ABSTRACT
Douglas-Peucker algorithm (Douglas & Peucker, 1973) are widely used to compress the linear vector data in many fields. It is easy to implement. However, there still exist the following shortcomings of this algorithm: (1) it will delete lots of important nodes by mistake if the distance from the middle node to the line segment (with the kept nodes as end points) is utilized as the criteria to judge whether it should be kept or not, which will make the simplified curves deviate from the original one and aect the subsequent analyses (Yang et al. , 2002; Ebisch, 2002; Wu & Marquez, 2003); (2) the time complexities of DPA is o n n( log ), which needs to be improved due to the demands for processing massive spatial data (e.g. vector data retrieved from remote sensing images) and real-time network transmission.
