How Stereo Vision works
The distance extraction in stereo vision is based on triangulation between two sensors whose baseline (B) and focal plane (f) is known. The disparity – a key parameter all triangulation methods - is calculated by processing the images of both sensors (rectification and matching algorithm) and extracting the correspondences.
The maximum detectable depth range is proportional to the baseline (B) and the sensor resolution. The final depth resolution is mainly limited by matching and calibration errors.