5

I'm very new in this field so my question may be very stupid. I apologize for this at first.

I'm trying to use a proper camera to run a SLAM algorithm. As suggested by some people, stereo camera may be a good choice. When I searched online, I found terms "stereo camera", "depth camera", "3d camera", and "rgb-d camera", and I feel confused. My understanding is "rgb-d camera" is more than "depth camera" because it has rgb channels while simple "depth camera" only makes white-black images. However, I have no idea about the difference from "stereo camera" and "3d camera". I know some camera options such as ZED, BumbleBee, Kinect, Intel RealSense, to me they are all "stereo camera".....

Anyone can give me some instructions? Thanks a lot in advance!

gladys0313
  • 153
  • 1
  • 1
  • 4

1 Answers1

11

3D/Depth-Camera:

The most inclusive terms, only means that the camera gives you 3d data.

Stereo-Camera:

One kind of 3D-Camera, consists (in most cases) of two gray-scale Cameras.

RGBD-Camera:

A sensor that gives you depth and color. In most cases, this refers to a Kinect-Style camera (or Primesense, Realsense), But a stereo-camera could also create colored pointclouds (if one camera is a rgb camera or if you have an additional camera).

You then also have the distinction of passive setups (like a standard stereo camera) or active cameras with some kind of projectors.

Active cameras can use a speckle projector to create a texture that is then captured by a camera (Kinect1, Primesense) or use flash and Time-based computations to measure the distance (Keyword is Time of Flight) (Kinect2, Picoflex).

FooTheBar
  • 1,375
  • 7
  • 18