Video alignment for difference-spotting - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2008

Video alignment for difference-spotting


We address the synchronization of a pair of videos sequences captured from moving vehicles and the spatial registration of all the temporally corresponding frames. The final goal is to fuse the two videos pixel–wise and compute their pointwise differences. Video synchronization has been attempted before but often assuming restrictive constraints like fixed or rigidly attached cameras, simultaneous acquisition, known scene point trajectories etc. which to some extent limit its practical applicability. We intend to solve the more difficult problem of independently moving cameras which follow a similar trajectory, based only on the fusion of image intensity and GPS data information. The novelty of our approach is the probabilistic formulation and the combination of ob- servations from these two sensors, which have revealed complementary. Results are presented in the context of vehicle pre–detection for driver assistance, on different road types and lighting conditions.
Fichier principal
Vignette du fichier
1569140042.pdf (264.58 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

inria-00326756 , version 1 (05-10-2008)


  • HAL Id : inria-00326756 , version 1


Ferran Diego, Daniel Ponsa, Joan Serrat, Antonio López. Video alignment for difference-spotting. Workshop on Multi-camera and Multi-modal Sensor Fusion Algorithms and Applications - M2SFA2 2008, Andrea Cavallaro and Hamid Aghajan, Oct 2008, Marseille, France. ⟨inria-00326756⟩


102 View
179 Download


Gmail Facebook X LinkedIn More