Authors:
Taynara Gonçalves
1
;
Andrea Bianchi
1
and
Glauco Yared
2
Affiliations:
1
Department of Computing, Federal University of Ouro Preto, Brazil
;
2
Department of Electrical Engineering at the Federal University of Ouro Preto, Brazil
Keyword(s):
Contour Segmentation, Automatic Detection, Computer Vision, Rail Inspection.
Abstract:
The railway is an essential part of the marketing chain. If it is considered efficient, safe, and competitive, on the other hand, the railways suffer from the enormous difficulty of maintenance due not only to their great extension, dispersion and lack of financial investments. Initiatives for automatic maintenance inspection, mostly done manually, require development and consolidation. Therefore, this work presents a method for identifying defects in sleepers based on analyzing images. We will use digital image processing techniques that will allow us to extract the contours of the sleepers and therefore analyze their curvature. The development is carried out with images from laboratory tests, not previously classified but subject to noise. The method is validated through analysis with an image bank with about 20 images of defective and flawless sleepers, with an average assertiveness of 94.24%. The detection, classification, and localization of faults in train tracks are then inves
tigated and discussed.
(More)