A vision-based automatic hotline tracking approach using unscented Kalman filter

Hotline inspection is a necessary, dangerous, and in many cases a very hard task. In this paper we present a vision-based approach, suitable for aerial inspection system, that automatically tracks a desired hotline to ease up the operation and would let the operator to focus on inspection of the lines. The operator flies the aerial inspection system and places it in the vicinity of a hotline. The aerial flying vehicle is equipped with a camera that is used to capture images of the desired hotline and finds the hotline in the image. Then the tracking system uses an Unscented Kalman Filter (UKF) to handle the disturbances that happens while the aerial system flies across the lines. The system has been simulated and actually tested which shows its capability to successfully and reliably follow hotlines.