Multi-function partitioned imitation of curve data compression
Wang Yan · Ha'erbin gongye daxue xuebao · 2002
This paper aims at finding a curve data compression method which does not damage the shape of curve, and enables the compressed data to be decompressed. A curve is partitioned by multiple curve functions with least square imitation. The function which is most appropriate for the curve or its partitions with given accuracy is found. The parameters of the function are substituted for vertex coordinates of the curve for the purpose of data compression. Curve compression experiments were made by a new method with artificial, natural curves and practical maps respectively. The curve data structure, compression and decompression algorithms, when the method is used to compress the curve data, are successfully developed. The experimental results are that the compression ratio is 1∶5.83～1∶9.8, the compression speed is 5.58～5.98 kb/s, decompression speed is 357.34～840.93 kb/s, there is no shape damage. It is indicated that the new method features high compression efficiency, quick decompression speed and free from shape damage. The storage and speed effects of the graphic or spatial systems are greatly improved when the method is successfully applied.