Resumo:
Periodically the electric power transmission lines need to be inspected to avoid future
problems or even to solve problems that have been noticed. In order to carry out these
inspections, companies have been trying to make it as safe as possible so that there is not
a very high expense and do not put any person at risk, for this they are increasingly using
drones to inspect, however these drones need to have people nearby operating navigation
and with that it often ends up not being as efficient as it ends up passing some problem.
To solve the problem of detecting a possible problem on the line I come to present a
solution using three drones that will make an autonomous flight over the line. The first
drone will initially travel to the first tower and with a camera, using a convolutional
neural network it will detect the tower and any possible problem that is found in it and
record the location of the problem tower for someone to do maintenance on it, the the
second drone will be responsible for checking the wires that conduct the electrical energy
and the third drone will be responsible for detecting if there is something under the line
that may be at risk or have come to cause future problems for the line.
The solution was implemented in the airsim simulator, where an environment already
available by the software for autonomous navigation tests with a high voltage line was
used. The navigation of the drones was made from the positioning of the towers where
the location of the latitude and longitude of each one is already known.
