The class implements an algorithm for 3D object's pose estimation from it's 2D coordinates obtained by perspective projection, when the object is described none coplanar points. The idea of the implemented math and algorithm is described in "Model-Based Object Pose in 25 Lines of Code" paper written by Daniel F. DeMenthon and Larry S. Davis (the implementation of the algorithm is almost 1 to 1 translation of the pseudo code given by the paper, so should be easy to follow).
Read 3D Pose Estimation article for additional information and samples.
Sample usage:
// points of real object - model Vector3[] positObject = new Vector3[4] { new Vector3( 28, 28, -28 ), new Vector3( -28, 28, -28 ), new Vector3( 28, -28, -28 ), new Vector3( 28, 28, 28 ), }; // focal length of camera used to capture the object float focalLength = 640; // depends on your camera or projection system // initialize POSIT object Posit posit = new Posit( positObject, focalLength ); // 2D points of te object - projection AForge.Point[] projectedPoints = new AForge.Point[4] { new AForge.Point( -4, 29 ), new AForge.Point( -180, 86 ), new AForge.Point( -5, -102 ), new AForge.Point( 76, 137 ), }; // estimate pose Matrix3x3 rotationMatrix; Vector3 translationVector; posit.EstimatePose( projectedPoints, out rotationMatrix, out translationVector );