ASSESSMENT OF DUALITY OF DEEM GENERATED FROM TOPOGRAPHIC MAPS AND GPS DATA

Abstract

Digital Elevation Models (DEM's) are digital representations of the elevation at regularly spaced intervals. DEMs thus are computer models of the surface of the earth and are being increasingly used for many tasks such as estimation of slope, gradient and aspect and operations such as analysing the hydrological flow paths on the surface. Digital elevation model may be generated by various means such as through digitization of contours and spot heights, (obtained from topographic map), GPS data, total station, theodolite and simple leveling, by the stereomatching of satellite images and aerial photographs, SAR Interferometry and Lidar data processing. Accuracy of DEM generated from different sources varies significantly. In this study, the accuracy of DEM have been assess generated from topographical map and GPS survey. A 15 x 15 km area in and around Roorkee has been selected as the study area. A number of control points have been identified in the area, which forms the basis of GPS surveys and subsequent DEM generation in ERDAS Imagine. A number of DEMs have been generated using different control point selection scenarios and different interpolation methods. The effect of variation in cell size on DEM has also been investigated. The quality of DEMs has been assessed with root mean square error computed from the check points. The quality of DEM can be improved by proper identification of ground control points. If no mark is given on the map for the spot height, the ground control point should be digitized in the middle of the spot height marked. For the DEM generation from GPS data, all the higher and lower point within the area must be properly identified and for good quality of DEM, GPS data must be on the smooth surface of the earth, neither it should be on the highly elevated artificial surface nor it should be in the deep river or narrow deep valley.