5G Edge Cloud Power Real-time Inspection Technology Based on YOLOV4-Tiny

2021 ◽  
Author(s):  
Jiaqi Song ◽  
Jing Li ◽  
Di Wu ◽  
Guangye Li ◽  
Jiaxin Zhang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
High Efficiency ◽  
Real Data ◽  
Detection Algorithm ◽  
Power Line ◽  
Edge Computing ◽  
Inspection Method ◽  
Computing Architecture ◽  
Low Efficiency

Power line corridor inspection plays a vital role in power system safe operation, traditional human inspection’s low efficiency makes the novel inspection method requiring high precision and high efficiency. Combined with the current deep learning target detection algorithm based on high accuracy and strong real-time performance, this paper proposes a YOLOV4-Tiny based drone real-time power line inspection method. The 5G and edge computing technology are combined properly forming a complete edge computing architecture. The UAV is treated as an edge device with a YOLOV4-Tiny deep- learning-based object detection model and AI chip on board. Extensive experiments on real data demonstrate the 5G and Edge computing architecture could satisfy the demands of real-time power inspection, and the intelligence of the whole inspection improved significantly.

scholarly journals REAL-TIME POWERLINE CORRIDOR INSPECTION BY EDGE COMPUTING OF UAV LIDAR DATA

2019 ◽  
Vol XLII-2/W13 ◽  
pp. 547-551 ◽  
Author(s):  
S. Pu ◽  
L. Xie ◽  
M. Ji ◽  
Y. Zhao ◽  
W. Liu ◽  
...  
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Point Clouds ◽  
Power Line ◽  
Edge Computing ◽  
Ground Vegetation ◽  
Lidar Data ◽  
Type I ◽  
Industry Standards ◽  
Potential Risks

<p><strong>Abstract.</strong> This paper presents an innovative power line corridor inspection approach using UAV LiDAR edge computing and 4G real real-time transmission. First, sample point clouds of power towers are manually classified and decomposed into components according to five mainstream tower types: T type, V type, n type, I type and owl head type. A deep learning AI agent, named “Tovos Age Agent” internally, is trained by supervised deep learning the sample data sets under a 3D CNN framework. Second, laser points of power line corridors are simultaneously classified into Ground, Vegetation, Tower, Cable, and Building types using semantic feature constraints during the UAV-borne LiDAR acquisition process, and then tower types are further recognized by Tovos Agent for strain span separation. Spatial and topological relations between Cable points and other types are analyzed according to industry standards to identify potential risks at the same time. Finally, all potential risks are organized as industry standard reports and transmitted onto central server via 4G data link, so that maintenance personal can be notified the risks as soon as possible. Tests on LiDAR data of 1000&amp;thinsp;KV power line show the promising results of the proposed method.</p>

scholarly journals Intelligent Highway Traffic Detection Algorithm based on Deep Learning

2021 ◽  
Author(s):  
ming ji ◽  
Chuanxia Sun ◽  
Yinglei Hu
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Traffic Congestion ◽  
Road Network ◽  
Vehicle Detection ◽  
Detection Algorithm ◽  
Recognition Algorithm ◽  
Traffic Data ◽  
Vehicle Recognition ◽  
Environment Recognition

Abstract In order to solve the increasingly serious traffic congestion problem, an intelligent transportation system is widely used in dynamic traffic management, which effectively alleviates traffic congestion and improves road traffic efficiency. With the continuous development of traffic data acquisition technology, it is possible to obtain real-time traffic data in the road network in time. A large amount of traffic information provides a data guarantee for the analysis and prediction of road network traffic state. Based on the deep learning framework, this paper studies the vehicle recognition algorithm and road environment discrimination algorithm, which greatly improves the accuracy of highway vehicle recognition. Collect highway video surveillance images in different environments, establish a complete original database, build a deep learning model of environment discrimination, and train the classification model to realize real-time environment recognition of highway, as the basic condition of vehicle recognition and traffic event discrimination, and provide basic information for vehicle detection model selection. To improve the accuracy of road vehicle detection, the vehicle target labeling and sample preprocessing of different environment samples are carried out. On this basis, the vehicle recognition algorithm is studied, and the vehicle detection algorithm based on weather environment recognition and fast RCNN model is proposed. Then, the performance of the vehicle detection algorithm described in this paper is verified by comparing the detection accuracy differences between different environment dataset models and overall dataset models, different network structures and deep learning methods, and other methods.

scholarly journals Real-Time Vehicle Detection Algorithm Based on Vision and Lidar Point Cloud Fusion

2019 ◽  
Vol 2019 ◽  
pp. 1-9 ◽  
Author(s):  
Hai Wang ◽  
Xinyu Lou ◽  
Yingfeng Cai ◽  
Yicheng Li ◽  
Long Chen
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Autonomous Vehicles ◽  
Point Cloud ◽  
Vehicle Detection ◽  
Detection Algorithm ◽  
Detection Methods ◽  
Depth Information ◽  
Detection Accuracy ◽  
Classification Rate

Vehicle detection is one of the most important environment perception tasks for autonomous vehicles. The traditional vision-based vehicle detection methods are not accurate enough especially for small and occluded targets, while the light detection and ranging- (lidar-) based methods are good in detecting obstacles but they are time-consuming and have a low classification rate for different target types. Focusing on these shortcomings to make the full use of the advantages of the depth information of lidar and the obstacle classification ability of vision, this work proposes a real-time vehicle detection algorithm which fuses vision and lidar point cloud information. Firstly, the obstacles are detected by the grid projection method using the lidar point cloud information. Then, the obstacles are mapped to the image to get several separated regions of interest (ROIs). After that, the ROIs are expanded based on the dynamic threshold and merged to generate the final ROI. Finally, a deep learning method named You Only Look Once (YOLO) is applied on the ROI to detect vehicles. The experimental results on the KITTI dataset demonstrate that the proposed algorithm has high detection accuracy and good real-time performance. Compared with the detection method based only on the YOLO deep learning, the mean average precision (mAP) is increased by 17%.

scholarly journals Real-Time Route Search by Locations

2020 ◽  
Vol 34 (01) ◽  
pp. 574-581
Author(s):  
Lisi Chen ◽  
Shuo Shang ◽  
Tao Guo
Keyword(s):  
Real Time ◽  
Data Streams ◽  
High Efficiency ◽  
Route Planning ◽  
Real Data ◽  
Location Based Services ◽  
Good Time ◽  
Efficient Route ◽  
Route Search ◽  
Insight Into

With the proliferation of GPS-based data (e.g., routes and trajectories), it is of great importance to enable the functionality of real-time route search and recommendations. We define and study a novel Continuous Route-Search-by-Location (C-RSL) problem to enable real-time route search by locations for a large number of users over route data streams. Given a set of C-RSL queries where each query q contains a set of places q.O to visit and a threshold q.θ, we continuously feed each query q with routes that has similarity to q.O no less than q.θ. We also extend our proposal to support top-k C-RSL problem where each query continuously maintains k most similar routes. The C-RSL problem targets a variety of applications, including real-time route planning, ridesharing, and other location-based services that have real-time demand. To enable efficient route matching on a large number of C-RSL queries, we develop novel parallel route matching algorithms with good time complexity. Extensive experiments with real data offer insight into the performance of our algorithms, indicating that our proposal is capable of achieving high efficiency and scalability.

Design and Implementation of Cholecystolithiasis Automatically Perfusion-Dissolved Instrument

2012 ◽  
Vol 241-244 ◽  
pp. 562-565
Author(s):  
Tian Shui Zhou ◽  
Zhen Bao Ling ◽  
Jun Wang
Keyword(s):  
Medical Treatment ◽  
Real Time ◽  
Pressure Sensor ◽  
High Efficiency ◽  
Constant Pressure ◽  
Pressure Data ◽  
Design And Implementation ◽  
Peristaltic Pumps ◽  
Low Efficiency ◽  
Control Quantity

In view of the difficulty to operate, low efficiency, and easy to leak of the manual perfusion method for dissolving cholecystolithiasis, an automatically perfusion-dissolved instrument is designed. By means of collecting and analysing the pressure data being outputted from pressure sensor in real time, and adjusting speed and direction of peristaltic pumps automatically to control quantity of dissolvent, so as to achieve the goal of litholysis under constant pressure in the gallbladder. The instrument has function of automatic heating and controlling temperature. Double controllers are used in the design, slave computer controls the running of each module of the instrument, and master computer is used to real-time display operation state and set parameters. Experimental results indicate that the system is stable and reliable, and with high efficiency for dissolving cholecystolithiasis, which provides a new and effective method for cholelithiasis medical treatment.

scholarly journals Distributed Secure Edge Computing Architecture Based on Blockchain for Real-Time Data Integrity in IoT Environments

Actuators ◽  
2021 ◽  
Vol 10 (8) ◽  
pp. 197
Author(s):  
Rongxu Xu ◽  
Lei Hang ◽  
Wenquan Jin ◽  
Dohyeun Kim
Keyword(s):  
Real Time ◽  
Strong Dependence ◽  
Data Integrity ◽  
Data Encryption ◽  
Edge Computing ◽  
Time Data ◽  
Computing Architecture ◽  
Security Issues ◽  
Multiple Data ◽  
Cloud Servers

The traditional cloud-based Internet of Things (IoT) architecture places extremely high demands on computers and storage on cloud servers. At the same time, the strong dependence on centralized servers causes major trust problems. Blockchain provides immutability, transparency, and data encryption based on safety to solve these problems of the IoT. In this paper, we present a distributed secure edge computing architecture using multiple data storages and blockchain agents for the real-time context data integrity in the IoT environment. The proposed distributed secure edge computing architecture provides reliable access and an unlimited repository for scalable and secure transactions. The architecture eliminates traditional centralized servers using an edge computing framework that represents cloud computing for computer and security issues. Also, blockchain-based edge computing-compatible IoT design is supported to achieve the level of security and scalability required for data integrity. Furthermore, we present the blockchain agent to provide internetworking between blockchain networks and edge computing. For experimenting with the proposed architecture in the IoT environment, we implement and perform a concrete IoT environment based on the EdgeX framework and Hyperledger Fabric. The evaluation results are collected by measuring the performance of the edge computing and blockchain platform based on service execution time to verify the proposed architecture in the IoT environment.

scholarly journals POWER LINE EXTRACTION AND ANALYSIS BASED ON LIDAR

2020 ◽  
Vol XLII-3/W10 ◽  
pp. 91-96
Author(s):  
X. H. Chen ◽  
J. Q. Dai ◽  
Y. R. He ◽  
W. W. Ma
Keyword(s):  
Transmission Line ◽  
Point Cloud ◽  
Spatial Information ◽  
Grid Method ◽  
Power Line ◽  
Airborne Lidar ◽  
Practical Significance ◽  
Inspection Method ◽  
Low Efficiency ◽  
Line Point

Abstract. The traditional electrical power line inspection method has the disadvantages of high labor intensity, low efficiency and long cycle of re-inspection. Airborne LiDAR can quickly obtain the high-precision three-dimensional spatial information of transmission line, and the data which collected by it can make it possible to accurately detect the dangerous points.It is proposed to use the grid method to divide the data into multiple regions for the elevation histogram statistical method to obtain the power line point cloud at the complex mountainous terrain. In the non-ground point data, part of the vegetation point cloud is separated according to the point cloud dimension feature, and then the power line point and the pole point are distinguished according to the density characteristics of the point cloud so as to realize the point cloud classification of the transmission line corridor. On this basis, the power line safety distance detection is carried out on the power line points and vegetation points extracted by the classification, and the early warning analysis of the dangerous points of the transmission line tree barrier is completed. The experimental results show that the method can classify the acquired power line corridor point cloud and extract the complete power line, which effectively eliminates the hidden dangers and has certain practical significance.

Sign in / Sign up

Export Citation Format

Share Document