Automatic detection of electricity pylons in aerial video sequences

Automatic power transmission line inspection with aerial video surveillance requires that the vehicle-mounted camera can automatically locate electricity pylons. In this paper, a new approach for locating electricity pylons within video sequences is proposed. Straight lines in each video frame are extracted with a two-dimensional separable infinite impulse response (IIR) filter and the Hough transformation. A technique for selection of lines representing the electricity pylons is also proposed. The experimental results from a set of real-world video images are shown in term of the receiver operating characteristic (ROC) curve to verify the algorithm´s detection performance. The ROC curve shows that the proposed algorithm performs very well and can very much help in the automatic electricity pylon detection system.