Research On Identification Of High Voltage Transmission Line Based On Aerial Image Of UAV

Power lines are an important part of the power grid,which plays an important role in the process of electric energy transmission,and is related to the safety and stability of the power grid.China’s power grid is not only large in scale and complex in structure but also has the problems of low efficiency and great limitations of traditional manual inspection.The rapid development of the information technology and civil unmanned aerial vehicles(UAV)industry has provided convenience for inspection of transmission lines.Its patrol inspection technology and method also meet the increasingly high requirements of intelligence and informationization.Based on the video image sequence collected during inspection,combined with machine vision,this thesis has carried out the research of transmission line identification and tracking.In this thesis,firstly the image of transmssion line is preprocessed.Combined with the characteristics of power transmission lines in aerial imagery,the image is converted from color space to grayscale image,and then the grayscale nonlinear transformation is performed to achieve the enhancement of image contrast.The denoising methods are analyzed and compared,and the improved Wiener filtering is used to reduce the noise of aerial images.Then this thesis introduces and analyzes Hough class straight line detection and a line segment detector(LSD)based on gradient method.Aiming at the defects that the two algorithms missed detection when detecting transmission lines,this thesis studies a power line detection algorithm based on principal curvature and principal direction.This algorithm uses the characteristic that the principal curvature difference between the background pixel and the linear pixel is large to screen potential linear pixels,construct a linear support area in the direction of the pixel,and realize the detection of the transmission line.The algorithm does not depend on the gradient edge segmentation of the image,which can effectively avoid the interference caused by the detection of the two-sided step edge feature based on the gradient side to the detection of the closer transmission lines.After detecting the potential transmission lines in the image,the target lines is screened out,and then the transmission lines are fitted and output.Finally,this thesis analyzes the movement of the transmission line target in the image when the UAV patrols the transmission line,and attempts to construct a simplified model of the transmission line target.At the same time,this thesis researches on the tracking method of Kalman filter,through Unscented Kalman Filtering(UKF)to track the region of interest(ROI)of the power line target,and realize the identification and tracking of the power line in the video image sequence.Simulation of the algorithm in this thesis is carried out on the MATLAB.The experiment proves that the method proposed in this thesis can effectively realize the identification and tracking of the transmission lines in the video image sequence.