A New Method for Computing the Principal Point of an Optical Sensor by Means of Sphere Images
Abstract: For some applications it can be preferable to use images of spheres in order to calibrate a 2D camera. All published sphere-based algorithms need the complete knowledge of the elliptic sphere image, i.e. 5 geometric parameters, in particular the ellipse orientation. Because sphere images tend to be close to circular shapes, this orientation is often very noise-sensitive. For example, it is common to compute the principal point as the intersection of the lines through the major axes of the elliptic images, but this procedure is quite unstable. We present a new method for computing the principal point by means of three sphere images, without making use of the ellipse orientation. By mean of simulations and real experiments we demonstrate that the proposed method is more accurate and stable in finding the principal point as compared to sphere-based calibration algorithms that use the complete ellipse geometry.
Loading