Classification of mangrove species in the north eastern of Brazil based on hybrid images of remote sensing

Abstract

Satellite images have been used to map, to monitor and quantify the quality of natural resources. The detailed mapping of mangrove vegetation is a growing demand because it is an important management instrument, maintenance and knowledge of the mangrove ecosystem related to human activities and/or natural. This research has combined multispectral data from the visible area and infrared LANDSAT-8 satellite with images of spectral bands of the microwave of RADARSAT-2 satellite in the segmentation and classification of some mangroves in Brazil’s Northeastern. The hybrid composition between the optical system and microwave have showed excellent performance in identifying the geoenvironmental unit and allowed greater emphasis on floristic and structural properties of mangrove species. The results allowed the delimitation of the general area of mangrove’s occupation approximately in 5.538 ha and revealed the occurrence of four main species structure conditions: Rhizophora mangle (high/conventional), Rhizophora mangle (low/dense), Avicennia schaueriana and regions of mixed species. R. Mangle is the most abundant mangrove specie in the studied area occupying approximately an area of 3.513 ha about 63% of all mangrove. Mixed regions of species occupies area of 1.142 ha, representing 21% of total. A.schaueriana occupies an area of 882 ha, about 16% of total. The results indicate that the difference in reflectance of mangrove species is not just influenced by chlorophyll content by the species, the prevailing environmental conditions, soil and bottom water, but mainly by spacing of crowns or density of canopy for each occurring physiognomy/specie. This research aims to attend the expectations for greater efficiency in lifting temporal space with high accurancy for monitoring the quality of mangrove ecosystems, highly sensitive to environmental changes, as subsidy for its preservation.