The Extraction of Spatial Information and Object Location Information from Video

摘要

This paper presents a extracting method for 3D spatial information of background and location information of subject from video. First to extract the 3D spatial information, we used rotating angle of frames from video. Second, we detected subjects from frames. Finally, we calculated angles from each frames. To find the relative position of subject in each frame, we mixed the information of surroundings and movement of subject. The image data can be calculated with the user-providing information on the first position of camera. Therefore the GPS coordinate can be estimated corresponding to the position of protagonist in each frame and many different techniques of enhanced-reality applied services could be developed.