Development of ‘3D City Models’ using IRS Satellite Data

Abstract

Urban models are computer-based solutions used for testing theories about spatial location and interaction between land uses and related activities. They also provide digital environments for testing the consequences of physical planning policies on the future form of cities. Earlier, 3D city models were prepared using stereo aerial photography, topographic data and conventional surveying. The geospatial information was extracted using either visual or digital classification techniques. Presently the 3D city models is prepared with the help of ALTM data along with very high resolution imagery, which is costly and time consuming. As alternatives, Cartosat-1 stereo data of the order of 2.5 m resolution, it becomes more practical and economical to generate digital surface model (DSM) with an accuracy of 4 m. Cartosat-II data with 1 m resolution fused with LISS-IV data give high resolution images for building footprint extraction. The availability of high resolution Cartosat-I stereo satellite data assists urban planners and municipal authorities to create a model and visualize the urban space in three dimensions. 3D visualization models have a variety of applications in urban studies. The studies includes urban morphological mapping, disaster management, updation, virtual reality and generation of fly through etc. The primary interest of the planners is not limited towards the extraction of urban features from satellite data. Planners also have interest in creation of 3D models of buildings and terrain surface for detailed urban planning. Therefore, to develop 3D city models 2.5 m Cartosat-1 stereo data fused with 1 m merged product of Cartosat-II and LISS-IV to create DSM and building footprints respectively. Keeping the significance of use of stereo satellite data, an attempt has been made to generate a 3D building model for Ahmedabad and Hyderabad cities. The 3D building model generated by employing geoinformatics techniques with Level of Details (LOD-1). The primary aim of this study was to develop geoinformatics techniques for the extraction of urban features based on object oriented classification techniques with level of details 1 (LOD-1) using photogrammetric DEM. This paper discusses all these techniques and their usefulness in 3D city modeling exercises.