Line Generalization Using Spectral Techniques

In this paper, a line generalization procedure is presented using filtering techniques and best fitting adjustment with conditions. The series of the original line coordinates is transformed into its slope values counterpart and through Fourier transform the spectrum of the slopes is computed. Applying a digital exponential filter and the inverse Fourier transform the filtered slopes are obtained. The filtered set of the slopes is transformed into a derived smooth line set which is then best fitted into the original line. The best fitting process assures conformality between the original and the derived line; the filter used is automatically selected according to the scales of the original and the derived map.