Aspects of Dem Generation From Uas Imagery

Abstract

ABSTRACT:Since a few years, micro UAS (unmanned aerial systems) with vertical take off and landing capabilities like quadro-or octocopter are used as sensor platform for Aerophotogrammetry. Since the restricted payload of micro UAS with a total weight up of 5 kg (payload only up to 1.5 kg), these systems are often equipped with small format cameras. These cameras can be classified as amateur cameras and it is often the case, that these systems do not meet the requirements of a geometric stable camera for photogrammetric measurement purposes. However, once equipped with a suitable camera system, an UAS is an interesting alternative to expensive manned flights for small areas. The operating flight height of the above described UAS is about 50 up to 150 meters above ground level. This low flight height lead on the one hand to a very high spatial resolution of the aerial imagery. Depending on the cameras focal length and the sensor's pixel size, the ground sampling distance (GSD) is usually about 1 up to 5 cm. This high resolution is useful especially for the automatic generation of homologous tie-points, which are a precondition for the image alignment (bundle block adjustment). On the other hand, the image scale depends on the object's height and the UAV operating height. Objects like mine heaps or construction sites show high variations of the object's height. As a result, operating the UAS with a constant flying height will lead to high variations in the image scale. For some processing approaches this will lead to problems e.g. the automatic tie-point generation in stereo image pairs. As precondition to all DEM generating approaches, first of all a geometric stable camera, sharp images are essentially. Well known calibration parameters are necessary for the bundle adjustment, to control the exterior orientations. It can be shown, that a simultaneous on site camera calibration may lead to misaligned aerial images. Also, the success rate of an automatic tie-point generation differs extremely between several photogrammetric software packages. In this article, the calibration results of a suitable camera system will be shown. For a small format consumer grade camera, the authors will give the proof of ability for photogrammetric measurements purposes. This includes the results of different processing approaches for DEM generation of environments showing high object height variations.

Inventory measurements based on non-metric aerial photos from the board of unmanned aerial vehicle (UAV)

“…Due to the large range of literature the topic in the review of this work focuses mainly on publications from 2013, this is the time when similar works were undertaken at the Poznan University of Technology. Unmanned aerial systems (UAS [5,9,10]) are called also remotely controlled systems (RPAS [2,7]), multirotor devices [1,13] or models of airplanes [5], sometimes limited to a wing only [2,14]. The main difference in their use is a different flight path -more flexible in the case of rotor's than wing vehicles.…”

“…There are used also other cameras recording in 'cropped-frame' CMOS technology (APS-C format). An alternative, due to the different way of building a sensor is Foveon, which in one cycle can acquire all color components, what eliminates many drawbacks of CMOS solutions in DSLR cameras [10]. Interesting for the purpose of remote sensing would be recording in near infrared band, however CMOS sensors typically do not provide such a possibility -short waves are optically filtered in the way to sensor.…”

“…Sieberth et al [20] drew furthermore attention, to a significant impact in the case of frequent drone blur images on the accuracy of matching, and later results of elaborations. The authors proposed their own self-correction algorithm, but Greiwe et al [10] argue that sharper images provide cameras with Foveon sensor. Cameras that are not built permanently into the structure of drone, are suspended by stabilizing module (preferably 2-or 3-axial gimbal), what allows to control the orientation of the camera, but also reduces the blur of images.…”

Modelling Steep Surfaces by Various Configurations of Nadir and Oblique Photogrammetry

“…Recent literature shows how UASs are becoming the most useful solution for pit monitoring. The topics covered are: camera and system calibration (Shahbazi, 2015), imagery bundle adjustment and orientation quality assessment (Rosnell, 2012;Greiwe, 2013Tong, 2015, point cloud generation (Sauerbier, 2011;Rosnell, 2012;Shahbazi, 2015), the attainable products (digital models and orthophotos), data validation (Niethammer, 2012;Bemis, 2014;. The use of oblique imagery is only initial (Greiwe et al 2013).…”