<p>Inspecting distribution lines is a meticulous and dangerous task, but it is essential to maintain the energy supply quality. With the advent of robotics, several alternatives have been proposed. Unmanned aerial vehicles are a powerful tool, allowing non-invasive inspections with operators acting safely from the high-voltage elements. Considering this type of application, it is crucial to think about developing autonomous algorithms, allowing drones to move over the lines so that the elements to be inspected are always in the camera’s field of vision during the mission. This paper presents a methodology for powerlines inspection, covering the stages of identification, alignment and following. It also presents a dataset for powerline detection and segmentation in urban scenario. The cables are detected using the YOLOv8 (You Only Look Once) object detection algorithm. Then, an image-based visual servoing control is carried out, which allows the drone to position itself at a certain distance and orientation relative to the cables. The detection stage was tested with real images of powerlines in various scenarios. Comparisons between YOLOv8 and its predecessor, YOLOv5, showed that the latest version can detect lines more accurately with a slight increase in processing time. The position control was implemented using the robot operating system framework and tested in a real scenario with the Bebop 2 drone. The results show that the drone could position itself according to the setpoint both laterally and concerning the orientation of the powerlines, making it possible to integrate what has been proposed into power grid monitoring routines.</p>