A depth camera is an impressive device, but not every developer or user has one and it has some limitations. Notably, a typical depth camera does not work well outdoors because the infrared component of sunlight is much brighter than the camera's own infrared light source. Blinded by the sun, the camera cannot see the infrared pattern that it normally uses to estimate depth.
As an alternative, we can use one or more normal cameras and we can estimate relative distances to objects based on triangulation from different camera perspectives. If we use two cameras simultaneously, this approach is called stereo vision. If we use one camera, but we move it over time to obtain different perspectives, this approach is called structure from motion. Broadly, techniques for stereo vision are also helpful in SfM, but in SfM we face additional problems...
