An effective method for interpolation

Interpolation is a significant problem in Geographic Information System and many other fields. Kriging interpolation is thought as optimal interpolation which has high quality but time-consuming at the same time. We use Quad-Tree method which is thought as the only suitable method to locate the pixel images in two-dimension to create spatial index. There are mainly three variants of kriging. We mainly do improvement for ordinary kriging which is the most useful of kriging interpolation. As we have many multi-core servers, we mix OPENMP and MPI to design our parallel kriging. We use MPI to design the communication among servers and use OpenMp to design the communication among processes in the same server. Results show that our proposed algorithm is effective and could save much time.