Automated deep learning system for power line inspection image analysis and processing: architecture and design issues

The continuous growth in the scale of unmanned aerial vehicle (UAV) applications in transmission line inspection has resulted in a corresponding increase in the demand for UAV inspection image processing. Owing to its excellent performance in computer vision, deep learning has been applied to UAV inspection image processing tasks such as power line identification and insulator defect detection. Despite their excellent performance, electric power UAV inspection image processing models based on deep learning face several problems such as a small application scope, the need for constant retraining and optimization, and high R&D monetary and time costs due to the black-box and scene data-driven characteristics of deep learning. In this study, an automated deep learning system for electric power UAV inspection image analysis and processing is proposed as a solution to the aforementioned problems. This system design is based on the three critical design principles of generalizability, extensibility, and automation. Pre-trained models, fine-tuning (downstream task adaptation), and automated machine learning, which are closely related to these design principles, are reviewed. In addition, an automated deep learning system architecture for electric power UAV inspection image analysis and processing is presented. A prototype system was constructed and experiments were conducted on the two electric power UAV inspection image analysis and processing tasks of insulator self-detonation and bird nest recognition. The models constructed using the prototype system achieved 91.36% and 86.13% mAP for insulator self-detonation and bird nest recognition, respectively. This demonstrates that the system design concept is reasonable and the system architecture feasible .

Prototype Monitoring System for Power Line Inspection by Means of a PandaBoard

The national energy supplier （Eskom in South Africa） supplies electricity through thousands-of-kilometers of overhead power lines. The current methods of inspection of these overhead power lines are infrequent and expensive. In this paper, the authors present the development of a prototype monitoring system for power line inspection in South Africa. The developed prototype monitoring system collects data （information） from the overhead power lines, is remotely accessible and fits into a power line robot. The prototype monitoring system makes use ofa PandaBoard （SBC） with GPS receiver and 5 MP camera to collect data. Hardware fatigue is the biggest problem faced on the overhead power lines and is captured by means of the 5 MP camera and is displayed on a website hosted by the PandaBoard via Wi-Fi. The monitoring system has low power consumption, is light weight, compact and easily collects data. The data obtained from the prototype monitoring system was satisfactory and provides an improved solution for monitoring power lines for Eskom in South Africa.

On-line full scan inspection of particle size and shape using digital image processing

An on-line full scan inspection system is developed for particle size analysis. A particle image is first obtained through optical line scan technology and is then analyzed using digital image processing. The system is composed of a particle separation module, an image acquisition module, an image processing module, and an electric control module. Experiments are carried out using non-uniform 0.1 mm particles. The main advantage of this system consists of a full analysis of particles without any overlap or miss, thus improving the Area Scan Charge Coupled Device （CCD） acquisition problems. Particle size distribution, roundness, and sphericity can be obtained using the system with a deviation of repeated precision of around ±1%. The developed system is shown to be also convenient and versatile for any particle size and shape for academic and industrial users.

Variable universe fuzzy control of walking stability for flying‐walking power line inspection robot based on multi‐work conditions

To address complex work conditions incredibly challenging to the stability of power line inspection robots,we design a walking mechanism and propose a variable universe fuzzy control(VUFC)method based on multi‐work conditions for flying‐walking power line inspection robots(FPLIRs).The contributions of this paper are as follows:(1)A flexible pressing component is designed to improve the adaptability of the FPLIR to the ground line slope.(2)The influence of multi‐work conditions on the FPLIR's walking stability is quantified using three condition parameters(i.e.,slope,slipping degree and swing angle),and their measurement methods are proposed.(3)The VUFC method based on the condition parameters is proposed to improve the walking stability of the FPLIR.Finally,the effect of the VUFC method on walking stability of the FPLIR is teste.The experimental results show that the maximum climbing angle of the FPLIR reaches 29.1°.Compared with the constant pressing force of 30 N,the average value of slipping degree is 0.93°,increasing by 35%.The maximum and average values of robot's swing angle are reduced by 46%and 54%,respectively.By comparing with fuzzy control,the VUFC can provide a more reasonable pressing force while maintaining the walking stability of the FPLIR.The proposed walking mechanism and the VUFC method significantly improve the stability of the FPLIR,providing a reference for structural designs and stability controls of inspection robots.