The Detection of Extra Matters on the Transmission Lines Based on the Filter Response and Appearance

Abstract

In this paper, a detection method of extra matters on the transmission lines is proposed. Our method can be divided into two steps: the detection of the transmission lines and the detection of the sky. To locate the lines, we design a set of simple and efficient filters to obtain the candidates of the lines. Compared with the previous work using the length of the lines to perform the transmission lines classification, we use the color and texture features to make it more robust to the variation of the background. To recognize the sky, we first over-segment the image. Then, we design the color and texture features for the detection of the sky. Finally, these features are used to train the classifier of the sky. After the transmission lines and the sky are detected, we confirm whether there is extra matter on the transmission lines. The experimental results indicate that our algorithm can recognize the extra matters on transmission lines fast and accurately.