Performance evaluation of spatial interpolation methods in the presence of noise

Abstract

This study aims to investigate the robustness of some spatial interpolation methods (SIM) in the presence of noisy data. SIM are used in different contexts such as the identification of hot-spots in brown fields, Geographical Information Systems (GIS), remote sensing and visualization/shape re-construction. In all contexts, the correct representation of a given surface is crucial. For instance, in a site characterization context identifying hot-spots, all contaminated areas should be determined in order to minimize health hazards in human use after the reclamation of the site. Here, we conduct a numerical survey on the performance of four spatial interpolation techniques using eight mathematical functions to represent domain-independent surfaces. Furthermore, we also investigate the effects of different sampling patterns on the performance of SIM.