Calibration

Weak calibration: 
    Let m1 = (x1,y1,1)T and m2 = (x2,y2,1)T be two corresponding (or homologous) points, that is, two points that are the images of a unique 3D point taken by two cameras.

    It has been proved that, according to the pinhole model, there is a 3x3 matrix, noted F, so that:

m2T.F.m1 = 0  (*) [Longu81]
    F is called the fundamental matrix and is defined up to a scale factor.

    F depends only on the relative positions of the sensors and their intrinsic parameters. It is said that F codes the epipolar geometry of the sensors.

    Relation (*) can be seen as follows: the point m2 lies on the line defined by F.m1; it lies on a line depending only on m1 and the epipolar geometry. This line is called the epipolar line associated with m1; the relation is then a weak version of the epipolar constraint.

    To weakly calibrate a sensor system is to estimate F.

    The weak calibration can be divided in two steps: