3D model by using oblique images from unmanned aerial vehicle (UAV)

Abstract

In recent years, 3D modelling had been a major outcome for obtaining a high-level detail structure of facade and building. The role of Unmanned Aerial Vehicle (UAV) as platform had been widely used as its accessibility to the location and geometry aspects such as terrain, mount and higher building. This paper aims to obtain highly-level-detail 3D model of building acquire from oblique images of UAV. Several parameters should be considered such as flight planning, minimum time to cover whole building, inclination of camera and ground control points to achieve complete information of the building. Oblique images acquire by a very close circular flight to the building. Some nadir images will be used for further tie and strengthen block. The images obtained were processed using Agisoft Photoscan Professional. Several steps were taken for image processing such as calibration, extraction, point cloud generation and orientation. The UAV data was validated with check points observed from GPS to obtain Root Mean Square (RMSE) values. RMSE value for planimetric accuracy is ±0.026m, ±0.034m and ±0.018m for X, Y and Z respectively. The average RMSE values obtained is ±0.037m. In conclusion, the value achieved centimetre accuracy and show that the use of UAV and processing software as an effective tool for surveying, reduce cost, acquisition of data more rapid and need for high level detail.