Image segmentation for vegetation types extraction using WorldView-2: a case study in parts of Dieng Plateau, Central Java

Abstract

Image segmentation is the most important stage on Geographic Object Based Image Analysis (GEOBIA). The result of segmentation affects the final accuracy of classification. One of the applications of image segmentation operations is to delineate vegetation objects. Further analysis of vegetation could be used for inventory of natural resources, agricultural, land cover, land use, etc. However, applying image segmentation for separating vegetation types is challenging due to their irregular shapes and various patterns and colors. This study aims to determine the optimum parameters of image segmentation for delineating vegetation types using a pan-sharpened WorldView-2 image (0.5 m pixel size) which was acquired on August 2018. Combinations of scale parameter and composition of homogeneity criterion (shape and compactness) were systematically simulated to obtain the best segmentation parameters. The result of segmentation was assessed quantitatively based on visually interpreted image map as a reference. This study found that application of shape and compactness simultaneously for vegetation extraction would produce rough segmentation result. The optimum parameters for segmenting vegetation types using WorldView-2 were using scale parameter of 5, shape of 0 and compactness of 0.5.