LAUSR

In quality assessment of digital elevation models (DEMs), geodetic field measurements have an important, but also a limited, role. They can achieve high accuracy, but the acquisition of data is time consuming, expensive and, in areas with a high-resolution DEM accessibility, non-effective. Therefore, field measurements are only performed at discrete selected points to evaluate the quality of the DEM used for further studies. The aim of this article is to show that differences in heights from field measurements and from a DEM can be used for height-deviation-surface modelling and for possible improvements in the available DEM. The procedure is related to situations where significant changes in the landscape occur. For surface modelling, this research included several radial basis function networks (RBFNs). From simulations, knowledge was acquired of appropriate results based on varying amounts of input data as well as on different neural network activation functions. This study indicates the potential use of geodetic field measurements in the improvement of a local DEM by RBFNs.