Forum Discussion
Movement using coordinate mapped panoramic stills
I'm keenly interested in this, can someone put me in touch with similar projects?
I'm currently working on an idea to map real space using a series of 3-D, 360 panoramic still photos.
The eventual goal is to recreate a real walkaround space without having to first rebuild that space with architectural software or a game engine, hopefully expediting the processing time and easing computer requirements. For instance, a recreation of a room in the Louvre, using real imagery by the methodical coordinate by coordinate photographing of the room. A simple coordinate navigation system would then allow one to move around anywhere in the room, viewing the actual room, in 3D, from the panoramic photo taken at that coordinate.
Help and critiques are welcome.
I'm currently working on an idea to map real space using a series of 3-D, 360 panoramic still photos.
The eventual goal is to recreate a real walkaround space without having to first rebuild that space with architectural software or a game engine, hopefully expediting the processing time and easing computer requirements. For instance, a recreation of a room in the Louvre, using real imagery by the methodical coordinate by coordinate photographing of the room. A simple coordinate navigation system would then allow one to move around anywhere in the room, viewing the actual room, in 3D, from the panoramic photo taken at that coordinate.
Help and critiques are welcome.
5 Replies
- Okay,
Either this is an absolutely ridiculus idea and I should immediately walk away, or I'm about to venture into uncharted territory. Was really hoping for a word in either direction. - Look at what Matterport is doing. http://matterport.com/
With their device take panoramas with depth data at enough location in a room so no backside are obscured.
The panos are then processed and stored on their cloud and you can either jump from pano to pano and have perfect clarity or wander around the room with less clarity. - Both Photosynth and Google Street View allow you to move around horizontally with panoramas by having multiple panoramas from different viewpoints.
If you can take video and move the camera at a constant speed, I don't see why you couldn't extract each panoramic frame, and allow the user to travel independently in small enough increments to make them feel as if they're actually walking on their own. - maybe this article on the "coverage" problem in robotics is relevant
http://dl.acm.org/citation.cfm?id=2542724
there are lots of articles under the References tab - These are all close to what I have in mind. Google street view comes closest. Imagine a room-sized version of google street view, but instead of a single path that street view offers, the photos taken are across a grid of the entire room, allowing one to move forward, backward, left and right. This completely eliminates the need to first build the room in a game engine, and also reproduces shadows, reflections, lighting and textures in a (literal) photo-realistic setting.
This technology would be best suited for museums and real estate.
