptive feature selection segmentation method and system | |
| CN114463257A (en) | Power equipment infrared image detection method and system based on deep learning | |
| Yang et al. | Research on subway pedestrian detection algorithms based on SSD model | |
| CN116846059A (en) | Edge detection system for power grid inspection and monitoring | |
| Dai | Uncertainty-aware accurate insulator fault detection based on an improved YOLOX model | |
| CN113988164B (en) | Lightweight point cloud target detection method for representative point self-attention mechanism | |
| Zhao et al. | FSDF: A high-performance fire detection framework | |
| US20210182615A1 (en) | Alexnet-based insulator self-explosion recognition method | |
| CN113536944A (en) | Distribution line inspection data identification and analysis method based on image identification | |
| CN113569672A (en) | Lightweight target detection and fault identification method, device and system | |
| CN111931577A (en) | Intelligent inspection method for specific foreign matters of power grid line | |
| CN116740516A (en) | Target detection method and system based on multi-scale fusion feature extraction | |
| Zhang et al. | Defect detection of anti-vibration hammer based on improved faster R-CNN | |
| Dongtao et al. | Traffic sign detection method of improved ssd based on deep learning | |
| Hao et al. | Based on Surf feature extraction and insulator damage identification for capsule networks | |
| Wang et al. | The image recognition based on restricted Boltzmann machine and deep learning framework |
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 |