Object-based analysis of WorldView-2 imagery of urban areas

In recent years, remote sensing has become an increasing data source to support urban planning and management due to availability of very high resolution (VHR) images having resolution of less than 0.5 m. Such images allow extracting of detailed information of various targets of urban areas with the help of object-based image analysis (OBIA) in contrast to traditional pixel-based methods [1]. The aim of the study is to investigate the capabilities and robustness of newly available 8-Band images of WorldView-2 (WV-2) for improving urban mapping capabilities with the help of object-based image analysis framework proposed by Rizvi and Mohan [2]. Within naturally occurring classes or within manmade categories it was found with 4-band Quickbird images that there were misclassification of certain objects like clear water and dense shadow [2]. The availability of better spectral capability of Worldview-2 offers greater discrimination capability along with object-based analysis. Therefore, main focus of this research is how this improved spatial and spectral resolution can contribute to extract urban features from the Worldview-2 image using an object-based analysis framework developed by the authors.