Laser Rangefinder
Laser rangefinders consist of a laser paired with a CCD/CMOS image sensor at a fixed distance (baseline) & angle relative to each other. On hitting something, the laser shows up as a dot in the image, but the location of that dot is offset a few pixels based on the distance between the rangefinder assembly and the object reflecting the laser. Simple equations solve the trigonometry, and the device gains a 1-sample depth map. If the rangefinder assembly is rotated on an axle, the depth map becomes a disc/plane with many samples.
Laser rangefinders as explained above, particularly those manufactured by the brands Sick and Hokuyo, have become standard tools for ground-based robotics teams. Cheap ones are present in robotic vacuums and expensive ones in Darpa Grand Challenge vehicles. An imaging capability is possible if the entire rotating assembly measured above is mounted on another rotating assembly, where the disk of measurements slowly sweeps out a sphere.
Inexpensive, small, fixed rangefinders, rather than rotating ones, can also be used for horizontal collision avoidance with large objects like walls.
[edit] Recent Developments
- Several teams in MAV competitions are using laser rangefinders as inputs for SLAM algorithms.
- The Microsoft Kinect, and the Primesense platform it is based on, represents a modification of the laser rangefinder using a "light coding" textured pattern, a 3D camera for depth-RGB fusion, and a MEMS-mirror laser projector instead of a doubly-rotating mechanical assembly.
- A paper was presented at ICRA 2008 on "A $30 Laser Rangefinder", detailing the simplicity of construction sufficient to produce a working product
- An importer of robotic vacuums, Neato Robotics, recently sponsored a contest to hack their laser rangefinders (*not* sold separately) in order to legally pioneer a new product: a laser rangefinder at the unheard-of price of $400, packaged in a free XV-11 vacuum.
