Power line image segmentation and extra matter recognition based on improved Otsu algorithm

Extra matters are easy to lead to power lines inter phase short circuit which not only endanger the pedestrians and cars, but also seriously affect normal operation of the power system. Firstly, an improved Otsu algorithm based on morphological approach is proposed. Then a new filtering method according to power lines characteristic is presented to remove tiny noise. Finally, accuracy power line recognition is realized by Hough transform, and the extra matter on power line can be identified by comparing the number of tested power lines and the number of local maxima in Hough transfer accumulator. The result shows that this method can effectively recognize the extra matter on power lines, providing reliability for the power system.