Open AccessThis article is
- freely available
Bridging Ridge-to-Reef Patches: Seamless Classification of the Coast Using Very High Resolution Satellite
Tokyo Institute of Technology, O-okayama 2-12-1-W8-13, Meguro-ku, Tokyo 152-8552, Japan
Institut des Sciences de la Mer, Université du Québec à Rimouski, Rimouski, QC G5L 3A1, Canada
USR 3278 CRIOBE CNRS-EPHE, BP 1013, Papetoai, Moorea 98729, French Polynesia
* Author to whom correspondence should be addressed.
Received: 28 May 2013; in revised form: 11 July 2013 / Accepted: 12 July 2013 / Published: 22 July 2013
Abstract: The structure and functioning of coral reef coastal zones are currently coping with an increasing variety of threats, thereby altering the coastal spatial patterns at an accelerated pace. Understanding and predicting the evolution of these highly valuable coastal ecosystems require reliable and frequent mapping and monitoring of both inhabited terrestrial and marine areas at the individual tree and coral colony spatial scale. The very high spatial resolution (VHR) mapping that was recently spearheaded by WorldView-2 (WV2) sensor with 2 m and 0.5 m multispectral (MS) and panchromatic (Pan) bands has the potential to address this burning issue. The objective of this study was to classify nine terrestrial and twelve marine patch classes with respect to spatial resolution enhancement and coast integrity using eight bands of the WV2 sensor on a coastal zone of Moorea Island, French Polynesia. The contribution of the novel WV2 spectral bands towards classification accuracy at 2 m and 0.5 m were tested using traditional and innovative Pan-sharpening techniques. The land and water classes were examined both separately and combinedly. All spectral combinations that were built only with the novel WV2 bands systematically increased the overall classification accuracy of the standard four band classification. The overall best contribution was attributed to the coastal-red edge-near infrared (NIR) 2 combination (Kappagain = 0.0287), which significantly increased the fleshy and encrusting algae (User’s Accuracygain = 18.18%) class. However, the addition of the yellow-NIR2 combination dramatically impacted the hard coral/algae patches class (Producer’s Accuracyloss = −20.88%). Enhancement of the spatial resolution reduced the standard classification accuracy, depending on the Pan-sharpening technique. The proposed composite method (local maximum) provided better overall results than the commonly used sensor method (systematic). However, the sensor technique produced the highest contribution to the hard coral thicket (PAgain = 30.36%) class with the coastal-red edge-NIR2 combination. Partitioning the coast into its terrestrial and aquatic components lowered the overall standard classification accuracy, while strongly enhancing the hard coral bommie class with the coastal-NIR2 combination (UAgain = 40%) and the green-coastal Normalized Difference Ratio (UAgain = 11.06%). VHR spaceborne remote sensing has the potential to gain substantial innovative insights into the evolution of tropical coastal ecosystems from local to regional scales, to predict the influence of anthropogenic and climate changes and to help design optimized management and conservation frameworks.
Keywords: coastal mapping; seamless; coral reefs; very high resolution; WorldView-2
Article StatisticsClick here to load and display the download statistics.
Notes: Multiple requests from the same IP address are counted as one view.
Cite This Article
MDPI and ACS Style
Collin, A.; Archambault, P.; Planes, S. Bridging Ridge-to-Reef Patches: Seamless Classification of the Coast Using Very High Resolution Satellite. Remote Sens. 2013, 5, 3583-3610.
Collin A, Archambault P, Planes S. Bridging Ridge-to-Reef Patches: Seamless Classification of the Coast Using Very High Resolution Satellite. Remote Sensing. 2013; 5(7):3583-3610.
Collin, Antoine; Archambault, Philippe; Planes, Serge. 2013. "Bridging Ridge-to-Reef Patches: Seamless Classification of the Coast Using Very High Resolution Satellite." Remote Sens. 5, no. 7: 3583-3610.