Next Article in Journal
Applicability of the Thermal Infrared Spectral Region for the Prediction of Soil Properties Across Semi-Arid Agricultural Landscapes
Previous Article in Journal
Simulating Forest Cover Changes of Bannerghatta National Park Based on a CA-Markov Model: A Remote Sensing Approach
Remote Sens. 2012, 4(10), 3244-3264; doi:10.3390/rs4103244
Article

Enhancing Coral Health Detection Using Spectral Diversity Indices from WorldView-2 Imagery and Machine Learners

*  and
Received: 16 August 2012; in revised form: 9 October 2012 / Accepted: 12 October 2012 / Published: 23 October 2012
View Full-Text   |   Download PDF [1680 KB, uploaded 19 June 2014]   |   Browse Figures
Abstract: The worldwide waning health of coral reefs implies an increasing need for monitoring them at colony scale over large areas. Relaying fieldwork considerably, the remote sensing approach can address this need in offering spectral information relevant for coral health detection with 0.5 m spatial accuracy. We investigated the potential of spectral diversity indices to achieve the discrimination of coral-dominated assemblages and health states from novel satellite imagery (WorldView-2, WV2). Both Equitability’s (E) and Pielou’s (P) operators were used to quantify the evenness of the corrected visible spectral bands (two times 26 combinations of five bands) corresponding to remotely sensed colonies. Three scleractinian corals (Porites lobata, P. rus and Acropora pulchra) that are primarily involved in Moorea’s reef building (French Polynesia) were examined in respect to their health state (healthy or unhealthy, referring to both bleached and dead coral, hereinafter). Using four classifiers, we showed that the Support Vector Machine (SVM) greatly discerned among the six coral classes based upon the five WV2 spectral bands (93%), thus surpassing the classification issued from the three traditionally used bands (80%). Coupling the WV2 dataset with Egreen-red, Eyellow-red or E“coastal”-blue-green allowed the SVM performance to attain 96%. On the other hand, adding the E“coastal”-blue to the WV2-dataset contributed to a substantially increase of the classification accuracy derived from the Random Forest classifier, stepping from 64% to 77%. Significant contributions of spectral diversity indices to surveying coral health were further discussed in the light of spectral properties of coral-related pigments. These findings may play a major role for the extensive monitoring of coral health states at a fine scale, and for the management and restoration of damaged coral reefs.
Keywords: coral health; spectral diversity index; WorldView-2; machine learner coral health; spectral diversity index; WorldView-2; machine learner
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Export to BibTeX |
EndNote


MDPI and ACS Style

Collin, A.; Planes, S. Enhancing Coral Health Detection Using Spectral Diversity Indices from WorldView-2 Imagery and Machine Learners. Remote Sens. 2012, 4, 3244-3264.

AMA Style

Collin A, Planes S. Enhancing Coral Health Detection Using Spectral Diversity Indices from WorldView-2 Imagery and Machine Learners. Remote Sensing. 2012; 4(10):3244-3264.

Chicago/Turabian Style

Collin, Antoine; Planes, Serge. 2012. "Enhancing Coral Health Detection Using Spectral Diversity Indices from WorldView-2 Imagery and Machine Learners." Remote Sens. 4, no. 10: 3244-3264.


Remote Sens. EISSN 2072-4292 Published by MDPI AG, Basel, Switzerland RSS E-Mail Table of Contents Alert