Research On Power Line Recognition And Tracking Algorithm Based On Machine Vision

Transmission lines play an important role in the transmission of electric energy in the power system,and as a consequence that regular inspection and maintenance of transmission lines is of great significance to the stable development of the State Grid.Due to problems such as low efficiency and high risk of the traditional manual power line inspection method,which makes it unable to meet the needs of current society for modernization and automation.With the help of unmanned aerial vehicle(UAV)technology such as strong flexibility,high security and intelligence as well as the continuous more mature development of image processing technology,the autonomous power line inspection technology by UAV based on machine vision has become an important research.The power line recognition for aerial images is the key and difficult part of this research field,and the result of power line recognition will be used as the control quantity of flight tracking,which assists UAV to complete the autonomous line inspection task.In this paper,therefore,we will take visible light aerial images as the research object to carry out the research of power line recognition algorithm under the background of complex ground objects.The main research contents of this paper are as follows:(1)Firstly,we investigate and compare the research of domestic and foreign scholars on the power line recognition methods based on UAV vision.Summarize the pros and cons of various algorithms at present and design the overall process of power line recognition and tracking algorithm based on the existing image processing technology theory.(2)Secondly,in terms of the research on the edge detection of power lines,an improved Ratio operator of edge detection is carried out to obtain the edge pixels of the image and the parameter relationship and edge detection feature value of the Ratio operator template are designed based on the original algorithm.Besides,the detection threshold value is given and a background-denoising algorithm based on contour features is proposed to filter most of the edge noise.(3)Thirdly,for the research on the edge extraction of power lines,a Hough transform algorithm with stronger anti-interference ability is used to obtain line primitives.Furthermore,an algorithm of line dynamic grouping-fitting and screening is proposed to complete the connection and edge extraction of linear primitives,and the power lines are completely identified at the end of this part.(4)Fourthly,the parameters of power line can be obtained from the results of power line recognition,then the deflection angle and offset of power line are calculated,which are used as the control value of UAV tracking line inspection to correct the route of flight timely and improve the efficiency of power line inspection.(5)Finally,we design simulation experiments of power lines recognition and UAV flight tests in actual scene.The simulation experiment results show that the proposed algorithm can effectively eliminate the noise interference and got a result with higher recognition accuracy and completeness rate.The test results show that the UAV can fly autonomously along the power line,which was further verified that the feasibility of the proposed algorithm and its higher application value for UAV power inspection.