A new interpolation model of convex hull in Delaunay triangulation

Abstract

Digital elevation model (DEM) based on Delaunay triangulation can better express the feature of the terrain surface and avoid significantly data redundancy. There is, however, litter research on the surface model of DEM based on Delaunay triangulation. The main surface model of DEM is linear interpolation function based on triangulated irregular network (TIN). Many researches show that the effect of smoothness and continuity is not ideal. Aiming at the problem of linear interpolation based on TIN of DEM that local part is substituted by plane, the whole surface is not smooth and there is great difference in the actual terrain, this paper put forward a new surface interpolation model based on Convex Hull (CH) in Delaunay triangulation. This method is that the known reference points can be well-distributed around the inserting point. The interpolation model is based on the effective dynamic subdivision of CH area of the triangle influenced set as the weight. This paper combines with the actual alpine terrain data and uses Root Mean Square Error (RMSE) and Maximum Error (ME) to evaluate and analysis the experimental results. The experiment shows that our method has more effective in structure and precision.