Research on remote sensing recognition algorithm of transmission line target based on deep learning

Abstract

In the innovation and development of artificial intelligence technology, the inspection of uav circuit components with deep learning algorithm as the core has become the main content of social technology discussion. It can realize the classification detection effect through training on the basis of collecting a large number of transmission line images. Due to the differences in the collected image information, the relative pixels of various objects are small, and the actual semantic information is not much, it is not good to detect and analyze the typical components of transmission lines only by using the traditional convolutional neural network. In this paper, a transmission line detection method based on YOLOv3 algorithm of Res2Net residual structure is proposed based on the understanding of deep learning and transmission line detection status. The final practice results show that this method can not only monitor the working status of transmission lines in real time, but also further improve the intelligent level of transmission line inspection, which meets the requirements of transmission line construction and management in the new era.