ALMA MATER STUDIORUM – UNIVERSITA’ DI BOLOGNA II FACOLTA‟ DI INGEGNERIA Dipartimento di Ingegneria delle Costruzioni Meccaniche, Nucleari, Aeronautiche e di Metallurgia DOTTORATO DI RICERCA IN DISEGNO E METODI DELL‟INGEGNERIA INDUSTRIALE Ciclo XXII Settore scientifico-disciplinare di afferenza: ING-IND05 DESIGN, ASSEMBLY AND TEST OF AN AIRBORNE AUTOMATED IMAGING SYSTEM FOR ENVIRONMENTAL MONITORING Presentata da: Ing. Nicola Melega Coordinatore Dottorato Relatore Prof. Franco Persiani Prof.
Abstract Remote sensing and photogrammetry are key technologies for several activities such mapping, agriculture, land use or soil and air pollution monitoring. In this study an airborne autonomous and fully automated system for photogrammetry and remote sensing purposes is presented. State of the art technologies in this field have been reviewed, in order to define a set of requirements needed to lead the develop of this new system.
Table of Contents Table of Contents Table of Contents .................................................................................................................. i List of Figures .................................................................................................................... iii List of Tables .....................................................................................................................vii Chapter 1. Introduction .........................................
Table of Contents 6.1 FASTER EM Hardware layout .................................................................. 60 6.1.1 Sensors And Data Logger ............................................................... 61 6.1.2 Visualization and data storage ........................................................ 64 6.1.3 Sensing Device ............................................................................... 66 6.2 Power Subsystem .........................................................................
List of Figures List of Figures Figure 1: Dalsa FTF2020M sensor ...................................................................................... 6 Figure 2: Dalsa demoboard with FTF2020M sensor support .............................................. 6 Figure 3: Acquisition method, e.g 6 bands image ............................................................... 7 Figure 4: E.F.L = 200 mm, Telephoto Ratio = 0.9, Overall Length = 180 mm, F/# = 4, B.F.L = 60 mm ..........................................
List of Figures Figure 25: Leica ADS40 sensor head and assembly [ ....................................................... 24 Figure 26: Direct Georeferencing System, reference frames [Ref. (1)] ............................ 26 Figure 27: FASTER selected cameras system geometry ................................................... 31 Figure 28: FASTER camera ground projections ............................................................... 32 Figure 29: Prosilica GE4000 [Ref.
List of Figures Figure 57: Canon EOS 450 and Canon EF28 f/2.8 [Re. (12)] ........................................... 66 Figure 58: Canon EOS 450D fixed in the rapid prototyping base; lens focus ring has been blocked to avoid ring movements during flight ................................................................. 67 Figure 59: M2-ATX-140W power supply [Ref. (14)] (a) and the DC-DC converter used to power the camera (b) ....................................................................................
List of Figures Figure 87: Brisighella area, planned vs HIL simulated trajectory ................................... 106 Figure 88: Brisighella area, distance between planned and HIL simulated trajectories, whole flight ...................................................................................................................... 107 Figure 89: Brisighella area, HIL simulated altitude profile vs planned, whole flight .....
List of Tables List of Tables Table 1: LCCES Microbolometer specifications ............................................................... 13 Table 2: FASTER system requirements ............................................................................ 29 Table 3:FASTER final model vs Engineering Model characteristics ................................ 36 Table 4: Datum ..................................................................................................................
List of Tables viii
Introduction Chapter 1. Introduction Since 2003, the Microsatellite Laboratory of II Faculty of Engineering of the University of Bologna has been involved in the design and realization of small satellite missions based on the ALMASat multipurpose platform. After the realization of the first microsatellite, ALMASat-1, it is currently under development a second Earth observation mission, named ALMASat-EO.
Introduction from the airplane (without fuselage modifications or the need of specific navigation instruments) in which is installed. Thus it will be possible to move the system from an airplane to another without affecting performances. This work is organized as follows: in Chapter 2 background experiences are presented introducing how the idea to realize such a system was conceived.
Introduction indispensable when using aircraft without navigation instruments. The interface monitors system status and offers the same information normally provided by a multi function display installed onboard commercial aviation aircrafts. The FASTER EM has been tested during three flight campaigns and their results are reported in Chapter 7. These early test campaigns were essentially a series of functional tests that verified the correct data exchange between different subsystem.
Background Chapter 2. Background In this Chapter a brief description of all the preparatory activities carried out during the three years PhD course is reported, emphasizing those that led to the definition of the subject of this work. In Section 2.1 the AMSC project is discussed in detail; this study allowed to deepen themes such as the development of optical systems for remote sensing purposes.
Background All these experiences led to the idea to develop a common test bed capable to replicate all the functions that will be available on the ALMASat-EO microsatellite in an operational environment different from a standard laboratory equipment. The idea has subsequently evolved in a more generic and expandable aerial platform, not only a test bed, able to manage remote sensing instrumentations and equipped with attitude and positioning sensors for high accuracy data georeferencing.
Background that best fits the implementation of a system for the acquisition of satellite images is a CCD (although CMOS based devices are currently the best choice in order to reduce power consumption). This stems mainly from the high sensitivity of silicon at wavelengths of less than 1.1 μm, ideal if one is interested in capturing images in the visible spectrum (0.4-0.7 μm) or nIR (near infrared).
Background tuning error is consistent with the one measured prior to the test. One of the main issues connected to the use of this kind of filter is the need of custom-designed optics instead of a commercial one; this is due to the filter causing a variation of the B.F.L. (Back Focal Length) which, in turn, results in a shift of the focus of the system. The filter has an aperture of 35 mm, thus matching the dimensions of the CCD (main diagonal of 34.
Background A complete set of analysis has been done on the selected optical layout in order to determine the expected performances of the lens in terms of both distortion and MTF as shown in the following figures. Figure 5: AMSC standard ray tracing analysis Figure 6: AMSC plotting of the MTF (b) Lens will be custom built by a specialized manufacturer (currently Gestione SILO is the selected company) and after the production phase a final test campaign on the flight model will be carried out.
Background system. The optical bench is currently used to test performances of the wide angle CCTV lens of the ALMASat-1 sun sensors. Procedures has been developed to retrieve the exact distortion function of optical system. In the following pictures some examples of the spot shape on the optical axis (a), at 30° of hFOV (b) and near the full field (c) of the 130° wide angle lens.
Background ADCS LCD Filter CCD FPGA JPEG2000 Hardware Compression (ADV212) ARM7 Microprocessor TX Storage (FLASH memories) Figure 8: Electronics Layout Block Diagram The result is a 50 Mbit image reduced down to 3 Mbit. JPEG2000 also provides quality features like region of interest coding (ROI) and different types of progressive transmission. The AMSC will guarantee a ground resolution of about 40 m from an altitude of 720 km. 2.
Background STARS Simulator Attitude and Orbit Simulator Camera Model Earth Model Image Sequqencies Generator Images Dataset Attitude Estimation Algorithm Figure 9: STARS simulation environment diagram The output of the model was used as a reference to test the image processing algorithms employed to retrieve the spacecraft attitude from Earth image pairs.
Background In particular the work that has been carried out at the II Faculty of Engineering is relative to the realization of a test-bed capable of test on really acquired images the algorithm previously selected.
Background Test-bed is made up of three main components: a DSLR camera (Digital Single Lens Reflex) which simulate the STARS sensing device, a referenced Al platform equipped with a fine pointing rotating device controlled via an RS232 serial interface and a target image. The same Landsat 7 image used for numerical simulations, was printed and used as a target for a 1 DOF platform.
Background Figure 12: LEO Conditions Figure 13: GEO conditions Figures above refers to the hFOV (Half Field of View) and lead to a maximum full aperture of 136.04° for LEO conditions and 20.2° for GEO.
Background order), astigmatism, distortion, MTF and PSF as reported in the following diagrams. In this approach lens coating hasn‟t been taken into account although in a more refined layout optimization it‟s important to properly simulate its effects.
Background The preliminary layout of the ES housing consist of a common electronics box both for LEO and GEO sensor, in front of an interchangeable LEO and GEO optical head. Since LEO optical assembly dimensions are quite larger than GEO ones, the preliminary housing layout has been dimensioned upon LEO optical assembly.
Remote Sensing and Photogrammetry, State of the Art Technologies Chapter 3. Remote Sensing and Photogrammetry, State of the Art Technologies In this chapter a survey on the state of the technologies used in the field of remote sensing and photogrammetry is presented.
Remote Sensing and Photogrammetry, State of the Art Technologies spatially referenced data (data could be remote sensed or retrieved through a photogrammetric process) named Geomatics [Ref. (2)]. It uses terrestrial, marine, airborne, and satellite-based sensors to acquire spatial and other data. It also includes the process of transforming spatially referenced data from different sources into common information systems with well defined accuracy characteristics.
Remote Sensing and Photogrammetry, State of the Art Technologies remote sensing applications. A possible classification of airborne sensors can be done based on the two principal architecture currently adopted [Ref. (4)]: Airborne digital cameras, that produce frame images; Airborne pushbroom line scanners, that produce continuous strip imagery of the terrain 3.1.
Remote Sensing and Photogrammetry, State of the Art Technologies Instead of using an RGB filter to achieve colour registration, cameras can also be equipped with three separate CCDs in conjunction with an optical beam splitter. This solution avoid the interpolation necessary using a Bayern filter enabling higher spectral performances. Systems made by Airborne Data System and Integrated Spectronics utilize this simultaneous method to acquire the Red, Green and Blue spectral bands.
Remote Sensing and Photogrammetry, State of the Art Technologies Medium format cameras that are in current use have been modified from existing film cameras built by Hasselblad, Rollei, Mamiya and Contax with the film magazine replaced by a modern digital back. Digital backs have been developed by several companies like Phase One, Imacon and Jenoptik and most of these devices uses the biggest CCDs available on the market with up to 65 Megapixels.
Remote Sensing and Photogrammetry, State of the Art Technologies Large format cameras are typically sets of multiple medium-format cameras coupled together to form an integrated unit, notwithstanding some military devices, used for reconnaissance, employs very large CCD array (up to 100 Megapixels) which are available in very limited quantities. The resulting images or sub images are then rectified and stitched together to form a single, large format, digital monochromatic image.
Remote Sensing and Photogrammetry, State of the Art Technologies for panchromatic images and 2800 for multispectral data, while the achievable GSD is 0.19 m. In the following image, geometries of multicamera systems are reported. Figure 24: Multiple cameras system geometries [Ref. (4), (8)] 3.1.
Remote Sensing and Photogrammetry, State of the Art Technologies The most diffuse airborne Pushbroom Line Scanner is the Leica ADS40. In this system each of the three lines consists of a pair of linear CCD arrays linked together in parallel, with each of the arrays being shifted laterally by half a pixel with respect to the other. In the same focal plane are mounted four additional single 12000 pixel liner CCD arrays that records the ground images in the individual RGB and nIR spectral bands.
FASTER Chapter 4. FASTER The review of the state of the art technologies presented in Chapter 3, allowed a better understanding of the major issues related to the photogrammetric field. In particular, it is clear that one of the most important innovations in this field was the introduction of the GPS and IMU sensors to determine, in a direct way, all parameters of the instrument external orientation1.
FASTER Very often, modified aircrafts that can be equipped with photogrammetric instruments, are located in a limited number of airports and long transferring flights are needed to reach the area of interest. It often happens that the flight time required for the transfer equals that for the photogrammetric campaign.
FASTER 4.2 Proposed Solution Before proceeding to the realization phase a list of requirements, arising from the analysis of the state of the art technologies, was prepared and reported in the following Table. Each requirement has been commented to clarify how it could be implemented in the proposed solution. RE.G.01 RE.F.01 RE.F.02 RE.F.03 RE.F.04 RE.F.
FASTER RE.F.06 RE.F.07 RE.F.08 RE.F.09 RE.F.10 RE.F.11 RE.P.01 RE.P.02 RE.P.03 RE.P.04 Images shall be acquired so that image mosaicing can be successfully accomplished during both real time processing or postprocessing operations The planning tool will take into account the selected along track and cross track overlap percentage as stated in Section 5.4. For the purpose of this work mosaicing operations will be done in a postprocessing phase.
FASTER RE.P.05 Georeferencing accuracy shall be in the same order of the GSD RE.P.06 The system shall be able to guarantee a 3 fps acquisition RE.P.07 The system shall be able to collect at least 20 GByte/hour RE.P.08 RE.I.01 RE.I.02 RE.I.03 RE.I.04 RE.I.05 At present this requirement cannot be fulfilled, neither applying postprocessed GPS differential corrections. Further developments will be focused to make the system compliant.
FASTER 4.
FASTER Figure 27: FASTER selected cameras system geometry Cameras are arranged in line, with the first camera tilted 15° on the right side and last one tilted 15° on the left side. The first and third cameras are equipped with a Bayer filter mosaic to obtain colored images while the central monochrome camera is equipped with a nIR filter with 720 < λ < 850 nm.
FASTER Figure 28: FASTER camera ground projections The Ground Sampling Distance (GSD) defined as [Ref. (14), (15)]: 𝐺𝑆𝐷 = 𝑃𝑖𝑥𝑒𝑙 𝑃𝑖𝑡𝑐 𝑓 where is the flight height and 𝑓 is the focal length, at a flight height of 150 m the GSD for the colour cameras, with the 28 mm lens, varies between 4.8 cm at the Nadir direction and 7.13 cm at the maximum swath width. The corresponding GSD for the nIR camera, equipped with a 21 mm lens, varies between 6.4 cm and 8.5 cm. Values are compliant RE.P.01. 4.
FASTER The Ground Segment infrastructure The Airborne internal management and computing segment The Airborne external POD, equipped with the sensing devices A detailed description of each unit is given in Chapter 6 where the FASTER Engineering Model is described. The major differences between the final system and the EM are the sensing device, and the positioning of the Inertial Measurement Unit and the GPS receiver.
FASTER The FASTER final system will be provided with the same Crossbow NAV420 IMU used in the EM but this time in full nav mode in order to avoid the use of an external GPS receiver. A complete 3D model of the FASTER external POD has been virtually assembled using Solidworks (Figure 30, Figure 31 and Figure 32); the CAD model is used to retrieve the relative distances between the camera reference systems and that of the IMU and between the IMU reference system and the GPS antenna phase center.
FASTER Figure 31: FASTER 3D model assembly, pc-box and SSDs mount Figure 32: FASTER 3D model canopy 35
FASTER Engineering Model 4272 x 2848 Pixels (3:2 form format) 5.2 µm APS-C (22.2 x 14.8 mm2) 40° cross track 29° along track 28 mm f/# 2.
FASTER Planning Software Chapter 5. FASTER Planning Software 5.1 Photogrammetric Aerial Mission Purpose of a photogrammetric aerial mission is systematically cover a portion of the Earth‟s surface.
FASTER Planning Software 𝑎= 𝑎′ 𝑐𝑓 Eq.2 where 𝑐 is a crop factor which takes into account the ratio between the 35 mm format and actual sensor diagonal. The effect is an apparent longer focal length (if 𝑐 > 1) that must be considered because directly effects the ground resolution (thus 𝑎). The base between two subsequent images in the same strip is expressed by Eq.3: 𝑏 =𝑎 1− 𝑜 100 Eq.3 in which 𝑜 is the selected along-track overlap.
FASTER Planning Software The image scale factor is given by the ratio /𝑓. For the purposes of this study flying height is limited to 500 ft (152.4 m) over the highest obstacle in a 3000 m range, because of the Italian regulation on Very Light Aircrafts flight.
FASTER Planning Software 5.2 FASTER planning software layout In the following diagram the FASTER planning software block diagram is presented. The idea is to use a standard platform, easy and quick to use, to define the target area. At present, different tools are available which use a three-dimensional representation of the Earth globe to access for example, satellite and aerial imagery [Ref (23)].
FASTER Planning Software The planning software has been divided into five different blocks and two of them are implemented using geobrowser features. It has been found that this solution is particularly suitable when the use of maps can be helpful to speed up the target area selection process and to better understand the obtained results. One of the most widely available solutions for geobrowsing is Google Earth; it has over 100 million users and all the characteristics listed above.
FASTER Planning Software The user GUI (Graphical User Interface) is shown in Figure 36, in the left frame address box, places and levels enabler are present.
FASTER Planning Software With the toolbar buttons, over the graphical window, is possible to draw paths or polygon over the Earth surface, carry out measurements, add images or placemarks for specific locations. Placemarks and polygons can be exported in the KML/KMZ format. 5.3.
FASTER Planning Software defined by a group of points in the space and their relative coordinates are reported in the same tag line in a repetitive sequence: longitude1,latitude1, altitude1 longitude2,latitude2,altitude2… longituden,latituden,altituden . Polygons, are stored in a Nx3 array, where N is the number of points that constitute the polygon.
FASTER Planning Software Figure 39: Imported KMZ in the Matlab environment Figure 39 is a plot of the Area_ridolfi array created by the read_kml.m script. Once the array is created in the Matlab workspace the FASTER_planner.m script is recalled. 5.4 FASTER Flight Planner The flight planner block shown in Figure 35 was also built in the Matlab environment and includes several scripts: read_kml.m, already described in Section 5.3.2; FASTER_planner.
FASTER Planning Software DEM_extraction.m which searches inside the ASTER DEM database and extracts the selected area altimetry data from a compressed archive; KML_export.m, produces a planned trajectory KML file. Once imported inside Matlab, the array corresponding to the area defined in Google Earth maintains the same reference system used by the KML standard, and geographical coordinates are still used.
FASTER Planning Software One of the advantages of the UTM projection is that geographic location are given in x and y coordinates in meters, using the meridian halfway between the two bounding meridians as the central meridian, and reducing its scale to 0.996 of true scale (point 4 of the previous list). This reduction was chosen to minimize scale variation in a given zone and the amount of distortion is held below 1/1000 inside each zone.
FASTER Planning Software The ellipsoidal Earth is used throughout the UTM projection system, but the reference ellipsoid changes with the particular region of the Earth. A list of reference ellipsoids (datum) is presented in Table 4. Datum NAD83/WGS84 GRS 80 WGS72 International (1924) Clarke 1866 Equatorial Radius [m] (a) 6378137 6378137 6378135 Polar Radius [m] (b) Flattening (a-b)/a Use 6356752.3142 6356752.3141 6356750.5 6378388 6356911.9 1/297 Global 6378206.4 6356583.8 1/294.
FASTER Planning Software 5.4.2 Trajectory Determination The FASTER_planner.m script is responsible of the trajectory determination. This script requires the following entries: Target area UTM coordinates, coming from the GEOtoUTM.
FASTER Planning Software Figure 42:Example of plotting: target area, surrounding square area, DEM area Figure 43:DEM extraction from the ASTER database, tile ASTGTM44N011E 50
FASTER Planning Software Figure 44: FASTER flight planner - computed flight height [m] To use the ASTER DEM database the DEM_extraction.m script has been created. It takes as input the DEM area coordinates generated by the FASTER_planner.m script. Longitude and latitude boundaries are evaluated in order to determine if all the DEM area corners belong to the same tile. Then tile file name is reconstructed using the ASTGTM_ prefix adding the coordinates already found.
FASTER Planning Software to different lens in order to have the same ground resolution when relative distance to ground varies. When along-track, 𝑜, and across-track overlap, 𝑝, are introduced 𝑎 and 𝑞 values can be calculated and the trajectory can be finally determined. The flight trajectory is a sum of different parts: acquisition stripes, re-alignment turns and the acquisition area entrance maneuver.
FASTER Planning Software The turn45.m script is recalled between each stripe calculation and waypoints and turns time are added during the iterations performed to cover the entire target area. In Figure 46 an example of planned trajectory where the green line is the waypoints (red dots) interpolation.
FASTER Planning Software Figure 47: FASTER_planner.
FASTER Planning Software Figure 49: Planned trajectory exported in a KML file and visualized inside Google Earth The final trajectory is saved in a trajectory.mat file that will be used in the FASTER Airborne internal management and computing unit to generate the guidance tunnel that will be displayed to the pilot during flight operations. Computed parameters are printed in a text file, values corresponding to the example used in the previous figures are summarized in the following table. FASTER_planner.
FASTER Planning Software Multiple areas can be managed inside the same flight plan but transfers between each areas are actually not computed by script. This because of the specific ULA regulation which doesn‟t allow these aircrafts to over flight densely populated areas or enter a controlled airspace. All these constraints are very difficult to implement in the planning software, so navigation outside the target area is currently a pilot responsibility.
FASTER Engineering Model Chapter 6.
FASTER Engineering Model RE.F.06 RE.F.07 RE.F.08 RE.F.09 RE.F.10 RE.F.
FASTER Engineering Model Despite these requirement relaxations, the result is compliant with the initial idea of a functioning test bed able to guarantee sufficient performances in order gain experience in the process of realization of the final system. In order to proceed to the definition of the system, a functional block diagram of the airborne internal management and computing system has been drawn and is shown in Figure 50.
FASTER Engineering Model Acquisition devices configuration (i.e. aperture and exposure time) is setup via the camera parameters controller and storage block which also includes the support for image data saving and download to the G.S; camera parameters are passed to this unit from the ground equipment which selects the appropriate values for aperture and exposure time.
FASTER Engineering Model Garmin GPS 18 5Hz RS232 serial interface G.P.I.O camera timing Canon EOS 450 RS232 serial interface PC/104 XPC Target RTOS Xbow Nav420 G.P.I.O arm switch C.F card Eth Ground Segment Eth 5V 5V Ethernet HUB VIA EPIA N15000 5, 12 V 3.3, 5, 12 V 12 V 80 Gb HD VGA connection Sunlight readable LCD 12 V USB touch screen 12 V DC/DC Converter USB camera control 8V Eth Camera Arm Switch Battery Charger 12 V 12 V Power Supply 6-32V M2-ATX-140 W [3.
FASTER Engineering Model connection s are used to send the state vector (which in addition to position and attitude data store also a reference time and the magnetic field components) to the EPIA EN15000 throughout the User Datagram Protocol (UDP). The UDP uses a simple transmission model and is often used in time-sensitive applications because dropping packets is preferable to waiting for delayed packets which may cause the system to not properly work.
FASTER Engineering Model parity bit) using the NMEA (National Marine Electronics Association) 0183 ASCII interface specification (GPGGA, GPRMC, GPVTG and PGRMV sentences are currently used). Figure 53: Garmin GPS 18x 5Hz and its mechanical drawing [Ref. (33)] The Crossbow NAV420 is a combined GPS navigation and GPS-Aided Attitude & Heading Reference system (AHRS) that utilizes high stability MEMS-based inertial sensors.
FASTER Engineering Model Missing GPS data prevent magnetic declination determination so the computed heading coming from magnetometers is relative to the magnetic north direction instead of the true north direction; true heading is thus provided by the Garmin GPS receiver. NAV420 accuracy is reported in the following table. NAV420 Accuracy Specification GPS disabled X,Y Velocity [m/s rms] N.S Z velocity [m/s rms] N.S Attitude angles [° rms] < 2.5 Bias: R,P,Y (EKF stabilized) [°/sec] < ±0.
FASTER Engineering Model The Canon EOS is connected to the EPIA via an USB port using the widely supported Picture Transfer Protocol (PTP) developed by the International Imaging Industry Association to allow the transfer of images from digital cameras to computers without the need of additional device drivers. In this case the protocol has been custom modified by Canon to fully support camera parameters control, replicating the same functions accessible from the hardware buttons on the camera.
FASTER Engineering Model connected via a USB port to the EPIA EN15000. The LCD matrix is mounted on a fiberglass support jointly with the AD5621GA control board (VGA and power connector), the LI0610A inverter board and the LID08A LED driving board [Ref (38), (39), (40), (41)]. Figure 56: Litemax 8.4" high brightness LCD display and its mechanical drawing [Ref. (42)] 6.1.
FASTER Engineering Model The camera has an all plastic structure so it is therefore necessary an adequate protection when installed outside the aircraft. It is fixed on a rapid prototyping plastic base which has been designed to perfect fit the lower part of the camera body, avoiding rotations once it is mounted.
FASTER Engineering Model modified, C++ based remote camera control software available from the Canon SDK support. 6.2 Power Subsystem As shown in Figure 51, the FASTER EM power subsystem includes the M2-ATX140W automotive power supply, the 12 to 8 V DC-DC converter and the 12 Ah-12V lead battery. The power supply covers a wide range of input voltages, 6 to 32 V and provides the standard outputs for digital electronics: +3.3, +5, +5SB, and ±12 V with maximum current values reported in Table 8.
FASTER Engineering Model Figure 59: M2-ATX-140W power supply [Ref.
FASTER Engineering Model During the development of the FASTER EM we had the opportunity to have at our disposal a Tecnam P92, which is one of the most popular ULA (Ultra Light Aircraft) in Italy. It is a single-engine high-wing aircraft which employs a monocoque tail cone section with the forward fuselage using sheet aluminum over steel tubing [Ref. (46)].
FASTER Engineering Model Figure 60: Tecnam P92 selected mounting area for the external POD 71
FASTER Engineering Model Figure 61: Canopy section After the definition of the canopy section profile, it has been extruded along the whole wood base in order to create a reference polystyrene solid to be used as a mold for fiberglass drafting.
FASTER Engineering Model Figure 63: FASTER EM fiberglass canopy installed on Tecnam P92 passenger side 6.4 Airborne Internal Management and Computing Unit Assembly The airborne internal management unit is entirely fit inside a 19 inches, 2U custom modified rack aluminum rack enclosure. Being the prototype an engineering model, there are some differences between this and the final FASTER system that will be compliant to all the requirements specified in Chapter 4.
FASTER Engineering Model are currently enabled to connect the camera arm switch, others can be employed to driver LEDs or to pilot an external alphanumeric display to monitor the system functioning status. At present there is no possibility to power off or reset single components in case of failure, but this function will be implemented in the upcoming version.
FASTER Engineering Model Figure 66: FASTER airborne internal and computing unit inside view of the rack box 6.5 FASTER EM Software Description To allow the proper functioning of all hardware components was necessary to develop appropriate software algorithms.
FASTER Engineering Model be created using the Real-Time Workshop and a C compiler; this will run on a compatible target PC in real-time mode using the same initial parameters from the Simulink model that were available at the time of code generation [Ref. (47)]. The target PC is booted using an xPC Target boot disk (or from a network boot image) that loads the xPC Target real-time kernel and then the target application can be downloaded from the host machine.
FASTER Engineering Model Figure 67: FASTER EM Simulink model 77
FASTER Engineering Model 6.5.1 Garmin GPS 18 5Hz Acquisition Block The Garmin GPS 18 5Hz block acquires data packets from the receiver decoding each sentence and the corresponding Simulink model is shown in Figure 68.
FASTER Engineering Model <10> <11> <12> meters Geoidal height, -999.9 to 9999.9 meters Null (Differential GPS) Null (Differential Reference Station ID) N N N Table 10: GPGGA sentence structiure Track Made Good and Ground Speed (VTG) Sintax <1> <2> <3> <4> <5> Used $GPVTG, <1>,T, <2>,M, <3>,N, <4>,K, <5>*hh True course over ground, 000.0 to 359.0 degrees (leading zeros will be transmitted) Magnetic course over ground, 000.0 to 359.
FASTER Engineering Model Figure 68: Garmin GPS 18 5Hz acquisition block sheme 80
FASTER Engineering Model 6.5.2 Crossbow NAV420 Acquisition Block The same structure used in the previous section has been adopted for the NAV420 acquisition block shown in Figure 69. While in the GPS acquisition block no commands can be sent to receiver , in this case configuration data can be transmitted using the bidirectional RS232 connection to set one of the three Data Packet Mode: Scaled Sensor Packet, Angle Packet and NAV Packet. As stated in Section 6.1.
FASTER Engineering Model 𝑡𝑒𝑚𝑝 = 𝑑𝑎𝑡𝑎 ∗ 100 215 Accelerations are measured in G‟s (actual measurement range is ±4 G), angular rates in deg/s (actual measurement range is ±200 deg/s), magnetic field components in Gauss and temperature in °C. The digital data representing each measurement is sent as a 16-bit number (data list in Table 14) and starts with a 2 byte header followed by the selected functioning mode.
FASTER Engineering Model BIT Message Definition 3 Turn detect 4 Comm Transit Error 5 Startup Rate Bias Check 6 GPS status 8 Algorithm Initialization 1 PPS Signal Lock 9 10 EEPROM integrity 11 Magnetometer Calibration Valid 12 User Port Comm Receive Error 14 Algorithm 15 Accuracy 0: Yaw rate magnitude < 0.
FASTER Engineering Model Figure 69: Crossbow NAV420 acquisition block scheme 84
FASTER Engineering Model 6.5.3 Magnetometer acquisition block This block was added because of the need to understand how the magnetic fields generated by the electronic equipment, placed near the I.M.U, can affect the integrated magnetometers measure. In order to do this a second 3-axial magnetometer (the same Applied Physics 539 used onboard ALMASat-1) will be placed on a boom fixed to the external POD, in order to be sufficiently far away from unwanted stray magnetic fields [Ref (50)].
FASTER Engineering Model In this case a surrounding quadrilateral or circular area is drawn around the selected target area, depending on the shape, and latitudinal and longitudinal boundaries are taken into account to define where the camera has to be enabled. In this case pilot doesn‟t interact with system at all. Values are calculated by the FASTER planning software described in the previous Chapter.
FASTER Engineering Model Figure 71: Camera shot signal generation Before reaching the parallel port output the correct waveform to control the shutter button release is generated. The shutter button is typically hold down for 0.5 s for frame rate ≥ 1. Greater frame rates could be reached simply keeping always down the shutter buttons; the camera enters the continuous driving mode and frame rates up to 3.5 fps can be achieved.
FASTER Engineering Model Figure 73: Canon EOS 450D Management Block Scheme 88
FASTER Engineering Model 6.5.5 Tunnel in The Sky Management Block The guidance tunnel is a critical tool for repeating the same planned trajectory during flight and constitutes one of the peculiarities of the FASTER EM. A complete description of the system is given in the next Chapter and here is presented only the Simulink block used for transmit all the state variables needed for attitude representation and the trajectory drawing.
FASTER Engineering Model Figure 74: Tunnel in the sky management block scheme 6.5.6 Virtual Tunnel Block The virtual tunnel block has been developed during a previous project aimed to realize an attitude visualization system for a remote piloted aircraft. The control and visualization system runs on a portable ground station and the pilot is able to control the aircraft even when the aircraft is not in sight.
FASTER Engineering Model currently can‟t be further modified) and read by the virtual tunnel block, but only a subset of 20 at a time is used to create a small portion of the trajectory which is transmitted via UDP connection to the Dynaworlds visualizer; Waypoints subset is refreshed every 5 seconds taking into account the actual GPS position. GPS fix is compared to the entire waypoint list selecting the nearest.
FASTER Engineering Model Figure 75: Virtual Tunnel block scheme 92
FASTER Engineering Model 6.5.7 Datalog Block The datalog block function is to save all the acquired information in a log file in which ancillary data for each image is contained (Figure 76). The log file structure is shown in the following table.
FASTER Engineering Model 6.5.8 Pilot Interface Block The pilot interface block has been added to control the FASTER status monitor which is a Java-based software. This block collects a series of variables and send it via UDP connection to the EPIA EN15000 which runs the FASTER status monitor.
FASTER Engineering Model Figure 78: FASTER Acquisition Display 6.6 Tunnel In The Sky Visual Interface The pilot visual interface is displayed on the LCD screen, which is placed over the cockpit dashboard in a central position not to obstructing pilot‟s view (Figure 79).
FASTER Engineering Model are visualized by colored markers: in particular, it is shown the prediction of the future position of the aircraft by means of emphasized tunnel sectors that, in the future, will be interested by the effective trajectory. Experimental tests have shown that the combination of the visual information coming from the guide tunnel and from the predictions reduces the oscillations around the reference trajectory during the mission [Ref. (52)].
FASTER Engineering Model Visualization layout was fixed and a poor screen resolution was used. During flight operations this screen configuration showed its limitation because of the reduced visibility due to the selected color palette. Characters were barely readable too and the distance from the next waypoint shown between the ground speed indication and the flight height was misleading.
FASTER Engineering Model To increase the system effectiveness, another tool called 're-entry tunnel' has been introduced (red tunnel in the previous images). This is a secondary tunnel that brings back the aircraft, compatibly with its ability to maneuver (turn rate is limited to standard 3 °/s) on the trajectory defined by the main tunnel, every time it goes off the course exceeding a defined error.
FASTER Engineering Model A screenshot of the camera control center software is provided in Figure 83. Figure 83: Camera control center screenshot As it can be seen the accessible camera functions are: AEMode, is the Automatic Exposure mode and during acquisitions is always set to manual in order to avoid different exposure setting between frames; the AEMode must be physically set on the camera using the AEMode wheel; Tv, is the exposure time, it can be set between 105 s and 2.
FASTER Engineering Model optimal value must be carefully evaluated. ISO can be set between 100 and 1600 but for APS-C sized detectors, like the ones used in the EOS 450D, ISO values higher than 400 should be avoided; Metering Mode, this setting refers to the way in which a camera determines the exposure but it is not applicable in manual mode; ImageQuality, refers to how the image is saved to storage support. Options available are JPEG (low, medium and high quality) and RAW mode.
Results of the test campaigns Chapter 7. Results of the test campaigns After completing the realization of the FASTER EM, the system was subjected to a series of tests to verify its correct operations and performance. The tests were carried out both in laboratory and in flight. 7.1 Laboratory Tests During the laboratory test phase each subsystem was thoroughly tested. Initially, PCbased boards were stressed to verify their limits in terms of cpu load and memory usage.
Results of the test campaigns The problem was corrected applying a files check routine which changes the file name if it already contains data. This action was taken because it had been verified that when the battery is almost discharged, small voltage variations can cause some systems to restart before the full discharge of the battery. Acquisition tests were conducted on the NAV420 and GPS receiver units, verifying expected data exchange rate.
Results of the test campaigns A cross check can still be made if the camera clock has been synchronized with the PC-board internal clock time, reading the exif3 data contained in each image where date and time are saved. Long runs of more than 4 hours of operations showed that each image was correctly acquired and saved.
Results of the test campaigns The FASTER block contains the same Simulink model used to built the xPC Target real time operating system which works directly in the host machine as detailed in Section 6.5. Finally, the same EPIA EN15000 is used to run the Dynaworlds tunnel in the sky guidance interface using the UDP Ethernet connection provided by the CIL workstation.
Results of the test campaigns The two emulators generates the same nmea strings for the GPS and the angle mode packet binary data output for the I.M.U. With the HIL setup more accurate simulations of the flight campaigns were made testing the functionalities of the updated version of the tunnel in the sky interface. The results of the third flight campaign simulation over the Brisighella area are reported in the following images.
Results of the test campaigns Once the coordinates of the two nearest waypoints are extracted, the straight line passing through that two points is found using the known equations: 𝑚𝑖 = 𝑦 𝑓𝑙𝑜𝑤𝑛 𝑖+1 −𝑦 𝑓𝑙𝑜𝑤𝑛 𝑖 , 𝑥 𝑓𝑙𝑜𝑤𝑛 𝑖+1 −𝑥 𝑓𝑙𝑜𝑤𝑛 𝑖 Eq.5 Eq.6 𝑞𝑖 = 𝑦𝑓𝑙𝑜𝑤𝑛 𝑖 − 𝑚𝑖 𝑥𝑓𝑙𝑜𝑤𝑛 𝑖 then, the distance between the straight line and the planned waypoint is calculated using Eq.7, and defining a index which gives an idea of “how close” are the two trajectories on the plane.
Results of the test campaigns Figure 88: Brisighella area, distance between planned and HIL simulated trajectories, whole flight Figure 89: Brisighella area, HIL simulated altitude profile vs planned, whole flight 107
Results of the test campaigns Figure 87 shows the HIL simulated trajectory in green, while the red dots represent the planned waypoints. The HIL simulation was done without simulating atmospheric conditions, thus in ideal conditions. Notwithstanding a small time lag between the given command and the aircraft response resulted in a non-optimal flight. The absolute mean distance between the planned and simulated trajectory for the whole flight, shown in Figure 88, is 13.
Results of the test campaigns Wires pass under the door and connect the two unit. The LCD screen has been installed on a metal support that fits above the compass over the dash board, so the screen is placed in a central position and can be easily seen by the pilot without head movements and maintaining into its field of view the outside environment (Figure 91).
Results of the test campaigns The camera arm switch, for the three campaigns, has been controlled by an operator which had the role of controlling the system functioning and intervene in case of need. The aircraft fully equipped is shown in Figure 92. 7.2.
Results of the test campaigns No further magnetometer calibration was applied with respect to that made inside the laboratory. The planned trajectory is shown Figure 93. The departure airport was the Verginese airfield near Ferrara. During the flight, the pilot, which was not trained before takeoff, had serious difficulties trying to follow the tunnel indications, probably because of an uncorrected bias in the tunnel geometry.
Results of the test campaigns Figure 95: First flight campaign, initial part of planned vs flown trajectory Figure 96: First flight campaign, distance between planned and flown trajectories 112
Results of the test campaigns Figure 97: First flight campaign, planned vs flown altitude Furthermore, because of the prolonged ground operations, during which the system was turned on, the battery went down before completing the entire flight. Although the guidance tunnel did not work correctly the FASTER airborne internal management and computing unit acquired all the information from the GPS receiver and the I.M.U, working, in flight, for more than 50 minutes.
Results of the test campaigns resolution Google Earth layer was created. The computed GSD in 2.9 cm/pixel, and the difference between the 1 m Google Earth database and the high resolution layer is clearly visible in Figure 98. Figure 98: First flight campaign, high resolution Google Earth layer created from an acquired image 7.2.2 Second Test Campaign The second test campaign was a closed circuit over the Verginese airfield.
Results of the test campaigns guidance tunnel. The position predictor, which provides a 5 s prediction, was activated utilizing a red dot.
Results of the test campaigns The pilot was trained for 30 minutes before takeoff to get used to the new interface; training was performed using the CIL workstation connected to a laptop pc. As shown in Figure 101, the new interface provides a position predictor shown as red dot, if the red dot remains inside the guidance tunnel, the tunnel structure is highlighted.
Results of the test campaigns Figure 102: Second flight campaign, selected waypoints of the flown trajectory used to compute the distance from the planned trajectory Figure 103: Second flight campaign, distance between planned and flown trajectory 117
Results of the test campaigns Figure 104: Second flight campaign, planned vs flown altitude The camera settings for the second test were: focal length 28 mm (equivalent to 46 mm for the APS-C format), ISO 100, aperture F/5 and exposure time of 4*10-3 s.
Results of the test campaigns 7.2.3 Third test campaign In the third test campaign, for the first time, the FASTER planning tool was used because of the presence of several adjacent flight stripes. In addition, the latest modifications to the waypoints management made to the FASTER EM management software were used. So the tunnel was not completely displayed since the beginning but it was updated every 5 s, taking into account the actual position (Section 6.5.6).
Results of the test campaigns difficult to keep the pre-computed heading. The flight analysis is presented in the following figures.
Results of the test campaigns Figure 107 and Figure 108 show the planned trajectory against the flown trajectory and the computed distance between them. The mean distance is below the tunnel boundaries (12.35 m), which for this flight were set to 15 m (30x30 m tunnel section). Distance peaks are relative to the course turnabouts and mainly due to the adverse meteorological conditions.
Results of the test campaigns In Figure 110, the distance between planned and flown trajectories is plotted for the first stripe, showing a mean distance of about 6.3 m. The same happens for the others stripes; the worst case is represented by the 2nd stripe which has a mean distance from the planned of about 9.6 m. The aircraft entered the target zone at 13:27 and exited at 14:04 for a total elapsed time of about 37 min against the 39 min computed by the planning software.
Results of the test campaigns Figure 112: Third flight campaign, Brisighella mosaic 123
Results of the test campaigns 124
Conclusions Chapter 8. Conclusions In this work a novel concept of photogrammetric and remote sensing instrument has been discussed. The system is a fully-digital, direct georeferencing system used to acquire images in different bands of the electromagnetic spectrum. To define the external orientation parameters, avoiding traditional image scanning and aerial resection, a combination of an inertial measurements platform and a GPS receiver has been used.
Conclusions The FASTER EM still needs to be fine tuned before the realization of the final system; other tests campaign, with the IMU installed inside the external POD, will be helpful in order to better evaluate the direct georeferencing accuracy of the system. In particular a flight over an airfield with calibration patterns installed could be the most appropriate solution.
Appendix-A Converting Geographical Coordinates to UTM Appendix-A Converting Geographical Coordinates to UTM In the following appendix the method used to convert Geographical coordinates to UTM coordinates is described. Relationship are used in the GEOtoUTM.m and UTMtoGEO.m scripts. The geometrical problem is shown in Figure 113.
Appendix-A Converting Geographical Coordinates to UTM 𝑒 2 3𝑒 4 5𝑒 6 3𝑒 2 3𝑒 4 45𝑒 6 𝑀 = 𝑎[ 1 − − − … 𝑙𝑎𝑡 − 3 + + … sin 2𝑙𝑎𝑡 4 64 256 8 32 1024 15𝑒 4 45𝑒 4 35𝑒 6 + + + ⋯ sin 4𝑙𝑎𝑡 − + ⋯ sin 6𝑙𝑎𝑡 + ⋯ )] 256 256 3072 Converting latitude and longitude to UTM, northing is defined as (𝑝 = 𝑙𝑜𝑛𝑔 − 𝑙𝑜𝑛𝑔0 ): 𝑦 = 𝐾1 + 𝐾2 𝑝2 + 𝐾3𝑝4 and considering that 𝑒 is the eccentricity of the Earth‟s elliptical cross-section, 𝑒′ = 𝑒2 (1−𝑒 2 ) and 𝜈 = 𝑎 , we have that (1−𝑒 2 𝑠𝑖𝑛 2 𝑙𝑎𝑡 ) 𝐾1 = 𝑀𝑘0 𝐾2 = 𝑘0 𝜈 sin 𝑙𝑎𝑡 𝐾3 = 𝑘0 𝜈
Bibliography Bibliography 1. Standalone Three-Axis Attitude Determination from Earth Images. Bevilacqua, Alessandro, et al. San Diego : XX AIAA Conference, 2010. 2. Wikipedia. Geomatics. [Online] http://en.wikipedia.org/wiki/Geomatics. 3. Characteristics of New Generation of Digital Aerial. Trinder, J. s.l. : GISdevelopment.net, 2005. 4. Airborne Digital Imagers: An Overview & Analysis. Petrie, Gordon e Walker, Stewart. Marne-La-Vallée, France : s.n., 2006. ISPRS Commission I Symposium. 5. Carl Zeiss AG.
Bibliography 15. An Optimally Integrated Direct Georeferencing and Flight Management System for Increased Productivity of Airborne Mapping and Remote Sensing. s.l. : USGS, 2004, Vol. WG I/4. 16. Prosilica. Prosilica GE4000/4000 C, User Manual. 2009. 17. Carl Zeiss AG. Distagon T* 2,8/28 ZF, Technical Specification. 2009. 18. —. Distagon T* 2,8/21 ZF, Technical Specification. 2009. 19. Prosilica. Prosilica - GigE cameras. [Online] http://www.prosilica.com/products/ge4000.html. 20. Carl Zeiss Ag.
Bibliography 35. —. NAV420 Navigation Aided IMU. [Online] http://www.xbow.com/Products/productdetails.aspx?sid=181. 36. VIA Technologies Incorporated. EPIA-EN User's Manual, Version 1.21. 2008. 37. VIA Technologies Inc. VIA Embedded - EPIA page. [Online] http://www.via.com.tw/en/products/embedded/ProductDetail.jsp?productLine=1&id=399 &tabs=1. 38. Litemax. Litemax AD5621GA AD Board - User Guide. 2007. 39. —. Litemax LI0610 Inverter - User Guide. 2007. 40. —. Litemax LID08A02 LED Driving Board - User Guide.
Bibliography 55. MIL Standards. MIL-PRF89020B, Performance specification Digital Elevation Data (DTED). 2000. 56. PC/104 Embedded Consortium. PC/104-Plus Specification Version 2.3. 2008. 57. VIA Technologies Incorporated. EPIA EN-Series Mini-ITX Mainboard Operating Guide. 2006. 58. A program for direct georeferencing of airborne and spaceborne line scanner images. Muller, Rupert, et al.
