Automatic Detection Method Of The Thickness Of Transmission Lines With Ice Under Complex Background

The operation and maintenance of transmission lines not only play an important role in the stable operation of the power grid,but also relate to the electricity experience and safety of millions of users.With the development of UHV and UHV transmission technologies in China,more and more power lines need to pass through heavily icy areas.Therefore,ensuring the safe operation of transmission lines and monitoring the degree of damage of ice-covered transmission lines play an extremely important role.In order to better adapt to the needs of the digital development of smart grids,this paper uses image processing methods to detect the icing of transmission lines.The edge detection algorithm provides a good model for detecting iced transmission lines.Canny edge detection algorithm is a more classic algorithm,which has the advantages of simple method and easy implementation.But it also has obvious shortcomings,for example,it is not only sensitive to salt and pepper noise but also easy to obtain the local optimal solution.Genetic programming is a commonly used optimization algorithm.It can be combined with Canny edge detection algorithm for edge recognition and extraction,which is not only beneficial to filter out pepper and salt noise,but also more accurately locate the power line edge point.Convolutional neural network can automatically detect the multi-directionality of image features,which is an innovative detection algorithm for edge detection.The improved CNN algorithm uses the overall nested model to predict and extract the edge of the iced transmission line,which improves the accuracy of edge detection.The main idea of the HED is to use the multi-level information of input images to extract edge features,and deconvolve the feature maps of different scales output by the middle layer to obtain images of the same size,and then merge these feature images according to different weights,and finally output the edge image of the target object.It can quickly and accurately obtain the outline of power lines.It has a high depth and multi-directionality.Finally,by comparing the pixel coordinates of the edges of the transmission line before and after ice coating,the thickness of ice coating can be calculated with the help of the relevant proportional relationship.The experimental results show that the HED model not only has a better detection effect for online monitoring of ice-covered transmission lines,but also has a larger scope of application.