A Comparison between Interpolation Methods for More Accurate Elevation Surface Using Gnss And Gis

Abstract

Elevation surface is a fundamental element of spatial data that can be employed to perform a variety of geostatistical and spatial analyses. On this basis, this paper presents the assessment of elevation surfaces interpolation methods such as Inverse Distance Weighting (IDW), Ordinary kriging, and Local Polynomial Interpolation (LPI). There are three scenarios for achieving this purpose by examining it in three different areas: mild slope area, steep slope area and combined case. The dimension of each tested area is decided to be 100*100 m with 121 survey points for each. The ellipsoidal height of survey points aremeasured by the Global Navigation Satellite System (GNSS) receiver exploiting the real-time kinematic (RTK) technique. All survey points are transferred to Arc GIS environment for generating elevation surfaces and conducting interpolations. Interpolated pixels of surfaces are then compared with the trusted data collected with GNSS receiver in RTK mode. The result revealed that the total Root Mean Square (RMS) error for kriging interpolation in the steep area is around 26 cm, whereas RMS for LPI interpolation in the flat area found to be approximately 8.0 cm. The evaluation outcomes can be utilized to understand the influence of the slope on interpolation methods as well as to select the most appropriate method according to the ground relief.