Font Size: a A A

UAV Line Survey Based On Route Planning And Image Recognition Technology

Posted on:2020-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:H Y ChouFull Text:PDF
GTID:2392330575988942Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
The power inspection drone is a new patrol operation and maintenance method for the power industry.It can complement helicopter inspections and manual inspections.It not only reduces the labor intensity of transmission line operation and maintenance personnel,but also improves the quality,efficiency and economic benefits.It is the direction of operation and maintenance management of transmission lines for UHV power grids in the future.This study relied on the support of the State Grid Yi Chun Power Supply Company's ?Automatic Inspection and Flight Test Project for UAV Lines?.The mission was to inspect the 110 kV Yi Bei transmission line in Yi Chun City.Among them,the focus is on the route planning technology and image recognition technology of the drone.The route planning of the power inspection drone is the basis for the autonomous flight execution of the drone.This paper studies the full coverage route planning and the tower route planning,and selects the convex polygon route planning as the method of this full coverage flight planning.And then,The Gauss-Kruger projection method is used to establish a plane coordinate system for the map projection method of the inspection mission,and the cattle farming method is used as the flight method for the drone inspection operation.The planning method of the tower inspection route is a conventional single-loop tower inspection procedure.The targets are insulator strings,ground wires,and wires.The results show that two route planning methods can complete the inspection task well and achieve maximum cost savings.The full coverage route plan selects a single frame image as the minimum processing unit.The ground station monitoring system completes the image sorting work.The geometric correction method of the image adopts image space coordinate transformation and grayscale resampling.In this experiment,HOG feature extraction method is used to extract features from the image,and the image is fused by weighted average method.The results show that the image can accurately identify the insulator strings,grounding wires and wires,and can fuse the local solid image of the full coverage flight.
Keywords/Search Tags:Unmanned helicopter, Image geometry correction, Image registration, Image feature extraction, Image fusion
PDF Full Text Request
Related items