Open jeromeetienne opened 9 years ago
The demo operates on panoramas captured in a synthetic scene created with Blender. I extracted the depth maps by simply saving the depth buffer on each render. A real world application would use, I imagine, depth maps captured by a remote sensing technology such as LIDAR. It has gotten quite good lately. Then, there are various computer vision techniques to extract geometric information from a set of images, though, as far as I know, these provide worse results than remote sensing technologies.
as i understand your algo you need a depth map, can you details how you get it ?