A Study on Curve Simplification Method Combining Douglas-Peucker with Li-Openshaw

Abstract

At present, there were two classical approaches commonly used for line simplification. One was Douglas-Peucker (D-P) algorithm and the other was Li-Openshaw (L-O) algorithm. Although the former was able to preferably reserve characteristic bending points of the curve and compress other non-feature points, the simplified result was excessively inflexible and sharp corners were also generated on feature points. As for the latter, not only can the corner of a line be smoothed, instead of becoming over inflexible, but feature point smoothing was also carried out by it for the line. Therefore, based on analyzing characteristics of such two algorithms, an improved algorithm was presented in this first place by means of combining the both together. To be specific, feature points of curves for generalized simplification were figured out by the D-P algorithm, while the L-O algorithm was used to perform curve processing by adjusting radius R of a circle SVO. In the end, real data were applied to carry out experimental verification contrasts for the modified algorithm and the existing two additional algorithms. Experimental results indicated that such a modified algorithm that combined them together exhibited their advantages and had the capability to reserve feature points and simplify other parts simultaneously. Moreover, it could be effectively applied in automated mapping.