Edge Detection on Pantograph Slide Image

Nowadays high-speed railway plays more and more important role in the traffic transportation. It is necessary to detect the wear on pantograph slide. This paper presents image processing algorithm to detect abrasion on pantograph slide. Discrete dyadic wavelet transform is applied to get the edges in the image, and Hough transform is used to extract the bottom and top edges of slide. It also utilizes edge point´s features to eliminate the pseudo-edges. Practical test shows that this approach can effectively detect the slide´ edges, and has high accuracy.