Evaluations on multi-scale camera networks for precise and geo-accurate reconstructions from aerial and terrestrial images with user guidance

Markus Rumpler, Alexander Tscharf, Christian Mostegel, Shreyansh Daftry, Christof Hoppe, Rudolf Prettenthaler, Friedrich Fraundorfer, Gerhard Mayer, Horst Bischof
2017 Computer Vision and Image Understanding  
During the last decades photogrammetric computer vision systems have been well established in scientific and commercial applications. Recent developments in image-based 3D reconstruction systems have resulted in an easy way of creating realistic, visually appealing and accurate 3D models. Affordable high quality digital consumer grade cameras and especially the increasing availability of unmanned aerial vehicles (UAVs) in conjunction with automated multi-view processing pipelines contributed to
more » ... these developments. With multi-copter UAVs it is possible to capture images with six degrees of freedom and thus generate a smooth transition from terrestrial images to nadir aerial images. High accuracy in the reconstruction and a precise georegistration is desired for photogrammetric applications such as mapping, recurrent topographic surveys or architectural and archaeological 3D documentations. In this paper we present a fully automated processing pipeline for precise, metric and geo-accurate 3D reconstructions of complex geometries using various imaging platforms. Our approach integrates automatic camera calibration and an online feedback method that allows assessment of the final reconstruction quality in terms of image overlap, ground sampling distance and completeness, and thus provides flexibility to adopt the image acquisition strategy already during image recording. Our approach allows for georeferencing of UAV imagery based on GPS-measurements as well as tie and control point information, which are integrated directly in the bundle adjustment to refine the georegistration and correct for systematic distortions of the image block. Our approach is furthermore suited for seamlessly matching and integrating images with different scales, from different view points (aerial and terrestrial as well as inside views), and different cameras into one single reconstruction. We evaluate our approach based on five different case studies for applications in mining, archaeology and urban environments and present several accuracy related analyses investigating camera calibration, georegistration, camera network configuration and ground sampling distance. 1
doi:10.1016/j.cviu.2016.04.008 fatcat:w34odookijb5bg3crzb4h46ewi