Extraction and accuracy assessment of DTMs derived from remotely sensed and field surveying approaches in GIS framework

Abstract

Generating a high precision continuous surface is a key capability required in most geographic information system (GIS) applications. In fact the most commonly used surface type is a digital elevation model (DEM). Recently, there are some sources of remote sensing data that provide DEM information such as; LiDAR, InSAR and ASTER GDEM which ranged from very high to low spatial resolution. However, new methods of topographic field surveying still highly on demand e.g. Differential GPS and Total station devices. In both method of capturing the terrain elevation the post processing need to be applied to create a continuous surface from point clouds. Geostatistical analysis were used to interpolate the taken sample points from site into continuous surface. In current research, we examined the height accuracy of LiDAR point clouds and total station dataset with three non-adoptive interpolation models including, invers distance weightage (IDW), nearest neighbour (NN) and radial basis function (RBF) based on referenced DGPS points. RMSE and R square regression analysis were conducted to reveal the most accurate approaches in pilot study area. The results showed Lidar surveying (less than 0.5 meter RMSE) has higher height accuracy compared to Total station surveying (above 1 meter in RMSE) to extract DTM in flat area; while consumed less computational processing time. Moreover, IDW was the best and accurate interpolation model in both datasets to generate raster cautious terrain model.