Immersive Photogrammetry in 3 D Modelling

Panoramic photography, on a massive scale, is widely used by Google Street View (exterior panoramas), Google Business View (company interiors) and Google Photo Sphere (Internet users create their own panoramic virtual tours using tablets and smartphones). These applications enable navigation outside and inside buildings. The navigation is possible by clicking arrows. The user clicks an arrow when they want to move in a specifi c direction. This method of moving, however, does not create a sense of continuity of movement which can be achieved when panoramas are presented as a video. However, in immersive video, when presenting a specifi c trajectory, it is not possible to change directions. A sequence of panoramas presented as a 360° video is defi ned as an immersive video or 360° video [7]. This form of presentation has not become very popular mainly due to large sizes of fi les (e.g. 1-minute raw video is a fi le whose size is about 1–2 GB; the fi le needs to be compressed in order to be presented on-line) [13]. Immersive video typically consists of 15–30 panoramas/second. So far, such a large amount of data has not been widely explored for photogrammetric purposes. Furthermore, no metric tools have been added to Google panoramic virtual tours and to immersive video. Immersive video is created by combining images generated by digital cameras that record in various directions. This video is becoming a new form of presenting art projects on interactive exhibitions, in fi lm clips and also in advertising [18]. 3D modelling from panoramic images is well know and described in [4, 14, 17, 19]. However, the potential of photogrammetric immersive video has not been examined so far and it is the subject of our experiments.