The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
1208 
Canada the UAV-system CropCam (http://www.cropcam.com/ 
index.htm) is in the market and in Europe a helicopter based 
system is under development, Annen und Nebiker, 2007. 
However current civilian and affordable UAVs are still not able 
to fulfil the previously listed requirements. In recent years 
UAVs have rarely been considered for civilian photogrammetry. 
The system design and the photogrammetric results largely 
depend upon the UAV and the sensor systems, such as the 
digital camera and the GPS/INS used, e.g. Jang, et al., 2004, 
Haarbrink & Koers, 2006. The accuracy of the GPS and the 
inertial measurement unit (INS) determines the degree of 
automation by means of aerotriangulation or direct 
georeferencing. Using mini or micro UAV-systems with 
consumer type digital camera for mapping and photogrammetry, 
several problems have to be mastered: 
• No vertical adjustment of the aerial camera, which results in 
tilted pictures caused by wind influences or instability of 
the platform. Furthermore to ensure complete photogram 
metric block configuration the end lap and the side lap have 
to be relatively high (70 / 70 %), compared to standard ae 
rial surveys. 
• Due to the small system size and the low-cost approach, 
small, inaccurate GPS-receivers and INS with a strong drift 
are used which do not allow for direct georeferencing. The 
quality is not even good enough to provide reasonable start 
ing EO-values for the aerotriangulation process. 
• Consumer grade cameras have an unknown or variable 
interior orientation of the camera. 
• Many images with small footprints, due to legal restrictions 
in Germany and elsewhere, RPVs and UAVs may fly only 
at altitudes of less than 300 m. In turn this leads to a small 
footprint of a single image. For instance a wide angle image 
will cover an area of only 200 * 300 m. 
• The number and distribution of the ground control points 
(GCP). Due to limited accuracy of the exterior orientation, 
the small footprints, critical overlaps due to winds and other 
factors a large number of GCPs are necessary. 
Beside the conventional approaches of aerotriangulation of 
image blocks, other procedures were developed in recent years 
which use existing ortho imagery of the area, e.g. the module 
Autosync/ERDAS Imagine, Jizhoua et al. (2004). Other 
approaches, e.g. Labe und Forstner (2005) automatically 
determine the relative orientation of overlapping images. There 
is however still a strong demand for further research. 
3. COMPARISON OF TWO MICRO UAVs 
The empirical tests of the photogrammetric potential with two 
different micro-UAVs were conducted within the Masters thesis 
of Engel, 2007. In the following sections the two different 
systems will be presented and the results of the empirical flight 
tests will be presented under the special focus of their 
photogrammetric performance. 
3.1 Micro-UAV “Carolo P330” 
Mavionics GmbH develops and sells autonomous aircrafts for 
different civilian fields of applications. One of the systems is 
called „Carolo P330“, which has been used for the practical 
tests, see (Figure 1). 
The UAV consists of a model plane, the Mavionics autopilot 
system, including data transmission, the mission control 
software (MCS) and a remote control for manual manoeuvres. 
The model aircraft is powered by a brushless DC-motor. The 
standard payload is a cheap off the shelf digital camera (Canon 
Powershot S60). The technical parameters of the UAV are 
compiled in Table 1. 
Figure 1: UAV “Carolo P330” (www.mavionics.de) 
The functionality and the necessary workflow for 
photogrammetric aerial surveys may be subdivided into three 
different steps:At the beginning a georeferenced map with the 
survey area is put into the Mission Control Software (MCS). 
The next step the MCS generates the flight strips with the 
necessary information such as distance between adjacent strips, 
flight height, flight speed. Every strip has a minimum of two 
waypoints at the beginning and the end of a strip. The turn to 
the next strip is flown in a constant radius. Thereby the 
minimum radius is limited to 30° in the roll angle by the MCS 
in order to avoid an instability of the aircraft. The GPS-module 
within the autopilot is important for the navigation and also 
delivers the coordinates / time of the perspective centres of the 
acquired images. The inertial sensor system of the autopilot 
which is tightly coupled with the GPS measures the rotation 
speed and the translatory inertia. Therefore approximate values 
of the exterior orientation are available. 
Starting and landing of the model plane is still done manually. 
After the start, once the aircraft is approximately at the first 
waypoint, the autonomous control of the UAV takes over. 
During the flight the course of the model plane may be visually 
controlled by a laptop. After the successful aerial survey the 
RC-pilot takes over again and lands the UAV safely. 
3.2 Low-cost Micro-UAV “SUSI” 
The micro-UAV called “SUSI” serves for the state forestry 
administration of Mecklenburg-Vorpommern and was formerly 
developed by the French company ABS Aerolight. “SUSI” 
assists the forestry administration in their every day operation. 
The UAV is based on a low-weight tubular frame with three 
low pressure tires. Within this frame the 4.2 KW two-stroke 
engine, the digital camera, the servo-mechanism, the batteries, 
petrol tank and an additional video camera are placed. During 
the flight a paraglider with a surface of 3.8 m 2 keeps the UAV 
in the air. The paraglider is responsible for a slow and non 
problematic handling and ensures a high reliability in case of an 
engine failure, Figure 2 
The core of the UAV is the digital camera within a gimbal- 
mounted platform in order to obtain near nadir looking images.