Classification of mangroves vegetation species using texture analysis on Rapideye satellite imagery

Abstract

Mangroves are unique ecosystem structures that are typically made up of salt tolerant species of vegetation that can be found in tropical and subtropical climate country. Mangrove ecosystem plays important role and also is known as highly productive ecosystem with high diversity of flora and fauna. However, these ecosystems have been declining over time due to the various kinds of direct and indirect pressures. Thus, there is an increasing need to monitor and assess this ecosystem for better conservation and management efforts. The multispectral RapidEye satellite image was used to identify the mangrove vegetation species within the Matang Mangrove Forest Reserve in Perak, Malaysia using texture analysis. Classification was implemented using the maximum likelihood classifier (MLC) method. Total of eleven main mangrove species were found in the satellite image of the study site which includes Rhizophora mucronata, Rhizophora apiculata, Bruguiera parviflora, Bruguiera cylindrica, Bruguiera gymnorrhiza, Avicennia alba, Avicennia officinalis, Sonneratia alba, Sonneratia caseolaris, Sonneratia ovata and Xylocarpus granatum. The classification results showed that the textured image produced high overall classification assessment recorded at 84% and kappa statistic of 0.8016. Meanwhile, the non-textured image produces 80% of overall accuracy and kappa statistic of 0.7061. The classification result indicated the capability of high resolution satellite image to classify the mangrove species and inclusion of texture information in the classification increased the classification accuracy.