Proceedings Article | 18 September 2009
Proc. SPIE. 7472, Remote Sensing for Agriculture, Ecosystems, and Hydrology XI
KEYWORDS: Image segmentation, Remote sensing, Ecosystems, Landsat, Image processing algorithms and systems, Satellite imaging, Satellites, Image classification, Earth observing sensors, Oceanography
The coralline reefs in Banco Chinchorro, Mexico, are part of the great reef belt of the western Atlantic. This reef
complex is formed by an extensive coralline structure with great biological richness and diversity of species. These
colonies are considered highly valuable ecologically, economically, socially and culturally, and they also inherently
provide biological services. Fishing and scuba diving have been the main economic activities in this area for decades.
However, in recent years, there has been a bleaching process and a decrease of the coral colonies in Quintana Roo,
Mexico. This drop is caused mainly by the production activities performed in the oil platforms and the presence of
hurricanes among other climatic events. The deterioration of the reef system can be analyzed synoptically using remote
sensing. Thanks to this type of analysis, it is possible to have updated information of the reef conditions.
In this paper, satellite imagery in Landsat TM and SPOT 5 is applied in the coralline reefs classification in the 1980-
2006 time period. Thus, an integral analysis of the optical components of the water surrounding the coralline reefs, such
as on phytoplankton, sediments, yellow substance and even on the same water adjacent to the coral colonies, is
performed. The use of a texture algorithm (Markov Random Field) was a key tool for their identification. This algorithm,
does not limit itself to image segmentation, but also works on edge detection. In future work the multitemporal analysis
of the results will determine the deterioration degree of these habitats and the conservation status of the coralline areas.
