Spectral Information Extraction from Worldview-2 Image for Urban Features Identification

Abstract

This study further explores the capability of very high-resolution Worldview-2 satellite image, which has four additional bands namely Coastal, Yellow, Red edge and Near-Infrared 2 in order to discriminate various types of urban land-cover features, as well as to determine the most potential band to extract features in the study area. By analyzing spectral information extract from statistical reflectance value, the potential spectral band that can be used to discriminate spectrally similar features can be determined. The output of this study is crucial as an information provider for further classification processing step via object-based image analysis approached. It is vital to have the information on the spectral characteristic of each spectral bands to extract specific urban features in order to set up appropriate ruleset as well as segmentation parameter later on based on result determine from this study.