Scribd Logo
Upload

Welcome to Scribd!

  • 81 views

    A Systematic High Resolution Remote Sensing Capture For The Magela Creek Catchment: A Case Study Using Worldview 2

    Uploaded by

    NT Spatial
    This document summarizes a study that used a WorldView-2 satellite to capture high resolution imagery of the Magela Creek Catchment in Australia's Northern Territory. Ground control targets and radiometric calibration targets were established to geometrically and radiometrically calibrate the imagery. The resulting imagery achieved sub-meter geometric accuracy and was useful for environmental monitoring applications in the region.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    A Systematic High Resolution Remote Sensing Capture For The Magela Creek Catchment: A Case Study Using Worldview 2

    Uploaded by

    NT Spatial
    81 views28 pages
    This document summarizes a study that used a WorldView-2 satellite to capture high resolution imagery of the Magela Creek Catchment in Australia's Northern Territory. Ground control targets and radiometric calibration targets were established to geometrically and radiometrically calibrate the imagery. The resulting imagery achieved sub-meter geometric accuracy and was useful for environmental monitoring applications in the region.

    Original Description:

    Presentation from NT Spatial 2011: The State of Play

    Original Title

    201110_RBartolo_ASystemHighResRemoteSensingCaptMagelaCrk

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes a study that used a WorldView-2 satellite to capture high resolution imagery of the Magela Creek Catchment in Australia's Northern Territory. Ground control targets and radiometric calibration targets were established to geometrically and radiometrically calibrate the imagery. The resulting imagery achieved sub-meter geometric accuracy and was useful for environmental monitoring applications in the region.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    81 views28 pages

    A Systematic High Resolution Remote Sensing Capture For The Magela Creek Catchment: A Case Study Using Worldview 2

    Uploaded by

    NT Spatial
    This document summarizes a study that used a WorldView-2 satellite to capture high resolution imagery of the Magela Creek Catchment in Australia's Northern Territory. Ground control targets and radiometric calibration targets were established to geometrically and radiometrically calibrate the imagery. The resulting imagery achieved sub-meter geometric accuracy and was useful for environmental monitoring applications in the region.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 28

    www.environment.gov.

    au/ssd

    A systematic high resolution remote sensing capture for


    the Magela Creek Catchment: A case study using
    WorldView 2
    Spatial Sciences & Data Integration Program
    Presenter: Dr R Bartolo#
    e: renee.bartolo@environment.gov.au
    # Program Leader, Spatial Sciences and Data Integration

    Supervising Scientist - working to protect the environment from the impacts of uranium mining
    Outline
    ■ Study Region

    ■ Problem statement and method question

    ■ Description of WorldView 2 satellite

    ■ Scheduling a capture
    – Specifications, costs and logistics

    ■ Ground Control Targets


    – Size & material
    – Positioning & layout

    ■ Radiometric Calibration Targets

    ■ Results

    ■ Lessons Learned

    Supervising Scientist Division 2


    Study Region: Alligator Rivers Region

    Darwin

    Jabiru

    Alligator Rivers Region


    28,000 km2

    Supervising Scientist Division 3


    Study Region: Alligator Rivers Region (KNP)

    Jabiru

    Kakadu National Park


    19,804 km2

    Supervising Scientist Division 4


    Study Region: Magela Creek Catchment
    Magela Creek Catchment
    1,640 km2

    Magela Floodplain
    236 km2

    M
    Jabiru ag
    el
    a
    Ranger Cr
    ee
    k

    Supervising Scientist Division 5


    Problem statement: Geometeric control
    ■ For the Alligator Rivers Region and in particular
    the Magela Creek Catchment & floodplain area,
    there has been a lack of geometric control for
    high resolution satellite data.

    Supervising Scientist Division 6


    ALOS AVNIR-2 2006 ALOS AVNIR-2 2007

    Jabiru

    X = Differentially corrected location


    100 Man
    Camp

    X X

    Supervising Scientist Division 7


    Problem statement: Calibration of data to reflectance
    values
    ■ How do we correct raw image data to reflectance values
    (removal of atmospheric effects)? Why do we need to do
    this?
    Atmospheric refraction

    After Jensen, 2000


    Supervising Scientist Division 8
    Atmospheric Scattering

    Type of scattering is a function of:

    1)Wavelength of the incident


    radiation

    2) Size of the gas molecule, dust


    particle &/or water vapour

    After Jensen, 2000


    Supervising Scientist Division 9
    Supervising Scientist Division 10
    Conversion Raw Data Æ Radiance Data Æ Reflectance
    Data

    ■ Raw DNs: surface reflectance, solar irradiance curve,


    atmospheric effects, variation in illumination resulting
    from topographic effects, instrument response.
    ■ Raw Data Æ Radiance Data removes instrument effects

    ■ Radiance Æ Reflectance
    L0(λ)=Lsun(λ) T(λ) R(λ) cos(θ) + Lpath(λ)
    – L0(λ) = observed radiance at sensor
    – Lsun(λ) = Solar irradiance above atmosphere
    – T(λ) = total atmospheric transmittance
    – R(λ) = surface reflectance Note:
    – θ = incidence angle Reflectance data scaled to get to
    – Lpath(λ) = path scattered radiance integer data (typically x 10,000)

    Supervising Scientist Division 11


    Why calibrate and correct data?
    ■ To compare multi-date images – some data sets even
    have different atmospheric properties across a scene

    ■ To compare data sets from different sensors

    ■ Needed for quantitative analysis, e.g., working with field


    data

    ■ When using band ratios such as EVI, NDVI, MNDWI

    ■ Reflectance data needed to compare data spectra with


    library reflectance spectra – helps in identifying materials
    based on their absorption features
    ■ To use spectral library to map materials, image must
    be in reflectance.

    Supervising Scientist Division 12


    Method Question:
    ■ How do we undertake a systematic remote
    sensing capture?
    – Logistics
    – What data do we acquire (format, bands etc)
    – Resources

    Supervising Scientist Division 13


    WorldView-2

    ■ Launch date: 08 October 2009


    – Estimated mission life is 7.25 years
    – Operating at full capability on 4 January 2010

    ■ Part of the DigitalGlobe constellation


    – WorldView 1 and QuickBird

    ■ Altitude: 770 kms

    ■ Revisit capability
    – 1.1 days at 1m GSD or less
    – 3.7 days at 20° off-nadir or less (52 cm GSD)
    Supervising Scientist Division 14
    The 8 Spectral Bands of WorldView-2
    4 New Spectral Bands Pan
    Coastal
    Coastal Band (400 – 450nm) - Blue
    can be used in vegetation Green
    identification & analysis Yellow
    (chlorophyll characteristics) Red
    Red Edge
    Yellow Band (585 – 625nm) - NIR1
    identify “yellow-ness” NIR2
    characteristics of targets, 350 450 550 650 750 850 950 1050
    Important for vegetation
    Wavelength (nm)
    applications
    Sensor Resolution (GSD)
    Red Edge Band (705 – 745nm) -
    analysis of vegetation condition. Directly Panchromatic: 0.46 meters GSD at nadir
    0.52 meters GSD at 20° off-nadir
    related to plant health via chlorophyll
    Multispectral: 1.84 meters GSD at nadir
    production. 2.08 meters GSD at 20° off-nadir
    (note that imagery must be re-sampled to 0.5
    NIR2 (860 – 1040nm) - overlaps NIR1 m (Pan) and 2m (mulit) for non-US Government customers)
    band but less affected by atmospheric
    influence. Used for vegetation analysis &
    Swadth Width:
    biomass studies.
    16.4 kms at nadir
    Supervising Scientist Division 15
    Scheduling a Capture
    ■ How do we undertake a systematic remote
    sensing capture?
    – Specifications
    – Costs
    – Logistics

    Supervising Scientist Division 16


    Setting up a back up base station

    Supervising Scientist Division 17


    Ground Control Targets

    • 27 3.5 x 3.5 m square silver–coloured


    tarpaulins positioned on the ground prior to
    the image acquisition window

    • 6 image objects (features such as cross


    roads evident in previous image data)

    • Centre of each of the tarpaulins and


    image objects were measured with a GPS
    & differentially corrected to within 12 mm
    x,y accuracy.
    Supervising Scientist Division 18
    Field Data: Panorama using MS-ICE &
    Photosynth

    Supervising Scientist Division 19


    Field Data: Google Earth

    Supervising Scientist Division 20


    Radiometric Calibration Targets & Recording
    Atmospheric Conditions

    Supervising Scientist Division 21


    Radiometric Calibration Targets

    Black synthetic upholstery

    Silver plastic weave tarpaulin

    White plastic weave tarpaulin

    Tyvec
    Supervising Scientist Division 22
    Pseudo-invariant Features: Golf Green

    Supervising Scientist Division 23


    Results: Geometric Correction

    Root Mean Square Error for the WV-2 Multispectral band

    Eriss = orthorecitification GCPs & RPC file DEM


    DigitalGlobe = Standard map ready product

    ERISS DigitalGlobe
    Site Standard
    Multispectral Multispectral
    25 2.6 4.6

    31 0.4 3.9

    32 3.2 6.4

    100 ManCamp 2.0 3.8

    ENRAD Bld 0.5 6.4

    Bush humpy 2.2 8.8

    Mean RMSE 1.8 5.6

    Supervising Scientist Division 24


    Results: Geometric Correction

    Root Mean Square Error for the WV-2 Panchromatic band

    Eriss = orthorecitification GCPs & RPC file DEM


    DigitalGlobe = Standard map ready product

    ERISS DigitalGlobe
    Site
    Panchromatic Panchromatic

    25 2.1 6.0
    26 0.8 7.2
    31 1.1 6.1
    32 2.7 6.3
    100 ManCamp 0.4 4.3
    Westcott Bld 2.6 6.5
    ENRAD Bld 0.4 4.4
    Bush humpy 1.4 8.8
    Mean RMSE 1.4 6.2

    Supervising Scientist Division 25


    Results: Conversion to Reflectance

    ■ Conversion to reflectance methods

    1) Scene derived (statistics): eg. Internal Average


    Relative Reflectance (IAR)

    2) Ground calibration methods


    Empirical Line Regression

    3) Radiative Transfer Models: eg FLAASH

    Supervising Scientist Division 26


    Results: Conversion to Reflectance

    Native Grass

    A total of nine field targets were used to derive the calibration


    model, resulting in a strong linear relationship (r2 = 0.99) for
    all 8 multispectral bands.

    Accuracy assessment of the calibration model was undertaken


    by comparing the image reflectance values against the
    surface reflectance values of 17 validation field targets.
    Supervising Scientist Division 27
    Some issues:
    • Majority (95%) of the areas of the three requested
    scenes were captured with the specified scene
    parameters (nadir angle of 13.8º and total cloud cover
    <2%) on 11 May.

    • The remaining five percent of one of the scenes was


    captured on 22 May (nadir angle of 11.6º). All 27
    ground control tarpaulins were visible in the imagery.

    Supervising Scientist Division 28

    You might also like