Distribution Line Detection In Complex Background Using UAVs

The low-voltage distribution network is a power supply distribution network facing the user side.The reliability of the power supply distribution network is increasingly demanded by industrial and domestic power.Regular inspection and maintenance of the distribution lines are the main means to ensure the normal operation of the distribution network.At present,the distribution line inspection can only be done through manual line inspection,which can no longer meet the development requirements of the power grid.Due to the better efficiency,higher quality and lower cost of automatic patrol inspection of drones,it has gradually become the mainstream of research.Distribution line detection is an important part of the automatic tracking technology in the autonomous patrol inspection of drones.Therefore,this thesis studies the algorithm and implementation of distribution line detection in a complex background.The research is mainly divided into two parts,namely,the distribution line detection method in aerial images and the distribution line detection method in depth images.The aerial image part mainly studies the image preprocessing algorithm and distribution line detection algorithm.The preprocessing operations include image gray processing and image enhancement.In order to suppress the interference factors of non-distribution lines in the image background,this paper proposes an improved segmented grayscale transformation enhancement method,and then compares the experimental results with the pulse neural network enhancement method and the Hessian matrix enhancement method.After preprocessing,the edge features of the image are extracted.After analyzing the Sobel algorithm,Canny algorithm and the adaptive threshold Canny algorithm,this paper selects the adaptive threshold Canny algorithm more suitable for linear targets,combined with morphological processing methods,further Strengthen the characteristics of distribution lines.Finally,by analyzing the characteristics of the distribution line in the image,a Hough transform method based on area screening is proposed to complete the distribution line target detection task.In the experiment,the detection effect and detection performance of the standard Hough transform and the probabilistic Hough transform made a comparison.In view of the situation of particularly messy scenes,this thesis also tried to use binocular cameras to collect power distribution line images.First,the preprocessing algorithm in the depth image is discussed,and then the distribution line detection method based on the depth information is proposed.Convert the acquired depth information to the actual distance,divide the distribution line area by the distance threshold,and then extract the distribution line target from it.This method can effectively remove diverse scene interference,and experiments show that this method has certain value.