A Head-Tail information break method oriented to multi-scale representation of polyline

It is proposed that a multi-scale representation model of geographic polyline based on Head-Tail break of amount of information. Firstly, a geographic polyline is transformed into a Fourier shape descriptors, and the amount of information in frequency domain is defined using Shannon’s information entropy theory. Secondly, Head-Tail breaks is introduced to determine the demarcation points of information content of the polyline, and a function is designed to estimate the Fourier truncating frequency corresponding to each demarcation point. Finally, the traditional radical model based on the frequency information amount of polyline is introduced to measure the key scales and to achieve hierarchical multi-scale representation of polylines. The experiments of contour and coastline demonstrate that the model proposed can effectively simplify geographic polylines according to the set scale, and the simplification results of different target scales exist good degree of differentiation and hierarchy. At the same time, under the same overlap ratio between the simplified result and the original geographic polyline, the model proposed is better than the traditional simplification algorithm.