Recent advances in structure from motion (SfM) and dense matching algorithms enable surface reconstruction from unmanned aerial vehicle (UAV) images with high spatial resolution, allowing for new insights into earth surface processes. However, accuracy issues are inherent in parallel-axes UAV image configurations. In this study, the quality of digital elevation models (DEMs) is assessed using images from a simulated UAV flight. Five different SfM tools and three different cameras are compared. If ground control points (GCPs) are not integrated into the adjustment process with parallel-axes image configurations, significant dome-effect systematic errors are observed, which can be reduced based on calibration parameters retrieved from a testfield captured with convergent images immediately before or after the UAV flight. A comparison between DEMs of a soil surface generated from UAV images and terrestrial laserscanning data show that natural surfaces can be very accurately reconstructed from UAV images, even when GCPs are missing and simple geometric camera models are considered.

No abstract

In many close‐range applications it is essential to obtain information about the geometry of the target surface as well as its chemical composition. In this study, close‐range hyperspectral imaging was integrated with terrestrial laser scanning to provide mineral and chemical information for geological field studies. The spectral data was collected with the HySpex SWIR‐320m sensor, which operates in the infrared spectrum between the wavelengths of 1·3 and 2·5 μm. This sensor permits surfaces to be imaged with high spectral resolution, allowing detailed classification and analysis to be carried out. Photogrammetric processing of the hyperspectral imagery was achieved using an existing geometric model for rotating linear‐array‐based panoramic cameras. Bundle block adjustment of multiple images resulted in the registration of the spectral images in the lidar coordinate system, with a precision of around one image pixel. Although the image and control point network was not optimised for photogrammetric processing, it was possible to recover the exterior camera orientations, as well as additional camera calibration parameters. With the known image orientations, 3D lidar models could be textured with hyperspectral classifications, and the quality of the registration determined. The integration of the hyperspectral image products with the terrestrial lidar data enabled data interpretation and evaluation in a real‐world coordinate system, and provided a reliable means of linking material and geometric information.

No abstract

Digital panoramic cameras represent a powerful tool for generating high resolution images of scenes. They generate images of up to 100 000 · 10 000 pixels and are especially suited for 360°recording of objects such as indoor scenes or city squares. The paper describes the development of a strict geometric model for rotating linear array panoramic cameras and the extension of the model by additional parameters adapting the camera model to the physical reality. The camera model has been implemented in a spatial resection and a bundle solution. The bundle solution also allows for the combined handling of panoramic and central perspective images. In several practical tests a potential accuracy of around ¼ pixel was demonstrated.