Characteristic Analysis of the Hough Transform: A Global View (Ⅱ)

The basic Hough transform is applied to the detection and extraction of ellipses. It was shown that for a centrosymmetric shape such as ellipses, after its linear Hough transformation under the normal parameterization, the mid-curve of the two boundary curves of the shape in the parameter space must be the transformed sinusoid of the shape’s center. Based on this property, circle and ellipse detection techniques are investigated and some key algorithmic steps are outlined and verified by simulated data.