A 3D generative of urban scenes using UAV image at Arau, Perlis / Zainundin Zainalwati

Urban 3D models are useful for geo-visualization tasks, they can help us to identify places in everyday navigation, assist autonomous driving via 3D road information systems and they can be used in movies and games. They enable the interactive measurements of 3D distances, surface areas and volumes without physical presence at the location. If the 3D model is augmented with semantic information, example it distinguishes between objects or regions like streets, sidewalks, vegetation, buildings, walls, windows, entrances, such information further simplifies the aformentioned tasks and can robustify other tasks, example knowledge of road areas contributes to vehicle detection and traffic monitoring, or locating entrances can robustify pedestrian detection and action recognition tasks. However, there is no single data that allows one to create full spherical coverage of all structures. The aim of the study is to creation of 3D surface models of urban scenes primarily of architecture from imagery. To achieve the aim of the study, the objective of the study are to generate the orthophoto using agisoft and to construct 3D modelling image. The location for this research study is at UiTM Arau, Perlis. The data that will be use is image from unmanned aerial vehicle (UAV). For the methodology for this study, UAV will be used to take the image at study area. Then the image can be process using Agisoft photoscan Profesional to build orthophoto image and also to construct the 3D mapping.