The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Articles | Volume XL-1/W2
https://doi.org/10.5194/isprsarchives-XL-1-W2-183-2013
https://doi.org/10.5194/isprsarchives-XL-1-W2-183-2013
16 Aug 2013
 | 16 Aug 2013

QUALITY OF 3D POINT CLOUDS FROM HIGHLY OVERLAPPING UAV IMAGERY

N. Haala, M. Cramer, and M. Rothermel

Keywords: Platforms, Point Cloud, Matching, DEM/DTM, Surface, High resolution, Three-dimensional, Performance

Abstract. UAVs are becoming standard platforms for photogrammetric data capture especially while aiming at large scale aerial mapping for areas of limited extent. Such applications especially benefit from the very reasonable price of a small light UAS including control system and standard consumer grade digital camera, which is some orders of magnitude lower compared to digital photogrammetric systems. Within the paper the capability of UAV-based data collection will be evaluated for two different consumer camera systems and compared to an aerial survey with a state-of-the-art digital airborne camera system. During this evaluation, the quality of 3D point clouds generated by dense multiple image matching will be used as a benchmark. Also due to recent software developments such point clouds can be generated at a resolution similar to the ground sampling distance of the available imagery and are used for an increasing number of applications. Usually, image matching benefits from the good images quality as provided from digital airborne camera systems, which is frequently not available from the low-cost sensor components used for UAV image collection. Within the paper an investigation on UAV-based 3D data capture will be presented. For this purpose dense 3D point clouds are generated for a test area from three different platforms: first a UAV with a light weight compact camera, second a system using a system camera and finally a medium-format airborne digital camera system. Despite the considerable differences in system costs, suitable results can be derived from all data, especially if large redundancy is available such highly overlapping image blocks are not only beneficial during georeferencing, but are especially advantageous while aiming at a dense and accurate image based 3D surface reconstruction.