How can I approximate distance errors of MultiSense cameras?


The equation below can be used to approximate distance errors in meters given a certain pixel error. As a reference, an S21B with 1/2 pixel of error, 4.8mm lenses, measuring a target at 5.0m, will have a distance error of up to 0.063m.

error (meters) = (D^2 * p) / (b * f + D * p)

p=pixels of error (+-1/2 pixel is our maximum allowed median error in our controlled test, less than 1/4 pixel is common) D=measurement distance (units of meters) f=focal length of lens (units of pixels) b=baseline (units of meters, 0.21 for S21B, 0.07 for S7). 

Focal lengths in mm, at full horizontal resolution, are converted to pixels below 4.8mm ~= 931.981 pixels 5.0mm ~= 970.814 pixels 6.0mm ~= 1164.976 pixels 6.5mm ~= 1262.058 pixels 8.0mm ~= 1553.302 pixels

Have more questions? Submit a request