This paper presents a real-time Hough-based algorithm for straight line segment extraction in complex multisensor images, which aims to avoid loss of spatial information as well as to eliminate spurious peaks and reduce discretization errors. A parameter space representation able to take into account spatial information during the voting phase is proposed. This representation allows the detection phase to be performed by focusing the algorithm on particular locations of the parameter space. The search space is consequently reduced, and a deeper decision strategy can be adopted, which takes into account the local distribution of segments along both a line and different lines, for comparable directions and positions. Experimental results on a large set of complex multisensor images (e.g. underwater images, low-light outdoor images, SAR images, etc.) are presented. The main advantages of the proposed method over both feature and image-space methods are evaluated in terms of computational efficiency, detection accuracy and noise robustness. (C) 2000 Academic Press.

A real-time Hough-based method for segment detection in complex multisensor images

FORESTI, Gian Luca
2000-01-01

Abstract

This paper presents a real-time Hough-based algorithm for straight line segment extraction in complex multisensor images, which aims to avoid loss of spatial information as well as to eliminate spurious peaks and reduce discretization errors. A parameter space representation able to take into account spatial information during the voting phase is proposed. This representation allows the detection phase to be performed by focusing the algorithm on particular locations of the parameter space. The search space is consequently reduced, and a deeper decision strategy can be adopted, which takes into account the local distribution of segments along both a line and different lines, for comparable directions and positions. Experimental results on a large set of complex multisensor images (e.g. underwater images, low-light outdoor images, SAR images, etc.) are presented. The main advantages of the proposed method over both feature and image-space methods are evaluated in terms of computational efficiency, detection accuracy and noise robustness. (C) 2000 Academic Press.
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/11390/673745
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 10
  • ???jsp.display-item.citation.isi??? 8
social impact