NI Vision 2019 for LabVIEW Help
»View Product Info
This section explains the basic principles of binocular stereo vision system for a simplified set-up.
A typical stereo vision system incorporates the following steps in order to compute 3D information:
Perform camera model calibration for each camera. A camera model calibration learns internal and external parameters for each
camera setup. Camera model calibration allows you to subsequently perform image correction to remove lens distortion and produce
Perform stereo calibration for the stereo vision system. Stereo calibration computes the relative spatial relationship between
Perform stereo image rectification. Stereo image rectification projects images acquired from the left and right cameras so
that the images reside in the same plane. The rows of rectified images align perfectly so that a point in the left image falls
on the same row in both left and right images.
Compute stereo image correspondence. Stereo correspondence establishes matches between the left and right rectified images
to produce a disparity map. A disparity map is a 2D image that uses grayscale values to indicate the disparity, or distance,
between features in the left and right image. Because disparity values indicate the relative depth of an object, a disparity
map is sufficient for many stereo vision applications.
Optionally compute 3D planes for applications that require precise depth information. 3D planes provide detailed depth information
which can be mapped to real-world coordinates.
Note NI Vision renders depth and disparity maps with respect to the left rectified image.
For best results, NI Vision prefers a horizontal baseline, or a system where the horizontal distance exceeds the vertical
distance between the two cameras.