Extraction of Water Bodies from Visible Color Satellite Images Using PCA Feature Map

Abstract

Water body extraction from remote sensing imagery is important in planning and monitoring of water resources. It is a complex task due to the diversity of water quality, and similarity of spectral characteristics with the background information. In this paper, an algorithm is proposed to extract the water bodies from the RGB visible color satellite images using principal component analysis (PCA) feature. First, the PCA features are extracted from the input image and select PCA component 2 feature and segmented it into binary classes by using threshold method. Further, this binary outcome is enhanced by morphological erode operation to obtain the final water regions. The qualitative analysis of experimental outcomes illustrate that the proposed algorithm extract most of the water regions with least false prediction rate.