Interpolation Technique for the Underwater DEM Generated by an Unmanned Surface Vessel

Abstract

High-resolution underwater digital elevation models (DEMs) are important for water and soil conservation, hydrological analysis, and river channel dredging. In this work, the underwater topography of the Panjing River in Shanghai, China, was measured by an unmanned surface vessel. Five different interpolation methods were used to generate the underwater DEM and their precision and applicability for different underwater landforms were analyzed through cross-validation. The results showed that there was a positive correlation between the interpolation error and the terrain surface roughness. The five interpolation methods were all appropriate for the survey area, but their accuracy varied with different surface roughness. Based on the analysis results, an integrated approach was proposed to automatically select the appropriate interpolation method according to the different surface roughness in the surveying area. This approach improved the overall interpolation precision. The suggested technique provides a reference for the selection of interpolation methods for underwater DEM data.