Range extraction in computer vision

Stereo cameras: use two cameras to take two pictures and compute distances by matching points on both images. It’s main disadvantage is that is a computationally expensive.
Light stripes: project some lines on the environment and with the cameras see how they distort to get information. It’s main disadvantage is that they perform okey on labs and closed environments, but not very well on the outside.
Laser ranging: use a laser beam to scan the environment to produce depth values, and use all this values to create a intensity/depth image. It’s main disadvantage are that laser can be quite expensive and are more susceptible to light reflection/absorption than when dealing with camera images.
Texture: sometimes, on certain tasks for example, on robotics 20210602191112, there are more simple and elegant solutions like detecting some textures on the environment to define depth.

Notes References

20210602191112 Cameras and Computer vision in Robotics - Basics

20210709195051 INDEX - Computer Vision

(Murphy 2000)

Murphy, Robin. 2000. Introduction to AI Robotics. Intelligent Robotics and Autonomous Agents. Cambridge, Mass: MIT Press.