Publications / 2017 Proceedings of the 34rd ISARC, Taipei, Taiwan

Automatic Generation of 3D Models from UAV-Captured Image Data for Immersive VR Applications

Nei-Hao Perng, Bo-Sheng Bai, Po-Han Chen, Jen-Yu Han, Ming-Yi Jiang, Jyun-Hao Huang, Cheng-Wei Su and Po-Yuan Chen
Pages 812-815 (2017 Proceedings of the 34rd ISARC, Taipei, Taiwan, ISBN 978-80-263-1371-7, ISSN 2413-5844)

There are some twenty-nine thousand bridges in Taiwan, managed respectively by Taiwan Railways Administration, Taiwan Area National Freeway Bureau, Directorate General of Highways, and local governments. According to the current bridge maintenance guidelines in Taiwan, bridge inspection is mainly carried out through visual inspection, supplemented with equipment inspection. Bridge inspectors would usually approach the bridge structure for visual inspection. However, for high-rise or cross-river bridges, it would be hard to carry out visual inspection due to the difficulty in accessing the bridge structure. In view of this, the Institute of Transportation under the Ministry of Transportation and Communications (MOTC) in Taiwan conducted a research project on “Enhancing the effectiveness of bridge inspection equipment” in 2014, and successfully carried out bridge inspection using a camera-mounted unmanned aerial vehicle (UAV). In this paper, the way to automatically convert UAV-captured image data to 3D models as well as the subsequent immersive virtual reality (VR) applications will be presented. It is expected that the proposed development would help identify the defects on the inspected bridge, and assist with the training of junior bridge inspectors.

Keywords: UAV, VR, 3D-Model, Automatic