lidarprimer - collins

Upload: k

Post on 01-Jun-2018

231 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/9/2019 LidarPrimer - Collins

    1/35

    Kenai GIS User Group LiDAR Workshop 1/35

     A LiDAR Primer

    Richard L CollinsGeophysical Institute and Department

    of Atmospheric SciencesUniversity of Alaska Fairbanks

    February 15, 2011

    Kenai GIS User Group

  • 8/9/2019 LidarPrimer - Collins

    2/35

    Kenai GIS User Group LiDAR Workshop 2/35

    Lidar studies in 1930’s using

    search lights. Use of lasers

    since the 1960’s.

    Used in a wide variety of both

    civilian and military

    applications;•  Biohazards

    •  Fisheries

    •  Forestry

    • 

    Fire

    • 

    Glaciology

    • 

    Mapping

    •  Meteorology

    •  Pollution

    •  Space Surveillance

    LiDAR – Light Detection And Ranging 

    Lidar Firsts – From “Airborne andSpaceborne Lidar” M. P. McCormick,2005.

    CALIPSO 2007

  • 8/9/2019 LidarPrimer - Collins

    3/35

    Kenai GIS User Group LiDAR Workshop 3/35

    LITE – LiDAR In-space Technology Experiment

  • 8/9/2019 LidarPrimer - Collins

    4/35

    Kenai GIS User Group LiDAR Workshop 4/35

    Contemporary LiDAR System Concept

    650 m widescan swath

    30°scanangle

    1200 mflight

    altitude

    1.5 m

    1.5 m

    IMU

    GPSgroundstation

    DifferentialGPS

    navigation

    GPS satellites

    Optech ALTM

    Airborne LiDAR

    ± 15 cm

    vertical

    accuracy

    30 cm widelaser

    footprint

     Topographic Laser Ranging and Scanning:Principles and Processing Shan, J, and C.K. Toth, CRC Press, 2009.

  • 8/9/2019 LidarPrimer - Collins

    5/35

    Kenai GIS User Group LiDAR Workshop 5/35

    Optech’s Gemini Airborne Laser Terrain Mapper System

    Range 150-4000 mWavelength 1064 nmElevation accuracy 5-35 cm

    PRF 33 - 167 kHzPosition GPS and GLONASS

    Scan Width 0-50°San Rate 0 - 70 Hz

    Range Capture < 4 returnsDivergence 0.25/0.8 mrad

  • 8/9/2019 LidarPrimer - Collins

    6/35

    Kenai GIS User Group LiDAR Workshop 6/35

    Hand-held Laser Scanners - Helimap

    Heigel laser scanning engine withHasselblad digital frame camera.Rigid carbon fiber frame withhandles.System IMU in box below thecamera. 

    Helimap system being operated from side ofAlouette III helicopter. System allowsaccurate and high resolution mapping ofsteep and narrow terrain. 

  • 8/9/2019 LidarPrimer - Collins

    7/35

    Kenai GIS User Group LiDAR Workshop 7/35

    LiDAR Comes of Age 

    Lidar systems were developed as research systems since the 1960’s.

    Terrestrial lidar systems come of age in the 1990’s as several enabling

    technologies mature;1.

     

    Fully solid-state lasers where solid-state lasers (e.g., Nd:YAG,

    Nd:YLF) pumped by laser diodes

    2. 

    High-speed electronics and computers

    3. 

    A mature Global Positioning System

    The following elements are found in contemporary lidar systems

    1. 

    Laser Ranging Unit

    2. 

    Optical Scanning Mechanism

    3. 

    Electronic and Computer Unit4.  Position and Orientation Unit

    5.  Software6.

     

    Imaging System

  • 8/9/2019 LidarPrimer - Collins

    8/35

    Kenai GIS User Group LiDAR Workshop 8/35

    Basic LiDAR System 

    Laser(s)

    Telescope

    Beam Expander

    Photodiode

    Optics -Collimation,

    FOV and

    BandwidthPhotodiode

    BS

    Lens

    Electronics -

    Counting/

    Timing and

    Threshold

    Computer

     

    Basic system composed of;•  Laser-based transmitter•  Telescope-based receiver•  High-speed electronics•  Optics• 

    Computer

    Lasers providehigh-intensitysmall-footprintnarrow-bandfrequency-stable

    beams for probing the environment.

  • 8/9/2019 LidarPrimer - Collins

    9/35

    Kenai GIS User Group LiDAR Workshop 9/35

    Echo Detection and Ranging 

    Knowing the speed oflight, the range fromthe LiDAR to a targetis determined by theround trip time.

    The echo timing isdefined by the

    leading edge of thelaser pulse and theecho pulse.

    Timing is everything,!R = (!t"v + t"!v)/2

    as !v/v is very small.

    The thresholddetection can dependon signal amplitude.

    Time

           R     a     n     g     e

    R1

    R2

    t

    1

    t

    2

    R = v x t

    v x t1 = 2 x R

    1v x t

    2= 2 x R

    2

     

    Time

         R    e    c    e     i    v    e      d

         S     i    g    n    a

          l

    t1

         T    r    a    n    s    m     i     t     t    e      d

         S     i    g    n    a      l

    t0

    Threshold

    Threshold

    !t = t1 - t0 

    T p 

    T p 

  • 8/9/2019 LidarPrimer - Collins

    10/35

    Kenai GIS User Group LiDAR Workshop 10/35

    d1 d2

     = d1 + R x !R

    Laser beams diverge with distance. Anyfinite beam inherently expands as it

    propagates.

    The Field-Of-View (FOV) of the receivermust match (or exceed) the divergence ofthe transmitter.

    •  The FOV determines the amount of

    background signal.•  The FOV places mechanical stability

    requirements on the lidar system.

    Beam Divergence and Reflection 

    Specular Lambertian Diffuse   Complex 

    Surfaces found in nature are rarely a smoothtransition between two homogenous media. 

  • 8/9/2019 LidarPrimer - Collins

    11/35

    Kenai GIS User Group LiDAR Workshop 11/35

    Energy Link Budget - 1 

    Ground

    dR

    AircraftET 

    Eig Erg

    ER

    Rg

    T T 

    !

     

    Consider a single laserpulse transmitted froman aircraft.

    Follow the round trip.

    1.  ET  

    2.  Eig = ET  " T

    3.  Erg = Eig " # 

    4.  ER = Erg " T " P

    where,

    P = " # $ # (dR /2)

    2

    2# $ # Rg2

     

  • 8/9/2019 LidarPrimer - Collins

    12/35

    Kenai GIS User Group LiDAR Workshop 12/35

    Energy Link Budget-2 

    The analysis yields a form of the Lidar Equation

    ER =   " # $ # (dR /2)2

    2# $ # Rg2

    #T 2% 

    ' ' 

    * * 

    #ET  

    Consider an airborne lidar, Rg = 1000 m (3281 ft, 0.6214 mile)

    dR = 11.3 cm (4.44 in)

    # = 0.5

    T = 0.8

    $  ER = 5.1 x 10-10 " ET  

    Consider a satellite lidar (CALIPSO)

    Rg = 705 kmdR = 1 m

    # = 0.5

    T = 0.8

    $  ER = 1.6 x 10-13 " ET  

    ET  = 110 mJPRF = 20HzT p = 10 ns% = 532 nm &

    1064 nm& = 130 µrad 

    ET  = 20 µJPRF = 100 KHz

    T p = 10 ns% = 1064 nm,

    1550 nm& = 0.2 – 1 mrad

  • 8/9/2019 LidarPrimer - Collins

    13/35

    Kenai GIS User Group LiDAR Workshop 13/35

    Energy Link Budget-3 

    80-90%

    ~50%

     F  r  e s  h  S  n  o w

     G  r  a s  s  , L  i   g h  t  s  o i   l  

    ~10%   ~15%~25%

     L  a k  e

     C  o n  c  r  e t  e

     A s  p h  a l   t 

    Reflectivities

    at 900 nm

    ~60%

    ~30%

     D  e c  i   d  u o u s  T  r  e e s 

     C  o n  i   f  e r  o u s  T  r  e e s 

     L  i   m e s  t  o n  e , C  l   a y

  • 8/9/2019 LidarPrimer - Collins

    14/35

    Kenai GIS User Group LiDAR Workshop 14/35

    Profiling 

    0

    Rg

      R

      a  n  g  e

    Aircraft

    Ground

      H  o  r  i  z  o  n  t  a  l   D  i  s  t  a  n  c  e

    0

    xf

    2xf

    3x

    fdg

    0

    va x T PRF

    va x 2 xT PRF

    v

    a

     x 3 xT 

    PRF

      T  i

      m  e

    dg

     

    Cessna 337 at 1000 m

    and 67 m/s (130 knots).

    Profiling (not scanning)

    PRF < 150 kHz.

    Divergence

    & = 0.2 mrad

    Footprintdg = 0.2 m

    xf > 0.5 mm

    Given divergence non-overlap occurs at

    xf > dg 

    PRF = 335 Hz

  • 8/9/2019 LidarPrimer - Collins

    15/35

    Kenai GIS User Group LiDAR Workshop 15/35

    Scanning

    The scanner extends the scope

    of the profiling system to yieldswaths in the transversedirection. A variety of scanningmethods have been implementedusing oscillating and rotatingmirrors.

    A Palmer scanner uses a nutatingmotion to yield a scan that yields an elliptical scan. Most ofthe measurement points arescanned twice, once in theforward and once in thebackward view. This redundancycan be used to calibrate thescanner and the position andorientation system.

    Topographic Laser Ranging and Scanning:

    Principles and Processing Shan, J, and C.

    K. Toth, CRC Press, 2009.

  • 8/9/2019 LidarPrimer - Collins

    16/35

    Kenai GIS User Group LiDAR Workshop 16/35

    Range Resolution and Range Discrimination-1

    The finite rise time of thelaser pulse, tR, determines theprecision with which thethreshold detector operates.

    Ideally tR should be as smallas ossible.

    The finite width of thelaser pulse, T P,determines theprecision with which

    the LiDAR candistinguish between

    closely spaced objects.

    Ideally T P should be assmall as possible.

    Time

          R    e    c    e      i    v    e      d

          S

          i    g     n    a      l

    t1

    Threshold

    t2

    RangeR2R1

    t2 - t1 = < T pR2 - R12 x v

    !R > 2 x v x T p

    0

    0

     

    Time

          R    e    c    e      i    v

        e      d

          S      i    g     n    a

          l

     Threshold

    0

    tR

     T P

    The need to minimize tR and T P pushes designers to employ high-

    speed (i.e., large bandwidth) analog circuits. 

  • 8/9/2019 LidarPrimer - Collins

    17/35

    Kenai GIS User Group LiDAR Workshop 17/35

    Range Resolution and Range Discrimination–2

    Threshold1

    Time

    !t1

     Threshold2 

    !t2

          R    e    c    e      i    v    e      d

          S      i    g     n    a      l

     

    Time

          R    e    c    e      i    v    e

          d

          S      i    g     n    a      l

    0   tS

    t

    R

     

    The slope of the leading

    edge may vary due to•  signal noise,•  pulse amplitude

    variations•  pulse spreading due to

    elevation variations in

    the footprint.

    Full-waveform samplingcaptures the shape of theentire signal, not just theleading edge.

    The approach supports more

    information at GREAT costin both (analog and digital)circuitry and data handling.

  • 8/9/2019 LidarPrimer - Collins

    18/35

    Kenai GIS User Group LiDAR Workshop 18/35

    Multiple Returns-1

    Topographic Laser Ranging and Scanning:

    Principles and Processing Shan, J, and C.

    K. Toth, CRC Press, 2009.

  • 8/9/2019 LidarPrimer - Collins

    19/35

    Kenai GIS User Group LiDAR Workshop 19/35

    Multiple Returns-2

    The lidar echo can contain much more information than is revealed by the multiple

    returns detected using a single threshold level.

    Full-waveform sampling captures the complete echo but increases the hardware

    data handling costs of the system.

  • 8/9/2019 LidarPrimer - Collins

    20/35

    Kenai GIS User Group LiDAR Workshop 20/35

    Terrain Mapping-1

    The first returns yield the Digital Surface Model (DSM). In vegetated areas the DSM isthe Canopy Top.

    The Digital terrain Map (DTM, or bare Earth Surface) is inferred from a spatial filteringthat identifies the lowest returns that define a continuous surface.

  • 8/9/2019 LidarPrimer - Collins

    21/35

    Kenai GIS User Group LiDAR Workshop 21/35

    Terrain Mapping-2 Commercial survey in Puget Sound. Lidarfootprint sub-meter diameter, with two

    foot rints er s uare meter, and four returns

    Topographic Laser Ranging and Scanning:

    Principles and Processing Shan, J, and C.K. Toth, CRC Press, 2009.

  • 8/9/2019 LidarPrimer - Collins

    22/35

    Kenai GIS User Group LiDAR Workshop 22/35

    Increasing PRF

    Multiple Pulses in Air "MPiA"

    t0

    Time , t0

     

    The maximum range

    (i.e., the range of thelast return) and the

    PRF of the LiDAR

    system are related

    as,

    T PRF

    =

    1

    PRF

    =

    2"Rmax

    v

     

    Thus as an aircraft

    flies at higher

    altitude for large

    area surveys, the PRF(and hence the

    sampling density)must decrease.Announced in 2006 and commercially fielded in 2007,

    MPiA is now a mainstream LiDAR technology.

  • 8/9/2019 LidarPrimer - Collins

    23/35

    Kenai GIS User Group LiDAR Workshop 23/35

    Interleaving Pulses-1

    Region of

    Interest

    Region of ?

    0

    Rc

    Rg

           R

         a     n     g     e

    Aircraft

     

    For an aircraft high above theground at an altitude of ~1000

    m may only be interested in

    the ~100 m above the ground

    “the region of interest”.

    The air between the ground

    and the region of interest

     yields no significant echoes.

    Is it possible to transmit

    multiple pulses spaced suchthat their echoes do not

    interfere with each other.

  • 8/9/2019 LidarPrimer - Collins

    24/35

    Kenai GIS User Group LiDAR Workshop 24/35

    Interleaving Pulses-2

    0 tc tg

    Recieved

    Signal

    0 tc tg

    Recieved

    Signal

     T  P  R  F 

     t  c +  T  P  R  F 

    Pulse #1

    Pulse #1 Pulse #2 Pulse #3

     t  g+  T  P  R  F 

     2 

     x T  P 

     R  F 

     t  c  +  2 

     x T  P  R  F 

     t  g+  2 

     x T  P  R  F 

     

    The time from 0 to tc is“dead time”, while the

    signal of interest is foundbetween tc and tg.

    In a single pulse system thesecond pulse must await thereturn of the first pulse.

    In an interleaved systemthe second pulse istransmitted “early” so thatit can return just after thefirst pulse without waiting

    for tc.

    Thus the PRF can bedoubled and possibly betripled or more.

    0 tc tg

    Recieved

    Signal

    Pulse #1 Pulse #2 Pulse #3

     T  P  R  F 

     t  c 

     +  T  P  R  F 

     t  g

    +  T  P  R  F 

     2 

     x T  P  R  F 

     t  c  +  2 

     x T  P  R  F 

     t  g+  2 

     x T  P  R  F 

    Pulse #4

     3 

     x T  P  R  F 

     t  c 

    +  3 

     x T  P  R  F 

     t  g

    +  3 

     x

     T  P  R  F  4 

     x T  P  R  F 

     t  c  +  4 

     x T  P  R  F 

     t  g +  4 

     x T  P  R  F 

    Pulse #5

     

  • 8/9/2019 LidarPrimer - Collins

    25/35

    Kenai GIS User Group LiDAR Workshop 25/35

    MPiA Performance

    The relationship between PRF and flightheight for the Leica ALS60 system.Very high PRFs can exceed thecapabilities of current lasers. Thus

    multiple laser systems are used toachieve very high PRFs (e.g., ~400 kHz). 

    The point density at a PRF of 400 KHz isshown as a function of ground speed andflight height for the Reigl LMS-Q680iAt 400 kHz the maximum single pulse

    range is 375 m. Thus there are 3 pulsesin the air when flying at 800 m.

  • 8/9/2019 LidarPrimer - Collins

    26/35

    Kenai GIS User Group LiDAR Workshop 26/35

    Current Airborne LiDAR Systems

    Toth, C. K., LARS, 2010.

  • 8/9/2019 LidarPrimer - Collins

    27/35

    Kenai GIS User Group LiDAR Workshop 27/35

    Kenai LiDAR Mappingin 2008-1

  • 8/9/2019 LidarPrimer - Collins

    28/35

    Kenai GIS User Group LiDAR Workshop 28/35

    Kenai LiDAR Mapping in 2008-2

    LiDAR mapping of the 4550 square miles (11,780 km2) of the Kenai Peninsula.

    Kenai LiDAR Mapping

    Location Western Lowlands Eastern Kenai Watershed

    Area 3295 sq miles (8530 km2) 1255 sq miles (3250 km2)

    Post Spacing 1.4 m 3.0 m

    Horizontal Accuracy 1 m 2.0

    BareEarth RMSE 18.5 cm 50 cm

    •  Contract to Aero-Metric for LiDAR data collection and processing.•  Kenai Watershed Forum contract to Alaska Satellite Facility (ASF) at the Geophysical

    Institute-University of Alaska Fairbanks (GI-UAF) to provide Quality Assurance.

    ASF Quality Assurance included;1.  Review of formatting and completeness of data deliverables2. Review of the completeness, clarity, and compliance of the metadata3. Review of the contractor-provided quality assurance reports4. Evaluation of the planimetric accuracy of the LiDAR data5. Evaluation of the height accuracy of the LiDAR data6. Identification and characterization of any systematic errors observed in the data

  • 8/9/2019 LidarPrimer - Collins

    29/35

  • 8/9/2019 LidarPrimer - Collins

    30/35

    Kenai GIS User Group LiDAR Workshop 30/35

    Planimetric Assessment of Road Intersection from GI-UAF Geodetic Control

    Courtesy Rick Guritz, ASF, GI-UAF.

  • 8/9/2019 LidarPrimer - Collins

    31/35

    Kenai GIS User Group LiDAR Workshop 31/35

    Planimetric Assessment of Lake Point from GI-UAFGeodetic Control

    Courtesy Rick Guritz, ASF, GI-UAF.

  • 8/9/2019 LidarPrimer - Collins

    32/35

    Kenai GIS User Group LiDAR Workshop 32/35

     Alaskan LiDAR Mapping

    Region of

    Interest

    0

    Rg

           R     a     n     g     e

     GroundGround

    Snow

    Ground

    Leaves

    Early-April Late-April Early-May

     

    In wooded terrain the trees may shadow the ground. A better ground map will be obtainedif the survey is conducted AFTER snow has melted and BEFORE trees have leafed out.

  • 8/9/2019 LidarPrimer - Collins

    33/35

    Kenai GIS User Group LiDAR Workshop 33/35

    LiDAR Researchers atUniversity of Alaska Fairbanks

    Researcher Primary Focus Contact

    Anthony Arendt Glacier Mapping [email protected] Collins Atmospheric Sciences [email protected] Cunningham Terrestrial Mapping [email protected]

    Javier Fochesatto Atmospheric Sciences [email protected]

    Rick Guritz Terrestrial Mapping [email protected] Sassen Atmospheric Sciences [email protected]

  • 8/9/2019 LidarPrimer - Collins

    34/35

    Kenai GIS User Group LiDAR Workshop 34/35

    Eruption of Mount Augustine in 2006

    PUFF Model Forecast

    Lidar Data

  • 8/9/2019 LidarPrimer - Collins

    35/35

    Kenai GIS User Group LiDAR Workshop 35/35

     Acknowledgements

    LiDAR research at the University of Alaska Fairbanks has been conducted with

    support from the following;•  State of Alaska

    •  US Department of Agriculture

    •  US Department of Defense

    •  US Geological Survey

    •  US National Aeronautics and Space Administration

    • 

    US National Oceanic and Atmospheric Administration•  US National Science Foundation

    •  Government of Japan

    •  Fulbright Commission of Germany

    LiDAR research at the University of Alaska Fairbanks depends on the active

    participation of undergraduate and graduate students.

    Thanks to Rick Guritz and Keith Cunningham for helpful discussions.