NOTHING SAYS MORE ABOUT A PERSON OR THEIR LIFESTYLE THAN THEIR WATCH.
CALLAWAY GOLF TIMEPIECES ARE KNOWN FOR INNOVATION, QUALITY AND ATTENTION TO EVERY DETAIL. From materials to movement, function to accent features, Callaway Golf Timepieces combine modern and colorful X Series and Sport styles with elegant, distinctive luxury styles from our Callaway Collection line. Each piece is as unique as the golfer wearing it. We believe its more than just a brand or productits a way of life.

COLLECTION

CY2019
Mens stainless steel, 3 piece case, black dial, chronograph movement, date function, butterfly clasp, 5 ATM

CY2042

Mens stainless steel, 3 piece case, silver dial, chronograph movement,

(a) (b)

date function, butterfly clasp, 5 ATM

CY2094

Mens stainless steel case, Silver dial, chronograph movement, butterfly clasp, 5 ATM

CY2095

Mens stainless steel case, black dial, chronograph movement, black croco strap, 5 ATM

CY2085

Mens stainless steel case Molded PU rubber strap, black dial with black dial ring, chronograph movement, 5 ATM

(c) (d)

CY2089
Mens stainless steel case, silver dual dial with chevron pattern on top dial, luminous hands and hour markers, black croco strap, 5 ATM

CY2096

Mens stainless steel bracelet with brown dual dial feature, date function, butterfly clasp, 5 ATM

CY2099

Mens stainless steel case, black dual dial feature, date function, black leather strap, 5 ATM

CY2098

Mens stainless steel two tone bracelet with silver dual dial feature, date function, butterfly clasp, 5 ATM

CY2097

Mens stainless steel case, brown dual dial feature, date function, brown leather strap, 5 ATM

S PORT

CY2059 CY2062 CY2058 CY2061 CY2060 CY2063
Mens stainless steel 3 piece case, blue dial, day/date function, rotating bezel, chevron logo bracelet, 5 ATM
Ladies stainless steel, 3 piece case, blue dial, day/date function, rotating bezel, chevron logo bracelet, 5 ATM
Mens stainless steel, 3 piece case, silver dial, day/date function, rotating bezel, chevron logo bracelet, 5 ATM
Ladies stainless steel, 3 piece case, silver dial, day/date function, rotating bezel, chevron logo bracelet, 5 ATM
Mens stainless steel, 3 piece case, silver dial, two-tone case and bracelet, day/date function, rotating bezel, chevron logo bracelet, 5 ATM
Ladies stainless steel, 3 piece case, silver dial, two-tone case and bracelet, day/date function, rotating bezel, chevron logo bracelet, 5 ATM

CY2068

CY2091 CY2092 CY2090 CY2093
Mens stainless steel case with black dual dial feature, luminous hands, black croco strap, 5 ATM
Mens stainless steel, 3 piece case, black dial, multi-function, black perforated leather strap, 5 ATM

CY2066

Ladies stainless steel, black dual dial feature, luminous hands, black croco strap, 5 ATM
Mens stainless steel, 3 piece case, white dial, smoky brown crystal, multi-function, brown perforated leather strap, 5 ATM
Mens stainless steel case with silver dual dial feature, luminous hands, white croco strap, 5 ATM

CY2075

Mens stainless steel, 3 piece case, silver dial, 3 hand/date function,

(a) (b) (a) (b)

black leather strap, 5 ATM
Ladies stainless steel, silver dual dial feature, luminous hands, white croco strap, 5 ATM

CY2076

Ladies stainless steel, 3 piece case, silver dial, 3 hand/date function, black leather strap, 5 ATM

CY2081 CY2082 CY2083

Mens stainless steel, 3 piece case, black dial with red accent on dial ring, black PU rubber strap, 5 ATM
Mens stainless steel, 3 piece case, blue dial with red accent on dial ring, blue PU rubber strap, 5 ATM

Mens stainless steel, 3 piece case, blue dial with yellow accent on dial ring, blue PU rubber strap, 5 ATM

X SERIES

CY2070
Mens stainless steel, 3 piece case, orange dial, chronograph function, orange PU rubber strap, 10 ATM

CY2069

Mens stainless steel, 3 piece case, green dial, chronograph function, white PU rubber strap, 10 ATM

CY2074

Mens stainless steel, 3 piece case, orange dial ring, black digital display, analog and digital function,
black leather strap, 10 ATM

CY2071

Mens stainless steel, 3 piece case, black dial, chronograph function, black PU rubber strap, 10 ATM
CY2102 CY2103 CY2101 CY2100 CY2088
Mens stainless steel, black sunray dial with IP top ring, chronograph movement, black PU rubber strap, 5 ATM
Mens stainless steel orange sunray dial, chronograph movement, brown PU rubber strap, 5 ATM
Mens stainless steel, black sunray dial with black IP top ring, chronograph movement with tachymeter, spring buckle, 5 ATM
Mens stainless steel, blue sunray dial, chronograph movement with tachymeter, spring buckle, 5 ATM
Mens stainless steel, 2 piece case, black dial with black IP plating, chronograph movement, 5 ATM

WOMENS

CY2104
Ladies stainless steel, two-tone bracelet with white MOP rectangle dial, Swarovski crystals at 12 and 6, 5 ATM

CY2105

Ladies stainless steel, blue MOP rectangle dial, Swarovski crystals at 12 and 6, 5 ATM

CY2107

Ladies stainless steel, two tone bracelet with white MOP oval dial, 5 ATM

CY2108

Ladies stainless steel, blue MOP oval dial, 5 ATM

Self Adjusting Links

CY2022
Ladies stainless steel, 2 piece case with Swarovski crystals in bezel, white MOP dial, 3 hand function, butterfly clasp, 5 ATM

CY2048

Ladies stainless steel, 2 piece case with Swarovski crystals in bezel, champagne dial, two-tone case and bracelet, 3 hand function, butterfly clasp, 5 ATM

CY2057

Ladies stainless steel and PU rubber, 3 piece case in IP pink plate and stainless
steel, pink MOP dial, chronograph function, rotating bezel, brushed steel and pink PU rubber bracelet, 5 ATM

CY2056

Ladies stainless steel and PU rubber, 3 piece case in IP white plate and stainless steel, white MOP dial, chronograph function, rotating bezel, brushed steel and white PU rubber bracelet, 5 ATM

CY2040

Stainless steel / plastic golf bag fob, plastic case base, stainless steel front plate, temperature function, gray nylon with metal golf bag clip, 3 ATM

MENS COLLECTION

PA G E

$160 $160 $185 $160 $160 $110 $135 $110 $135 $110
$155 $155 $180 $155 $155 $105 $130 $105 $130 $105
$150 $150 $175 $150 $150 $100 $125 $100 $125 $100
$145 $145 $170 $145 $145 $95 $120 $95 $120 $95
CALLAWAY GOLF CUSTOMIZED VIRTUAL COMP
CY2042 CY2094 CY2095 CY2085 CY2089 CY2096 CY2099 CY2098 CY2097

GENERAL INFORMATION

SPORT SERIES
CY2059 CY2062 CY2058 CY2061 CY2060 CY2063 CY2091 CY2092 CY2090 CY2093 CY2068 CY2066 CY2075 CY2076 CY2081 CY2082 CY2083
$95 $95 $95 $95 $95 $95 $110 $110 $110 $110 $110 $110 $85 $85 $95 $95 $95
$90 $90 $90 $90 $90 $90 $105 $105 $105 $105 $105 $105 $80 $80 $90 $90 $90
$85 $85 $85 $85 $85 $85 $100 $100 $100 $100 $100 $100 $75 $75 $85 $85 $85
$80 $80 $80 $80 $80 $80 $95 $95 $95 $95 $95 $95 $70 $70 $80 $80 $80
CY2070 CY2069 CY2074 CY2071 CY2102 CY2103 CY2101 CY2100 CY2088
$110 $110 $110 $110 $135 $135 $160 $160 $210
$105 $105 $105 $105 $130 $130 $155 $155 $205
$100 $100 $100 $100 $125 $125 $150 $150 $200
$95 $95 $95 $95 $120 $120 $145 $145 $195

WOMENS COLLECTION

CY2104 CY2105 CY2107
$95 $95 $95 $95 $110 $110 $135 $135
$90 $90 $90 $90 $105 $105 $130 $130
$85 $85 $85 $85 $100 $100 $125 $125
$80 $80 $80 $80 $95 $95 $120 $120
ART REQUIREMENTSART REQUIREMENTS
Vector art is requested due to the need our designers have when working with logos, illustrations, type and graphics that are typically scaled to different sizes. Vector graphics, which Adobe Illustrator and InDesign create, also called draw graphics, consist of smooth lines that retain their clarity when scaled up or down to size. If no vector art can be provided, 300 dpi camera ready art or art on disc will ensure production of the highest quality.
CY2108 CY2022 CY2048 CY2057 CY2056

G O L F B A G WAT C H

DECORATED GIFT PACKAGING
(run charge) $1.00 ( c ) per color (set up) $100 ( c ) per color Lead time 2 weeks
*no charge for blank packaging
CONTACT INFORMATIONCONTACT I
Phone: 800.699.9580 Fax: 972.699.6995 www.callawaygolfwatches.com PRICING
Pricing includes up to a 2 color dial imprint Additional colors add $1.00 (c) (run charge) Engraving $5.00 (c) MOQ 15
Set up: $100.00 Set up: $100.00 Set up: $100.00

(c) (c) (c)

No LTM for decorated product. Production lead time 2 weeks from computer design approval (based on availability). Same for reorders. Virtual samples on e-mail are free of charge.

ASI55145 PPAI196495

Callaway Golf, X Series, Callaway Collection and the chevron device are trademarks of Callaway Golf Company. Sold by an official licensee of Callaway Golf Company. www.callawaygolfwatches.com

I List of Figures Figure 7-12: DSM2 (April 2009) image data generated with SAT-PP.. 158 Figure 7-13: Comparison of the DSM1 and DSM2 of the gravel pit.. 158 Figure 7-16: Schema for the GIS-database.. 164 Figure 7-17: Top (Experiment A) and bottom (Experiment B) figures show a screenshot of the 3D-model using the Swissimage orthoimage
(swisstopo) draped over the DHM25 (swisstopo) combined with the orthoimage of the maize field and the position of the sampled receptor plants generated from the UAV images. Yellow donor areas are grey emphasized; the remaining areas are those of the white receptor maize. 166 Figure 8-1: UAVs: Ready for new tasks... 169

II. LIST OF TABLES

Table 1-1: Table 3-1:
Features of Aerial, close range and UAV photogrammetry.. 3 Classification of UAVs according to the classes unpowered and powered, as well as lighter or heavier than air... 34

Table 3-2:

Pro and cons of the different type of UAVs (0: Lowest value; +: Middle value; ++: Best)... 35
Table 3-3: Table 3-4: Table 3-5:
Classification of the CASA for UAVs... 36 Categorization with respect to price and payload of UAV systems. 38 Classification of UAVs regarding to the type of georeferencing, real time capability and application requirements... 38
Table 3-6: Table 3-7: Table 3-8: Table 3-9:
Regulation for UAVs in Switzerland (BAZL, 2001)... 39 Overview of societies and conferences related to UAVs.. 40 Overview of OM-class systems... 42 Overview of a selection of existing commercial quad- and multicopter MClass systems.... 44
Table 3-10: Given accuracy measures of the three wePilot1000 systems which have been used in our studies (weControl, 2009).. 52 Table 3-11: Overview of fixed-wing, single rotor and coaxial M- and L-class systems. 53 Table 4-1: Flight parameters for the mission planning of photogrammetric aerial flights.... 63 Table 4-2: Flight tools for UAVs showing the specification and possible application of the tool.... 65 Table 4-3: Attributes used for a waypoint in the weGCS software.. 71
II List of Tables Table 4-4: Table 4-5: Attributes used for a waypoint in the mdCockpit software.. 75 The mean value (xM), the mean corrected standard deviation Xdiff and the RMSE of the offset between GPS and tachymeter position at the same time (UTC)..... 90 Table 5-1: Table 6-1: Table 6-2: Table 6-3: Table 6-4: Comparison of the software packages... 97 Main features of the Riegl laser scanner LMS-Z420i.. 113 Overview of original data and derived products of the site Pinchango. 124 Results from the comparison of the different DSM datasets.. 125 Main parameter of the flight planning defined for area I (East and west court) and II (Main plaza).... 131 Table 6-5: Accuracy values of the image orientation of the east court image block of Copn.... 134 Table 7-1 : Parameter for the rotation of the DSM... 150 Table 7-2: Results of the orientation of the Helimap images in ISDM (ZI/Intergraph, 2009)..... 152 Table 7-3: Orientation values of the quadrotor images taken from the gravel pit in February and April 2009.... 156 Table 7-4: Table 7-5: Table 7-6: Parameters for the flight planing... 162 Number of aerial images used for photogrammetric processing.. 162 Comparison of the image acquisition points of the flight in 2006: Planned vs. Metadata (wePilot1000) and image orientation (oriented) vs. Metadata. 165

III Abbreviations ROA RPA RPV SAT-PP SLR camera TBD TIN SNF UA UAS Remotely Operated Aircraft Remotely Piloted Aircraft Remotely Piloted Vehicle Satellite Image Precision Processing Single-lens reflex camera To be defined Triangulated irregular network Schweizer National Font Unmanned Aircraft Unmanned Aircraft System consists of an Unmanned Aircraft (UA), a Control System (CS) -usually a Ground Control System (GCS) - and a communications data link between the UA and the CS. Unmanned Aerial Vehicle Block Triangulation of image data acquired from an UAV Digital Surface Model which was generated out of image data taken from an UAV Ultra High Frequency United States Universal Transverse Mercator Unmanned Vehicle Unmanned Vehicle System Virtual Reality Modeling Language Vertical takeoff and landing The Wallenberg Laboratory for research on Information Technology and Autonomous Systems World Geodetic System 1984 Extensible Markup Language Zeiss Intergraph
UAV UAV BT UAV-DSM UHF U.S. UTM UV UVS VRML VTOL WITAS WGS84 XML Z/I

INTRODUCTION

Figure 1-1: Model helicopter flying over the Campus Hnggerberg (ETH Zurich).

1.1 Definition of UAVs

UAVs are to be understood as uninhabited and reusable motorized aerial vehicles. states van Blyenburgh, 1999. These vehicles are remotely controlled, semi-autonomous, autonomous, or have a combination of these capabilities. Comparing UAV to manned aircraft, it is obvious that the main difference between the two systems is that on the UAV no pilot is physically present in the aircraft. This does not necessarily imply that an UAV flies by itself autonomously. In many cases, the crew (operator, backup-pilot etc.) responsible for a UAV is larger than that of a conventional aircraft (Everaerts, 2008). The term UAV is commonly used in the Computer Science, Robotics and Artificial Intelligence, as well as the Photogrammetry and Remote Sensing communities. Additionally, synonyms like Remotely Piloted Vehicle (RPV), Remotely Operated Aircraft (ROA) or Remotely Piloted Aircraft (RPA) and Unmanned Vehicle Systems (UVS) can also infrequently be found in the literature. RPV is a term to describe a robotic aircraft flown by a pilot using a ground control station. The first use of this term may be addressed to the United States (U.S.) Department of Defense during the 1970s and 1980s. The terms ROA and RPA have been used by National Aeronautics and Space Administration (NASA) and Federal Aviation Administration (FAA) in the U.S. in place of UAV. Furthermore, the term Unmanned Aircraft System (UAS) is also being used (Colomina, et al., 2008). The FAA has adopted the generic class UAS, which was originally introduced by the U.S. Navy. Common understanding is that the terminology UAS stands for the whole system, including the Unmanned Aircraft (UA) and the Ground Control Station (GCS).

3.1.1 Classification of UAVs
The definition of UAVs encompasses fixed and rotary wings UAVs, lighter-than-air UAVs, lethal aerial vehicles, decoys and targets, alternatively piloted aircrafts and uninhabited combat aerial vehicles. Sometimes, cruise missiles are also classified as UAVs (van Blyenburgh, 1999). Furthermore, UAVs can be categorized using the main characteristics of aircrafts like unpowered or powered, lighter than air or heavier than air and flexible, fixed or rotary wings (see Table 3-1). Table 3-1 shows a classification of the existing UAVs, which can be used for photogrammetric applications. Table 3-2 gives pro and cons of the in Table 3-1 classified systems regarding their range, endurance and weather, wind dependency and maneuverability.
Table 3-1: Classification of UAVs according to the classes unpowered and powered, as well as lighter or heavier than air.

Lighter than air

Heavier than air Flexible wing Fixed wing Gliders Rotary wing Rotor-kite

Unpowered

Balloon
Hang glider Paraglider Kites

Powered

Airship

Paraglider

Propeller Jet engines
Single rotors Coaxial Quadrotors Multi-rotors
Rotary wing UAVs, also known as vertical takeoff and landing vehicles (VTOL), can be further classified into single-, double-, four- and multi-rotor systems (see Table 3-1). Singlerotor systems have one main rotor and a tail rotor. The main rotor supplies lift and thrust; the tail rotor is used to counteract the yaw motion and the torque. Double-rotor systems, so called 34
3 UAV-Systems coaxial systems, differ mainly from single-rotor systems because they have an increased payload and they are able to operate at higher absolute altitude for the same engine power. In addition, they are more easily controllable and have a reduced noise level. These systems have approximately 30 percent more degree of efficiency, since all of the available engine powers is devoted in lift and thrust. However, coaxial systems have one main disadvantage: Using two main rotors mounted in the same rotor shaft, which need to rotate in opposite directions, results in an increased mechanical complexity of the rotor hub. In general, the existing single and double-rotor systems have more power than four- and multi-rotor systems. Accordingly, these kinds of systems are able to carry more payloads, which correspond into the number, size and weight of the sensors mounted on the UAV-system.

3.3.2.2 The Flight Control System wePilot 1000
The wePilot1000 is intended for professional applications. The wePilot1000 (see Figure 3-10) is a so called on-demand system, which means that the pilot controls the helicopter with directional commands (forward/lateral/vertical/heading velocity) from the joystick connected to the ground control station or by the geographic coordinates given by the flight path 48
3 UAV-Systems (waypoints). A combination of both operations is possible in order to interrupt, modify, or continue a specific mission. The wePilot1000 features the following main characteristics: Altitude stabilization and velocity control, Position and RC transmitter sticks interpreted as velocity commands, Integrated GPS/INS system, Barometric altimeter, Compass, Payload intensive flight controller, Built-in data logger and telemetry capability, Programmable hardware for rapid customization, Flight control software weGCS and Embedded computer system.
Figure 3-9: Scheme for the flight control system wePilot1000.
The wePilot1000 also provides safety enhancements in case of loss of data communications or failure of the ground control station. Specifically, the wePilot1000 supports a fully autonomous homing process to a previously defined home location. Furthermore, the flight control system can stabilize the helicopter under severe weather conditions, including persistent strong wind and wind gusts. The stabilization performance of the wePilot1000 is an important feature, which allows the use of the system for aerial photography, aerial video recording, or infrared night operations.
3.3 Micro & Mini UAV Systems: M-Class The flight control system interacts with the helicopter as shown in Figure 3-9. The helicopter and its integrated sensors provide the navigation unit with input. Using an extended Kalman Filter the system calculates the position, orientation, speed and acceleration of the system online (Eck, 2001). These values are transmitted to the Controller, which checks the differences to the given values. Identifying the differences, the Controller sends new action commands to the helicopter to correct the actual values. In addition the navigation data are transmitted to the Guidance, which checks the position to the predefined flight path, which is or can be defined and modified by operator of the system. The Guidance unit gives reference commands to the controller as well, which are transferred to the helicopter again as actuator commands. Moreover, the computer board allows data communication with various payloads such as gimbals, video and photo cameras. It also includes black box data recording capability for post-flight data processing in case of detailed flight analysis. The wePilot1000 electronics is also well isolated against high frequency engine vibrations.

The single terms are defined as follows:
4 Project workflow and image data acquisition Object point in object coordinate system (Global or local coordinate system) Ob. GPS/INS position in the object coordinate system Ob. Vector between INS and the projection center of the camera, defined in the INS coordinate system. Vector between GPS (antenna phase center) and INS, defined in the INS coordinate system. Rotation of the INS coordinate system into the object coordinate system Ob. The rotation angles are defined through the GPS/INS modul. Rotation of the image coordinate system P into the INS coordinate system.
4.2.2.3 WeGCS weControl Ground Control Station
As input for weGCS, a text file with 3D coordinates of the acquisition points, parameters for flying velocity and the definition of point status (stop, crossing or turning point) is generated. These coordinates can be transferred into the weGCS-Mission file (XML-file). For each way point the attributes are defined (Table 4-1). The first and the second way point define the start and way point respectively, while the following numbers stand for a mission point. The class identifies the point absolutely or relatively with respect to the start point. Behavior of the point stands for cruising or stop, while the payload enables or disables the payload function.
Table 4-3: Attributes used for a waypoint in the weGCS software.
Waypoint Number Class Behavior Speed Payload RelativeData (Northing, Easting, Height) AbsoluteData (Longitude, Latitude, m a.s.l. Height)
4.2 Flight Planning The RelativeData and AbsolutData are the X, Y, Z or Lo, La, Hortho coordinates of the way point (See Table 4-3). The maximum speed vmax is limited through vmax (UAV) and vmax (w). vmax is defined as: 4-11
The AbsoluteData can be calculated with respect to a starting point P1 using the following parameters 4-13
The RealtiveData of a point can be calculated from the difference of the coordinates of the start point and the AbsoluteData. 4-14
The acquisition of the images using the wePilot1000 is possible in different modi: The images can be acquired through manual triggering on a stop point during the manual, as well as the autonomous, flight (see Figure 4-9). The image acquisition can also be automatically enabled after arriving a predefined stop or cruising point.

4.3.1 Example model helicopter
In the flight test in Sweden 2003 the main focus was on the integration of a differential GPS into the UAV system (Eisenbeiss, 2003). Given our current study, has focused on the importance of using autonomous flights for appropriate data processing, the outcomes of the comparison of manual and autonomous flight in Eisenbeiss (2003) are described below.
Figure 4-14: Predefined flight path showing the flight lines, runway markings (grey blocks) and control points (Test Motala, Eisenbeiss, 2003).
Figure 4-15: Image cover of the first flight line (I) shown in Figure 4-14.
Figure 4-16 shows an example of a manual controlled flight under extreme wind conditions. An experienced pilot controlled the system, attempting to fly two flight lines (see Figure 77
4.3 Manual versus autonomous flight 4-14). Figure 4-15 and Figure 4-16 show clearly that without stabilization, it is impossible to fly a precise block configuration under such wind conditions. Therefore, the results of the image orientation were not successful.
Figure 4-16: Trajectory of the manually controlled flight (Test Motala, Eisenbeiss, 2003).
After this test, the helicopter was stabilized using a GPS/INS system from C-MIGITS II (Rockwell, 1996), which enabled the data acquisition to be in an optimal block configuration (see Figure 4-17).
Figure 4-17: Trajectory of the autonomous controlled flight (Test Motala, Eisenbeiss, 2003).
Figure 4-17 shows that the helicopter was able to follow the predefined flight path in the autonomous mode, while the starting and landing phase was controlled manually. Moreover, 78
4 Project workflow and image data acquisition the turbulent part of the flight trajectory indicate the transition between manual and autonomous control, where the helicopter tried directly to move towards the desired position. However, at the time of this test, the action commands of the helicopter for transition areas were not defined probably, which resulted into the turbulent trajectory.
4.3.2 Manually controlled Kites and Zeppelin 4.3.2.1 Kites
The system Susi described in chapter 3.2.2 is already used for the production of orthoimages of forest areas in the forestry department Mecklenburg-Western Pormania (Germany). An example of a typical block configuration is shown in Figure 4-18. The images are distributed irregular over the area of interest, while also the image cover varies significantly from one to next image. Furthermore, the orientation and the shape of the images are changing in almost all images, which forebode the change of the orientation angle. Thus, the block configuration is not sufficient for an optimal image orientation.

Figure 4-18: Typical block configuration of the motorized kite Susi over a forestry area (Source: Forestry department Mecklenburg-Western Pormania).
However, for economic reasons the data acquisition and the analysis of these data are worthwhile. Quite often, the areas of interest are too small for commercial photogrammetric flights, since the cost-value ratio is not arguable. In addition, frequently during the relevant time frame for forestry analysis, the weather does not allow manned flights, while in the area of interest the weather conditions had been sufficient enough. The manual measurement 79
4.3 Manual versus autonomous flight effort, which is conditional through the non ideal block configuration, is maintainable. Even though, since the photogrammetric flight companies quite often change their operating system. Finally, the establishment of new workflows for the new systems, also results into more manually work for. The same system is also currently used at the Hydrology and Water Resources Management group at ETH Zurich. The system is applied in the CCES project RECORD (REstored COrridor Dynamics; Record, 2009). The goal of this work is on the influence of vegetation roots on the morphodynamics processes that shape the river morphology as the restoration progresses. A detailed DEM is need at any time there, when relevant changes due to big floods occur. The purpose of having a manual controlled UAV was to check if this kind of image data is a valuable low cost alternative to LiDAR and large format aerial images in such an environment (i.e., in the presence of flowing water). The acquired images of the first flight were oriented using the software Leica Photogrammerty Suite (LPS, see section 5.1). The significant changes of HG (see Table 4-1) and the Yaw angle showed that the systems are strongly influenced by local wind conditions (see Figure 4-19). Apparently, from the image configuration, problems in the image orientation and matching occur due to the fact that half of the images are covered from the water surface.
Figure 4-19: Block configuration of the motorized kite Susi used in the RECORD project at ETH Zurich (Hagenbach and Wanner, 2009).

4.3.2.2 Zeppelins

Figure 4-20: Example of a Zeppelin used at Samoilov island (Source: AWI, 2009).
A Zeppelin has a similar concept to the balloon. The main difference between the systems is the shape of a zeppelin. It is more like an airship and, thus, is easier to control and keep stable on a flight trajectory. Usually just one wire is necessary for manual control and the flight trajectory is easier to keep on line. The system used at Samoilov island was constructed from two Zeppelins. Two Zeppelins were necessary to carry the payload (Ries and Marzolff, 2003). Because of the inaccessibility of Samoilov island, it was not possible to define flight lines for a complete coverage of the area. The connecting flight lines had also only minimal overlap, which meant the orientation of the whole image block was not possible at once (see Figure 4-21). These circumstances prolonged the processing of the data significantly. However, it was possible to orient the images taken from the Zeppelin at Samoilov island in separate image blocks, and to produce orthoimages of the area of interest (the workflow of the image orientation and the generation of photogrammetric products using UAV-images are described in Chapter 5). Since the height differences of the area were not significant, the DSM was reduced to a planar surface using the average height of the terrain.

Figure 4-26: Measured GPS velocity values of a flight in Copan (Honduras).
The results presented above show clearly the differences between the flight modi of our UAV system Copter 1B. For independent analysis of the flight trajectory we analyzed the trajectory by tracking the position of the system through a tachymeter. The method and experimental results are described in the following section.
4.4.3 Tracking tachymetry
The precise reconstruction of the 3D-trajectory of mini-UAVs in real time can be determined with several geodetic measurement methods. In order to achieve real-time 3D measurement data with accuracy better than 1 cm and high temporal resolution (5Hz) , only tracking total stations can be used (Stempfhuber and Kirschner, 2008). This kind of technology is particularly well suited for applications, which require millimeter accuracy in planimetry and height for object distances up to 300 m, which is a normal distance of an UAV image strip. The polar coordinate measurement system determines, using the 3D-distance as well as the horizontal and vertical angle, the XYZ-position of a moving 360 prism with respect to a certain time stamp. In the last years, all serious manufacturers of geodetic measurement systems worked on the optimization of kinematic observations of total stations. GNSS, Lasertracker, INS-systems, Interferometer and LiDAR-Systems however do not meet the 87
4.4 Analysis of the trajectory of autonomous UAV flights requirements for this particular task. Additionally, the tracking of the helicopter using high speed cameras could be applied in small areas, while for larger areas, due to the limitation of the field of view of the camera, a new construction for the automatic target tracking would be necessary. Thus, we decided to use the available tracking tachymeter. The tracking tachymeter is mainly used for the control of construction machines and surveying in the one-man mode. It is also used for kinematic measurements in road construction, where grading, milling off existing layers, compaction and, above all, installation of road surfaces has to be carried out to a level of precision of a few millimeters. Applications such as tunnel and airport construction require a comparable precision (Kirschner and Stempfhuber, 2008). Considering the influence of different parameters like: - the maximum tracking speed, the precision of the 360-prism, the dynamic behavior of the integrated two-axis tilt sensor, the synchronization of the angle and distance measurements, as well as the latency and dead times and the behavior of the target tracking,
it is possible to achieve an absolute accuracy of few millimeters. This is not particularly required for the analysis of the flight trajectory of image data acquisition. However, the potential accuracy would help to analyze the flight trajectory of an UAV with an implemented LiDAR system in future work. More parameters and detailed information about their influence on accuracy are described in Kirschner and Stempfhuber, 2008.

5.2 Photogrammetric products If more than two images are available, the MPGC procedure can use them simultaneously and the matching results are more robust. Here, the resulting DSM from an image pair can be used as an approximation for the MPGC procedure. Through the quality control procedure, e.g. using the local smoothness and consistency analysis of the intermediate DSM at each image pyramid, the analysis of the differences between the intermediate DSMs, and the analysis of the MPGC results, blunders can be detected and deleted. For each matched feature, a reliability indicator is assigned. This is based on the analysis of the matching results from cross-correlation and MPGC. This indicator is used for assigning different weights for each measurement, which are used when a regular grid is interpolated (Zhang, 2005). During this study, the in-house developed software SAT-PP was intensively evaluated and for the processing of UAV-images additional parameters of the interior orientation were integrated: for example the 10 Brown (Brown, 1971) parameters implemented in BUN and LPS and the 12 Ebner parameters (Ebner, 1976) used in Z/I. The software SAT-PP allows similar to LPS, the definition of excluded areas, while the adaptive parameters are generated automatically during the matching process. So far, the output DSM formats are limited to the ASCII formats ArcASCII (ESRI) and XYZ.
5.2.1.3 NGATE and Match-T
During the most recent years, more and more commercial software packages, which combine various matching strategies for the DSM generation, have become available on the market. In the following section we will briefly describe two promising software packages. Next-Generation Automatic Terrain Extraction (NGATE) is a tool implemented into SocetSet (BAE-Systems, 2009) and provides automatic generation of elevation models by multi-imagematching and correlation. NGATE combines area and feature based matching algorithms. The feature based matching also includes edge matching, similar to SAT-PP. In addition to NGATE, also Match-T DSM from Inpho (Inpho, 2009b) combines different matching techniques in a multi-image-matching approach. Several features like an intelligent multi-image matching through on-the-fly selection of the locally best suited images for DTM generation or the consideration of pre-measured morphological data (breaklines, 2D and 3D
5 Photogrammetric data processing exclusion areas, borderlines) and the elimination of outliers, e.g. trees, houses, by robust finite element interpolation are available with this software (Inpho, 2009b). The practical experiments shown in the following chapters have been mainly processed using the software packages SAT-PP and LPS. In addition, one practical experiment using NGATE is given in section 6.2.5.1.

7.1.4.3 Helimap images

The orientation of the oblique images was done in LPS and ISDM. Tie points were automatically and manually measured as well as visually checked in LPS. From the GPS/INS measurements initial exterior orientation parameters were available. They even allowed stereo viewing of the images with a y-parallax, which results from the accuracy level of the GPS/INS-system. Therefore, by using only the given exterior orientation the processing of the images was not possible. The GCPs were generated by precise geodetic measurements through the engineering company Geomatik AG (Geomatik, 2009 and Figure 7-6). 150
Figure 7-6: Overview of the available GCPs at the site Randa.
With LPS it was not possible to calculate an exterior orientation, neither with GCPs alone nor with supplementary exterior orientation information. The bundle block adjustment did not only terminate but also yielded very erroneous results (e.g. negative RMSE values). In our opinion the optimisation of the bundle adjustment or some basic implementation of it cannot cope with the oblique geometry that also is not constant over the whole block. The orientation angles change a lot between single images as the helicopter trajectory is not as stable as the one of a fixed-wing airplane and the helicopter position was adapted to the mountain (e.g. change of orientation to see otherwise occluded parts of the rock-face). Unfortunately, in LPS no relative orientation can be calculated separately, therefore a relative orientation with ORIMA was attempted. Here, only single models could be calculated, the relative orientation of the whole block failed. Also tests with exterior orientation parameters, GCPs, free-network adjustment or minimal datum were unfruitful. Therefore, a piece of software (interface) was written to transfer the tie points from the LPS image coordinates (in Pixel) to ISDM image coordinates (in m). In this step, also different orientations and centers of the image coordinate system had to be taken into account. SAT-PP uses the top-left corner of a pixel as origin and the coordinate axes facing down and right whilst all these properties can be chosen freely in ISDM. Finally, using the software package 151
7.1 The rockslide Randa (Switzerland) ISDM, the absolute orientation of the 54 images with 4 well identifiable GCPs was computed. The results of the orientation are shown in Table 7-2.
Table 7-2: Results of the orientation of the Helimap images in ISDM (ZI/Intergraph, 2009).

Block UAV Helimap

0 [Pixel] 4.07 6.07
RMSE of GCPs X [m] / 0.04 Y [m] / 0.04 Z [m] / 0.03
RMSE of image observation x [m] 1.98 2.99 y [m] 2.01 3.79
7.1.5 Comparison and data analysis
Using the acquired UAV-images (4-5cm GSD), a DSM of the lower part of the Randa rockslide with 15cm resolution was generated (see Figure 7-7). The complete site was documented with the Helimap system, with a pixel size of 6cm to 8cm. In comparison to the flight defined for the mini UAV-system, the Helimap flight was controlled by a pilot, who followed the flight lines with an accuracy of 30m. For the whole area the scale varied depending on the complexity of the terrain. It was also found that the Helimap system captured the cliff with an oblique field of view, which resulted in gaps in the data set. For the laser scan, the final point density was approximately 3pt/m2 (see Figure 7-7). The visual comparison of the extracted UAV-DSM and the LiDAR-DSM shows clearly that the fine structure of the cliff could be modeled from the UAV-DSM, while the LiDAR-DSM had large holes and less resolution (see Figure 7-7).

8 Conclusions and Perspectives on the flight performance, which is related to the accuracy of the 3D trajectory of the UAV. Since the accuracies given by the system manufacturers are generally not verified, we set up a test field for the evaluation of UAVs at the Campus Hnggerberg (ETH Zurich) and applied a new automated independent method for the tracking of the 3D trajectory of UAVs. These results confirmed the accuracy values of our UAV system as specified by the system manufacturer. Continuing the workflow for the photogrammetric data processing, we evaluated a number of different commercial photogrammetric software packages used for the orientation and DSM generation from UAV images. The study showed that not all software packages could be used in all possible applications of UAV photogrammetry. Quite often, only the combination of several packages enabled us to completely process the data set. These limitations of commercially available software packages result from the fact that they are designed for standard aerial images and cannot handle arbitrary image configurations. In section 5.3 we showed results of a UAV-borne laser scanning using a LMS-Q160 (Riegl, 2009) integrated in a large scale UAV Scout B1-100 (Aeroscout, 2009) platform. The first results are quite promising for upcoming applications, while the 3D trajectory of the LiDAR data has to be improved and investigated further in future studies. With the proposed workflow for the data acquisition and processing of UAV images, it is possible to firstly work in similar applications as in aerial and close range photogrammetry. Secondly, associated with the developments in automated flight planning tools and the improvement of the automated flight of UAVs, it is also possible to document areas faster and cheaper than with aerial and close range photogrammetry. Thirdly, using UAVs it is possible to fly closer (up to few meters) to objects than with manned systems, while the aerial view allows the data acquisition of objects which cannot be documented using terrestrial photogrammetry. In contrast to the traditional aerial photogrammetry, UAV systems can operate autonomously. The coordinates in the flight planning are frequently defined relative to the start point, which allows flight planning independently from existing maps and coordinate systems. Finally, the capability of using UAVs in unaccessible and dangerous areas, and the improvements in the data processing open up new applications in photogrammetry.

8.1.3 Applications

As examples for the suitability of UAVs for a wide range of applications, particular in archaeological, cultural heritage, and environmental applications and monitoring of a rock slide were conducted in this thesis. 171

8.2 Perspectives

It was possible to generate dense and accurate elevation models with 5cm resolution for Copn, 10cm for Pinchango Alto and the Maize field study and ~15cm for the rockslide Randa from off-the-shelf SLR images. The generated elevation data was compared to manual measurements in stereoscopic images, terrestrial and air-borne laser scanning. Depending on the resolution of the image and laser data the generated elevation data are comparable, while the automated methods (image matching and laser scanning) provide denser and more accurate data than manual measurements. The main difference in our study between the laser DSM and the DSM produced using image matching was due to the locations of the applied sensors (aerial or terrestrial). The example castle Landenberg showed the improvement of the particular flight planning for single object acquisition. In this example it was possible to use the flight planning tool circle successfully, which allowed us to generate images which covered the house roof and faade in one image strip. Therefore, it was possible to reduce the number of images which were necessary for the modeling and texturing of the castle. Furthermore, in our case studies we could underline the great potential of low-cost UAVs for the fast data acquisition by non-experts. Using an easily controllable system, such as a quadrotor under good weather conditions, it is possible to acquire images using the assisted or automated operation mode, for the automated extraction of surface models of small areas like in our gravel pit project (60m x 60m). The variety of applications also showed that UAVs are ready to be used under different environmental conditions worldwide. However, our tests showed that the maximum flying height, which depends on the type of UAV system and the environmental conditions (temperature, air pressure and wind), is so far an important limiting factor for some of the systems.

8.2 Perspectives trajectory can be further improved and a DSM can be generated both from the LiDAR and image data. In addition to the combined systems also OM, M and L-class image-based systems can be used for precise measurements. It can be expected that differential GPS / GNSS systems and other sensors are getting more compact and the cost for the systems, related to the growing number of applications in the future, will become cheaper. So far for most of the existing UAV-systems the integrated low-cost INS is the limiting factor for high accuracy. Finally, it can be stated that more UAVs will fly autonomously in the future, the 3D trajectory can be generated with higher accuracy and more systems will be stabilized. These improvements will turn UAV photogrammetry in to an almost autonomous photogrammetric measurement tool. These developments will give an important input for the further progress in the developments of the Digital Photogrammetry techniques and processes.

BIBLIOGRAPHY

AAAI, 2008. American Association for Artificial Intelligence corporate website, http://www.aaai.org, (accessed 04.16. 2008). Abbeel, P., Coates, A., Quigley, M. and Ng, A. Y., 2007. An application of reinforcement learning to aerobatic helicopter flight, In: NIPS, 19. Abdulla, Q. A., 2004. Photogrammetric platforms, In: Manual of Photogrammetry, Ed.: McGlone, J. C., American Society for Photogrammetry and Remote Sensing, Bethesda, Maryland, USA, 677-730. Aber, J. S., Aber, S. W. and Pavri, F., 2002. Unmanned small-format aerial photography from kites for acquiring large scale, high-resolution, multiview-angle imagery, In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Pecora 15/Land Satellite Information IV/ISPRS Commission I/FIEOS, Denver, CO, USA. Adobe, 2009. Adobe Photoshop CS4 corporate website, http://www.adobe.com/de/products/photoshop/compare/?sdid=DYWMG, (accessed June 6 2009). Aeroscout, 2009. Aeroscout corporate website, http://www.aeroscout.ch/, (accessed 31 March 2009). Aerotopol, 2009. Aerotopol corporate website, http://www.aerotopol.com/, (accessed May 12 2009). Agouris, P., Doucette, P. and Stefanidis, A., 2004. Automation and digital photogrammetric workstations, In: Manual of Photogrammetry, Ed.: McGlone, J. C., American Society for Photogrammetry and Remote Sensing, Bethesda, Maryland, USA, 949-981. Ahlfeldt, J., 2004. On Reconstructing and Performing Maya Architecture Temple 22, Copan, Honduras (AD 715)&#8221, Unpublished dissertation, Department of Art History and Archaeology, Columbia University, p. 596. Akca, D., 2007. Least Squares 3D surface matching, Dissertation, Institute of Geodesy and Photogrammetry, ETH Zurich, Zurich, p. 92. Altan, M. O., Celikoyan, T. M., Kemper, G. and Toz, G., 2004. Balloon photogrammetry for cultural heritage In: International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XX ISPRS Congress, Istanbul, Turkey, XXXV-B5, 964-968. ASL, 2009. Autonomous System Lab (ASL) corporate website, http://www.asl.ethz.ch/, (accessed June 5 2009).