• DocumentCode
    632094
  • Title

    eSPACE: Emergency spatial pre-SCAT for Arctic Coastal Ecosystem

  • Author

    Demers, Anne-Marie ; Banks, S. ; Duffe, Jason ; Carriere, Melanie ; Torontow, Valerie ; Chaudhary, Bhavana ; Laforest, Sonia

  • Author_Institution
    Landscape Sci. & Technol. Div, Environ. Canada, Ottawa, ON, Canada
  • fYear
    2013
  • fDate
    April 29 2013-May 3 2013
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    The eSPACE project (emergency Spatial Pre-SCAT for Arctic Coastal Ecosystems; SCAT: Shoreline Cleanup and Assessment Technique) is focused on developing capacity to enhance our state of preparedness for emergencies in the event of an oil spill in Canada´s North. Baseline coastal information is required for operational prioritization, coordination of on-site spill response activities and wildlife management. Earth Observation data from satellites such as RADARSAT-2 can potentially be used to identify and map shoreline characteristics, coastal habitats and resources at risk. The commonly used shoreline mapping method consists of manual interpretation of oblique videos to segment and classify existing shoreline vectors (based on 1:50k maps). In the summers of 2010, 2011 and 2012, geotagged video imagery was collected along the coastline for six selected pilot sites in the Arctic. RADARSAT-2 data was simultaneously acquired over these sites. An object-based classification scheme was developed using RADARST-2 data, SPOT-4/-5 optical data as well as ancillary data such as a digital elevation model. This paper presents the general methods used for all eSPACE sites, however only the results for the Richards Island pilot site are presented and discussed. Using ground observation data for validation, the classification reached an overall accuracy of 75% for Richards Island. Although the vectors produced from the standard method carry more information on the substrate types and on-site accessibility, the object-based classification allows for a spatial representation (area) of the shoreline classes as opposed to a line (vector), and the representation of all the shoreline types, not only the ones covered by the helicopter surveys. As well, given the vastness of Canada´s Arctic region, manual interpretation of the shorelines won´t be feasible.
  • Keywords
    artificial satellites; emergency services; geophysical image processing; helicopters; image classification; image representation; image segmentation; oceanographic techniques; remote sensing by radar; video signal processing; Canada; RADARSAT-2; SPOT-4/-5 optical data; ancillary data; artificial satellites; baseline coastal information; eSPACE; earth observation; emergency spatial Pre-SCAT for arctic coastal ecosystem; geotagged video imagery; ground observation data; helicopter; manual oblique video interpretation; object-based classification scheme; oil spill; onsite spill response activity coordination; operational prioritization; shoreline mapping method; shoreline representation; shoreline vector classification; shoreline vector segmentation; spatial representation; wildlife management; Accuracy; Arctic; Helicopters; Sea measurements; Sediments; Silicon; Videos;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Radar Conference (RADAR), 2013 IEEE
  • Conference_Location
    Ottawa, ON
  • ISSN
    1097-5659
  • Print_ISBN
    978-1-4673-5792-0
  • Type

    conf

  • DOI
    10.1109/RADAR.2013.6586151
  • Filename
    6586151