Grid pattern based residential area detection from Hyperion data

Satellite hyperspectral imagery has been widely used in geological and agricultural studies. However, its use for an urban environment is limited due to low spatial resolution. Few methods in the literature attempt to use both spectral and spatial information in hyperspectral imagery. In this paper, a new method of performing spectral correlation and detecting urban signatures by analyzing discrete Fourier components of a correlation image is presented. The results from the recognition of residential building areas are demonstrated.