Video Mosaicing of Planer Scenes using Extended Kalman Filter

Bati E., GÜRBÜZ Y. Z. , ALATAN A. A.

21st Signal Processing and Communications Applications Conference (SIU), CYPRUS, 24 - 26 Nisan 2013 identifier identifier


In this work, the large planar scene is reconstructed from small images. Small images can be consecutive video frames or sequence of photographs. In the problem, called mosaicing, instead of using widely used optimization based methods, probabilistic methods are used in the proposed method. Simultaneous Localization And Mapping(SLAM) techniques are adapted for video mosaicing. Probabilistic measures for the landmark locations are used to merge small images to create large scene. Experimental tests give promising results if the performance-complexity is considered at the same time.