On high- and medium-voltage power lines, signalling beacons are key elements in preventing birds from colliding with the cables. Traditionally, these beacons are installed and removed manually, requiring an operator to work directly on the cable itself, which entails operational risks and safety limitations.
With the aim of mitigating these risks, SAPREM has developed and operates pioneering robotic systems for the installation and removing of signalling beacons. This equipment reduces the direct exposure of staff to working at height and on the cable, although the manoeuvres continue to be controlled by an operator on the ground, who supervises the process based on the visual information provided by the system.
Until now, this approach still required a high level of human intervention, making the process dependent on the operator’s experience and limiting its degree of automation and scalability.
The client was looking to move towards a more autonomous and intelligent system, capable of automatically identifying the beacon and its surroundings and executing the removal manoeuvre with precision, even in variable lighting and orientation conditions.
NAIR Center carried out the selection, adaptation and advanced integration of Artificial Intelligence algorithms and image processing, optimised specifically for the system’s operating context, enabling:
The solution is based on segmentation models trained from real data and fine-tuning techniques, combined with image processing to extract relevant geometric information.
Increase in the level of automation in operations on power lines.
SAPREM I Sociedad Anónima de Preformados Metálicos
Navarre (Spain)