DEVELOPMENT OF MANGROVE ZONATION PATTERN MAP USING OBJECT BASED IMAGE ANALYSIS (OBIA) FOR DENSE MANGROVE COVER

摘要

Tolerability to the specific ranges of environmental parameters form a limiting factor to the mangrove species distribution. This characteristic is translated into the formation of mangrove zones in a mangrove forest. Mapping the mangrove zonation pattern is a challenging task especially when the mangrove cover is thick and dense. The aim of this study is to identify and map the mangrove zonation pattern at Pulau Kukup, Johor using a high resolution WorldView-2 satellite data and Object Based Image Analysis (OBIA) system, SPRING 5.2. Mangrove extent was extracted from multispectral and panchromatic images using region growing segmentation method. Several thresholds were used to identify the best-fit segmentation parameters. Eleven plots of 100m transects were established in the study area to sample the representative mangrove trees. The type of mangrove species, tree height, Diameter at the Breast Height (DBH), elevation, and coordinate location were collected during the field survey based on Point-Centre-Quadrate Method (PCQM). The general characteristics of mangrove tree were investigated and sampled from 186 mangrove trees. Rhizophora apiculata and R.mucronata were found dominating the outer part of the mangrove island facing the seaward area. From the field analyses, the relative density of Rhizophora apiculata and R.mucronata are 41.9% and 16.7% respectively. The mangrove coordinate locations were then recorded and used as a reference point for classification process using the Bhatacharyya distance. Three multispectral band combinations were tested for classification. Classified map with the highest accuracy level (94.05%) was constructed using the band combination 7 (near infrared-1), 6 (red-edge) and 5 (red) for the study area. From the study, it was found that the application of OBIA technique in mangrove mapping had successfully classified the dense mangrove cover. It concluded that the OBIA technique offers an alternative way to map the dense mangrove distribution and can potentially be adopted to other ecosystem.