Prelaunch assessment of worldview-3 information content

Abstract

The upcoming WorldView-3 satellite is designed to collect unique data by combining very-high spatial resolution (VHR) with observation bands in the short wave infrared (SWIR) in addition to the visible and near-infrared (VNIR) multispectral and panchromatic bands currently available on the VHR WorldView-2 system. These SWIR bands were specifically selected in order to target unique reflectance and absorption features presented by various surface materials and should, therefore, significantly improve the platforms information content for many image mining applications. This presentation explores the information content available to the WorldView-3 platform in two ways. First, second-order statistics and mutual information estimates are utilized to measure the spectral content of simulated WorldView-3, WorldView-2, and QuickBird data relative to AVIRIS hyperspectral imagery. Then, WorldView-3 supervised classification performance is explored relative to that of hyperspectral imagery for both urban and agricultural data sets. Results suggest that the additional spectral content of the WorldView-3 platform provides a competitive information source to that of hyperspectral for broad applications in agricultural, mineral exploration, or urban monitoring.