A Combined Method For Digital Terrain Models Generation

Abstract

Recent studies have shown that the various methods of generating Digital Terrain Model (DTM) from its discrete form to a continuous one using simple functions suffer from the influence of terrain complexity and the density of the data points that can introduce significant errors in the final generated model. The interpolation of the grid node is generally carried out using some functions. These functions are based on the assumption that the measured height values are autocorrelated i.e. the adjacent points on the surface are expected to have the same value. This assumption is only true when the terrain surface is smooth. However, in the case of a variable and mountainous terrain, the assumptions made about the underlying variation (correlation) of the measured height values can be critical. For a successful application of the simple functions in these terrain the density of the measured sample data must be increased. The acquisition of this data by ground survey methods, photogrammetric method or digitization is time consuming and financially constitutes a major cost in the overall process of DTM generation. It can be seen from the above that the use of the simple functions in DTM genearation in a variable terrain surface would be of no meaning if the number of measured height values is not reduced. In this thesis a new method that can effectively be used in the generation of DTM in a variable terrain surface is proposed. The new method which is known as a combined method uses a few number of discrete sample points in the interpolation of grid nodes.