Obstacles detection systems are essential to achieve a higher level of safety on railways. Such systems should have the ability to contribute to the development of automated guided trains. Even though some laser equipments have been used to detect obstacles, short detection distance and low accuracy on curve zones make them not the best solution. In this paper, computer vision combined with prior knowledge is used to develop an innovative approach. A function to find the starting point of the rails is proposed. After that bottom-up adaptive windows are created to focus on the region of interest and ignore the background. The whole system can run in real time thanks to its linear complexity. It performs well in different conditions and it can work both on online and offline recorded video.

Video Analysis for Improving Transportation Safety: Obstacles and Collision Detection Applied to Railways and Roads

Damiani, Lorenzo;Giribone, Pietro;Revetria, Roberto;RONCHETTI, GIACOMO
2017

Abstract

Obstacles detection systems are essential to achieve a higher level of safety on railways. Such systems should have the ability to contribute to the development of automated guided trains. Even though some laser equipments have been used to detect obstacles, short detection distance and low accuracy on curve zones make them not the best solution. In this paper, computer vision combined with prior knowledge is used to develop an innovative approach. A function to find the starting point of the rails is proposed. After that bottom-up adaptive windows are created to focus on the region of interest and ignore the background. The whole system can run in real time thanks to its linear complexity. It performs well in different conditions and it can work both on online and offline recorded video.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/891136
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact