Národní úložiště šedé literatury Nalezeno 2 záznamů.  Hledání trvalo 0.00 vteřin. 
Visual Geolocation of a Stationary Camera
Šimurda, Pavel ; Španěl, Michal (oponent) ; Herout, Adam (vedoucí práce)
In this thesis we address a problem of geo-locating a camera using only visual information of the scene. The aim of the thesis is to study existing methods of geo-location, discuss and show their limitations and suggest possible improvements. We propose two different methods of geo-localization. The first method is based on detection of sunrise and sunset in the image sequence. Its main benefit is that this method can be successfully applied for any outdoor camera and it does not require any specific object in the scene. The second is based on the analysis of shadows in the scene. This method allows estimation of the  Sun's position a single image using two objects and their shadows. It is required to have at least two images captured at different time for the correct estimation of the geographical position. We evaluated proposed methods and introduced discussion and possible improvements. Both methods can be successfully used for the geo-location determination. In addition, we created an extensive dataset of time-lapse images from public accessible web cameras.
Visual Geolocation of a Stationary Camera
Šimurda, Pavel ; Španěl, Michal (oponent) ; Herout, Adam (vedoucí práce)
In this thesis we address a problem of geo-locating a camera using only visual information of the scene. The aim of the thesis is to study existing methods of geo-location, discuss and show their limitations and suggest possible improvements. We propose two different methods of geo-localization. The first method is based on detection of sunrise and sunset in the image sequence. Its main benefit is that this method can be successfully applied for any outdoor camera and it does not require any specific object in the scene. The second is based on the analysis of shadows in the scene. This method allows estimation of the  Sun's position a single image using two objects and their shadows. It is required to have at least two images captured at different time for the correct estimation of the geographical position. We evaluated proposed methods and introduced discussion and possible improvements. Both methods can be successfully used for the geo-location determination. In addition, we created an extensive dataset of time-lapse images from public accessible web cameras.

Chcete být upozorněni, pokud se objeví nové záznamy odpovídající tomuto dotazu?
Přihlásit se k odběru RSS.